Advances in Immunology Volume 35

ADVANCES IN Immunology VOLUME 3 5 CONTRIBUTORS TO THIS VOLUME DENNIS J . BEER STEPHENT. CREWS RICHARDDOUGLAS PATRIC...

Author: Author Unknown

19 downloads 1311 Views 17MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES IN

Immunology VOLUME 3 5

CONTRIBUTORS TO THIS VOLUME

DENNIS J . BEER STEPHENT. CREWS RICHARDDOUGLAS PATRICIA J. GEARHART LEROY HOOD DAVID R. JACOBY NELSON JOHNSON STEVENM . MATLOFF NADINENIVERA ROBERTJ. NORTH LARSB. OLDING MICHAELB. A. OLDSTONE ROGERM. PERLMUTTER Ross E. ROCKLIN JOHN ROGERS GREGSORENSEN HANSL. SPIEGELBERG RANDOLPHWALL

ADVANCES IN

Immunology BY

EDITED HENRY G. KUNKEL

FRANK J. DIXON

The Rockefeller hiversify New York, N e w York

Scripps Clinic and Research Foundation La Jollo, California

VOLUME 35

1984

ACADEMIC PRESS, INC. (Harcourt Brace Jovonovich, Publishers]

Orlando Son Diego Son Francisco New York Toronto Montreal Sydney Tokyo

London

Slo Poulo

COPYRIGHT @ 1984,

BY

ACADEMIC PRESS,INC.

ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TIWNSMXTTBD IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

Orlando, Flonda 32887

United Kingdom Edition publislied by ACADEMIC PRESS, INC. (LONDON) LTD. 24\28 Oval Road, London N W l IDX

LIBRARY OF

CONGRESS CATALOG CARD

NUMBER:6 1 - 1 705 7

ISBN 0-12-022435-6 PRINTED IN THE UNITED STATES OF AMERICA 84 85 86 87

9 8 7 6 5 4 3 2 1

CONTENTS

CONTRIBUTORS ....................................................... PREFACE ............................................................ HENRYG K U N K E L. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix xi

xv

The Generation of Diversity in Phosphorylcholine-Binding Antibodies

ROGERM. PEHLMUTIXR,STEPHENT. CREWS,RICHAHI)DOCJGLAS, SORENSEN, NELSONJOHNSON, NADINENIVERA, IIOOD PATRICIA J. GEARHAR'I', AND LEROY

GREG

I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . .

11. Restricted Nature of the Anti-PC Response. . 111. Clonotypes of Anti-PC Antibodies . . . . . . . . .

IV. Hidden Diversity of the Anti-PC Response. . . . . . . . . . . . . . . . . . . . . . . . V. T h e Structure of the Variable Regions of Antibodies Which Bind Phosphorylcholine and the Molecular Basis of Their Diversity . . . . . . . ...................... VI. Heavy Chain Variable Regions VII. Somatic Diversification of Heavy Chain Variable Regions. . . . . . . . . . . . VIII. Somatic Diversification Can Be Extensive an Boundaries of the VII G e n e . . . . . . . . . . . . . . . IX. Somatic Diversification Probably Occurs by a ..................... Mechanism. . . . . . . . . . X. Soniatic Mutation Prob That Is Localized in and around t h e V,, G e n e . . . . . . . . . . . . . . . . . . . . . XI. Light Chain Variable Regions . . . . . . . . . . . . . XII. Somatic Diversification of Light Chain G e n e s . . . . . . . . . . . . . . . . . . . . . XIII. T h e Pattern of Variation by Somatic Hypermutation ......... XIV. Soniatic Mutation Is Correlated with Innnunoglobuli XV. V H , D, and J I I Segments a n d Junctional Diversity XVI. Diversity in the VII S e g m e n t . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . ...................... XVII. Diversity in the JtI Segment XVIII. Diversity in the D Sekment.. . . . . . . . . . . . . . , ......................................... XIX. T h e N Region XX. J. G e n e Segments and Junctional Diversity . . . . . . . . . . . . . . . . . . . . . . XXI. Summary of Diversity of Antibodie XXII. Selection of Variant Antibodies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XXIII. Molecular Basis of the T I 5 Idiotype. . XXIV. T h e T15 VIIG e n e Family . . . . . . . . . . . . , . . . , . . . . . . . . . . . . . . . . . . . . . ..................... XXV. Evolution of the T15 G e n e Fariiily XXVI. Structural Diversity in Antihodies to P C . . . . XXVII. Future Research ......................................... ...., .... References . . . . . . . . . . . . . . . . . . . . . . . .

.

V

6 7 7 15 16

17 18 19 20 22 22 23 23 25 25 26 26 27 29 29 30 31 33 35

vi

CONTENTS

Immunoglobulin RNA Rearrangements in 6 Lymphocyte Differentiation

JOHNROGERSAND RANDOLPH WALL I. 11. 111. IV. V. VI. VII. VIII. IX.

.

................................... Introduction. . . . . . . . . . Immunoglobulin Structrir ................................... Innnunoglobdins as Antigen Receptors on B Cells . . . . . . . . . . . . . . . . . Two mRNAs with Different 3' Ends Encode Membrane and Secreted p Heavy Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . All Immunoglobulin Heavy Chain Genes Have Membrane Gene Segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Model for the Transmembrane Insertion ............. Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Membrane and Secreted Heavy Chain mR Complex Transcription Units. . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . Coexpressed p and 6 mRNAs Are Coded by a Very Complex Transcription Unit. . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary: Developmental Regulation of Heavy Chain Gene Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. .

.

................................................

39 40 42 43 46

47 50 52 55 56

Structure and Function of Fc Receptors for IgE an Lymphocytes, Monocytes, and Macrophages

HANSL. SPIECELBERC Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Affinity for IgE and Structure of FceR on Lymphocytes and M+ . . . . . . Rosette Assays for Detection of FcsR. . . . . . . . . . . . . . . . . . . . . . . . . . . . . FcsR+ Cultured Lymphocytes and MI$ and FUR+ Leukemic Lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. FcsR+ B and T Cells in Nonatopic Healthy Humans. . . . . . . . . . . . . . . . VI. FcsR+ B and T Cells in Atopic Patients. . . . . . . . . . . . . . . . . . . . . . . . . . . VII. FcsR+ Lymphocytes in Normal and Parasitically Infected Rats and

I. 11. 111. IV.

...................................

VIII. FcsR+ Monocytes in Nonatopic and Atopic Humans . . . IX. FceR+ Rat and M O L I SM+. ~ . . .. . . . . . . . , . . . .. . . . . . . . . . . . . . . . . . . . . X. FcsR+ Eosinophils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI. Conclusions: Induction and Function of FcsR on Lymphocytes and . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

61 62 65

68 69 71 74

77 79 80 81 85

The Murine Antitumor Immune Response and Its Therapeutic Manipulation

ROBERTJ. NORTH I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Analysis of Antitumor Immunity by Adoptive Immunization against Established Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. T h e Meaning of the Adoptive Immunization Assay . . . . . . . . . . . . . . . . .

..

. .

.

.

89 92 107

CONTENTS

IV . Analysis of Concomitant Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Tumor Innuunotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

vii 122 133 142 153

Immunologic Regulation of Fetal-Maternal Bolance

DAVIDR . JACOBY. LARSB. OLDING. AND MICHAELB. A . OLDSTONE 1. Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I1 . Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Proposed Mechauisms for Maintenance of the Fetus . . . . . . . . . . . . . . . . IV . Fetal Expression of Histocompatibility Antigens . . . . . . . . . . . . . . . . . . .

V . Immunologic Basis of Lymphocyte Interactions between Mother and Fetus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Human Maternal and Neonatal Lymphocyte Interactions . . . . . . . . . . . . VII . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..........................................

157 158 159 170 172 180 199 202

The Influence of Histamine on Immune and lnflammotory Responses

DENNIS J . BEER. STEVEN M . MATLOW. A N D Ross E . ROCKLIN

I . Histamine as an Autacoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Histamine Modulation of Polymorphonuclear Inflammatory Cells . . . . . I11. Histamine Modulation of Iniinune Effector Cells . . . . . . . . . . . . . . . . . . . IV. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . INDEX ............................................................... CONTENTS OF PREVIOUS VOLUMES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

209 215 223 262 263 269 273

This Page Intentionally Left Blank

CONTRIBUTORS

Nrimbers in parentheses indicate the pages on which the authors' contributions begin.

DENNIS J. BEER(209), Pulmonary Medicine Section, Evans Memorial Department of Clinical Research, Boston University School of Medicine, Boston, Masscichusetts 02118 STEPHENT. CREWS~ (l),Division of Biology, California Institute of Technology, Pasadena, Californici 91 125

RICHARDDOUCLAS~ (l),Division of Biology, California Institute of Technology, Pasadena, California 91 125 PATRICIA J. GEAHHART' (I),Divi-sion of Biology, California Institute of Technology, Pnsadenci, California 91 125 LEnoY HOOD(l),Division of Biology, California lnstitute of Technology, Pasadena, California 921 15 DAVID R. JACOBY (157),Department of Immunology, Scripps Clinic and Research Foundation, La Jolla, California 92037

NELSONJOHN SO^ (I), Division of Biology, California Institute of Technology, Pasadena, Cnlifornia 91125 STEVENM , MArmtw (209), Allergy Division, Department of Medicine, New England Medical Center, Tufts University School of Medicine, Boston, Massachusetts 02111 NADINE NIVERA~ (l),Division of Biology, California Institute of Technology, Pasadena, Californici 91 125 ROBERTJ. NoRm (89), Trudeau Institute, Inc., Saranac Lake, New York 12983

' '

Present address: Department of Pathology, Stanford University, Stanford, California 94305. Present address: Integrated Genetics, Framingham, Massachusetts 01707. -iPresent address: Department of Biochemistry, Johns Hopkins University School of Hygiene and Public Health, Baltimore, Maryland 21205. Present address: Department of Biochemistry, State University of New York at Stony Brook, Stony Brook, New York 11790. Present address: Deparhnent of Biochemistry, Johns Hopkins University School of Hygiene and Public Health, Baltiniore, Maryland 21205.

'

ix

CONTRIBUTORS

X

LARSB. OLDING(157), Department of Pathology, University of Goteborg, Goteborg, Sweden MICHAELB. A. OLDSTONE (157),Department of Immunology, Scripps Clinic and Research Foundation, La Jolla, California 92037 ROGER M. PERLMUTTER (l),Division of Biology, California Institute of Technology, Pasadena, Calqornia 91 125

Ross E. ROCKLIN(209), Allergy Division, Department of Medicine, New England Medical Center, Tufts University School of Medicine, Boston, Massachusetts 021 11

ROGERS(39), MRC Laboratory of Molecular Biology, University Medical School, Cambridge CB2 2QH, England

JOHN

GREGSORENSEN (l),Division of Biology, Cal$ornia Institute of Technology, Pasadena, California 91 125 HANSL. SPIEGELBERC (61),Depnrtment of Immunology, Research l n stitute of Scripps Clinic, La Jolla, California 92037

RANDOLPHWALL(39), Molecular Biology Institute, and Department of Microbiology and Immunology, UCLA School of Medicine, LOS Angeles, California 90024

PREFACE

Progress in the field of inmunology continues at an ever-increasing rate and at every level of investigation. The once mystifying maneuvers of DNA as a prelude to antibody formation and the manipulation of RNA in the course of carrying out orders from the immunologic genome are now reasonably well understood. No longer do we need to be primarily concerned with the basis of antibody diversity nor with the mechanisms of translating genetic information into antibody molecules. The complicated events underlying the manifestations of immunologic diseases are becoming better understood in terms of the cell types involved, their regulation provided largely by products of immunocytes and their effector mechanisms. The complex interrelationship between a host and a partially incompatible graft in the form of either a conceptus or a neoplasm is also being elucidated. The most effective defense of a fetus against an immunologically sensitized mother appears to be conducted by fetal suppressor T cells, which fight their battle in the trenches of the placenta. As we learn more about the nature of antitumor immune responses and the reasons for their relative ineffectiveness, possibly strategies may be devised that can influence the outcome of the host-tumor struggle. These are the subjects addressed in this volume, and they represent exciting excerpts from the broad spectrum of immunologic research. Drs. Roger Perlmutter, Leroy Hood, and associates have chosen the murine antibodies directed against phosphorylcholine as a model system for studying the generation of antibody diversity, a subject to which they have been major contributors. In the first article they review the various elements contributing to diversity including the combinatorial association of heavy and light chains and joining of germline gene segments, the variable joining within gene segments, the appending of additional nucleotides to the D segment, and finally the somatic hypermutation operating coordinately on all V, J, and D gene segments. Together these mechanisms can generate a diverse repertoire of similar but distinct antibody specificities from a single germline V gene. The operation of these different diversity producing events is considered in the general context of B cell maturation, plating both molecular and cellular events in perspective. The role of posttranscriptional RNA processing in the regulation and differentiation of B lymphocytes is reviewed in the second article by Drs. John Rogers and Randolph Wall. Discovery of the membrane gene segment established the p heavy chain gene as the first example of a complex transcriptional unit in chromosomal DNA. This unit proxi

xii

PREFACE

duced two heavy chain p mRNAs with different 3' spliced structures coding for secreted and membrane bound forms. The authors, who were key movers in this development, predicted and later found a similar complex transcriptional unit responsible for secreted and membrane forms of all immunoglobulin heavy chains. The coexpression of IgM and IgD on the surfaces of early B cells was also found to involve a complex transcription unit encoding both p and 6 mRNAs. This transcription unit was developmentally regulated by the choice of multiple polyadenylation sites and by selective recognition and use of RNA splicing sites. Thus, posttranscriptional processing appears to be intimately involved in the changes in p and 6 expression by maturing B cells. Recent demonstration of similar mechanisms operating in species as different as yeast and rat would seem to establish the generality of posttranscriptional RNA processing in eukaryotic gene regulation. Although the binding of IgE to basophils and mast cells has been recognized for some time, the association of IgE with lymphocytes, monocytes, and macrophages is a more recent discovery. In the third article, Dr. Hans Spiegelberg summarizes the available data on the structure and function of Fc receptors for IgE on various immunocytes. Not only is the chemical nature of receptors for IgE on immunocytes quite different from that on mast cells but the strength of binding to the former is several magnitudes lower. The function of IgE receptors on immunocytes is not Entirely certain. However, the number of receptor positive cells, and probably receptors per cell, parallels the levels of extracellular IgE, suggesting that they are a part of the IgE response. On macrophages and monocytes, IgE receptors promote phagocytosis and killing of IgE coated targets and, in the presence of IgE complexes, induce release of phlogogens. On lymphocytes the role of IgE receptors is less clear, but there is some evidence for the hypothesis that receptor positive T cells may be involved in down regulating IgE synthesis by B cells. One of the great challenges in the field of immunology is the development of means to enhance host antitumor immunity. I n spite of a few promising but inconsistent leads, there is no generally successful antitumor immunostimulatory measure. One of the major difficulties in this field is our lack of precise knowledge of the immunologic hosttumor interaction during oncogenesis and tumor growth. In the fourth article Dr. Robert North proposes concomitant antitumor immunity, i.e., the development of transient, early T cell antitumor immunity that is soon negated by the generation of suppressor T cells, as a rational model for the analysis of natural tumor immunity and for the development of appropriate therapeutic manipulations. H e presents convincing evidence for the existence of such an antitumor response

PREFACE

xiii

in mice and then proposes means of potentiating or facilitating it to achieve the elimination of established syngeneic tumors. Several factors probably contribute to the persistence of a histoincompatible fetus during a long gestation period in an immunocompetent maternal host. However, since the mother clearly becomes sensitized to a variety of fetal histocompatibility antigens during pregnancy, and since maternal immunocompetence is not systemically suppressed, it seems likely that the mechanisms primarily responsible for fetal maintenance act locally at the placenta inhibiting the action of sensitized maternal lymphocytes. In the fifth article Drs. David Jacoby, Lars Olding, and Michael Oldstone review this field, focusing on the potent suppressor effects of fetal lymphocytes, a subject to which they have been leading contributors. Apparently fetal lymphocytes, via their suppression of maternal immune functions at the site of placentation, are the major protectors of the conceptus during gestation. In addition to its long recognized role as a vasoactive amine producing symptoms of allergic disease, histamine is now considered, together with prostaglandins and beta-mimetic catecholamines (the autacoids), as a regulator of both immune and inflammatory events. In the final article Drs. Dennis Beer, Steven Matloff, and Ross Rocklin review this field that has largely developed within the past decade. Histamine can be derived not only via the interaction of antigen with specifically sensitized mast cells, as in IgE reactions, but also by stimulation of sensitized effector T cells to make histamine releasing factor, which may provide a source of histamine in the absence of IgE mediated responses. Once available, histamine may act to modulate the immune response by activation of either or both suppressor and contrasuppressor cells with the result depending on the ratio of these two cell types activated. The effects of histamine on inflammation can also b e pro or anti. Its phlogogenic effects are achieved at least in part via T cells; these are stimulated to produce chemoattractant and migration inhibitory lymphokines that attract and hold lymphocytes and eosinophils at sites of inflammation. The antiinflammatory effects of histamine are achieved both by directly suppressing the action of cytotoxic T cells, natural killer cells, neutrophils, and eosinophils and indirectly via suppressor T cells. The latter may augment the production of prostaglandins by macrophages and monocytes, resulting in inhibition of effector T cells and thereby dampening cell-mediated immune reactions. As always, the editor wishes to thank the authors, who have given generously of their time and effort, and the publisher, whose staff does much to ensure a volume of high quality. FKANK J. DIXON

HENRYG. KUNKEL (1916- 1983)

HENRY G. KUNKEL (1916-1983)

Henry Kunkel’s untimely death has left a void in the field of immunology which will be felt in many ways. Among the people who will miss his advice and guidance most will be those of us involved with the Advances in Immunology, a series he coedited since 1967. As an editor, his ability to recognize the most significant movements in immunologic research, to identify those most expert in the area, and then to prevail upon them to write scholarly reviews was unexcelled. Much of the success enjoyed by this series is owed to his efforts. It is appropriate that in this volume of Advances we present a close and rather personal view of this remarkable man’s scientific career, and Dr. Hans Miiller-Eberhard, a long time associate and friend of Dr. Kunkel, has joined in its preparation. To a large extent Henry Kunkel was a self-made immunologist and clinical investigator. H e had no formal training in immunology or biochemistry, nor did he have any clinical specialty training. H e started his career at the Rockefeller Institute for Medical Research in 1945 in the field of liver disease. After the untimely death of his laboratory chief, he continued these investigations for several years making fundamental contributions to the diagnosis, prognosis, and treatment of liver cirrhosis. His interest in y-globulin, which was to continue to the end of his life, originated with the study of this disease. To acquire expertise in protein separation by electrophoresis and to prove himself worthy of appointment to senior rank at the Institute, he took a leave of absence and worked at the Biochemical Institute in Uppsala under Arne Tiselius. After a most successful year with Tiselius, Henry Kunkel was promoted to full Member of the Rockefeller Institute and was given a laboratory of his own at age 36. Henry Kunkel’s art of conducting science was to establish a fact by simple technology. His laboratory was austere, containing only the necessary basic equipment. The intellectual input was all that counted. He was exceedingly well read in the biomedical science literature and had a penetrating and critical mind. By association of seemingly unrelated facts and by informed intuition he was able to identify potential breakthroughs in immunology. In the first decade of his career as an immunologist, he showed that myeloma proteins are immunoglobulins, that 7 S and 19 S y-globulins are related but immunologically and chemically distinct proteins, and that rheumatoid facxv

xvi

HENRY G. KUNKEL

tor is an autoantibody to IgG. H e discovered idiotypy of human antibodies and, in an interlude between phases of strict immunologic research, he described henioglobin A2 and its relationship to thalassemia. All these early advances were accomplished merely employing precipitin techniques and starch block electrophoresis and ultracentrifugation as the only high-technology tool. Henry Kunkel was an ingenious mentor of his research trainees and associates. Particularly in the earlier years h e set an example skillfully experimenting at the bench. H e was gifted in inspiring his people through long and frequent discussions conducted individually. He was a proponent of training in the philosophy of research which, he felt, involved questions concerning discipline of thought, intellectual integrity, respect for the written word, and the ethics of research work itself. He was masterful in creating an atmosphere in the laboratory in which fellows were compelled to go forward to eventual success or hopelessly fall behind. Tension in the laboratory was high at times and the admonishing reprimand “they will beat you to it” was a hard experience for a beginner in research and meant longer hours at the bench and greater mental effort. For some it meant humiliation and anguish. But good work, the exciting results of a “key experiment” that would “advance the field significantly,” were always met with a beaming face and eventually would lead to true recognition, respect, and often lasting friendship. Henry Kunkel trained many young physicians and scientists and he did so with a phenomenal success rate. It seems a fair estimate that at least twenty senior professors of leading medical schools and research institutions, among them one Nobel Laureate and four members of the National Academy of Sciences, trace back the beginnings of their careers to Henry Kunkel’s laboratory. Henry Kuiikel wrote lucidly, often pondering for a considerable time over the precise formulation of a sentence. His numerous publications attest to his talent as an author of scientific prose. Yet his spoken word could be sketchy, even vague, as though he was expecting the other person to know what he was talking about or to read his mind. He was impatient with ignorance, especially in relation to publications from his own laboratory. He was indignant with “shoddy” work published in the literature since his own high standards of performance did not allow publication of work unless it was thoroughly substantiated and documented: “one of our finest traditions in sci,, ence, h e wrote, “concerns the sanctity of the written word and the special pride involved in the avoidance of error. We should preserve it at all costs.”

HENRY G . KUNKEL

xvii

Henry Kunkel became a formidable leader and pioneer in the investigation of immune complex and autoimmune diseases in man. His abiding interest in antibody structure, function, and genetics which led to the elucidation of much of what is known today in this field, was later extended to studies of B cell-associated immunoglobulins and recently to the T cell antigen receptor. In recognition of his many fundamental contributions to immunology and medicine he received numerous awards and honors. He held an endowed chair at the Rockefeller University, the Abby Rockefeller Mauze Professorship, and he had been president of two learned societies, the American Society for Clinical Investigation and the American Association of Immunologists. Yet neither the honors bestowed on him nor his natural dignity and high self-esteem prevented him, when the occasion arose, from joining his associates and friends in merry socializing. He delighted in playfully poking fun at them and being the target of their humorous attacks. Such situations revealed the engaging warmth and the humanness of his personality. Henry Kunkel will be remembered as the gifted teacher and scientist he was, endowed with the drive and ability to be creative and to be productive throughout his life. He was dedicated to and excited by science. As he put it, “scientific inquiry is a sort of opiate that once experienced is not readily shaken off.” Those who knew him well in the scientific community, his students, colleagues, and friends, will behold his memory with the admiration and deep affection they had for him. HANSJ. MULLER-EBERHARD FRANK J. DIXON

This Page Intentionally Left Blank

ADVANCES IN

Immunology VOLUME 3 5

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 35

The Generation of Diversity in Phosphorylcholine-Binding Antibodies ROGER M. PERLMUTTER,' STEPHEN T. CREWS,2 RICHARD DOUGLAS,3 GREG SORENSEN, NELSON JOHNSON: NADINE NIVERA? PATRICIA J. GEARHARTt5AND LEROY HOOD Division of Biology, California Institute of Technology, Pasadena, California

I. 11. 111. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. XV. XVI. XVII. XVIII. XIX. XX. XXI. XXII. XXIII. XXIV.

Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . Restricted Nature of the Anti-PC Response. . Clonotypes of Anti-PC Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . Hidden Diversity of the Anti-PC Response. . . . The Structure of the Variable Regions of Anti Phosphorylcholine and the Molecular Basis o Heavy Chain Variable Regions. . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . Somatic Diversification of Heavy Chain Variable Regions Somatic Diversification Can Be Extensive and Extends Be Boundaries of the VII Gene. Somatic Diversification Prob Somatic Mutation Probably Occurs by a Hypermutational Mechanism That Is Localized in and around the VIIGene . . . . . . . . . . . . . . . . . . . . . . Light Chain Variable Regions. . . ........................ Somatic Diversification of Light es . . . . . . . . . . . . . . . . . . . . . . The Pattern of Variation by Somatic Hypermutation . . Somatic Mutation Is Correlated with Immunoglobulin Vll, D, and Jll Segments and Junctional Diversity.. . . . . . . . . . . . . . . . . . Diversity in the VII Segment.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diversity in the J11 Segment ........................ . ......... Diversity in the D Segment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The N Region. ............................................. J. Gene Segments and Junctional Diversity . . . . . . . . . . . Summary of Diversity of Antibodies That Bind PC . . . . . Selection of Variant Antibodies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .................. Molecular Basis of the T15 Idiotype . ............ The T15 VII Gene Family . . . . . . . . . . . . . . . . . . . . .

6 7 7 15 16

17 18

19 20 22 22 23 23 25 25 26 26 27 29 29

Recipient of New Investigator Award AI-18088 from the National Institute of Allergy and Infectious Diseases. Present address: Department of Pathology, Stanford University, Stanford, California 94305. Drs. Crews and Perlmutter contributed equally to this review. Present address: Integrated Genetics, Framingham, Massachusetts 01707. Present address: Department of Biochemistry, State University of New York at Stony Brook, Stony Brook, New York 11790. Present address: Department of Biochemistry, Johns Hopkins University School of Hygiene and Public Health, Baltimore, Maryland 21205. 1 Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-0224356

2

ROGER M. PERLMUTTER E T AL.

XXV. Evolution of the T15 Gene Family.. .............................. XXVI. Structural Diversity in Antibodies to P C . . ......................... XXVII. Future Research., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ....................................................

30 31 33 35

1. Introduction

The ability of higher vertebrates to mount an immune response to a seemingly infinite variety of distinct antigens has attracted the attention of biologists for decades. I n particular, immunologists have struggled to explain the extraordinary diversity of antibody molecules. During the early part of this century, Landsteiner’s classic analysis documented the exquisite specificity of immunoglobulins, which, for example, could clearly distinguish between identical chemical structures substituted at different positions on a phenol ring (Landsteiner, 1945). These early results prompted “instructionist” theories which viewed antigen as a template around which antibodies would fold. Beginning around 1960, structural analysis of antibody polypeptides defined the kappa, lambda, and heavy chain families, and the sequence studies of Hilschmann and Craig (1965) and Putnam (reviewed in Putnam et al., 1971) identified light chain variable (V) and constant (C) regions. Viewing these data, Dreyer and Bennett suggested with admirable foresight that antibody heavy and light chains are encoded by more than one gene, thus anticipating the noncontiguoils nature of eukaryotic genes and the DNA rearrangements which are central to the formation of antibody coding regions (Dreyer and Bennett, 1965). More detailed structural analysis revealed that the amino terminal variable regions of both heavy and light chains contain three short segments of hypervariability (Wu and Kabat, 1970; Capra and Kehoe, 1974). These hypervariable regions were shown by X-ray crystallography to comprise the antibody combining site (Padlan et al., 1973; Amzel et al., 1974), whereas the remaining portions of the variable region are relatively invariant in structure and hence are called “framework” regions. In the early 1970s the problem of antibody diversity was approached at the nucleic acid level. Hozumi and Tonegawa (1976), using cDNA probes for murine kappa light chains, were the first to show that the constant and variable region-encoding segments of antibody genes are separated by intervening DNA in the germline but are more closely juxtaposed during B cell differentiation. Four separate coding regions, leader (L), V, J, and C, for lambda genes were identified (Brack et al., 1978) and a similar analysis was carried out on

DIVERSITY IN

PC-BINDINGANTIBODIES

3

murine kappa chains (Seidman et al., 1979; Sakano et al., 1979).Hood and co-workers demonstrated that heavy chain genes include an additional gene segment, the D or diversity segment (Schilling et al., 1980) which encodes a portion of the third hypervariable region and thus two separate rearrangement events are required to generate a functional heavy chain gene (Early et al., 1980; Sakano et al., 1981). In addition, the expression of multiple isotypes with the same variable region structure was found to involve another type of DNA rearrangement, the class switch, which juxtaposes a fully assembled VH-D-JH unit with a 6, y, a,or E heavy chain constant region gene located 3’ to the p heavy chain gene on chromosome 12 (Davis et al., 1980). Thus germline DNA, typically isolated from sperm cells, contains multiple dispersed gene segments which undergo rearrangement in B lymphocytes to yield a functional immunoglobulin coding region. Cells of the lymphocyte lineage are the only known cells which demonstrate immunoglobulin gene rearrangements. Examination of antibody genes clearly demonstrated the elegant strategies which permit a limited number of coding sequences to direct the synthesis of millions of different antibodies. Protein sequence and DNA or “Southern” blotting experiments [which permit the identification of restriction endonuclease-digested DNA fragments containing immunoglobulin genes via electrophoretic separation and subsequent hybridization with appropriate 32P-labeledprobes (Southern, 1975)] indicate that only a few hundred light and heavy chain variable region gene segments probably exist in germline DNA. Each variable region gene is constructed through the combinatorial joining of the V and J, in the case of light chains, or the V, D, and J gene segments, in the case of heavy chains. Thus combinatorial joining of gene segments further expands the germline repertoire. Flexibility in the site of joining of the V, D, and J gene segments provides a second mechanism by which the germline information encoding antibodies is amplified (Max et al., 1979; Sakano et al., 1979; Weigert et al., 1980). Finally, the combinatorial association of thousands of different heavy and light chains provides an additional level of information amplification in the formation of antibody combining sites. The significance of these strategies for the generation of antibody diversity is outlined in Table I. Differentiation of pleuripotent lymphocyte stem cells into pre-B cells is marked by rearrangement of heavy chain genes, the earliest event in the formation of B cells (Maki et al., 1980). Light chain gene rearrangement occurs subsequently with K chain rearrangement apparently preceding A chain rearrangement (Hieter et al., 1981). The

4

ROGER M. PERLMUTTER ET AL.

TABLE I GERMLINE AND COMBINATORIAL STRATEGIES FOR THE GENERATION OF ANTIBODY DIVERSITY Germline gene segments -250 V, 4 J. 2 VA 3 JA -250 VII -10-20 D 4 Jii Combinatorial joining 250 V, X 4 J x = 1000 V, genes 250 VII x 10 D x 4 J I I = 10,000 VIIgenes Combinatorial Association

1000 K x 10,000 H = lo7 antibody molecules

resultant B cells, once stimulated with antigen, may terminally differentiate into plasma cells secreting large quantities of antibody protein. The structural analysis of antibody genes was achieved for the most part using murine myeloma tumors, generally induced by intraperitoneal administration of mineral oil (Potter, 1972), as a source of cells “frozen” at the level of plasma cell differentiation. More recently, hybridomas generated through the fusion of a nonsecreting plasmacytoma with hyperimmune spleen cells have served as a ready source of monoclonal antibodies (Kohler and Milstein, 1976). In order to examine the generation of antibody diversity in greater detail, we studied a model immune response to a simple, well-characterized immunogen. The murine antibodies directed against phosphorylcholine (PC) which we discuss in this article proved an ideal system for delineating the fundamental mechanisms which generate the immunoglobulin repertoire of higher vertebrates. II. Restricted Nature of the Anti-PC Response

Phosphorylcholine is the immunodominant determinant in vaccines derived from certain rough pneumococcal strains (e.g., R36A, Leon and Young, 1971). Early serologic analyses of murine antibodies raised against pneumococcal vaccine showed that the heterogeneity of anti-PC antibodies is restricted in several respects: (1) greater than 90%of the induced antibody is IgM (Lee et al., 1974), (2)the affinities

DIVERSITY IN

PC-BINDINGANTIBODIES

5

of these antisera for PC are nearly homogeneous as measured by hapten inhibition of plaque formation (Claflin and Davie, 1974a), and (3) the majority of induced anti-PC antibodies bear idiotypic determinants related to those present on a prototype PC-binding BALB/c plasmacytoma protein, T15 (Lieberman et al., 1974). Furthermore, neonatal administration of anti-T15 idiotypic antibodies completely abrogates the anti-R36A pneumococcal response in mice (Cosenza and Kohler, 1972). Taken together, these data suggested that the murine response to PC is essentially monoclonal. 111. Clonotypes of Anti-PC Antibodies

Using more refined analyses based on idiotypy and isoelectric focusing of antibody light chains, Claflin, Davie, and others were able to define several related clonotypes of anti-PC antibodies in mice immunized with pneumococcal vaccine or with PC coupled to protein carriers via a hydrophobic spacer (e.g., PC-KLH, Gearhart et al., 1975a; Claflin and Rudikoff, 1976). Lieberman et al. (1974) generated an allogeneic anti-T15 serum which identified the majority of anti-PC antibodies derived from Ig-la allotype mice (e.g., BALB/c) but which failed to recognize anti-PC antibodies raised in mice of other allotype groups (the allotypes are serologically defined polymorphisms in heavy chain constant region genes). Close linkage of this immunoglobulin heavy chain variable region marker with a well-defined heavy chain constant region marker reinforced the speculation that the T15like antibodies were encoded in germline DNA. A complementary antiserum was later raised against anti-PC antibodies from Ig-1'' mice (Lieberman et al., 1981). Other idiotypic markers include a hapteninhibitable xenoserum which identifies most anti-PC antibodies in all mouse strains (Claflin et al., 1974b) further emphasizing the structural homogeneity of these molecules. Four serologically distinct PC-binding plasmacytomas have been isolated, and immunization of mice with pneumococcal vaccine yields low levels of anti-PC antibodies resembling three of these, M511, M167, and M603, as well as the majority of induced anti-PC antibodies which resemble T15 (Ruppert et al., 1980). Light chain isoelectric focusing provided further evidence of a limited clonal repertoire of anti-PC antibodies encoded in the germline (Claflin and Rudikoff, 1976). IV. Hidden Diversity of the Anti-PC Response

Although hapten inhibition profiles of murine anti-PC antisera were consistent with the thesis that the PC antibodies constitute a highly

6

ROGER M. PERLMUTTER ET AL.

restricted family of antibodies, studies of monoclonal anti-PC antibodies generated using a splenic focus technique revealed a large number of different antibodies distinguishable by affinity for PC and PC analogs and by idiotype (Gearhart et al., 1975a). Under these circumstances, anti-PC antibodies of IgGl and IgGz, subclasses could be detected as well (Gearhart et al., 197513) although anti-PC antisera contain mainly IgM and some IgG3 antibodies (Perlmutter et al., 1978). Thus while serologic analyses indicated that a small family of murine antibodies encoded by germline genes formed the entire antiPC response, limited clonal analysis suggested that the total antibody repertoire was rather heterogeneous and perhaps in part generated by somatic mutation (Gearhart et al., 1975a). In this context, we chose to apply protein and DNA sequencing strategies to anti-PC hybridomas in hopes of elucidating the structural basis of antibody diversity. Determination of the complete variable region structure of M603 (Rudikoff and Potter, 1976), a PC-binding myeloma protein, and the construction of a high resolution electron density map of the Fab fragment of this antibody (Padlan et al., 1973) added further impetus to our study since the sites of antigen-antibody interaction could be determined. Specific contact residues in the hypervariable regions of the heavy and light chains (situated at the loops of the immunoglobulin fold) interact with the charged phosphoryl and choline moieties of the PC hapten. Of particular importance are tyrosine, glutamic acid, and arginine residues at positions 33, 35, and 52 of the heavy chain. Changes in antibody primary structure can be correlated to some degree with alterations in the affinity of PC binding of variant antibody molecules. V. The Structure of the Variable Regions of Antibodies Which Bind Phosphorylcholine and the Molecular Basis of Their Diversity

Further understanding of the structural diversity of PC-binding antibodies required the isolation and sequencing of monoclonal antibody variable regions and the isolation and sequencing of the genes which encode these antibodies. Fortunately, monoclonal antibodies that bind PC are available in abundance from myeloma and hybridoma sources. In addition, recent advances in recombinant DNA technology coupled with efficient methods for sequencing DNA have expedited the structural characterization of both the germline gene segments as well as the rearranged expressed genes encoding anti-PC antibodies.

DIVERSITY IN

PC-BINDINGANTIBODIES

7

VI. Heavy Chain Variable Regions

Initially, the sequences of nine heavy chain variable regions from myeloma immunoglobulins that bind PC were determined by our group and by others (Rudikoff and Potter, 1976; Hood et al., 1977a,b). As shown in Fig. 1, four of these sequences are identical to a prototype sequence found in the heavy chain of the T15 myeloma protein. The remaining five sequences have from 1 (H8) to 14 (M167) amino acid substitutions compared with T15, half of which (15/29) are located in a region encoded by the D and JH gene segments; however, there is also considerable diversity scattered throughout the V H segment which encodes position 1-101 in the protein sequence. While five of the VH segments are identical, four have substitutions from the T15 prototype sequence. The variants differ from one another as well and include from one to eight substitutions compared with T15. Much of the diversity localized between residues 100 and 115 probably results from the flexible mechanism of V-D-J joining which we discuss in more detail below. In the following section, we focus on the mechanism that generates substitutions within VH segments.

VII. Somatic Diversification of Heavy Chain Variable Regions

Two general explanations might account for the observed sequence differences in VH segments: the variants might each be derived from distinct germline gene segments, or they might result from somatic diversification operating on one or a small number of germline gene segments. In order to distinguish between these alternative explanations, it was necessary to isolate and sequence all of the germline gene segments which might encode these variant protein sequences (Crews et al., 1981). Our laboratory constructed a cloned copy of the heavy chain variable region mRNA from the S107 myeloma tumor, the immunoglobulin of which shares an identical primary sequence with T15 (Early et al., 1979). Since all of the PC-binding VH segment protein sequences are quite similar, utilizing this clone as a hybridization probe we could detect all of the germline genes that could encode these VkI segments. A Southern blot of BALB/c sperm DNA digested with EcoRI and probed with labeled S 107 cDNA revealed four dominant hybridizing bands (Fig. 2). We utilized this S107 probe to screen a recombinant library of BALB/c sperm DNA cloned into a bacteriophage vector and

-J-D

V

mmm TI5

1

11

ZI

3

40

I

I

I

I

I

?

W

n

m

a

IW

111

I

I

I

1

I

I

121 I

EVKLVESCGCLVOPCGSLRLSCATSGFTF~FY~~UVR~PPGKRLE~I~SRNKANDYTTEYSASVK~FIVSRDTSOSILYLOMNALRAEDTA YYGSS I Y Y C AYUYFDV,UGAGTTVTVSS ~

TI5 s63

I

I

I

I

I

I

I

I

I

I I

I

I

I

I I

1 1

I

I

1

I

1

I

L I

I

.

.

s1m

I

I

1

1

1

I

1

IB

I

1

I

I

I

I

I

I

1683

u3a.J 1611 11167 -v

G-K

I

1

,

I

I

t

1

I

1

I

I

I

I t

1

' R-H-s'1

;w1;

rn

1

-l -KID

I

:

-V-L

:

'GD1-TJ-AD-0-YFC-

:

-

l

FIG. 1. Complete amino acid sequences of nine heavy chains from murine PC-binding myeloma proteins. [Data derived from Rudikoffand Potter (1976)and Gearhart et al. (1981).]The sequences encoded by VH,D, and JH gene segments are delineated by arrows. Dashed vertical lines denote the first, second, and third hypervariable regions [HVl, HV2, HV3 (Kabat et al., 1983)l. Numbering of residues is sequential from the amino terminus relative to the longest sequence, M167. At each position, residues identical to the prototype T15 sequence are indicated by a solid line and all sequences are shown in single letter code (Dayhoff, 1976).

DIVERSITY IN

PC-BINDING ANTIBODIES

9

FIG.2. Southern blot analysis at the T15 VII gene family. At right is shown the pattern obtained when 10 pg of BALB/c sperm DNA is completely digested with EcoRI and the resulting fragments analyzed by Southern blotting (Southern, 1975) with a 32Plabeled probe complementary to the T15 VIIgene segment. A similar analysis using EcoRI digests of recombinant phage clones corresponding to the four major genomic bands is shown in the left lane. Band sizes are indicated in kilobases and the k clones containing each band are identified at the far left of the figure (modified from Crews et az., 1981).

were thus able to obtain at least one clone for each of the four major bands observed on the genomic Southern blot. Figure 2 also shows the comigration of the EcoRI fragments homologous to the S 107 probe in our phage clones when compared with the bands in the genomic blot. Comparison of the intensity of hybridization of known amounts

10

ROGER M. PERLMUTTER E T AL,

of the cloned DNA to the intensity of the bands on the genomic blot indicated that each genomic band contained at most one or two homologous VII gene segments. The complete nucleotide sequences of the cloned VH gene segments identified with the S107 probe were determined and the translated protein sequences of these were compared with our previously obtained myeloma heavy chain sequences (Fig. 3). One gene segment, labeled V1, encodes a protein that is identical to T15, thus confirming that this prototype heavy chain variable region is indeed a germline sequence (Early et al., 1980). Three additional, independently isolated clones with identical restriction maps also were found to contain identical V1 DNA sequences. This result adds support to the conviction that each hybridizing band on the genomic Southern blot contains only a single VH sequence complementary to the S107 VH probe. The remaining three germline gene segments isolated from the BALB/c library (designated V3, V11, and V13 in order of their characterization) are each greater than 13% different in amino acid sequence from T15. None of the variant myeloma heavy chain sequences is identical to any of these three VH gene segments and all are much closer to V 1 than to the other three gene segments (Fig. 3). Since we are confident that we have cloned all of the germline gene segments that could encode these VH regions, we conclude that the variant antibody segments are the result of somatic diversification processes operating on the germline V1 gene segment. Studies to be discussed subsequently have verified this conclusion. A more complete picture of the murine immune response to PC was developed through the analysis of a large number of anti-PC hybridomas (Gearhart et al., 1981).Unlike the mineral oil-induced plasmacytomas which produce mainly IgA antibodies, the hybridoma anti-PC antibodies include both IgM and IgG classes. Figure 4 shows the complete amino acid sequences of 11 hybridoma heavy chain variable regions in addition to the previously examined myeloma sequences. N-terminal sequences of a total of 42 heavy chains from anti-PC antibodies are shown in Fig. 5. All but one of these heavy chains are clearly encoded by the V 1 gene segment and those that differ from the germline sequence have undergone minor somatic diversification. As shown in Fig. 6, the single exception is HPCG15 which is similar to the translation of the V11 gene sequence and probably resulted from somatic diversification of that gene segment, differing from V11 at five positions of 90 residues positively identified. Thus although V1 is the principal VH gene segment utilized in the anti-PC response, with sufficient somatic alteration the V11 gene segment can encode a PC-

HV I

HV2

PROTOTYPE ~ 1 5EVKLVESGG~LVOP~GSLRLSCATSGFTFS~FYNE~VROPPGKRLEUIMSRNKANDYTTEYSASVK~FIVSRDTSOSILYLOMNALRAEDTAIYYCARD I

1

I

I I

I

I

I

1

1

I

VH GENE SEGMENTS

Vl

v3 V11 v13

A -. -S-

N

~

Y

TLY-S~-SP-L-C

EA--

M603 6 11 11167

wcte WCtlJ

~-

'

I

! 1

I 1

I

I

!

I

I

1

;

,I .

-V

;

-

I

I

LI _

T-S-S-T-

S-A-L&I1-S-1-

r - I

115 VARIANT VH SEGMENTS

u3a7

I

I I

T:Y--NI-A--XL-LI-~-I-~~-LTI--~

--+

-A

- T -N-N- I T !G- I

HPCGlZ _

- !

_

_

_

~

I I

AS-T-K-

I

I

I I I

I I

j

I

I

G-K

N

I

I

F

I

I -S - T ~ s - H - R

II

;

I

I

T

; I

V

I

I

I

I

-N

I

1

HPCGI4 ~ - - _ _ _ _ _ II > HPCGll -___ I I 1

!

I I

I

T-T-

I I

F

I

VY

I

-

=

F I

I

FIG. 3. Genes encoding PC-binding heavy chains. Shown is a comparison of the sequences of the four germline VI, segments identified in Fig. 2, translated into amino acid code, with the sequences of variant heavy chains from Fig. 1. The sequence of the amino-terminal 101 residues of the T15 myeloma heavy chain is listed at the top. Dashed lines demarcate the positions of first (HV1) and second (HV2) hypervariable regions (Kabat et al., 1983).

I

I0

zp

am

40

5a

80

I

I

I

I

I

I

I

?n I

a

91

I

I

! I I

EVYLVESC~CLVOPC~SLRLSCATSCFTF~FY~~UVREPPCKRLEUI~SRWKAWDYTTEYSASVK~FIVSRDTSESILYLE~WALRAEDTA~YYCARD I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I I I I

1 I I 1

I I

I

I I

I I

I

I

I

I

I

I

I

1

I I

I

tpa8

154

m 1 3 m l 4

-

L+--+---T-vY I

m 11

1

:

m 1 2

1

:

I

I

:

I

l

l

I I I

I

!

I

I I I

I

1

I

I

I

1

'

FIG.4. Complete amino acid sequences of VHregions from PC-binding hybridoma proteins. Sequences derived from 10 hybridoma proteins are compared with those of the PC-binding myeloma heavy chains shown in Fig. 1 and with the germline T15 V13 ( V l ) sequence shown at the top. Sequences are grouped by heavy chain class at left. Dashed lines demarcate the positions of HV1 and HV2. Numbering and residue identification are as in Fig. 1.

DIVERSITY IN

PC-BINDING ANTIBODIES

13 4

HPCMl

1

-

20

; 1

?O E V K L V E S G G G L V Q P G G S L R L S C A T S G F T F S D F Y M E W

HQCn.2 HP0.13

m 4 HP0.15 H

M

m

7

H W 5

ww7

HPC16 HPc35 HPC52 HPCl.26

wCG8 HPCG9 HPCGlO HPCGll wffi12

1

Hpu;18 HPffi32

WCG13 HPCGl4 HPCG15

-17

wcG20 HPCG21 HPC622 HPCG23 HPCG24 HPCG26 Hpa;28 HPCG29

-31 M603 W3207 M167 M511 S63 Y5326 H8

"

5107 115

FIG.5. Amino-terminal sequences of 42 heavy chains from PC-binding myeloma and hybridoma proteins, Solid lines indicate sequence identity with the HPCMl sequence shown at the top. Gaps indicate unidentified residues. The position of HV1 is indicated with arrows.

V11

EVKLVESGGGLVOPGGSLRLSCATSGFTFTDYYMSWVRPPNGYTT~YSASVKGR~TISRDNSOSILYLOMNTLRAEDSATYYCA

M47A

HPCGIS

e1--5

A-<7)

(

7

)

FIG.6. The V11 germline VF1segment likely encodes the heavy chain of HPCG15. Shown is a comparison of the amino-terminal sequence of the heavy chain of HPCGl5 with the sequence of the V11 germline VH gene segment translated into amino acid code. The published sequence of the heavy chain of M47A, a myeloma protein of unknown specificity, is also shown (Robinson and Appella, 1979).

V

VH

DJH

JH2

SWITCH REGION

FIG.7. The distribution of somatic mutations in rearranged Vl1 genes. Hatch marks indicate the positions of mutations from the germline (sperm) sequence in heavy chain genes of two PC-binding myelomas, M167 and M603 (modified from Kim et al., 1981). The structural elements of the genes are shown boxed. Distances from the VI,-D joining point are noted in kilobases at the bottom of the figure.

DIVERSITY IN

PC-BINDING ANTIBODIES

15

binding heavy chain. HPCG15 has a low affinity for PC and is probably unimportant in the overall anti-PC immune response. We have also obtained nucleotide sequence data for two of the rearranged myeloma heavy chain genes, M603 and M167 (Early et aZ., 1980; Kim et aZ., 1981). These sequences are compared with the germline V1 sequence in Fig. 7. With this large body of structural information available for analysis, important features of the somatic diversification process have become apparent. VIII. Somatic Diversification Can Be Extensive and Extends Beyond the Boundaries of the VH Gene

Examination of the protein sequences in Figs. 4 and 5 reveals that the number of substitutions in the VfI segment of somatic variants of the V1 gene ranges from 1 to 8. M167 is the most extreme variant in this regard with 44 nucleotide substitutions in the VbI gene and the regions immediately surrounding the gene (Fig. 7, Kim et al., 1981). Furthermore, whether comparing all of the protein sequences (Figs. 4 and 5 ) or a single sequence (e.g., M167), somatically generated substitutions are scattered throughout the variable region and are found in all three framework regions and both hypervariable regions of the VH segment. Additional substitutions occur in the JH segment and probably in the D segment as well (see below). A small cluster of substitutions occurs in the second hypervariable region at position 56 (see Section XXII). It is worth noting that the antibodies which we have analyzed are highly selected to retain a functional PC-combining site. Therefore, somatic diversification may have the potential for producing sequences which diverge even more dramatically from the V1 prototype but which escape our attention because their PC-binding function was disrupted. Examination of the protein sequences in Figs. 4 and 5 also shows that in general each variant is unique. Only rarely do two variants share even a single substitution (Ile at position 28 in HPCG11 and HPCG12, Thr at position 40 in HPCGl3 and M167, and Asn at position 101 in M603 and W3207). Thus sequence diversity appears virtually unlimited, but it occurs within the context of a single germline gene. Figure 8 shows hybridoma heavy chain sequences derived from fusion products of a single mouse spleen. Five of the six sequences are different, demonstrating that each individual mouse can express multiple somatic variants. These data correspond to the idiotypic diversity seen at the clonal level in the response to PC-KLH (Gearhart et aZ., 1975a) and provide an explanation for the extraordinary diver-

16

ROGER M. PERLMUTTER E T AL.

-

m1-

10

HPCMl HPCG13 HPCGl4 HPCG9 HPCGlO HPCGll

30

20

E V K L V E S G G G L V Q P G G S L R L S C A T S G F T F S D F Y M E W L

n

-I

I

FIG.8. Somatic mutation in hybridomas from a single mouse spleen. Shown are the amino-terminal sequences of six heavy chains from PC-binding hybridomas derived from a single fusion. Solid lines indicate homology with the HPCMl sequence shown at top.

sity of private idiotypes superimposed on a common cross-reactive idiotype in the antiarsonate response (Lamoyi et al., 1980). IX. Somatic Diversification Probably Occurs by a Mutational Mechanism

Two alternative processes might explain the observed somatic diversification, either recombination or mutation. The somatic recombination model is illustrated in Fig. 9. Recombination or gene conversion occurring as the result of unequal crossing over between two different VH genes could lead to the formation of a hybrid gene (Edelman and Gally, 1967; Seidman et al., 1978). By comparing our protein sequences with the sequences of the four genes of the T15 family, we sought to determine whether the sequences of the somatically derived variants could be the result of recombination between the V1 gene and another member of the family. Only one residue of 24 somatic substitutions (position 95 of M167) could be generated in this fashion (Fig. 9). Although it might be argued that the V1 gene could recombine with a V H gene other than V3, V11, or V13 to generate a novel sequence, this mechanism could not explain substitutions in the JH1region or the 3’ intervening sequence (Fig. 7) where multiple different sequences are not present in the germline. It thus seems clear that the observed somatic diversification in PC-binding antibody heavy chains must occur by a mutational mechanism. The sequence of a heavy chain from a PC-binding antibody from a CBA mouse has been presented as an example of a gene conversion event operating within the T15 family (Clarke et al., 1982). However, until the elements of the CBA germline T15 family are isolated and sequenced, this hypothesis cannot be rigorously evaluated. Recombination events clearly do not contribute significantly to the diversity of antibodies which bind PC.

PC-BINDING ANTIBODIES

DIVERSITY IN

A'

A

A"

1

1

I Homologous Unequal V Gene Alignment

,.

ONR PROTEIN

'<

I

A'

IN2

Mn A E+JY

17

'

I IL'.

5

A

1

:;,

!

'5-d

e

j r

!

k -T*-TI-&lI

I I I

sY F K GY H S

R

I

-F-

T-T+i

I

FIG.9. The somatic recombination model for V gene mutations. Homologous but nonidentical variable region genes could undergo a recombination event, yielding a novel hybrid V gene shown to the right of the figure. Alternatively gene conversion could result in alterations of short sequence regions as shown at left (Baltimore, 1981).A complete tabulation of all substitutions from the germline sequence found in the V3, V11, and V13 VIrgene segments (DNA) and in the 42 complete and N-terminal protein sequences is shown at the bottom of the figure.

X. Somatic Mutation Probably Occurs by a Hypermutational Mechanism That Is Localized in and around the VH Gene

Analysis of the nucleotide substitutions of the M603 and M167 variable region genes demonstrates that silent mutations occur in addition to replacement mutations (Early et al., 1980; Kim et al., 1981). Both the M167 and M603 genes also have mutations in the intervening sequence 3' to the variable region gene and the M 167 gene has mutations within the intervening sequence between the leader and VfI exons (Fig. 7). Sequences 5 kb 5' and 3' to the variable region are unmutated (Kim et al., 1981). Thus mutations are sharply localized in and around the variable region gene. Since the silent mutations presumably confer no selective advantage and since the mutations are

18

ROGER M. PERLMUTTER ET AL.

highly localized, it seems likely that the multiple mutations seen in the M167 and M603 VH genes occurred as one or a few clustered genetic events and not as a series of single base mutations occurring individually over time. This process which produces a group of clustered substitutions is referred to as hypermutation. Although relatively little is known about mutational mechanisms in non-antibody genes in somatic cells, it is likely that hypemutation is uniquely associated with antibody genes and B cell proliferation. XI. Light Chain Variable Regions

All of the light chains of PC-binding antibodies that have been studied are of the kappa type. Analysis of the available complete and amino terminal sequence data (Figs. 10 and 11)reveals that there are three major groups of variable region sequences named after the myeloma prototype in each case, T15, M603, and M167 (Gearhart et al., 1981).Potter has somewhat arbitrarily defined subgroups as sets of VH sequences differing at three or more positions in the first 23 residues (Potter et al., 1982). The HPCG26 V, region falls into a fourth, separate light chain group which is typically associated with a distinct class of antibodies having a higher affinity for PC analogs than for PC itself (Todd et al., 1983). In general, V, subgroups defined by protein sequencing appear to identify V, gene segments-three or more differences in the amino-terminal protein sequence generally reflect the expression of a distinct germline V, gene segment. It has been repeatedly shown that a single germline V, gene segment encodes all of the observed sequences in the M167 group (Selsing and Storb, 1981; Gershenfeld et al., 1981), and it is likely that the same is true of the M603 light chain group. Most of the T15 V, light chain group sequences are identical to the T15 prototype (14/17); however, three variants all share a glutamic acid substitution at position 17 suggesting either that parallel somatic mutations have occurred or that these V, +mq8107 115 I4603 Y167 W511

D

I

V

I I

~

T

O

S

P

T

5 5 5 5 GE W OELSNP SGES S D E L S K P SGES 5

I

10

I

20

F

L

A

KS 0 RS K AS K

Crn+ V

T

A

S

K

K

V

T

l

S

C

T

A

S

E

S

L

Y

S

S

K

~

K

V

H

Y

L

I

F

Y

A bV ONDH L N GNOKN 0 G P 1 ES LYKDG TVLNWFL RPGOSPOL I S L M S T R A S G V S D R F S G S G S R T D F T L E I S R V K A E D V O V Y COOLVEYP LYKDG TYLNWFL RPGOSPOL I Y L W S T R A S G V S D R F S G S G S G T O F T L E I S R V K A E D V G V Y COOLVEVP

I

lo

I

4a

I

m

I

w

I m

I

P

I

w

FIG.10. Complete sequences of five V, regions from PC-binding myelomas. Residues are numbered sequentially from the amino terminus. The S107 sequence is from Barstad et d.(1974) and Kwan et d.(1981), T15 from Kwan et al. (1981), M603 from Rudikoff et al. (1981), M167 from Rudikoff and Potter (1978), and M511 from Appella (1980). Numbering is sequential from the amino terminus. The positions of hypervariable regions are marked at the top of the figure.

S

Y

P

DIVERSITY IN

PC-BINDING ANTIBODIES

19

1 10 HPm1 HPW HPCM5 HPO.16 HPO.17 HPC52 HPCG~

20

3?

D 1 V M T Q S P T ; L A V T A S K K V ; l S C T A S E S L Y S S K H K V

c

HPCGll HPCG12 HPCG14 HPCG20 HPCG21 S63 Y5236 H8 5107 T15

-

T

P F

HP0.13 HPCMO HPC126 WCM25 HPCG15 W3207 M603

D I V M T Q S P S S L S V S A G E K V T M S C K S S Q S L L N S G N Q K

HPCl9 HPClOO HPC16 HP0.127 HPCG9 HPCGlO HPCG32 HPCG13 HPCGl7 HPCG22 HPCG23 HPCG24 HPCG28 HPCG31 M511 M167

D I V I T Q D E L S N P V T S G E S V S I S C R S S K S L L Y K D G K T

HPCG26

A

D

T

a

nlD

Y b!

E

D V L M T Q T P L S L P V S L G D Q A S I S C R S S Q S L L N S N G N T

FIG.11. Amino-terminal sequences of 42 light chains from PC-binding hybridonia and myeloma proteins. The sequences are organized i n four groups related to the T15 light chain, the M603 light chain, the M167 light chain, or HPCG26 proceeding from top to bottom. The position of the first hypervariable region is delineated with arrows. Solid lines indicate homology with the prototype sequence, gaps indicate unknown residues.

regions are encoded by another closely related germline V, gene. All of the completely sequenced light chains utilize a single germline J K segment, JK5. XII. Somatic Diversification of Light Chain Genes

Variants of the M167 light chain protein sequence have been studied as examples of the somatic diversification process already described in heavy chains. Here again, the weight of evidence suggests

20

H167 H511 WCG9

ROGER M. PERLMUTTER ET AL.

-

wCG22 wCG10 WCG13

FIG. 12. Somatic mutation in PC-binding light chains defined by nucleotide sequence. The complete sequences of six light chains from PC-binding myelomas or hybridomas were obtained by DNA sequencing of the cloned rearranged genes (Gearhart and Bogenhagen, 1983).The translated amino acid sequences are shown in this figure, numbered sequentially from the amino terminus. The positions of the first and second hypervariable regions are delineated with arrows.

that mutation and not recombination is responsible for the variant sequences. In addition to the protein sequences, shown in Figs. 10 and 11, there are complete variable region gene sequences for six members of the M167 light chain group (Fig. 12; Gearhart and Bogenhagen, 1983). Two of these, HPCGS and HPCG22, are identical in sequence to the germline V, and J, genes. In the approximately 1200 nucleotides examined including the variable region gene, M511 had four single-base substitutions and one 3-base deletion, HPCGlO had six substitutions, HPCGl3 had 10 substitutions, and M167 had 11 nucleotide substitutions. These alterations were found in both the V, and J, genes and in the 5' and 3' intervening sequences. However, no substitutions were observed in the C, gene segment. Thus here, as in the heavy chain genes, somatic mutation is localized in and around the variable region gene segment. In addition, four of the eight substitutions found in the coding region are replacement changes and four are silent substitutions, again suggesting that a hypermutation mechanism may operate initially in the absence of selection. XIII. The Pattern of Variation by Somatic Hypermutation

The mechanism of somatic hypermutation is unknown. As expected from the diverse distribution of the substitutions, no conserved sequences appear to be recognition sites for mutation. In Table I1 are tabulated the characteristics of the observed nucleotide substitutions from the available VH and VL gene sequences as compared to their germline equivalents. In addition, we have included an analysis of the deduced nucleotide sequences reverse translated from all of the

TABLE I1 OF SOMATIC MUTATIONALEVENTSIDENTIFIED BY DNA SEQUENCING OR CLASSIFICATION EXTRAPOLATED FROM PROTEIN SEQUENCE DATAFOR PC-BINDING HEAVY AND LIGHTCHAINS Transitions A+G H-DNA H-Prot L-DNA L-Prot Totals

G+A

5 3 5 0 13

Total transitions = 60.6 Deletions = 8b

7 3.3" 2 1 13.3

C+T

Transversions T+C

6 6 5 1

10 1.3" 5

18

16.3

0

A+C

A+T

G+C

1 2 2 0

1 5 2 1

1.3" 0

5

9

G+T

C+A

C+G

0

3 1.3" 0 1

1 2 1 0

4 1 1 0

1.3

5.3

4

6

0

Total transversions = 49.2 Insertions = 2b

Fractional values reflect unknown codon usage in reverse translation from protein sequence data. Deletions and insertions identified in nucleotide sequences only.

T+A

7 1.3" 2 0.5" 10.8

T+G 0

2.3" 4

1.5" 7.8

22

ROGER M. PERLMUTTER E T AL.

known variant VH and VL segment protein sequences. No really surprising features emerge from this analysis. Transition and transversion mutations occur at equal frequency although on a random basis transversions should be twice as frequent as transitions. The mutational process exhibits no clear base preference or codon position asymmetries. Base deletions and insertions may also occur. XIV. Somatic Mutation Is Correlated with Immunoglobulin Class

Yet another striking feature of somatic mutation in the PC system is its correlation with antibody class switching (Crews et ul., 1981). IgG and IgA antibodies are frequently altered, while IgM antibodies are invariably germline. This pattern can best be seen in Fig. 4 where the heavy chain sequences are grouped by antibody class. Five IgM heavy chain sequences are identical to the T15 prototype whereas all five IgG VH segments are mutated. Only four of nine IgA heavy chain sequences are identical to the germline sequence. Figure 5 further substantiates the correlation. Here 12 of 29 IgG and IgA VH segments have undergone somatic mutation while all 13 IgM VIf sequences are identical to the germline sequence. Both IgG, and IgG3 subclasses can undergo somatic mutation. There is no information available for the IgGz, or IgGzi, subclasses or for IgD or IgE anti-PC antibodies although a somatic variant IgGz, antibody has been reported in the NP family (Bothwell et al., 1981). Somatic mutation of light chain variable region genes is also correlated with the class of antibody produced: IgG and IgA anti-PC antibodies undergo somatic mutation, IgM antibodies do not (Figs. 10 and 11).There may be a correlation between the quantitative degree of mutation observed in heavy and light chains from the same antibody. For the three antibodies where complete VH and V, sequences are available, the number of VH segment substitutions is as follows: M167,9; HPCG13,4; and M511, 1;and the total number of somatic mutations in and around the light chain variable gene is M167, 11; HPCG13, 10; and M511, 5. Although preliminary, this correlation suggests graded activity of the hypermutation mechanism acting on both heavy and light chain gene segments and associated at least circumstantially with the heavy chain class switch. This observation raises the possibility that a common mechanism may diversify both VH and V, genes. XV. VH, D, and JH Segments and Junctional Diversity

As previously discussed, the majority of replacement substitutions in the variable regions of sequenced anti-PC antibody heavy chains

DIVERSITY IN

PC-BINDINGANTIBODIES

23

occur in the third hypervariable region. In Fig. 13 we attempt to reconstruct these variant sequences by analyzing the germline segments which might potentially encode them. In dissecting these sequences we have relied insofar as is possible on germline sequences to explain the data, recognizing that there are alternative, though usually more complicated, possible constructions. We assume that the germline D segments utilized are those isolated and sequenced by Kurosawa and Tonegawa (1982) and leave as unresolved the question of whether additional D segments exist in germline DNA. Additional nucleotides or mutations required to generate each specific variant are also listed in Fig. 13.

XVI. Diversity in the VH Segment All but one of the heavy chain variable regions utilize the V1 gene segment, however, the recombination point for D joining within the V1 gene can vary. Sixteen of 18 sequences have an aspartic acid at residue 101 which is encoded by the V1 gene segment. The two sequences that differ, M603 and W3207, may reflect somatic mutation converting the aspartic acid to asparagine. There is one example, HPCG13, in which the joining event occurs 3’ to codon 101 to yield an alanine at position 102 and another example of this has been described in anti-PC antibodies of C57BL/6 origin (Clarke et al., 1983). Thus it is clear that flexibility in the rearrangement site of the V1 gene can generate diversity at the V-D boundary. XVII. Diversity in the JH Segment

All of the BALB/c anti-PC heavy chain variable regions utilize the same germline JH gene segment, JH1. This may reflect a restriction on the gene segment joining mechanism such that the V1 gene segment and the appropriate D gene segments join preferentially to the J H 1 gene segment. More likely, use of the Jill gene segment may be important in the formation of antibodies with affinity for PC, thus the association arises by virtue of selection for PC binding. Although it is true that the only Jki contribution to PC binding as determined by Xray crystallography is a van der Waals component associated with the tryptophan at position 110 (a residue that is mutated to glycine in M167), Scharff has described a culture-derived mutant of the S107 myeloma selected for loss of binding to PC-KLH which had a single replacement substitution at position 113 in the Jkrsegment (Cook et al., 1982). Thus JH substitutions can alter antigen binding in the antiPC antibodies and selection is probably responsible for the invariant

24

ROGER M. PERLMUTTER E T AL. 715 PROmPlPE (5107): A

v1

R

101 D Y

Y

G

S

S

Y

W

Y

MAAMGAT

D CFL16.1)

TAC TAC GGT PGT kX TAC TGG TAC

JH1

~~

i-Fo6:

v1

A

R

101 D Y

Y

D

Y

P

-uC TGG TAC A

R

101 D F

Y

R

Y

D

T?C TAT &T

D (SP2.3)

v1

TAC GAC

IEX A

R

101 D Y

Y

G

S

R

A

R

Y

Y

G

S

Y

W

TIC TAC GGT LET TAC TGG TAC T

v1

R ? A

D

Y

G

N

S

Y

D CFL16.1)

TAC GGT

*;lr

LUW A

R

101 D V

Y

Y

G

Y

D

TGG TAC A

R

101 D Y

Y

G

S

N

A GU

R

Y

G

S

T

D (FL16.1)

Y

TAC TAC GGT PGT $C TG6 TAC

101 N Y

A R MA W L T

Y

K

TAC TAT

Y

D

L

QC

A

R

101 D G

D

W

Y

T K 0s

JH1

J"1

W

m LT

D (FL16.11

611: v1

Y

TAC TGG TAC 101 N Y

JH'

v1

W

TAC TAC GGT K.T A%

D (FLl6.l)

w3207:

Y

MAkXGAT

JH1

v1

Y

TC TAC TAT GGT TAC GAC

D (FL16.1)

603:

W

6 U AM GAT G

D CSP2.3)

Ha:

6

Y

PGC TIC

JHl

v1

F

AM GAT GC&

JH1 WCGlL!:

Y

MAKAGATGCA

D (FL16.1)

v1

Y

TAC TGG TAC 101 D A

JH1

f4l61:

W

TAC TAC GGT PGT &C

JH1

v1

Y

MAAMGAT

D (FL16.1)

HF'CG13:

G

MAWGAT

JH1 WGB:

Y

TAC TAT GAT TAC

JH'

v1

W

MAKAGAT

D CSP2.2)

Wc44 :

H

Y

G

S

S

TGG TAC Y

W

Y

GCA AM GAT ffiJ &4C TAC GGT PGT PGC TAC TGG TAC

TAC

DIVERSITY IN

PC-BINDINGANTIBODIES

25

use of the JH1 gene segment in PC-binding antibodies. Clarke et aZ. (1983)have reported a CBA/J anti-PC hybridoma which uses a different VH segment and JH3. Thus the constraints on JH gene segments in PC-binding antibodies are not absolute. Figure 13 shows that the site of D-J joining can vary by as many as four nucleotides within the J Hregion. ~ Junctional diversity in the D-J joining site of the M603 sequence causes a deletion of one codon that shortens the M603 variable region by a single amino acid. XVIII. Diversity in the D Segment

At least three different germline D segments are probably utilized in the heavy chain variable region of PC-binding antibodies. The DFL16.1 gene segment is used in 13/17 variable regions. The four remaining proteins utilize either DSP2.2 or a member of the DSP2.3 D segment family (DSP2.3, DSP2.4, DSP2.6), all of which encode identical amino acids. As in the case of the VH and JH gene segments, the recombination point of the germline D segments is variable. This variation may occur at both the 5‘ and 3’ ends of the D segment and may involve up to six nucleotides. There is also evidence for somatic mutation-generated diversity within D segments of PC-binding antibodies. For example, M167 appears to have undergone a mutation of the 5’-most serine residue of DFL16.1 to an aspartic acid residue and W3207 and HPCM4 have mutations at position 105 yielding a different residue in each case. XIX. The N Region

Many of the heavy chain variable region sequences contain extra nucleotides both 5‘ and 3’ to the D segment which are not encoded in any known germline sequence (Kurosawa and Tonegawa, 1982; Clarke et al., 1983).These extra sequences, termed “N regions” by Alt and Baltimore (1982), span from one (e.g., M603) to four (e.g., M167) nucleotides and vary in nucleotide sequence. Eight of the sequences FIG.13. Junctional diversity in heavy chains from PC-binding antibodies. For each protein for which complete sequence data are available, the amino acid sequence is listed at the top of each box, and the likely nucleotide sequences derived from VII (Vl), D, and JII segments are listed below. The D segment nomenclature conforms to that used by Kurosawa and Tonegawa (1982).An “X” indicates a variable position reflecting degeneracy in the genetic code. Bases that likely resulted from somatic hypermutation are marked with an asterisk (*). Bases which do not appear to derive from any germline element and which are localized to V1I-D or D-JII junctions (N regions, see text) are underlined. The position of residue 101 is indicated at the top of each box.

26

ROGER M. PERLMUTTER ET AL.

do not include an obvious N region. In the remainder, it is clear from Fig. 13 that this is a potentially significant source of diversity in the third hypervariable region of PC-binding antibodies. Four alternative hypotheses have been proposed to explain the phenomenon of N region nucleotides: (1) joining between D segments (Kurosawa and Tonegawa, 1982), (2) addition of nucleotides during V-D-J joining by a mechanism perhaps involving terminal transferase (Alt and Baltimore, 1982), (3) the existence of additional germline D segments which have not yet been isolated, and (4) dramatic and localized hypermutation at the appropriate ends of the VH , D, and J H sequences. Data from the PC-binding heavy chains do not permit us to distinguish between these possibilities, however, the existing repertoire of D segments cannot satisfactorily explain the observed sequences even if D-D joining is permitted. Thus we believe that either additional D segments must exist (Kaartinen et d., 1983; Perlmutter et al., 1984) or the process of antibody gene rearrangement must include the addition or mutation of extra nucleotides at the boundaries of the D segments. XX. J, Gene Segments and Junctional Diversity

For the M167 light chain group, only the JK5gene segment is utilized, the variable regions are all of the same length and the sequence surrounding the V-J joining site is invariant and does not differ from the expected germline result. This may reflect the importance of the leucine at position 101 of the J region which is a contact residue for binding PC. XXI. Summary of Diversity of Antibodies That Bind PC

As expected from serologic and electrophoretic analyses, most PCbinding antibodies are remarkably similar. Nevertheless, all of the described mechanisms which contribute to antibody heterogeneity are utilized by this family. Beginning with a germline repertoire comprised of one VH, three D, one J H , three V,, and one J, elements, a large somatic repertoire of PC-binding antibodies is generated through a series of events which serve to amplify the germline information.

1. Combinatorial association of heavy and light chains. The PCbinding antibodies provide proof that a particular heavy chain can productively associate with multiple light chains. For example T15

DIVERSITY IN

PC-BINDINGANTIBODIES

27

and HPCM3 have identical heavy chain variable regions but utilize light chains from different variable region groups. 2. Combinatorial joining of germline gene segments. Although V1 and JH1 are used almost without exception in PC-binding heavy chains, these gene segments are associated with at least three different D segments (see Fig. 13). 3. Junctional diversity. Examination of the heavy chain variable region sequences of PC-binding antibodies reveals that the site of VD and of D-J joining may vary within each gene segment. Thus the alanine residue at position 102 in HPCG13 is generated by a joining event which occurs two or three nucleotides 3’ to the usual V-D recombination site at position 101. Such shifts can also alter the length of the variable region (e.g., M603, Fig. 13). 4. Diversity may also be generated by a mechanism which appends additional nucleotides not encoded in the germline DNA to the D segment (e.g., M167). The nature of this mechanism is obscure. 5 . Finally the PC-binding antibodies provide strong support for the existence of somatic hypermutation operating coordinately on VH , D, J N , V,, and JKsegments. This is a major feature of diversity in the immune response to PC. Structural analysis of the anti-PC antibodies and the genes which encode them illuminates the multiple strategies utilized by higher vertebrates to amplify a limited amount of genetic information in order to permit the expression of at least 10 million different antibodies (Klinnian, 1980).Of the thousands of different germline antibody gene segments, fewer than a dozen are employed in PC-binding antibodies.

Mil.

Selection of Variant Antibodies

The progenitor lymphocyte for all B cells ultimately secreting antiPC antibodies is an IgM-bearing cell which synthesizes heavy and light chains identical to those encoded by the germline gene segments. Subsequent development yields a population of B cells, some of which secrete variant molecules. Selection of specific clones of variant B cells might well depend upoii the affinity of the variant antibodies for antigen. This scheme would provide the basis for the phenomenon of affinity maturation and is referred to as antigen-driven selection. Using diazophenyl-phosphorylcholine (DPPC), a PC analog thought to more closely resemble the haptenic determinant of PCderivatized protein conjugates, it has been possible to show that variant antibodies isolated from PC-KLH primed fusion products in

28

ROGER M. PERLMUTTER ET AL.

TABLE I11 INCREASED AFFINITY OF VARIANT ANTIBODIES FOR DIAZOPHENYLPHOSPHORYLCHOLIN

HPCM27 HPCGS HPCGlO HPCG13

Class

Sequence

K x 105 ( ~ - 1 )

IGM IgG3

Germline Germline Variant Variant

4.7 3.4 20.6 25.6

k c 3

IgGi

" Data from Rodwell et al. (1983).

general show a higher affinity for DPPC than does the germline-encoded molecule, T15 (Rodwell et al., 1983). Table I11 shows the measured dissociation constants for germline and variant monoclonal antibodies binding DPPC. The variant antibodies bind DPPC five times more avidly than germline antibodies using light chains from the M167 group. It is interesting to note that the antigen contact residues of the variant heavy chains remain largely unmutated while a cluster of substitutions occurs around position 56 (Fig. 4). Here selection may bias the distribution, number, and types of substitutions observed in somatic variants. Thus antigen-driven selection may indeed be responsible for the recruitment of specific variant clones and hence for the fine tuning of the antibody response. Recruitment of specific variant clones by antigen may also underlie the observed association of somatic mutation with antibody class: those clones most likely to undergo a class switching event are also those most exposed to the selective pressure of antigen. In this regard, it is interesting to note that no correlation between somatic mutation and antibody class has been found in antiarsonate antibodies bearing a cross-reactive idiotype which appear to be encoded by a single VHgene segment (Siekevitz et al., 1983; Sims et al., 1982). In addition, Kocher et al. (1980) have reported an amino-terminal sequence of an IgM anti-PC heavy chain which may have resulted from somatic mutation of the V1 gene, implying that somatic mutation can occur in IgM-secreting B cells. Another set of variant clones with high affinity for DPPC utilizes light chains similar to HPCG26 associated with a very different heavy chain that is not encoded by any member of the T15 gene family (Chang et al., 1984; Todd et al., 1984). These antibodies bind PC only weakly, are present in relatively low concentration in PC-immune sera, and in contrast to the T18like antibodies are largely restricted to the IgGl subclass (Chang et al., 1982). Thus two subpopulations of PC-binding B cells can contribute to the generation of a diverse antiPC response and antigen may select from these subpopulations vari-

DIVERSITY IN

PC-BINDING ANTIBODIES

29

ant clones, products of somatic mutation, with increased binding affinity. An alternative formulation invokes idiotype-specific T cells which cause the selective expansion of T15 idiotype-positive B lymphocytes (Bottomly et al., 1978).The appearance of variant clones later in the anti-PC response is then attributed to the action of suppressor T lymphocytes and/or antiidiotypic antibodies which provide variant molecules with a selective advantage. XXIII. Molecular Basis of the T15 ldiotype

Although the T15 idiotype has been identified as a potential site for regulation of the immune response to PC, attempts to correlate primary structure data with serologic reactivity of anti-T15 reagents reveal that the T15 idiotype is a complex structure. This analysis is further complicated by heterogeneity in existing anti-T15 sera which may not recognize identical determinants (Perlmutter and Davie, 1977; Clarke et ul., 1982). Linkage of the T15 nonbinding site marker to the Ig-la allotype suggests that the T15 idiotype localizes to the heavy chain (Lieberman et al., 1981). However, since T15 and HPCM3 share identical heavy chain variable regions and differ in light chain sequence, and HPCM3 is T15 idiotype negative, the T15 idiotype must reflect contributions from both the light chain and the heavy chain. XXIV. The T15 VH Gene Family

The four VH gene segments of the T15 VH gene fiamily constitute a small group of sequences greater than 88% homologous. As we have discussed, the V1 gene segment encodes all of the heavy chains of anti-PC antibodies with a single exception which is probably encoded by V11 (PCG15, Fig. 6). The translated V11 sequence is identical to the VH region sequence of the M47A myeloma heavy chain (Kabat et al., 1983). The V 3 gene segment contains two in-phase termination codons and an aberrant 3' recognition sequence and so is a pseudogene (Huang et al., 1981). The V13 gene segment has never been seen in an expressed antibody sequence but there is no obvious feature of its nucleotide sequence that would preclude it from being functional. The arrangement of the members of this gene family along chromosome 12 is currently under investigation. The V3 gene segment has been located 16 kb 5' to the V 1 gene segment, thus two members of this family are closely linked in the genome. Although we suspect that

30

ROGER M. PERLMUTTER ET AL.

the V11 and V13 gene segments are also linked 5’ to the V3 gene segment, the precise location of these coding regions remains to be determined. The small size of the T15 gene family parallels the highly restricted nature of the antibodies which this family encodes-only one VH gene is responsible for virtually all PC-binding heavy chains. Antibodies binding group a streptococcal carbohydrate (GAC) are considerably more diverse and several closely homologous VH gene segments contribute to the total antibody repertoire (Perlmutter et al., 1984). XXV. Evolution of the T15 Gene Family

Anti-PC antibodies are highly protective against in v i m challenge with viable pneumococci in mice (Briles et al., 1981) and antibodies bearing the T15 idiotype are most protective (Briles et al., 1982). Since all of these antibodies must include heavy chains encoded by the V1 gene segment, we expect that there is strong selection to maintain this structure in the germline. In support of this contention, Clarke et al. (1982) have provided clear evidence for a V1-equivalent allele in Ig-1” mice which differs from its BALB/c Ig-la counterpart at four positions in the variable region. An N-terminal sequence of a human myeloma with PC-binding activity has also been determined and shows strong homology with T15 including four of five residues identical in the first hypervariable region (Riesen et al., 1976). Clarke et aZ. (1983) have reported the complete heavy chain variable region sequences of five PC-binding antibodies of C57BL/6 (Ig-ll’) origin and five of CBA/J origin. As shown in Fig. 14, the VH structure is well conserved in these mouse strains with apparent polymorphisms in the V1 gene at positions 14, 16,40, and 44 of the protein sequence. One of the CBA/J antibodies, 6G6, utilizes a heavy chain gene distinct from -”-.-L -!

w W I C C518L

z?

l!

715 293

~

L

V

E

~

L

~

o! S

L

R

L

S

~

T

1c2

G

F

T

F

S

~

~

R

L

~

I

9p ~

~

~

lf‘.

l?O Y

~

~

‘TO

F WGSS I V WSF l ?~W 2T& l l S V l U~S I N-

23169 2851 2313 GF9

7C6 666

S

Bp

T?

-I--

1613

CB4

40

5?

b.0

s

Ai )-G-

N--06

r.

+-)c-----

TV TL -H -lT-)

Ho

4-4

x2

FIG.14. Conservation of the TI5 VH sequence in other inbred mouse strains. Amino acid sequences for heavy chains derived from five PC-binding hybridomas raised in C57BL6/J mice and five PC-binding hybridomas raised in CBA/J mice are compared with the sequence of the BALB/c T15 heavy chain (modified from Clarke et al., 1983). Residues are numbered sequentially from the amino terminus. The positions of VI1,D, and JII-encoded regions are delineated with arrows.

L

Y

~

A

DIVERSITY IN

PC-BINDING ANTIBODIES

31 t . HV1+

1

10

30

20

115

E V K L V E S C C G L V Q P G C S L R L S C A T S C F T F S D F Y M E W V

FR

-Q-D Y

Rat GPyl

A

-4 - 1 Q -1-4

CPy2 -Q

N-0-

A - L - RE A E

VA-CS-1-NY-D-

E

V A-C

5-1-NY

W-D-

FIG.15. Sequences of PC-binding heavy chains from several species. The aminoterminal sequence of the heavy chain of FR, a human myeloma protein with PCbinding activity (Riesen et al., 1976),is compared with the amino-terminal sequence of the murine T15 heavy chain. Also shown is the N-terminal sequence of heavy chains from pooled anti-PC antibodies raised in Lewis rats (Braciale et ul., 1977) and the sequences of y l and y2 subclasses of anti-PC antibody heavy chains raised in outbred Hartley guinea pigs. The position of the first hypervariable region is marked with arrows.

V1 which is perhaps the result of complicated gene conversion or recombination events within the T15 family although this has not been proved (Clarke et al., 1982). Although the strong conservation of the T15 VH sequence at the protein level may reflect selection pressures favoring this structure, polymorphism in the laboratory mouse may be quite limited (Ferris et ul., 1982). We have performed amino-terminal sequence analysis on PC-binding antibodies from hyperimmune rats and guinea pigs. Figure 15 compares the available heavy chain sequences with the mouse germline T15 sequence. The framework positions of mouse, human, and rat anti-PC heavy chains are quite closely conserved; however, the guinea pig sequences are more divergent. It is worth noting that guinea pigs make very little antibody in response to PC immunization (Schroer and Davie, 1977).The light chain sequences of rat, human, guinea pig, and mouse anti-PC antibodies are much less similar than the heavy chain sequences (Fig. 16), although in some respects the guinea pig and rat sequences resemble a hybrid between mouse T15 and M167 light chains. Examination of the T15 V H gene family in different mouse strains and other species can be expected to yield important information about the selective pressures which operate on heavy chain variable gene segments. WI. Structural Diversity in Antibodies to PC

The PC-binding antibodies which we have analyzed are encoded by only nine germline gene segments and yet only half exhibit a

32

ROGER M. PERLMUTTER E T AL.

1 715

10

c--- Hv1

20

30

40

D I V M T Q S P T F L A V T A S K K V T I S C T A S E S L Y S S K H K V H Y L A ~

M603

S ~ - S - S - C E ~ H - K S - Q - - L N - G N P K N -

M167 - 1 - D E L S N P - S C E S - S - R S - K - L Y K D C - I Y L N W F

-V - LIN-

FR Rat

GPy1

L-P-LCEPAS-9-RS-Q-VYRNCNIYLNWF AS-T-S-GE-M-KS-Q?-LY-E 1 - L S L S - P G E P A V - R l - Q R - L P A N D C O T N F Y R

CPy2 - l - L S L S - P C E P A V - R T - Q R - L P A N D C D ? N F Y

FIG. 16. Sequences of PC-binding light chains from several species. The aminoterminal sequences of the light chains of the T15, M603, and M I 6 7 murine myelomas are compared with those obtained for FR, a human PC-binding myeloma protein (Riesen et al., 1976) and for light chains from pooled Lewis rat, Hartley guinea pig yl, and Hartley guinea pig 72 anti-PC antibodies. The position of the first hypervariable region is shown at the top. Residues are numbered sequentially from the amino terminus.

germline sequence. Of 42 N-terminal sequences available for scrutiny, 16 different VH and 13 different VL sequences are represented and in five of these antibodies only one chain differs from the germline; thus 21 different antibody sequences were found. Table IV lists the number of mutational events attributable to each mechanism employed in the generation of antibody diversity. Since new IgG PCbinding variants have been steadily accruing without evidence of repTABLE IV AMINOACID SUBSTITUTIONS IN PC-BINDING ANTIBODIESCLASSIFIED ACCORDINGTO SITE AND PROBABLE UNDERLYING MECHANISM^ Heavy chain HV3h

Total

Different D segments Recombination site shift Somatic mutation N region diversity Somatic mutation or N region

8 5 3 10 6

8 5 26 10 6

Totals

32

55

Light chain total

20

Total number of substitutions = 75 Includes 19 complete VI, sequences, 23 N-terminal Vlr sequences, 6 complete V,, sequences, and 36 N-terminal VL sequences. Third hypervariable region.

DIVERSITY

IN

PC-BINDING ANTIBODIES

33

etition, we anticipate that the potential repertoire of PC-binding antibodies must be very large indeed. Extrapolating this result to the entire humoral immune system, it is clear that the number of possible antibody sequences is virtually unlimited. XXVII. Future Research

Critical examination of the gene structures involved in generating anti-PC antibodies has permitted us to identify the mechanisms responsible for antibody diversity. Clearly, in this model immune response, information encoded in multiple germline elements is expanded through combinatorial joining of gene segments, somatic hypermutation, and combinatorial association of heavy and light chains to generate a diverse repertoire of similar but distinct antibody specificities. Figure 17 diagrams the differentiative pathway of B lymphocytes and the points at which these mechanisms of diversification operate. There is now considerable evidence that there exists a developmental hierarchy of V gene expression such that the total antibody repertoire of each individual is acquired in a programmed fashion in fetal and early neonatal life (Klinman, 1980). The experimental tools are now in hand to test this hypothesis and to evaluate the mechanism which directs the sequential activation of different V genes. Understanding the biology of this process will likely illuniinate many aspects of differentiation. A second area of intense investigation will be the enzymology of DNA rearrangements and of somatic hypermutation. A prerequisite for these studies will be the development of cell lines which can be induced to rearrange V genes and activate the class switch mechanism in a coordinated fashion. This will be a challenging task for cellular immunologists in the future. A third area likely to assume increasing importance will be the study of the evolution of the antibody gene families which should provide insight into the forces acting to mold the genome. The strategies for information amplification utilized in the generation of the immunoglobulin repertoire are likely employed in other complex systems, for example in T cell recognition structures, in directing embryonic development and perhaps in the nervous system. Assembly of cell surface molecules utilizing gene rearrangements could provide individual cells or cell subsets with a specific identity, thus enabling cell-cell recognition (Hood et al., 1977a,b). The elucidation of these processes will aid biologists in obtaining a detailed understanding of normal and abnormal development.

34

ROGER M. PERLMUTTER ET AL.

MECHANISM OF DIVERSITY

-NUCLEIC ACID REARRANGEMENT

GERMUNE REPERTOIRE VHFORMATION COMBINATORVJ. JOINING JUNCTIONAL DIVERSIP(

STEM CELL

L

@

N REGIONS

COMBINATORUL JOINING

@

PRE-B CELL

V FORMATION

JUNCTIONAL

DMRSlTY

@

COMBINATORUL AS!socIATION

HYPERMUTATION

B CELL

ISOTYPE SWITCH

PLASMA CELLS

FIG.17. The generation of antibody diversity. Each of the different mechanisms which contributes to antibody diversification (shown at left) is associated with specific nucleic acid rearrangements (center) and occurs at specific times during B cell ontogeny (indicated schematically at right). The action of these different mechanisms results in a population of plasma cells with extraordinarily diverse binding properties in multiple effector classes.

ACKNOWLEUGMENTS We thank Drs. J. Schroer, G . der Balian, and J. M . Davie for providing pooled rat and guinea pig antibodies, Dr. Mitch Kronenberg and Gerald Siu for critical reading of the manuscript, and Connie Katz and Bernita Larsh for expert secretarial assistance. This work was supported by grants from the NIH.

DIVERSITY IN

PC-BINDING ANTIBODIES

35

REFERENCES Alt, F., and Baltimore, D. (1982).Proc. Nutl. Acud. Sci. U.S.A. 79,4118-4122. Amzel, L., Poljak, R., Saul, F., Varga, J., and Richards, F. (1974). Proc. Nutl. Accid. Sci. U.S.A. 71, 1427-1430. Appella, E. (1980). Mol. Immunol. 17, 711-718. Baltimore, D. (1981).Cell 24, 592-594. Barstad, P., Rudikoff, S., Potter, M.,Cohn, M.,Konigsberg, W., and Hood, L. (1974). Science 183, 962-964. Bothwell, A. L. M., Paskind, M., Reth, M., Imanishi-Kari, T., Rajewsky, K., and Baltimore, D. (1981). Cell 24, 625-637. Bottomly, K., Mathieson, B., and Mosier, D. (1978).J.E x p . Med. 148, 1211-1227. Braciale, V., der Balian, G., and Davie, J. M.(1977). Biochemistry 16, 5303-5308. Brack, C., Hirama, M., Lenhard-Schuller, R., and Tonegawa, S. (1978). Cell 15, 1-14. Briles, C., Nahm, M., Schroer, K., Davie, J. M.,Baher, P., Kearney, J., and Barletta, R. (1981).J. E x p Mecl. 153, 694-705. Briles, D., Forman, C., Hudak, S., and Claflin, J. L. (1982).J.E x p . Med. 156,1177-1185. Capra, J. D., and Kehoe, J. M. (1974). Proc. N o t / . Acud. Sci. U.S.A. 71, 845-849. Chang, S . P., Brown, M.,and Rittenberg, M. B. (1982).J. Imrnunol. 128, 702-710. Chang, S. P., Perlmutter, R. M., Brown, M., Heusser, C. H., Hood, L., and Rittenberg, M. B. (1984).J. Immunol., in press. Claflin, J. L., and Davie, J. M. (1974a).J . Immurao/. 133, 1678-1684. Claflin, J. L., and Davie, J. M. (1974b).J.Zmrnunol. 114, 70-75. Claflin, J. L., and Rudikoff, S. (1976). Cold Spring Harbor Symp. Quunt. Biol. 41, 725734. Clarke, S. H., Claflin, J. L., and Rudikoff, S. (1982). Proc. Natl. Acud. Sci. U.S.A. 79, 3280-3284. Clarke, S. H., Claflin, J. L., Potter, M., and Rudikoff, S. (1983).J.E x p . Mecl. 157,98-113. Cook, W., Rudikoff, S., Ginsti, A., and Scharff, M. (1982). Proc. Nutl. Acud. Sci. U.S.A. 79, 1240-1244. Cosenza, H., and Kohler, H. (1972).Proc. Natl. Acud. Sci. U.S.A. 69, 2701-2705. Crews, S., Griffin, J., Huang, H., Calame, K., and Hood, L. (1981). Cell 25, 59-66. Davis, M. M., Calame, K., Early, P. W., Livant, D. L., Joho, R., Weissman, I. L., and Hood, L. (1980). Nature (London) 283,733-738. Dayhoff, M. 0. (1976).“Atlas of Protein Sequence and Structure.” National Biomedical Research Foundation, Bethesda, MD. Dreyer, W. J., and Bennett, J. C. (1965). Proc. Natl. Acud. Sci. U.S.A. 54, 865-869. Early, P. W., Davis, M . M., Kaback, D. B., Davidson, N., and Hood, L. (1979).Proc. Nutl. Acad. Sci. U.S.A. 76, 857-861. Early, P. W., Huang, H., Davis, M. M., Calame, K., and Hood, L. (1980). Cell 19, 981992. Edelman, G. M., and Gally, J. A. (1967). Proc. Nutl. Acud. Sci. U.S.A. 57, 353-358. Ferris, S. D., Sage, R. D., and Wilson, A. C. (1982). Nuture (London) 295, 163-165. Gearhart, P., and Bogenhagen, D. (1983). Proc. Nutl. Acud. Sci. U.S.A. 80,3439-3443. Gearhart, P., Sigal, N., and Klinman, N. (1975a).J . E x p . Med. 141, 56-74. Gearhart, P., Sigal, N., and Klinman, N. (197511).Proc. Natl. Acud. Sci. U.S.A.72, 17071711. Gearhart, P., Johnson, N., Douglas, R., and Hood, L. (1981). Nature (London)291, 2934. Gershenfeld, H., Tsukamoto, A., Weissman, I. L., and Joho, R. (1981). Proc. Nutl. Acad. Sci. U.S.A. 78, 7674-7678.

36

ROGER M. PERLMUTTER E T AL.

Hieter, P. A., Korsmeyer, S. J., Waldmann, T. A., and Leder, P. (1981).Nature (London) 290,368-372. Hilschmann, N., and Craig, L. C. (1965). Proc. Natl. Acad. Sci. U.S.A. 53, 1403-1409. Hood, L., Huang, H. V., and Dreyer, W. J. (1977a).J.Supramoh. Struct. 7,531-559. Hood, L., Loh, E., Hubert, J., Barstad, P., Eaton, B., Early, P., Fuhrman, J., Johnson, N., Kronenberg, M., and Schilling, J. (197713).Cold Spring Harbor Symp. Quant. Biol. 41,817-836. Hozumi, N., and Tonegawa, S. (1976). Proc. Natl. Acad. Sci. U.S.A. 73,3628-3632. Huang, H., Crews, S., and Hood, L. (1981).J. Mol. Appl. Genet. 1,93-101. Kaartinen, M., Griffiths, G., Markham, A. F., and Milstein, C. (1983).Nature (London) 304,320-324. Kabat, E. A,, Wu, T. T., Bilofsky, H., Reid-Miller, M., and Perry, H. (1983).“Sequences of Proteins of Immunological Interest.” U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda, MD. Kim, S., Davis, M. M., Sinn, E., Patten, P., and Hood, L. (1981). Cell 27, 573-581. Klinman, N. (1980). In “Immunoglobulin Genes and B Cell Differentiation” (J. R. Battisto and K. L. Knight, eds.), pp. 193-205. Elsevier North Holland, Amsterdam. Kocher, H. P., Berek, C., Schrier, M. H., Cosenza, H., and Jaton, J. C. (1980). Eur. J. lmmunol. 20,264-271. Kohler, G., and Milstein, C. (1976). E u r . ] . lmrnunol. 6,511-519. Kurosawa, Y., and Tonegawa, S. (1982).J. E x p . Med. 155,201-218. Kwan, S.-P., Rudikoff, S., Seidman, J. G., Leder, P., and Scharff, M. D. (1981).J. E x p . Med. 153,1366-1370. Lamoyi, E., Estess, P., Capra, J. D., and Nisonoff, A. (1980).J. lmmunol. 124, 28342840. Landsteiner, K. (1945). “The Specificity of Serological Reactions,” 2nd ed. Harvard Univ. Press, Cambridge, Mass. Lee, W., Cosenza, H., and Kohler, H. (1974). Nature (London) 247,55-57. Leon, M. A., and Young, N. M. (1971). Biochemistry 10,1424-1429. Lieberman, R., Potter, M., Mushinski, E., Humphrey, W., and Rudikoff, S. (1974).J. Enp. Med. 139,983-991. Lieberman, R., Rudikoff, S., Humphrey, W., Jr., and Potter, M. (198l).J.Immunol. 126, 172-176. Maki, R., Kearney, J., Paige, C., and Tonegawa, S. (1980). Science 209, 1366-1369. Max, E. E., Seidman, J. G., and Leder, P. (1979).Proc. Natl. Acad. Sci. U.S.A.76,34503454. Padlan, E. A., Segal, D. M., Spande, T. F., Davies, D. R., Rudikoff, S., and Potter, M. (1973). Nature (London) New B i d . 245, 165-167. Perlmutter, R. M., and Davie, J. M. (1977).J. lmmunol. 118, 769-774. Perlmutter, R. M., Hamburg, D., Briles, D. E., Nicolotti, R. A,, and Davie, J. M. (1978). J. Immunol. 121,566-572. Perlmutter, R. M., Klotz, J. L., Bond, M. W., Nahm, M., Davie, J. M., and Hood, L. (1984).1.E x p . Med., in press. Potter, M. (1972). Physiol. Reo. 52,631-719. Potter, M., Newell, J. B., Rudikoff, S., and Haber, E. (1982). Mol. Immunol. 19, 16191630. Putnam, F. W., Shimizu, A., Paul, C., Shinoda, T., and Kohler, H. (1971).Ann. N.Y. Acad. Sci. 190, 83-103. Riesen, W., Braun, D., and Jaton, J.-C. (1976). Proc. Natl. Acad. Sci. U.S.A.73, 20962100.

DIVERSITY IN

PC-BINDING ANTIBODIES

37

Robinson, E. A., and Appella, E. (1979).J.Biol. Chem. 254, 11418-11430. Rodwell, J., Gearhart, P., and Karush, F. (1983).J . Zmmunol. 130, 313-316. Rudikoff, S., and Potter, M. (1976). Proc. Natl. Acad. Sci. U.S.A.73, 2109-2112. Rudikoff, S., and Potter, M. (1978). Biochemistry 17,2703-2707. Rudikoff, S., Satow, Y., Padlan, E., Davies, D., and Potter, M. (1981).Mol. Zmmunol. 18, 705-71 1. Ruppert, V. J., Williams, K., and Claflin, J. L. (1980).J . Immunol. 124, 1068-1074. Sakano, H., Huppi, K., Heinrich, G., and Tonegawa, S. (1979). Nature (London) 280, 288-294. Sakano, H., Kurosawa, Y., Weigert, M., and Tonegawa, S. (1981).Nature (London) 290, 562-570. Schilling, J., Clevinger, B., Davie, J. M., and Hood, L. (1980).Nature (London)283,3540. Schroer, J., and Davie, J. M. (1977).J.Zmmunol. 118, 1987-1993. Seidman, J. G., Leder, A., Nau, M., Norman, B., and Leder, P. (1978). Science 202, 1117. Seidman, J . G., Max, E. E., and Leder, P. (1979).Nature (London) 280,370-375. Selsing, E., and Storb, U. (1981). Cell 25, 47-58. Siekevitz, M., Huang, S. Y., and Gefter, M. L. (1983). Eur.J. Zmmunol. 13, 123-132. Sims, J., Rabbitts, T. H., Estess, P., Slaughter, C., Tucker, P. W., and Capra, J. D. (1982). Science 216, 309-311. Southern, E. M. (1975).J . Mol. Biol. 98, 503-517. Todd, I., Chang, S. P., Perlmutter, R. M.,Aebersold, R., Heusser, C. H., Hood, L., and Rittenberg, M. (1984).J . Zmmunol., in press. Weigert, M., Perry, R., Kelley, D., Hunkapiller, T., Schilling, J., and Hood, L. (1980). Nature (London) 283,497-499. Wu, T. T., and Kabat, E. A. (1970).J . E x p . Med. 132,211-250.

This Page Intentionally Left Blank

ADVANCES I N IMMUNOLOGY, VOL. 35

Immunoglobulin RNA Rearrangements in B Lymphocyte Differentiation JOHN ROGERS' AND RANDOLPH WALL' 'MRC Laboratory of Molecular Biology, Univedy Medico1 School, Cambridge, England and tMolecular Biology Institute and Department of Microbiology and Immunology, UCLA School of Medicine, Los Angeles, California

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............. 11. Immunoglobulin Structure .................... ............. 111. Immunoglobulins as Antigen Receptors on B Cells. . . . . . . . . . . . . . . . . . . . IV. Two mRNAs with Different 3' Ends Encode Membrane and Secreted p Heavy Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. All Immunoglobulin Heavy Chain Genes Have Membrane Gene Segments . . . . . . . . . . . . . . . . . . . . . . . . . . ......................... VI. A Model for the Transmembrane Insertion of Heavy Chain M Regions. . . VII. Membrane and Secreted Heavy Chain mRNAs Are Coded by Transcription Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Coexpressed p and 6 mRNAs Transcription Unit . . . . . . . . . . IX. Summary: Developnrental Reg1 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note Added in P r o o f . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39 40 42 43

46 47

50 52 55 56 59

I. Introduction

The genetics and diversity of immunoglobulin genes have long fascinated molecular biologists. The applications of the powerful techniques of recombinant DNA cloning and nucleic acid sequencing have now revealed that immunoglobulin light and heavy chain genes are subject to an amazing range of dynamic processes at the DNA level (reviewed by Honjo, 1983).DNA rearrangements generate transcriptionally active light and heavy chain genes (i.e., transcription units). These DNA rearrangements are unique among the higher eukaryotic genes studied to date. In conjunction with multiple germline variable region genes and as yet unresolved mechanisms which generate somatic mutations in variable regions, these DNA joining events contribute significantly to the immense diversity of antibodies. These active mechanisms for generating diversity are also unique among the eukaryotic genes now studied. This article reviews recent findings which reveal the importance of RNA splicing and processing rearrangements in the expression and regulation of immunoglobulin heavy chain genes. RNA rearrange39 Copyright D 1984 b y Academic Press, Inc. A11 rights of reproductlon in any form reserved. ISBN 0-12-022435-6

40

JOHN ROGERS AND RANDOLPH WALL

ments greatly increase the versatility of immunoglobulin heavy chain genes by generating alternate forms of mRNA, with different functions, from one transcription unit. Both secreted and membrane mRNA species for all heavy chain classes are generated by alternative RNA splicing pathways. Coexpressed p and 6 heavy chains in surface IgM and IgD are derived from a single large transcription unit by a complex array of alternative RNA processing pathways. While such alternate modes of RNA processing are known for viral genes, immunoglobulin heavy chain genes are the first chromosomal genes whose expression has been shown to be developmentally regulated by these posttranscriptional RNA processing mechanisms. II. Immunoglobulin Structure

The immunoglobulin molecule can exist in two very different environments: in the lymphocyte cell membrane as a surface antigen receptor, and in the circulation as a secreted antibody. Until recently, all detailed information on the sequence and structure of immunoglobulin molecules referred to secreted antibodies. Most secreted immunoglobulin molecules consist of two identical light chains and two identical heavy chains. Each chain is composed of a series of homologous domains, the NH2-terminal one being the variable (V) region and the remainder forming the constant (C) region. The V regions are responsible for antigen binding and appear to have thousands of different sequences. The C regions embody the various effector functions of the molecule and have only a few alternative sequences. The various classes of immunoglobulins with different biological functions (IgM, IgD, IgG, IgE, IgA) are distinguished by different heavy chains (p, 6, y, E , a ) defined by their C regions (Cp, C8, C,, C,, CJ. The light chains in an immunoglobulin molecule can be either K or A class (containing C, or C k regions). For most immunoglobulin classes, the four-chain molecules are secreted as monomers. However, in IgM and IgA, the heavy chains carry a short COOH-terminal extension which permits them to be linked into secreted pentamers or dimers, respectively. These immunoglobulin molecules, secreted as multimeric complexes, are covalently linked through their COOH-terminal segments to J (or joining) chain molecules (reviewed in Koshland, 1975; McHugh et al., 1981). Not surprisingly, the COOH-terminal segments of p and a chains are much more homologous in sequence than their other domains (Kehry et al., 1980; Robinson and Apella, 1980).

IMMUNOGLOBULIN

RNA

41

REARRANGEMENTS

The early development of antibody-producing B cells is distinguished by the placement of the p heavy chain. As a lymphoid stem cell differentiates to a pre-B cell, p heavy chains (but not light chains) are expressed in the cytoplasm (Levitt and Cooper, 1980; Burrows et al., 1979). As the pre-B cell develops to a B cell, light chains are expressea and IgM (pZL2)monomers are displayed on the cell surface as antigen receptors (Vitetta et al., 1971). Next, IgM and IgD (p2L2) are coexpressed on the B lymphocyte surface (reviewed in Goding, 1978). Coexpressed p and 6 heavy chains in IgM and IgD, respectively, have the same V region (Coffman and Cohn, 1977; Raschke, 1978) and their expression exhibits allelic exclusion (Herzenberg et al., 1976). Subsequently, the B cell can differentiate to become a plasma cell which actively secretes IgM molecules (Melchers and Anderson, 1974). These are secreted in pentameric form through disulfide linkages between the p chains and J chains which generate the ( P ~ L ~ ) ~complexes -J (Della Corte and Parkhouse, 1973; Koshland, 1975; Mather et al., 1981). These early stages of B cell development correspond to the primary immune response which culminates in secreted IgM molecules (Fig. 1). In the secondary immune response, B cells express different heavy C,, C,) linked to the V region first associated with chain C regions (C7, p chains in surface IgM. This process is called class switching (reviewed in Hood et al., 1980; Honjo, 1983). Resting lymphocytes bearing surface immunoglobulins with C regions other than p are postu-

IMMUNOGLOBULIN CHAINS PRODUCED

STEM CELL

PRE-B CELL

-

II

RESTING B CELL

B CELL

I m U N f f i L O B U L I N SECRETING PLASMA CELLS

Y. L

II. L

p + 6 , L

P.L

E. L a. L

FIG.1. A simplified scheme of immunoglobulin heavy and light chain gene expression in mouse B cell development.

42

JOHN ROGERS AND RANDOLPH WALL

lated to be memory cells (Okumura et al., 1976). Unfortunately, cell lines corresponding to this stage are not available. However, tumors of the B cell lineage corresponding to the other development stages have proven to be of great value in the analysis of immunoglobulin gene expression. The pre-B cell is represented by Abelson virus-transformed cells (Siden et al., 1979). The B cell is represented by B cell lymphomas (Warner et al., 1979; Rashke et al., 1979), and the plasma cell by plasmacytomas or myeloma tumors. The recent development of continuous cloned lines of normal mouse lymphoid cells and their precursors is an important breakthrough for future studies on B cell development (Whitlock and Witte, 1982). 111. Immunoglobulins as Antigen Receptors on B Cells

Immunologists have long known that the secretion of immunoglobulins in response to antigen is triggered by the binding of the antigen to IgM displayed on the surfaces of B cells. Membrane IgM has the same basic structure as monomeric secreted IgM (pZL2). In early studies, there was some disagreement as to whether membrane p (p,) and secreted p ( p s )chains were different or identical. Peptide mapping studies indicated that the V regions and C domains from p chains in cell lines making both secreted and membrane IgM appeared to be highly similar in sequence (Yuan et al., 1980; Raschke et al., 1979). However, other studies showed that p,,, and p, chains exhibited slight differences in size and in physical properties. Comparison of pmand ps chains by SDS-polyacrylamide gel electrophoresis and by ultracentrifugation showed that p, chains are slightly larger than pschains (Melchers and Uhr, 1973; Bergman and Haimovich, 1978). Because both the p m and ,uschains are glycosylated, these small differences in apparent size could have reflected different degrees of glycosylation. However, comparison of p , and p, polypeptides synthesized in the presence of tunicamycin to inhibit glycosylation confirmed that the molecular weight differences were due to differences in amino acid sequence (Vassalli et al., 1979; Sibley et al., 1980). Translation of p, and p, mRNA isolated from B lymphoma cells also showed that p, polypeptides are larger than p, polypeptides (Singer et al., 1980). Several lines of evidence suggested that p m and pschains contained different COOH-terminal sequences which accounted for their differences in size and physical behavior. As would be expected for integral membrane proteins, the p,,, chains exhibited hydrophobic properties. Membrane IgM molecules were solubilized from the cell surface by detergents (Melchers et al., 1975). Furthermore, both membrane IgM

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

43

molecules and isolated p m chains bound small amounts of detergent, whereas secreted IgM and p\ chains did not (Melcher and Uhr, 1977; Vassalli et al., 1979; Parkhouse et uZ., 1980).These results suggested that pnlchains (but not ps chains) possessed a sequence capable of interacting with the hydrophobic plasma membrane. This hydrophobic region was subsequently localized to the Fc segment and more definitively to the COOH-terminus of pmchains by peptide mapping (Yuan et aZ., 1980; Kehry et al., 1980). Carbohydrate differences in p m and p5 chains were also localized to the COOH-terminal segments of p m and p5chains (Yuan et al., 1980; Parkhouse et ul., 1980; Williams et al., 1978). The most complete structural comparison of pm and ps chains by peptide mapping, microsequencing, and carboxypeptidase COOH-terminal analysis has confirmed that these two chains are identical in sequence up to their COOH-terminal segments (Kehry et al., 1980). The complete amino acid sequence of p5 chains from the mouse myeloma line, MOPC 104E, has been determined (Kehry et al., 1979). However, it has not been possible to purify pmchains sufficiently to obtain a complete amino acid sequence for the COOH-terminal segment. The amino acid sequence of the p m COOH-terminal segment in pnlchains was directly established by molecular cloning and nucleotide sequencing of pnlRNA (Rogers et al., 1980). This sequence was entirely consistent with the characterization of the pill COOH-terminal segment by Kehry et al. (1980). IV. Two mRNAs with Different 3’ Ends Encode Membrane and Secreted p Heavy Chains

Several models were advanced to account for the different COOHterminal segments of pill and p, chains.

1. The shorter p, chains could be generated from longer pill chains by proteolytic cleavage which removed a hydrophobic COOH-terminal sequence. 2. The p, and ps chains could be encoded by two p genes which differed only in their COOH-terminal coding sequences. 3. The pm and p, chains could be encoded by a single p gene in which the COOH-terminal coding segments were rearranged during B cell development. 4. The mRNAs coding for p,ll and p, chains could be generated from a single large nuclear RNA precursor by alternative RNA splicing or processing events.

44

JOHN ROGERS AND RANDOLPH WALL

The last alternative was found to be the correct one. Analysis of p mRNAs and the p heavy chain gene established that p, and p, chains are encoded by two distinct p mRNA species containing different COOH-terminal coding sequences (Rogers et al., 1980; Early et al., 1980; Alt et al., 1980). Various B lymphoma cells and myeloma cells were found to contain multiple p cytoplasmic RNA species, including two prominent p mRNAs at 2.7 and 2.4 kb (kilobases)' (Alt et al., 1980; Perry and Kelley, 1979; Rogers et al., 1980; Singer et al., 1980; Kemp et al., 1980). These two p mRNA species were shown to code for p membrane (p,) and secreted (p,) chains (Rogers et al., 1980; Alt et al., 1980). Cloning and sequencing of p cDNA clones made from myeloma cells containing both p mRNA species revealed that the p, and p, mRNAs were identical through the end of the C,4 constant region coding sequence but thereafter contained very different sequences coding for alternative COOH-terminal segments and 3'-untranslated region (3'-UT) sequences (Rogers et al., 1980). The ps chain sequence had a hydrophilic segment encoding 20 residues after the C,4 domain which corresponded to the p, COOH-terminal amino acid sequence determined by Kehry et al. (1979). The finding that the p, COOH-terminal segment was terminated with a stop codon eliminated the possibility that p, chains might be generated from pn,chains by proteolytic cleavage. The p, cDNA clone had a COOH-terminal segment encoding 41 residues (designated the M or membrane region) with the properties of a transmembrane protein (see Section VI). The four domains of the C, gene segment are encoded in four separate exons bounded by RNA splicing sites (Calame et al., 1980; Kawakami et al., 1980; Liu et al., 1980). The C, gene was found to be present in a single copy per haploid genome (Calame et al., 1980) and was not rearranged in cells making p, or p, chains. The first clue to the origin of the membrane COOH-terminal segment in p, mRNAs came from R-loop electron microscopy of the cloned p gene with mRNA from myeloma cells containing both p, or p, mRNAs (Rogers et al., 1980).A minor fraction of the p genes which showed R-loops of the C region domains contained another R-loop located 3' to the C,4 domain. Isolated restriction fragments from this region of the p gene Hereafter, the 2.7 and 2.4 kb p,,, and ps species will be cited as 2.35 and 2.1 kb, respectively. This reduction is based on rRNA size standards of 1.84 kb (18 S), 5.0 kb (28 S), 6.35 kb (32 S), and 11.8 kb (45 S) which were revised to conform with the known coding sequences of 16 S (1541 nucleotides) and 23 S (2904 nucleotides) E . coli rRNAs (Brosius et al., 1978, 1980). These reduced sizes are more consistent with the known nucleotide sequence lengths of p m and pqmRNA (Rogers et al., 1980).

IMMUNOGLOBULIN

RNA

45

REARRANGEMENTS

clone (called the membrane or M exon) hybridized specifically with al., 1980). The locations of the COOH-terminal segments of pmor p, mRNAs were directly established by nucleotide sequencing of the cloned p gene (Early et al., 1980). The secreted COOH-terminus coding sequence and the 3' untranslated region of psmRNA were derived from p gene sequences contiguous with 3' end of the C,4 domain. The pm cDNA clone did not include these nucleotides, but instead contained a new COOH-terminus and 3'-untranslated region encoded in two exons located 1850 base pairs 3' to the C,4 sequence. Sequence comparison of the p gene and p cDNA clones revealed that, in p, mRNA, an RNA splice removed 1850 nucleotides and joined a site in the coding sequence at the end of C,4 with a site at the beginning of the first M exon. A second RNA splice of 118 nucleotides joined the first and second M exons (Fig. 2). These RNA splices replaced the p s COOH-terminus and 3'-untranslated region sequences with the M exons and generated pmmRNA. The codons at the 3' end of the C,4 domain where the pmand p, cDNA sequences diverged have the sequences G/GTAAA encoding Gly-Lys. This sequence in the p chromosomal g e n e c a l a m e et al., 1980; Early et al., 1980) was similar to p, mRNA but not with p s mRNA in RNA gel blots (Rogers et

P

V

c1

c2

c3

c4

M1

M2

C3

M1

M2

H

I.IP

V

c1

H

C2

I=>

J-Y+-=#

3;

5

V

-\

3m

membrane mRNA

1 \ secreted mRNA

Amputated Transcript

FIG.2. The structures of complex transcription units encoding secreted and membrane heavy chain mRNAs. Active p and y heavy chain genes are depicted schematically (not to scale) showing the exons (as blocks), introns (as lines), and untranslated regions (as shaded blocks). The polyadenylated nuclear RNA precursors and alternative RNA splicing patterns for secreted and membrane mRNAs are shown below. The amputated transcript is a novel polyadenylated RNA species predicted in the text to be the product of heavy chain complex transcription units.

46

JOHN ROGERS AND RANDOLPH WALL

the consensus sequence for an “upstream” RNA splicing site, G/GTAAG (Seif et al., 1979; Rogers and Wall, 1980; Lerner et ul., 1980). The underlined GT has been universally found at the exon/ intron juncture of “upstream” splicing sites in eukaryotic genes (Breathnach and Chambon, 1981). V. All Immunoglobulin Heavy Chain Genes Have Membrane Gene Segments

Because all heavy chains contained the sequence Gly-Lys at the end of the last domain encoded by a DNA sequence, G/GTAAA or G/GTAAG, which constituted a potential site for RNA splicing into an M F n e segment, we predicted (Rogers et ul., 1980) that all other immunoglobulin heavy chain genes would also contain M gene segments. Distinct heavy chain polypeptide species presumed to b e membrane forms have been reported for y2;, (Oi et ul., 1980; Word and Kuehl, 1981), for 6 (Vassalli et ul., 1979) and recently for a chains (Kikutani et al., 1981). By analogy with M exons in the p gene, we expected to find other heavy chain M exons encoded downstream to the final C region domain. We further expected to find an M 1 exon with the coding sequences for an acidic polypeptide segment followed by a hydrophobic segment. The M I exon and the final C region domain would be bounded by RNA splice sites fitting the known consensus sequences (Seif et ul., 1979; Lerner et al., 1980; Rogers and Wall, 1980) and located in the same coding phases as in the C region domains in the p gene. We also expected to find an M 2 exon which would contain codons for a positively charged polypeptide tail, followed by a 3’ untranslated region (Fig. 2). All B cell lines screened which produced either y1, yza, Y2b, or y3 secreted heavy chains also made a minor y mRNA species with a spliced structure analogous to that of the pm mRNA (Rogers et al., 1981; Tyler et al., 1982). Mapping and sequencing of the y1, yza,Y2b, and 7 3 M gene segments (Rogers et al., 1981; Tyler et al., 1982; Yamawaki-Kataoka et al., 1982; Komaromy et al., 1983) located the M exons in C, genes as predicted and revealed an unexpectedly high level of sequence conservation in the transmembrane peptide sequences and in the nucleotide sequences of the M 1 exons (Fig. 3, see also Section VI). Membrane and secreted forms of a heavy chain mRNA have been reported (Kikutani et al., 1981; Word et al., 1983). The sequences of the M exons of the a (Word et al., 1983) and E (Ishida et al., 1982) genes have been published. B lymphocyte cells coexpressing surface IgM and IgD contained membrane 6 m m A s with spliced COOH-terminal sequences like pm mRNAs (Liu et al., 1980; Tucker et al., 1980; Mushinski et al., 1980;

IMMUNOGLOBULIN

I

u

RNA

47

REARRANGEMENTS

4 I 4 4 44

EGEVNAEEEGFENLWTTASTFIVLFLLSLFYSTTVTLFKVK

GIVNTIQHSCIMDEqSDSYMLEEENGLWPTMCTFVALFLL~LYSGFVTFIKVK

6

y3

ELELNGTCAEA~DGELDGLITITIFISLFLLSVCYSASV~FKVKWIFSSVVQVKQTAIPDYRNMIGQ~

5

GLQLDETCAEAQDGELDGLWTIT I FI SLFLLSVCYSAAVTLFKVKWIFSSVVELKQTLVPEYKNMIGQAF'

2' b

GLDLDDICAEAKDGELDGLWTTITIFISLFLLSVCYSASVTLFKVKWIFSSVVELKQKISPDYGNMIGQGA

y2a E

a

GLDLDDVCAEAQDGELDGLITITIFISLFLLSVCYSASVTLFKVKWIFSSVVELKQTISPDYRNMIGQ~ ELDLQDLCIEEVEGEELEELWTSICVFInFLLSVSYGATVTVLKVKlLSTPMQDTPqTFqDYANILQTRA ERQEPLSYVLLDQSqDILEEEAPGASLWPTTVTFLTLFLLSLFYSTALTVTTVRGPFGSKEVPQY A C I D I C S E G M E N T L T R A N S M E M B R A N E SEGMENT-I-INTRACELLULAR

SEGMENT

FIG.3. Comparison of the sequences of heavy chain M regions. The amino acid sequences shown are derived from the coding sequences of the M gene segments of p (Rogers et d.,1980; Early et al., 1980, (Cheng et nl., 1982), y , , y2.%,y21, (Rogers et a/., 1981; Tyler et al., 1982; Yamawaki-Kataoka et ul., 1982), y3 (Komaromy et al., 1983), a (Word et al., 1983), and E (Ishida et al., 1982).The arrows denote seven contact residues in the dimer model described in the text. The one-letter amino acid code is A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G , Gly; H, His; I, Ileu; K, Lys; L, Leu; M, Met; N, Asn; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

Moore et d , ,1981; Maki et ul., 1981; Cheng ct ul., 1982; Fitzmaurice ct al., 1982; Knapp et al., 1982). The M exon sequences (Cheng et al., 1982) coding for the M region of ti,,,chains also exhibited a relatively high degree of homology to other heavy chain M 1 exon sequences (Fig. 3 ) .A more detailed consideration of the organization and expression of the 6 gene is presented in Section VIII. VI. A Model for the Transmembrane Insertion of Heavy Chain M Regions

The amino acid sequences encoded by the heavy chain M 1 and M2 exons aligned in Fig. 3 are divided into three structurally distinct segments. The M regions of the all heavy chains Ilegin with negatively charged sequences of variable length. There is little evident homology between the various amino acid sequences in this segment (i-e.,the acidic segment).Only the acidic segments of the y subclasses are significantly homologous with 11 of 17 (65%) identical residues. Strikingly, all these acidic segments from different heavy chain classes retain a similar overall charge distribution with a net charge of -6, except for E . The following 26 amino acids in all cases are all uncharged and include a very hydrophobic 11-residue core, surrounded by two uncharged but hydrophilic stretches containing serine and threonine residues. This stretch of 26 uncharged amino acids (the hydrophobic

48

JOHN ROGERS AND RANDOLPH WALL

segment) in the M region is presumed to be the portion of cell surface immunoglobulin which anchors the heavy chains in the cell membrane. In constrast to the acidic segments, the 26 amino acid hydrophobic stretches of all heavy chain genes are strongly conserved (Fig. 3). The hydrophobic segment of 6 is identical to that of p in 14 of 26 amino acids (54% homology). The hydrophobic segments of the y subclasses share 16-17 amino acids (62-65% homology) with the p segment, in contrast to the conservation of only about 27% of the amino acids between p and any of the y chains in the constant region domains. The y 3 , y1, Y2b, and yza hydrophobic segments differ from each other by only a single amino acid. When analyzed for hydrophobicity according to the method of Segrest and Feldman (1974), all of these heavy chain M segments exhibited hydrophobicities typical of known transmembrane segments. Their lengths are also typical of known transmembrane peptides (Tomita and Marchesi, 1975; Nakashima and Konigsberg, 1974). The intracellular segments of all heavy chains (except a)begin with the residues Lys-Val-Lys. This short, positively charged sequence constitutes the complete intracellular segments of p and 6 heavy chains. The 71, Y2b, yza, and y3 chains have intracellular segments of 28 residues including the Lys-Val-Lys sequence. The intracellular segments of y chains are positively charged and are the most homologous (71-89%) in sequence. The a and E intracellular segments contain 14 and 28 amino acids, respectively. The conservation of sequence in the heavy chain transmembrane or hydrophobic segments was unexpected, since transmembrane segments in other genes seem to have nothing in common except length and hydrophobicity (Burnstein and Schecter, 1978; Gething et al., 1980). However, the immunoglobulin molecule on the lymphocyte surface is composed of two identical heavy chains linked together in the hinge region, suggesting that one possible function of the highly conserved transmembrane sequence might be to form a dimer within the membrane. A dimer might be favored on energetic grounds because dimerization would permit association of some of the relatively hydrophilic amino acid side chains of this segment. In collaboration with David Eisenberg (UCLA Molecular Biology Institute), we searched for a feasible dimer configuration by constructing molecular models (Rogers et al., 1981). We assumed (1) that each of the two transmembrane segments is an a-helix, since in this configuration all backbone hydrogen bonds are satisfied, and an a-helix of 26 residues (about 39 %.) can span the apolar region of the lipid bilayer Tanford, 1978); (2) that these two transmembrane helices (about 30 %.;

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

49

cross at some angle between 20 and 60°, as is common for crossing ahelices (Crick, 1953; Chothia et d., 1981) because this allows the side chains to interdigitate, and the tilted helices would still be able to span the membrane; and (3) that the two chains are related by a twofold axis perpendicular to the membrane. With the aid of these assumptions and molecular models, we were able to find a plausible structure for the paired immunoglobulin M regions (Fig. 4).In this model, the two a-helices cross at an angle of 60", with seven conserved residues from each chain (indicated on Fig. 3) forming contacts. These include potential hydrogen bonds between two pairs of

LIPID BILAYER

C-terminus

C-kn;lhx6

FIG.4. A model for the insertion of paired immunoglobulin heavy chains through the lymphocyte cell membrane. This scale diagram shows an edge-on view of the cell nienrbrane in which the transniemhrane segments ol' the nreinbrane-l,oiiiid irnmunoglohulin niolecule are envisaged ;is a pair of a-helices crossing at an aiigle of -60". T h e nearer of the two a-helices (bold outline) has heen cut away to show the contact residues. T h e proposed contact resithies are sketched, and irre Id~elcdwith the one-letter code such that the lal,els lie on the a-carbons. All seven contact residues are conscrved in p, y l , y , ,yz., and y21,chains. Six ol'the seven contact residues are identical in 6 chains with ir conservative serine to threonine change at the altered residue. Ilytlrogen 1)oiids can form a m o n g the forrr serines and between the tyrosine-tlireoniiie pairs.

50

JOHN ROGERS AND RANDOLPH WALL

serines, and between a threonine and tyrosine from each chain. There are additional contacts between a pair of valine residues and between two pairs of phenylalanine residues; these bulky residues delimit the contact region on the COOH-terminal and NH2-terminal sides. These residues in potential contact are very highly conserved in all the heavy chain transmembrane segments. Several threonine and serine residues do not make contact, but some of these are probably near enough to the membrane surface to hydrogen-bond to water, and others may hydrogen-bond to the backbone carbonyl groups (Engleman and Steitz, 1981). The noncontact residues in the y chains also include a pair of cysteines, which would not be in a position to form a disulfide bond. However, there is another pair of cysteines in the extracellular, acidic part of the y M segments, and this may form a disulfide bridge between the two chains. The complete M region serves both as a membrane anchor and as a transmembrane signaling device to stimulate B cell proliferation and differentiation upon binding of antigen. In the p and 6 M sequences, the intracellular segments are composed of only three residues which are very basic (Fig. 3). The intracellular segments of the y M regions contain these three amino acids and extend for an additional 25 residues. Tyler et aZ. (1982) have postulated that the additional intracellular residues on membrane IgG may generate a different cell signal on binding antigen than that sent through receptor IgM or IgD. However, the mechanism by which antigen binding is transmitted through surface immunoglobulin receptors remains to be established. VII. Membrane and Secreted Heavy Chain mRNAs Are Coded by Complex Transcription Units

Most eukaryotic genes now studied (including immunoglobulin light chain genes; Wall et d., 1980; Perry, 1981)are simple transcription units which code for a single mRNA (reviewed in Darnell, 1979, 1982). Simple transcription units contain a single poly(A) addition site, and the pattern of RNA splicing is invariant except in unusual cases where alterations in splicing sites generate aberrant splicing patterns (Seidman and Leder, 1980; Choi et aZ., 1980).Complex transcription units encode multiple mRNA species with different 5' or 3' ends which are generated though the use of multiple poly(A) addition sites or alternative RNA splicing pathways (Darnell, 1979, 1982). Complex transcription units are well documented in adenovirus, SV40, polyoma virus (reviewed in Darnell, 1979; Ziff, 1980), and retroviruses (reviewed in Temin, 1981, 1982). The presence of complex

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

51

transcription units in viral genomes is consistent with an evolutionary adaption for obtaining maximum expression of a limited amount of coding sequence. The mapping of membrane and secreted mRNAs established immunoglobulin heavy chain genes as the first-known nonviral complex transcription units. Immunoglobulin heavy chain transcription units contain multiple poly(A) addition sites (reviewed in Word et al., 1983).Nuclear RNA precursors which have poly(A) added at the first site yield secreted heavy chain mRNAs. Nuclear RNA precursors which have poly(A) added at a site following the M exon coding sequences undergo additional RNA splices which generate membrane heavy chain mRNAs. Because poly(A) addition is a very rapid processing event which precedes RNA splicing (reviewed in Darnell, 1979, 1982), we have proposed that the processing pathways leading to secreted or membrane heavy chain mRNAs are determined by the choice of poly(A) addition sites (Rogers et ul., 1980; Early et al., 1980; Wall et al., 1980). After the addition of poly(A), RNA splicing acts on all available exons flanked by complete splicing sites (Fig. 2). Several laboratories have attempted to define the primary transcripts and the nuclear RNA processing pathways for secreted and membrane heavy chain mRNA species (reviewed in Wall and Kuehl, 1983). Large nuclear RNA species with sizes expected for the primary transcripts of the p heavy chain genes have been observed. However, because of the low levels of the presumptive primary transcripts and difficulties in reproducibly detecting splicing intermediates in these studies, the differential processing pathways leading to membrane and secreted heavy chain mRNAs have not been resolved. Accordingly, we searched for an alternative way of confirming our proposal that membrane and secreted heavy chain mRNAs are processed from the transcripts of a complex transcription unit. As a general rule, it appears that poly(A) addition occurs at points of cleavage in nuclear RNA molecules rather than through termination of transcription (Nevins and Darnell, 1978; Hofer and Darnell, 1981). Indeed, transcription apparently continues some distance past poly(A) addition sites (several kb) before cleavage of the nascent transcript occurs (Nevins and Darnell, 1978; Hofer and Darnell, 1981). Assuming that poly(A) addition occurs at points of cleavage in immunoglobulin heavy chain gene transcripts, we presumed that transcription past the poly(A) addition site for secreted mRNA precursors would produce a novel polyadenylated species which we have called the “amputated transcript” (Fig. 2). This predicted RNA species should have

52

JOHN ROGERS AND RANDOLPH WALL

the following properties. First, it should contain intron sequences between the secreted poly(A) site and the membrane exons. Second, since the M 1 and M2 exons are separated by complete splicing signals in the amputated transcript, it is likely that this second intron would be spliced out (Fig. 2). Finally, the amputated transcript should have a 5' terminus beginning in the sequence following the secreted poly(A) addition site and should not contain any heavy chain exons (V or C regions) preceding the secreted COOH-terminus poly(A) addition site. Discrete nuclear RNA species with precisely these predicted properties have now been detected in both p and y2b producing cells (Rogers and Wall, 1984). Experimental confirmation of the amputated transcript as predicted establishes that heavy chain genes are complex transcription units and reaffirms that poly(A) is added by a mechanism involving RNA cleavage. The aaputated transcript is a polyadenylated RNA molecule which appears to be preferentially retained in the nucleus, possibly because it contains an intron which cannot be removed by RNA splicing. VIII. Coexpressed p and 6 mRNAs Are Coded by a Very Complex Transcription Unit

After displaying membrane IgM, most B lymphocytes coexpress membrane-bound IgM and IgD. A number of experimental findings strongly suggested that the coexpressed p and 6 heavy chains in surface IgM and IgD might be generated from a single large complex transcription unit. Coexpressed p and 6 chains appeared to have the same variable region (Fu e t al., 1975; Goding and Layton, 1976; Stern and McConnell, 1976). Furthermore, studies on allotypic markers showed that coexpressed p and 6 chains exhibited allelic exclusion and are encoded on the same chromosome (Herzenberg et al., 1977; Bourgois et al., 1977). That p and 6 were coexpressed from a single complex transcription unit was also made plausible by studies of mouse sperm DNA clones containing the C, gene, which revealed that the Cs gene is separated from the p M exons by only approximately 2 kb (kilobases) of DNA (Liu et al., 1980; Moore et al., 1981; Mushinski et al., 1980; Tucker et al., 1980; Maki et al., 1981). This model for p plus 6 expression strongly predicted that the 6 gene would not be rearranged in lymphocytes simultaneously producing both chains. Southern blot analysis of DNA from p plus 6 producing mouse lymphomas revealed that the Cs gene was in the germ-line configuration (Moore et al., 1981; Maki et al., 1981; Knapp et al.,

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

53

1982). Restriction enzyme digestion mapping of C, and Cs genomic clones isolated from a DNA library of spleenic p+6+ B cells also gave no evidence for DNA rearrangement in the region between the C, and Cg genes (Moore et al., 1981). Analysis of genomic DNA cloned from two different p+6+ lymphoma cells lines revealed, in each case, a single rearranged V region located 5‘ to the C, gene segment (Maki et al., 1981; Knapp et al., 1982). These data clearly indicated that IgD expression in p+6+ B cells did not involve a VH to Ca DNA switch rearrangement. Thus, the simultaneous expression of C, and Cgwith a single VH region appeared to be mediated by alternative routes of RNA processing of a very large primary nuclear transcript which contained the VH, C,, and C6gene segments (Fig. 5). The Cg gene is organized differently from other heavy chain genes in that it contains an extended hinge and lacks an internal CH2domain (Tucker et al., 1980). Unlike all other heavy chain genes where the secreted C-terminal sequences are continuous with the final C region domain, the 6 gene has both its secreted and membrane exons in separate non-contiguous gene segments 3’ to the Cg3domain (Tucker et al., 1980; Moore et al., 1981; Maki et al., 1981).Mouse B lymphoma cells coexpressing membrane p and 6 chains contain both p, and p, mRNA as well as a predominant 6 mRNA species reported to be 2.72.9 kb, and a minor 6 inRNA at 1.8-2.1 kb (Moore et al., 1981; Mushinski et al., 1980; Maki et al., 1982).2Both these mRNA species code for membrane chains and contain M exon sequences located 6.5 kb to the 3‘ side of the Cg3 domain (Maki et al., 1981; Cheng et al., 1982). These two species apparently terminate at different poly(A) sites and contain identical coding sequences but different 3’-UT. An analogous situation has been reported for two mRNA species for dihydrofolate reductase (Setzer et al., 1980). There is no evidence for functional differences between these two 6, mRNAs. IgD-secreting myeloma cells contain a single prominent 6 mRNA species reported to be about 1.6 kb2 (Moore et al., 1981; Maki et al., 1981; Mushinski et al., 1980; Fitzmaurice et al., 1982) which is presumed to be 6, mRNA. This 6, mRNA contains 3’ sequences which are located 4.7 kb from the Cs3 domain (Cheng et al., 1982). Those rare instances of IgD secretion by myeloma cells in mice do not occur through RNA splicing, but rather appear to result from a DNA rearrangement which brings the rearranged VDJH region into proximity with the Cs gene with the deletion of the C, gene segments (Moore et Based on our revised RNA size standard, these 6 , mRNA species are estimated to be

2.35 and 1.6 kb. The 6, mRNA on this basis is calculated to be 1.4 kb.

54

JOHN ROGERS AND RANDOLPH WALL

B 5'

v

3'

FIG.5. IgD expression occurs through two different mechanisms. (A) The coexpression o f p and 6 mRNAs in membrane IgM and IgD involves alternative RNA processing mechanisms acting on the transcripts of the p 6 complex transcription unit (not shown to scale). The polyadenylated nuclear RNA precursors and the differential RNA splicing events which generate p and 6 mRNAs bearing the same V region are shown below the schematic diagram of the large p + 6 complex transcription unit (Z denotes secreted COOH-terminal gene segments). The structure of the p gene segments is based on Early et al. (1980). The structure of the 6 gene segments is adapted from Tucker et al. (1980), Moore et al. (1980), Maki et al. (1981) and Cheng et al. (1982). The exon structure of the M gene segment has not been fully resolved in relation to the spliced structures of 6m mRNAs. (B) The transcription unit coding for secreted 6 chains in secreted IgD is generated by a DNA rearrangement which deletes the C, gene segments and brings the V region into proximity to the Cs gene segments (Moore et al., 1981; Maki et al., 1981). The RNA splicing events which generate secreted 6 mRNA are depicted below the rearranged 6 transcription unit.

+

al., 1981; Maki et al., 1981).Furthermore, the 6, mRNA is not detect-

able in B lymphomas making p,, p,,,, and a,, mRNA species. These findings indicate that two different mechanisms are employed in 6 gene expression and the production of IgD. RNA rearrangements in B cells produce membrane 6 mRNAs along with p mRNAs from a single large complex transcription unit, whereas a rare DNA rearrangement in plasma cells produces secreted 6 mRNA (Fig. 5).

IMMUNOGLOBULIN

RNA

REARRAh’GEMENTS

55

+

The p 6 complex transcription unit (Fig. 5) is approximately 25 kb long, with five known alternative polyadenylation sites and more than a dozen potential exons for splicing. This is almost as complicated as the adenovirus late transcription unit, which is 32 kb long with five polyadenylation sites and numerous exons for splicing viral mRNAs. As in heavy chain transcription units for membrane and secreted mRNAs, the choice of poly(A) sites in the p + 6 transcription unit presumably determines the mRNA to be produced. While RNA splicing operates on all available introns in making membrane and secreted heavy chain mRNAs, selective RNA splicing must occur in making 6 mRNA. The splicing enzymes apparently skip over RNA splice sites in the C, exons when splicing the V H region to the Cs exons in making 6 mRNAs. We have repeatedly screened nuclear RNA blots from IgM and IgD producing lymphoma cell lines for large precursors to p and 6 mRNA. We have not detected any p or 6 specific nuclear RNA species approaching the 25 kb size estimated for the p + 6 primary transcript. This is presumably not due to nuclear RNA degradation during isolation as the isolated 32 S and 45 S pre-rRNA species were intact. A complete primary transcript for the major late adenovirus transcription unit was never detected directly. Instead confirmation of this major transcription unit was obtained by indirect means including pulse-labeling studies and UV transcript mapping (reviewed in Darnell, 1979). Such indirect means appear likely to be required for formally confirming the large p 6 complex transcription unit. How exons are chosen for splicing remains a mystery. A model for RNA splicing has been proposed in which U-1, a ubiquitous small nuclear RNA in eukaryotic cells, basepairs with both ends of an intron and aligns them precisely in register for cutting and splicing (Rogers and Wall, 1980; Lerner et al., 1980). Yang et al. (1981) have obtained preliminary evidence in support of this model. However, this model does not explain how the correct pairs of splice sites are distinguished from incorrect pairs in selective RNA splicing. This problem is particularly evident in the p 6 complex transcription unit where an exon in one splicing pathway is an intron in another (see Fig. 5). It seems plausible that the secondary structure of the nuclear RNA molecule might direct the course of RNA splicing, but this hypothesis is not yet amenable to experimental testing.

+

+

IX. Summary: Developmental Regulation of Heavy Chain Gene Expression

The discovery of the membrane gene segment (containing the M exons) established the p heavy chain gene as the first example of a

56

JOHN ROGERS AND RANDOLPH WALL

complex transcription unit in chromosomal DNA. This complex transcription unit produced two p mRNAs with different 3’ spliced structures coding for secreted and membrane p chains. The secreted COOH-terminal sequence is encoded contiguous to the C,4 domain while the membrane COOH-terminal sequence is encoded in two separate exons located 3’ to the C,4 domain. Exactly the same structure was predicted and found in the other immunoglobulin heavy chain genes. The 6 heavy chain gene has a similar structure except that both the secreted (6,) and membrane ( 6 , ) COOH-terminal sequences are encoded separately from the C6 region exons. The complex transcription unit model for production of membrane and secreted heavy chains was expanded to account for the coexpression of p and 6 chains in IgM an IgD on the lymphocyte cell surface. The expression of the complex transcription unit encoding p and 6 mRNAs is developmentally regulated by the choice of multiple polyadenylation sites and by the selective recognition and use of RNA splicing sites. Thus, posttranscriptional processing mechanisms play a major role in regulating the changes in p and 6 mRNA species seen in early B cells making surface IgM, in later B cells coexpressing surface IgM and IgD, and in myeloma cells actively secreting IgM. Complex transcription units and posttranscriptional control of chromosomal genes are not unique to immunoglobulin heavy chain genes, Two recent publications indicate that the cellular genes for yeast invertase (Carlson and Botstein, 1982) and rat calcitonin (Anasa et al., 1982) are complex transcription units encoding functionally distinct enzyme forms and tissue-specific hormones respectively. These findings clearly establish the importance of posttranscriptional RNA processing mechanisms in regulation of genes in eukaryotic cells. ACKNOWLEDGMENTS

Our work in these studies has been supported by grants from the National Institute of Health (CA12800, AI13410) and the National Science Foundation (PCM 79-24876). We gratefully acknowledge the assistance and advice of M. Komaromy, J. Kobayashi, and M. O’Connor in the preparation of this manuscript. We thank 0.Witte for his comments on reading the manuscript.

REFERENCES Alt, F. B., Bothwell, A. L. M., Knapp, M., Siden, E., Mather, E., Koshland, M., and Baltimore, D. (1980). Cell 20, 293-301. Amara, S., Jonas, V., Rosenfeld, M. G., Ong, E. S., and Evans, R. M. (1982). Nature (London)(in press). Bergman, Y., and Haimovitch, J. (1978). Eur.J. Zmmunol. 8, 876-880. Bourgois, A., Kitajima, K. Hunter, I. R., and Askonas, B. A. (1977). Eur. J. Zmmunol. 7, 151-156.

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

57

Breathnach, R., and Chambon, P. (1981). Annu. Rev. Biochem. 50, 349-383. Burnstein, Y., and Schecter, I. (1978). Biochemistry 17,2392-2400. Burrows, P., Le Jeune, M., and Kearney, J. F. (1979). Nature (London) 280, 838-840. Calame, K. Rogers, J., Early, P., Davis, M., Livant, P., Wall, R., and Hood, L. (1980). Nature (London)284,452-455. Carlson, M., and Botstein, D. (1982). Cell 28, 145-154. Choi, E., Kuehl, M., and Wall, R. (1980). Nature (London) 286, 776-779. Chothia, C., Levitt, M., and Richardson, D. (1981)./. M o l . H i d . 145, 215-250. Coffman, R. L., and Cohn, M. J. (1977).J. Zmmunol. 118, 1806-1815. Crick, F. H. C. (1953). Acta Crystallogr. 6, 689-697. Darnell, J. E. (1979). Prog. Nuc. Acids Res. Molec. Biol. 22, 327-353. Darnell, J. E. (1982). Nature (London) 297,365-371. Della Corte, E., and Parkhouse, R. M. E. (1973). Bi0chem.J. 136, 597-604. Early, P. W., Davis, M. M., Kaback, D. B., Davidson, N., and Hood, L. (1979).Proc. Natl. Acad. Sci. U.S.A. 76, 857-861. Early, P., Rogers, J., Davis, M., Calame, K., Bond, M., Wall, R., and Hood, L. (1980).Cell 20,313-319. Engleman, D. M., and Steitz, T. A. (1981). Cell 23, 411-422. Fitzmaiirice, L., Owens, J., Blattner, F. R., Cheng, Ii., Tucker, P. W., and Mushin\ki, J . F. (1982). Nature (London)296, 459-462. Gething, M. J., Bye, J., Skehel, J., and Waterfield, M. (1980).Nature (London)287,301306. Goding, J. W. (1978). Cont. Topics Zmmunobiol. 8, 203-243. Herzenberg, L. A., Herzenberg, L. A.; Block, S. J., Loken, M . R., Okumura, K., Van der Loo, W., Osborne, B. A., Hewgill, D., Coding, J., Gutman, G., and Warner, N. L. (1977). Cold Spring Harbor Symp. Quant. Biol. 41,33-45. Hofer, E., and Darnell, J. E. (1981). Cell 23, 585-593. Honjo, T. (1983). Ann. Reo. Zmmunol. 1, 499-528. Honjo, T., Obata, M., Yamawaki-Kataoka, Y., Kataoka, T., Kawakami, T., Takahashi, N., and Mano, Y. (1979). Cell 18, 559-568. Hood, L., Davis, M., Early, P., Calame, K., Kim, S., Crews, S., and Huang, H. (1980). Cold Spring Harbor Symp. Quant. Biol. 45, 887-898. Ishida, N., Ueda, S., Hayashida, H., Miyata, T., and Honjo, T. (1982).EMBOJ. 1, 11171123. Kawakami, T., Takahashi, N., and Honjo, T. (1980). Nucleic Acids. Res. 8, 3933-3938. Kehry, M., Sibley, C., Fuhrman, J., Schilling, J., and Hood, L. E. (1979). Proc. Natl. Acad. Sci. U.S.A. 76,2932-2936. Kehry, M., Ewald, S., Douglas, R., Sibley, C., Raschke, W., Fanibrough, D., and Hood, L. (1980).Cell 21, 393-406. Kemp, D. J., Harris, A. W., and Adams, J. M. (1980). Proc. Natl. Acad. Sci. U.S.A.77, 7400-7474. Kikutani, H., Sitia, R., Good, R. A,, and Stavnezer, J. (1981).Proc. Natl. Acad. Sci. U.S.A. 78,6436-6440. Knapp, M. R., Liu, C., Newell, N., Ward, R. B., Tucker, P. W., Strober, S., and Blattner, F. (1982). Proc. Natl. Acad. Sci. U.S.A. 79,2996-3000. Komaromy, M., Clayton, L., Rogers, J., Robertson, S., Kettman, J., and Wall, R. (1983). Nucleic Acids Res. (in press). Koshland, M. E. (1975). Adv. Zmmunol. 20,41-69. Lerner, M. R., Boyle, J. A., Mount, S. M., Wolin, S. L., and Steitz, J. A. (1980). Nature (London) 283,220-224.

58

JOHN ROGERS AND RANDOLPH WALL

Levitt, D., and Cooper, M. D. (1980). Cell 19,617-625. Liu, C., Tucker, P. W., Mushinski, J. F., and Blattner, F. R. (1980). Science 209, 13481352. McHugh, Y., Yagi, M., and Koshland, M. E. (1981).In “B Lymphocytes in the Immune Response: Functional, Developmental, and Interactive Properties” (N. Klinman, D. Mosier, I. Sher, and E. Vitetta, eds.), pp. 467-474. Elsevier/North-Holland, Amsterdam. Maki, R., Roeder, W., Traunecker, A., Sidman, C., Wabi, M., Raschke, W., and Tonegawa, S. (1981). Cell 24, 353-365 Mather, E. L., Alt, F. W., Bothwell, A. L. M., Baltimore, D., and Koshland, M.E. (1981). Cell, 23. 269-378. Melchers, F., and Anderson, J. (1974) 4, 181-188. Melchers, U., Eidels, L., and Uhr, J. W. (1975). Nature (London) 258,434-435. Melchers, U., and Uhr, J. W. (1973).J . E x p . Med. 138, 1282-1287. Moore, K. W., Rogers, J., Hunkapillar, T., Early, P., Nottenburg, C., Weissman, I., Bazin, H., Wall, R., and Hood, L. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 1800-1804. Mushinski, J. F., Blattner, F. R., Owens, J. D., Finkelman, F. D., Kessler, S. W., Fitzmaurice, L., Potter, M., and Tucker, P. W. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 7405-7409. Nakashima, Y., and Konigsberg, W. (1974).J. Mol. Biol. 88, 598-600. Nevins, J. R., and Darnell, J. E. (1978).Cell 15, 1477-1493. Oi, V. T., Bryan, V. M., Herzenberg, L. A., and Herzenberg, L. A. (1980).J. E x p . Med. 151, 1260-1274. Okumura, K., Julius, M. H., Tsu, T., Herzenberg, L. A., and Herzenberg, L. A. (1976). E u r . J. Immunol. 6, 467-472. Parkhouse, R. M. E., Lifter, J., and Choi, Y. S. (1980). Nature (London) 284, 280-281. Perry, R. P. (1981).J. Cell Biol. 91, 28s-38s. Perry, R. P., and Kelley, D. E. (1979). Cell 18, 1333. Raschke, W. C. (1978). Curr. T o p i c s Microbiol. Zrnrnunol. 81, 70-76. Raschke, W. C., Mather, E. L., and Koshland, M. E. (1979).Proc. Natl. Acad. Sci. U.S.A. 76,3469-3743. Reyes, A. A., Schold, S., Itakura, K., and Wallace, R. B. (1982). Proc. Natl. Acad. Sci. U.S.A.79, 3270-3274. Robinson, E. A., and Appella, E. (1980). Proc. Natl. Acad. Sci. U.S.A. 77,4909-4913. Rogers, J., and Wall, R. (1980). Proc. Natl. Acud. Sci. U.S.A.77, 1877-1879. Rogers, J,, and Wall, R. (1984). In preparation. Rogers, J., Early, P., Carter, C., Calame, K., Bond, M., Hood, L., and Wall, R. (1980).Cell 20,303-312. Rogers, J,, Choi, E., Souza, L., Carter, C., Word, C., Kuehl, M., Eisenberg, D., and Wall, R. (1981). Cell 26, 19-27. Sakano, H., Rogers, J. H., Huppi, K., Brack, C., Traunecker, A., Maki, R., Wall, R., and Tonegawa, S. (1979). Nature (London) 277,627-633. Schibler, U., Marcu, R. B., and Perry, R. P. (1978). Cell 15, 1495-1509. Segrest, J. P., and Feldman, R. J. (1974).J. Mol. Biol. 87, 853-858. Seidman, J. G., and Leder, P. (1980). Nature (London) 286,779-783. Seif, I., Khoury, G., and Dhar, R. (1979). Nucleic Acids Res. 6,3387-3398. Setzer, D., McGrogan, M., Nunberg, J. H., and Schimke, R. T. (1980).Cell 22,361-370. Sibley, C. H., Ewald, S. J., Kehry, M. R., Douglas, R. H., Raschke, W. C., and Hood, L. E. (1980).J. Immunol. 125,2097-2105. Siden, E., Baltimore, D., Clark, D., and Rosenberg, N. E. (1979). Cell 16, 389-396.

IMMUNOGLOBULIN

RNA

REARRANGEMENTS

59

Singer, P. A., Singer, H. H., and Williamson, A. R. (1980). Nature (London)285, 294300. Temin, H. M. (1981). Cell 27, 1-3. Temin, H. M. (1982).Cell 28, 3-5. Tomita, M., and Marchesi, V. T. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 2964-2968. Tucker, P. W., Liu, C., Mushinski, J. F., and Blather, F. R. (1980). Science 209, 13531360. Tucker, P. W., Slightom, J. L., and Hlottnc.r, F. 11. (1981). Proc. N u t / . Actid. Sri. II.S.A. 78,7684-7688. Tyler, B. M., Cowman, A. F., Gerondakis, S. D., Adams, J. M., and Bernard, 0. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2008-2012. Vassdi, P., Tedghi, K., Lisowski-Hcriisteiit, B., l‘artakof’f; A., and Jaton, J. C. (1079). Proc. Natl. Acutl. Sci. U.S.A. 76, 5515-5519. Vitetta, E. S., Baur, S., and Uhr, J. W. (1971).]. Exp. Med. 134, 242-264. Wabl, M. R., Johnson, J. P., Haas, I. G., Tenkhoff, M., Meo, T., and Inan, R. (1980).Proc. Natl. Acad. Sci. U.S.A. 77, 6793-6796. Wall, R., and Kuehl, M. (1983). Annu. Reo. Zmmunol. 1, 393-422. Wall, R., Choi, E., Carter, C., Kuehl, M., and Rogers, J. (1980). Cold Spring Harbor Symp. Quant. Biol. 45, 879-885. Warner, N. L., Leary, J. F., and McLaughlin, S. (1979). In “B Lymphocytes in the Immune Response” (M. Cooper et al., eds.), pp. 371-378. ElseviedNorth Holland, Amsterdam. Whitlock, C., and Witte, 0. N. (1982). Proc. Natl. Acad. Sci. U.S.A. 79,3608-3612. Williams, P. B., Kubo, R. T., and Grey, H. M. (1978).]. Immunol. 121,2435-2439. Word, C. J., and Kuehl, M. W. (1981). Molec. Zmmunol. 18, 311-322. Word, C. J., Mushinski, J. F., and Tucker, P. W. (1983). EMBO]. 2, 887-898. Yamawaki-Kataoka, Y., Kataoka, T., Takahashi, N., Obata, M., and Honjo, T. (1980). Nature (London) 283, 786-789. Yamawaki-Kataoka, Y., Miyata, T., and Honjo, T. (1981). Nucleic Acids Res. 9, 13651381. Yamawaki-Kataoka, Y., Nakai, S., Miyata, T., and Honjo, T. (1982).Proc. Natl. Acad. Sci. U.S.A. 79, 2623-2627. Yuan, D., Uhr, J. W., and Vitetta, E. S. (1980).]. Immunol. 125, 40-46. Ziff, E. B. (1980). Nature (London) 287,491-499.

NOTEAI)I)EI) I N PHOOP.The p nriclc:tr IiNAs, including the “antputatetl transcript,” have also been ex;uninecl by N ~ I S O IrtI ul. ( M o l c c . Cell. H i o l . 3, 1317- 1332, 1983). Howc~cv,there are too niany R N A specic.s prc+xnt to prwnit iin:iml)igtious resolution of the various possiblr processing pathways

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 35

Structure and Function of Fc Receptors for IgE on Lymphocytes, Monocytes, and M acrophages’ HANS L. SPIEGELBERG Department of Immunology, Research Institute of Scripps Clinic, La Jolla, California

..................... unians..

.................

IX. FcER+Rat and Mouse ,M4.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. FcsR+ Eosinophils.. . . . . . . . . . . . ............................. XI. Conclusions: Induction and Function of F ~ E R on Lymphocytes and M+ . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

65

69

79 80 81 85

1. Introduction

After Ishizaka et ul. (1966) discovered IgE as the class of immunoglobulins (Ig) responsible for most allergic diseases (atopic diseases2), the interaction of IgE with basophils and mast cells was studied extensively in many laboratories. This research established that the Fc fragment of IgE binds with high affinity to specific cell surface receptors (FcsR) on basophils and mast cells (Ishizaka et ul., 1970; Kulczycki and Metzger, 1974; K. Ishizaka and T. Ishizaka, 1978) and that cross-linking of FcsR by antigen-IgE antibody complexes (Ishizaka and Ishizaka, 1968; Dembo et ul., 1978), aggregated IgE (KageySobotka et al., 1981), or anti-IgE receptor antibodies (T. Ishizaka and

’

To my collaborator and friend, Dr. B. G. Fishkin, in gratitude for his help and encouragement with respect to my research over the last 15 years. He provided me with sera from over 1500 patients with multiple myeloma, including serum from the third known patient with IgE myeloma (Fishkin et al., 1972), whose IgE myeloma protein was used to detect binding of IgE to lymphocytes for the first time. The term “atopic” is used meaning presumed to be caused by IgE antibodies. PBL, peripheral blood lymphocytes; MI$, macrophage; FcER, FcyR, Fc@R,FcaR, Fc receptors for IgE, IgG, IgM, IgA, respectively; TE, Ty, T cells bearing Fc receptors for IgE, IgG; ADCC, antibody-dependent cellular cytotoxicity; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BDB, bis-diazotized benzidine; mIg, membrane-bound cell surface immunoglobulin.

61 Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-022405-6

62

HANS L. SPIEGELBERG

K. Ishizaka, 1978) induces histamine release from these cells. For some time it was believed that IgE interacts only with FcsR on basophils and mast cells, and little effort was directed toward investigating binding of IgE to receptors on other cell types. Therefore, immunologists were somewhat surprised when Lawrence et al. (1975) found that radiolabeled aggregated IgE binds to human lymphocytes and Capron et al. (1975)described an IgE-dependent killing of schistosomules by rat macrophages (M4), two observations which suggested the presence of FcsR on lymphocytes and M+. Subsequent research by Dr. Capron, us, and several other researchers on the interaction of IgE with lymphocytes and MC#Jestablished that these cells carry specific IgE Fc receptors, albeit of a different nature than those on basophils and mast cells. Whereas FcsR on basophils and mast cells bind IgE with a high affinity, FcsR on lymphocytes and M 4 bind IgE with a low affinity, and, in particular, monomeric IgE rapidly dissociates from the FcsR of these cells, which is one of the reasons that FcsR on lymphocytes and M 4 were not detected earlier. Another reason is that the expression of FcsR on lymphocytes and M+ depends on the activity of the IgE immune system. Patients with atopic disorders have higher percentages of FcsR+ lymphocytes (Spiegelberg et al., 1979) and monocytes (Melewicz et al., 1982a) than nonatopic humans, and infection of rats with the IgE-inducing parasite Nippostrongylus brusiliensis (Jarrett and Bazin, 1974) results in a dramatic increase of FcsR+ lymphocytes in the mesenteric lymph nodes (Yodoi and Ishizaka, 1979a). Another cell type on which FcsR were not easily demonstrated in the past is the eosinophil, presumably because such cells also carry low-affinity FcsR whose expression depends on IgE formation (Capron et al., 1981).However, by employing methods involving multiple IgE-FcER interactions, such as rosette formation with IgEcoated erythrocytes (Gonzalez-Molina and Spiegelberg, 1977), cells expressing low-affinity FcsR can be detected relatively easily. Information thus obtained regarding the nature and function of FcsR+ lymphocytes and M+ has led to new approaches toward understanding the regulation of IgE formation, IgE-induced release of inflammatory mediators from M+, and IgE-mediated defense mechanisms. II. Affinity for IgE and Structure of FcsR on Lymphocytes and M$

The binding properties of radiolabeled IgE to FcsR+ human cultured lymphoblastoid cells (Ishizaka and Levy, 1980; Spiegelberg and Melewicz, 1980; Melewicz et al., 1982a), mouse lymphocytes (Vander-Mallie et al., 1982), cultured human U937 MC#J(Anderson and

Fc

RECEPTORS FOR

IgE

63

Spiegelberg, 1981), and rat MC#J(Capron et al., 1977; Dessaint et al., 1979a; Finbloom and Metzger, 1982) have been reported. Ishizaka and Levy (1980) determined an association constant K1 of 3.5 x lo4 M-I sec-' , a dissociation constant K 1 of M-' sec-', and an association equilibrium constant K , of 6.3 x lop7M-' for binding of monomeric IgE to 2.4-3 x lo5 binding sites on cultured human RPMI-8866 cells. Similar values were obtained in our laboratory; in particular, we noted that IgE rapidly dissociates from FceR+ cultured lymphocytes (Spiegelbergand Melewicz, 1980) and U937 M$I (Anderson and Spiegelberg, 1981). Vander-Mallie et al. (1982) studied binding of IgE to lymphocytes from normal and N . brusiliensis-infected mice. They detected only 5.1 X lo3 IgE-binding sites per IgE-rosetting lymphocyte from normal BALB/c mouse spleens, and these cells bound with a K , of 0.95 x 108 M-l. In contrast, lymphocytes from N . brusiliensis-infected mice bound much more IgE per FcsR+ cell, indicating a larger number of FcsR. However, the number of receptors could not be estimated accurately, because binding was not saturable, and the lymphocytes from the infected mice bound IgE with a much lower affinity. Therefore, the authors concluded that FcsR on mouse lymphocytes are heterogeneous and that N . brusiliensis-induced F ~ E R bind IgE with a lower affinity than FcsR on lymphocytes of normal mice. Melewicz et al. (1982a), studying binding of IgE to human lymphoblastoid cells and U937 cultured M+, did not obtain straight Scatchard plots even after prolonged incubation of the cells when equilibrium was reached between bound and unbound IgE. This finding could also be interpreted as indicating FcsR heterogeneity. However, until more information becomes available about the structure of F ~ E R on lymphocytes, the binding studies must be interpreted with caution; the high dissociation rate as well as technical difficulties may be responsible for the presumed structural heterogeneity of FcsR on lymphocytes and monocytes. Whatever the explanation for the somewhat unusual binding properties of IgE to lymphocytes, data from many laboratories indicate that the binding of IgE to lymphocytes is of much lower affinity (K,, lo7 M - ' ) than that of IgE to mast cells and basophils. The K , values reported for rat (Kulczycki and Metzger, 1974; Conrad et al., 1975) and mouse (Sterk and Ishizaka, 1982) basophils and mast cells are on the order of lo9 to 1O'O A4-I. Furthermore, IgE dissociates extremely slowly from mast cells and basophils with K l values on the order of 1-2 x lo5 M-l sec-'. Binding of radiolabeled IgE to MC#Jwas first demonstrated by Capron et ul. (1977)and Dessaint et ul. (1979a), who showed transiency of binding at 3TC, with a maximum at 10 minutes, after which IgE

-

64

HANS L. SPIEGELBERG

apparently dissociated from the cells. This finding is unexplained; perhaps aggregates in the radiolabeled IgE preparations caused capping and shedding of FceR. When Finbloom and Metzger (1982)reinvestigated binding of radiolabeled IgE to rat M+, they could not accurately measure binding of monomeric IgE to rat M+ because of the rapid dissociation rate. Nevertheless, they estimated a K , for monomeric IgE on the order of lo7 M-' and for dimeric IgE loxM-' to 4-5 x los binding sites per cell. These values are similar to those obtained by Anderson and Spiegelberg (1981) for the binding of human IgE to cultured U937 M+, which bound monomeric IgE with a K , of 4 X lo7 M-' to 5-9 x lo5binding sites per cell. At present it is not possible to determine whether the binding characteristics of monomeric IgE to FCERon M+ are identical or different from those on lymphocytes. As mentioned previously, the rapid dissociation rate does not allow accurate measurements of the association equilibrium constant K , or the number of binding sites per cell. The precise structure of the FceR on lymphocytes and M+ is still unknown. Treatment of the cells with trypsin (Gonzalez-Molina and Spiegelberg, 1977; Meinke et al., 1978; Melewicz and Spiegelberg, 1980) abolishes IgE rosette formation. In contrast, FceR on basophilic leukemia cells are trypsin resistant (Metzger et al., 1976), suggesting that FceR differ on individual cell types. Meinke et al. (1978) used NP-40 lysed 12sIsurface-labeled WiI-2WT and RPMI-8866 cells and studied peptides from them that bound to IgE-immunoadsorbent columns. When analyzed by SDS-PAGE, three radiolabeled peptides with molecular weights (MW) of 86,000, 47,000, and 23,000 were observed. The 47,000 MW protein was found most consistently; the 86,000 band was more prominent with RPMI-8866 cells than Wil2WT cells. The 23,000 MW peptide was often absent, which suggests that it may be a breakdown product of one of the larger peptides. A goat antiserum to these IgE-binding peptides was prepared and used to study immunoprecipitation and histamine release from peripheral blood leukocytes (Meinke et al., 1978; Melewicz et al., 1982a). As shown in Table I, the antiserum inhibited IgE rosette formation on normal and cultured B lymphocytes, monocytes, alveolar M+, and cultured M+. It also inhibited rosette formation on eosinophils from donors with hypereosinophilia (Capron et ul., 1984). In contrast, it did not inhibit IgE rossette formation on basophils and failed to induce histamine release from peripheral blood leukocytes. These findings indicate that FceR on lymphocytes, M+, and eosinophils are antigenically related and differ from those on basophils. Immunoprecipitation of solubilized, radiolabeled surface proteins from U937 cells showed two peptide bands with molecular weights of

Fc

RECEPTORS FOR

65

IgE

TABLE I INHIBITION OF IgE AND IgG ROSETTE FORMATION ON DIFFERENT HUMAN CELLTYPES CARRYING FROM WIL-2WT LYMPHOBLASTOID CELLS FceR BY A GOATANTISERUMTO FceR ISOLATED IgC rosettes (%)

IgE rosettes (%)" Inhibition Cell type

Control

aFceR

(%)

Control

aFceR

Inhibition (96)

Lymphocytes6 WiI-2WT Monocytes Alveolar MrjJ U937 MrjJ Eosinophils' Basophils*

7.1 93.4 10.6 7.4 80.3 37.7 35.1

0.5 1.0 1.7 1.4 3.5 6.7 37.8

93 99 84 81 92 85 0

52.7 2.1 90.2 65.1 82.0 32.4 83.5

55.6 2.5 92.0 65.9 89.0 32.0 80.6

0 0 0 0 0 0 4

" Mean of three experiments: control preimmune serum 1 : 10, anti-FceR serum (aFceR) absorbed with FcsR- Raji and Molt4 cells 1 : 10. Normal lymphocytes depleted of E rosetting T cells (B cells and null cells). Eosinophils from three donors with hypereosinophilia. Basophils from a patient with chronic myelogenous leukemia.

approximately 45,000-50,000 and 20,000-25,000, similar to those precipitated from RPMI-8866 lymphocytes. Whether FcsR on lymphocytes and MC#Jare structurally identical has not been established; however, the antigenic and structural analyses suggest that they are closely related and that they clearly differ from FcsR on basophils (Hempstead et al., 1979). Finbloom and Metzger (1983) undertook structural analyses of FcsR on rat M+. By using cross-linking reagents to stabilize the interaction between IgE and its receptors, two peptides binding to IgE were detected. One had a molecular weight of 40,000-70,000 and was on the cell surface, whereas the other had a molecular weight of 33,000 and was not accessible for cell surface iodination. Finbloom and Metzger (1983) suggested that FcsR on MC#Jmay consist of a larger (Y chain and smaller p chain analogous to the FcsR on rat basophilic leukemia cells (Metzger et al., 1982), a pattern that may be common to all FcR. The function of the p chain would be to anchor the a chain that carries the binding site for the Fc fragment in the memlirane. 111. Rosette Assays for Detection of FCER

As described earlier, monomeric IgE dissociates rapidly from FcsR on lymphocytes and M+. Therefore, assays involving multiple IgEFcsR interactions such as rosette formation or binding of aggregated IgE are most suitable for detecting the low-affinity FcsR. Since the rosette assay is convenient and easy to perform, we modified the one

66

HANS L. SPIEGELBERG

FIG.1. IgE rosetting lymphocyte (A), monocyte (C), and basophilic granulocyte (D), and IgC rosetting lymphocyte (B).

conventionally used for detecting IgG Fc receptors (FcyR) (Hallberg et al., 1973) for IgE. For the IgG rosette assay, fresh ox erythrocytes are sensitized with a subagglutinating dose of rabbit IgG anti-ox erythrocyte antibodies, and the sensitized erythrocytes are then added to white cells. White cells that have FcyR bind the ox erythrocytes forming a “rosette” (Fig. 1B). Since FceR are highly species specific and no human IgE antibodies are presently available for sensitizing erythrocytes, we coated fixed ox erythrocytes with an IgE myeloma protein (Gonzalez-Molina and Spiegelberg, 1977) according to a modified procedure described by Hirata and Brandriss (1968). Ox erythrocytes are treated with trypsin and fixed with pyruvic aldehyde and formaldehyde. At pH 5.0, such treated cells adsorb IgE myeloma proteins (or other myeloma proteins and normal IgG) in a manner that allows the IgE to react with FceR and form rosettes with lymphocytes, M+, and basophils (Fig. lA, C, and D). To control for specificity of the rosette formation, we routinely incubate cells at 2 mg/ml of an IgE

Fc

RECEPTORS FOR

IgE

67

myeloma protein to inhibit the rosettes. We have since slightly modi-

fied the original method by using a bis-diazotized benzidine (BDB)treated IgE myeloma protein (Ishizaka et al., 1967) rather than the native IgE for coating the fixed ox erythrocytes. Although most of the IgE was not aggregated after the BDB treatment (15p g BDB/mg IgE), it seemed to coat the fixed ox erythrocytes more efficiently, and maximum percentages of rosettes were obtained more consistently and reproducibly. Fixed ox erythrocytes obtained from some steers spontaneously formed nonspecific rosettes with lymphocytes and M+. If this phenomenon occurs, it is necessary to switch to erythrocytes from another animal. Aldehyde fixed ox erythrocytes cannot be used as targets for IgE antibody-dependent cytotoxicity assays (ADCC) because they cannot be lysed. Therefore, fresh ox erythrocytes (Spiegelberg and Dainer, 1979) or chicken erythrocytes (Melewicz and Spiegelberg, 1980) were coated with conjugates consisting of rabbit Fab‘ antierythrocyte fragments covalently coupled to IgE or IgE Fc fragments according to a technique described by Strausbach et al. (1970). Erythrocytes coated with Fab’-IgE conjugates yield percentages of IgE rosetting cells similar to those obtained with IgE-coated fixed erythrocytes; however, the latter test can be performed with much less IgE, and this material is not readily available. Murine IgE monoclonal antibodies specific for the haptens dinitrophenyl (DNP) and trinitrophenyl (TNP) have been developed (Liu et al., 1980; Bottcher et ul., 1980; Rudolph et al., 1981); the clone developed by Rudolph et al. (1981) is commercially available through the American Type Culture Collection. The monoclonal mouse IgE anti-DNP antibodies have been used to sensitize TNP-conjugated sheep erythrocytes for rosette assays to detect FceR on mouse lymphocytes (Chen e t al., 1981; Vander-Mallie et al., 1982) and M+ (Boltz-Nitulescu et al., 1982). In our laboratory, the percentages of rosetting lymphocytes and M+ were similar whether monoclonal IgE anti-DNP sensitized fresh erythrocytes or monoclonal IgE-adsorbed fixed erythrocytes were used. As has been described for all rosette asays, the optimal concentrations for sensitization or coating of the erythrocytes must be determined in preliminary experiments. A quantity of IgE must b e used that yields maximum ( “ p l a t e a ~ ” percent) ages of rosettes with cells whose percentage of the FceR+ population is known. We use cultured human lymphoblastoid cells for standardization of the indicator erythrocytes because they represent a source of relatively homogeneous cells that form >90% rosettes with optimally coated indicator cells.

68

HANS L. SPIEGELBERG

IV. FcER+ Cultured Lymphocytes and MI$ and FcER+ Leukemic Lymphocytes

Established cultures are a valuable source of abundant, relatively uniform cells that can be used for characterizing membrane proteins such as Fc receptors. Therefore, we analyzed established lymphocyte (Gonzalez-Molina and Spiegelberg, 1976) and M+ cell lines of both human and murine origin for FcR by using rosette assays. The human lymphoblastoid cell lines could be divided into three major groups: FceR+, FcyR- B cells; FceR-, FcyR+ B cells; and T cells that neither formed IgE nor IgG rosettes (Table 11). Of 36 B lymphoblastoid cell types tested, only one line (Victor) showed both FcsR+ and FcyR+ cells, whereas >80% of all of the other lines expressed only one class of FcR. Interestingly, 64% of the B cell lines expressed FcsR, an unexpectedly high frequency considering that IgE is only a minor component of the total immunoglobulin in serum. Perhaps Fc&+ B cells are preferentially transformed by the Epstein-Barr virus, or the culture conditions stimulate expression of FcsR rather than FcyR. To investigate whether IgE would induce FcsR on FcsR- lines, we cultured FceR- Raji cells in the presence of l, 10, and 100 pglml IgE. However, no induction of FcsR expression occurred. PERCENTAGES OF IgE

AND

TABLE I1 IgG ROSETTE-FORMING CULTURED LYMPHOCYTES AND MC#J Rosetting cells (%)”

Cell type Human cells B cells RPMI-8866 B cells Wil-2WT B cells GM-3163 B cells Daudi B cells Raji B cells NC-40 T cells Molt-4 T cells Jurkat T cells CEM MC#JU937 Murine cells MI$ P388Dl M+ WR19M.1 MC#JWEHI3

FcER+ 98 85 98 0

0 0 0 0

0 65 93 55 16

FcyR+ 0 3 0

98 35 42 0 0 0 82 98

97 83

Average of three or more experiments; the cells were analyzed at the end of the experimental growth phase, where maximum numbers of rosette-forming cells were observed.

Fc

RECEPTORS FOR

IgE

69

The human cultured M 4 line U937 (Melewicz et nE., 18824 and six murine M 4 lines (Boltz-Nitulescu et al., 1982) were also analyzed for FcR. All M+ lines that were tested demonstrated both FcsR+ and FcyR+ cells, and usually >50% of the cells formed IgG and IgE rosettes, indicating that individual cells carried both receptors (Table 11).This is in contrast to the lymphoblastoid cells, which expressed either FcsR or FcyR. Lymphocytes from patients with chronic lymphatic or acute leukemia were analyzed for IgE, IgG, and IgM FcR (Spiegelberg and Dainer, 1979). Lymphocytes from 35% of 30 patients with B cell leukemia had FcsR+ cells, whereas no significant percentages of FcsR+ leukemic cells were found in two patients each with Sezary syndrome and acute leukemia. In contrast to cultured B lymphoblastoid cells and peripheral blood B cells, leukemic B cells expressed two or more FcR on the same cells. Tumor cells from a histiocytic lymphoma of an aged rat (Woda et al., 1981) that developed into M4-like cells after short-term culturing produced >80% IgE and IgG rosettes. Apparently, transformed human, murine, and rat M4-like ceIls express both F~ER and FcyR. V. FcER+ B and T Cells in Nonatopic Healthy Humans

Peripheral blood lymphocytes (PBL) (Gonzalez-Molina and Spiegelberg, 1977) and lymphocytes from several lymphoid organs (Hellstrom and Spiegelberg, 1979) were analyzed for IgE and IgG rosetting lymphocytes (Table 111). Because a fraction of monocytes PERCENTAGES OF IgE

AND

TABLE 111 IgC ROSETTE-FORMING LYMPHOCYTES IN NONATOPIC HEALTHY HUMANS Rosetting cells

Cell type

F~ER+

(%)O

FcyRt

Lymphocytes

Blood (n = 20) Cord blood (n = 6) Tonsildadenoids (n = 11) Spleen (n = 4)

1.1 2 0.4 3.0 2 1.3 4.9 2 4.2 5.6 ? 4.7

Fractionated peripheral blood lymphocytes <0.1 T cells (mIg+ depleted) 7.1 2 1.6 B cells (E' depleted) 1.6 2 0.2 FcyR+-Depletedlymphocytes Mean 2 SD.

16.1 2 40.2 2 12.5 2 40.2

*

4.3 16.4 7.5 6.9

7.6 2 3.8 52.7 t 8.6 0.8 2 0.3

70

HANS L. SPIEGELBERG

also form IgE rosettes, phagocytic cells were removed by incubating the lymphocytes with colloidal iron (Perlmann et al., 1976), which usually results in preparations containing <0.2% nonspecific esterasepositive cells. In our first study (Gonzalez-Molina and Spiegelberg, 1977), we reported a mean ? SD of 4.3 1.6% FceR+ lymphocytes for normal donors; however, in all subsequent analyses, the mean of FceR+ PBL in nonatopic donors was approximately 1%. Most likely, a population of donors selected for high percentages of FceR' cells was studied initially. As shown by examining isolated T and B cells and by employing a mixed rosette assay, more than 90% of the FcsR+ PBL were mIgM+/mIgD+ B cells that lacked FcyR. Yodoi and Ishizaka (19794 confirmed that most human FceR+ PBL are B cells. Similarly, Thompson et al. (1983) did not detect more than 0.1% IgE rosetting cells reactive with the monoclonal pan-T cell antibody Lyt-3 (Kamoun et al., 1981) in nonatopic donors. Because FcsR+ PBL are mostly B cells, it is not surprising that the percentages of Fc&R+cells are higher in tonsil and spleen lymphocyte preparations, which contain more B cells than PBL. Additionally, cord blood lymphocytes contain 3% of Fc&R+lymphocytes, an unexpectedly high percentage. The FcsR on PBL react specifically with the native Fc fragment of IgE. Neither reduced and alkylated nor 56°C-heated IgE inhibits or induces IgE rosette formation. IgE Fab fragments and human myeloma proteins of all four IgG subclasses, the IgA, IgM, and IgD classes, and rat IgE fail to inhibit IgE rosettes. Similarly, IgG rosettes on human PBL cannot be inhibited with IgE, IgA, IgM, or IgD myeloma proteins. This contrasts with IgG rosette formation on rat lymphocytes, which can be inhibited with rat IgE (Yodoi and Ishizaka, 197913). The effect of culturing on the quantity of Fc&R+cells was analyzed to determine whether incubation in fetal calf serum increases the number of FceR+ cells in PBL as is the case with FcpR+ T cells (Moretta et al., 1975). We also questioned whether adding IgE to the medium induces FcsR as has been shown for rat lymphocytes (Yodoi et al., 1979). However, the percentage of FcsR+ human lymphocytes did not increase under either condition, but usually decreased. Treatment of the cells with acetate buffer, pH 4.0 (Ishizaka and Ishizaka, 1974), also did not increase the percentage of FceR+ cells. These experiments indicate, first, that IgE bound to cells in vivo does not interfere with subsequent IgE rosette formation presumably because it is washed off during cell isolation, and, second, that IgE cannot induce FcsR in vitro on human lymphocytes, as Ishizaka and Sand-

*

Fc

RECEPTORS FOR

IgE

71

berg also found (1981).However, these authors did induce F ~ E R in vitro when they incubated lymphocytes from atopic donors with both IgE myeloma protein and ragweed antigen E. Because humans cannot be injected with IgE myeloma protein to investigate whether IgE induces Fc&+ PBL in viuo, we injected two monkeys whose PBL formed rosettes with human IgE-coated ox erythrocytes. Sufficient IgE myeloma protein was injected to yield an IgE serum level of approximately 2 p g (50,000 IU)/ml for 2 weeks (Spiegelberg et al., 1979). However, these injections did not significantly change the percentage of Fc&+ PBL in either monkey. Although the reason for this negative result may be that human IgE has a lower affinity for F ~ E on R monkey cells and therefore does not induce FceR+ PBL, this experiment at least indicates that Fc&R+lymphocytes are no more easily induced in living monkeys than in cultured human lymphocytes. VI. FceR' B and T Cells in Atopic Patients

Patients with atopic disorders such as allergic rhinitis (hay fever), asthma, or atopic dermatitis (eczema) produce more IgE antibodies than nonatopic individuals, usually manifested by an elevated IgE serum level. Therefore, it was of great interest to study FcsR+ B and T cells in atopic donors (Spiegelberg et al., 1979; Spiegelberg and Simon, 1981; Thompson et al., 1983). Atopic patients were divided into three categories, the first consisting of patients with allergic rhinitis, food anaphylaxis, or asthma and with moderately elevated IgE levels of 100-2000 IU/ml. Patients in the second category had severe atopic dermatitis and highly elevated IgE levels of 2000-30,000 IU/ml. Some of the patients with atopic dermatitis also had severe atopic asthma and were treated systemically with corticosteroids. Because the percentages of Fc&+ lymphocytes in the latter patients differed markedly from untreated patients, they were classified separately into a third category. As shown in Table IV, patients with mild allergic disease analyzed at a time when they had no allergic symptoms did not differ significantly in percentages of FcsR' PBL from those of the nonatopic control donors. In contrast, patients with severe atopic dermatitis had significantly elevated numbers of FceR+ PBL ( p < 0.01). Interestingly, the patients with severe atopic dermatitis and asthma treated systemically with corticosteroids had significantly lower percentages ( p < 0.02) of Fc&+ PBL than both nonatopic and atopic donors. Although the patients with severe atopic dermatitis who had

72

HANS L. SPIEGELBERG

PERCENTAGES OF IgE

AND

TABLE IV IgG ROSE-ITE-FORMING PERIPHERAL B~.oooLYMPHOCYTES IN ATOPICPATIENTS"

Fc&R+ Description of patients Donors Nonatopic ( n = 12) Mild atopic in remission ( t i = 12) Severe atopic dermatitis ( t i = 6) Severe atopic dermatitis and asthma treated systemically with corticosteroids ( n = 4)

(%)

FceR+/mm3

1.2 f 0.5 1.6 f 0.9 7.2 2.5

41 24 59 f 43 198 f 73

*

0.2

* 0.1

Seasonal changes in grass pollen-sensitive atopic patients 1.3 f 1.1 Nonatopic controls preseason ( n = 10) Nonatopic controls pollen season (n = 10) 1.7 1.9 Nonatopic controls postseason ( n = 10) 1.4 f 0.9 Atopics preseason ( n = 7) 2.0 f 3.1 Atopics pollen season ( n = 7) 4.7 f 1.2 Atopics postseason ( n = 7) 2.1 f 1.9

*

Mean

*

12

f6

31 f 24 40 f 46 26 f 13 48 +- 52 134 69 62 82

*

FcyR+ (%)

*

18.7 7.1 20.1 f 5.9 13.8 2 3.8

25.8 f 10.8 15.7 f 7.8 8.0 2 8.1 14.8 f 7.5 17.6 f 12.6 11.9 5.9 14.4 4.7

* *

* SD.

the highest IgE levels also had the highest percentages of FcsR+ PBL, in individual patients the percentage of FcsR+ PBL did not correlate with the IgE serum level. To determine whether the percentages of FcsR+ PBL correlate with the activity of atopic disease, we analyzed atopic patients suffering from seasonal allergic rhinitis (Spiegelberg and Simon, 1981). As shown in Table IV, these patients had significantly more Fc&R+lymphocytes during the grass pollen season when they suffered from allergic rhinitis and presumably formed IgE anti-pollen antibodies. In 8 of 10 nonatopic control donors, no such seasonal change was observed. However, two healthy control donors having neither a history of allergic disease nor any symptoms of allergic rhinitis at the time of testing, also showed an increase of Fc&R+PBL during the grass pollen season. One of these donors tested repeatedly during two summer and three winter periods showed elevated percentages in the summer months but never in the winter when the pollen counts were low. She had a very low IgE serum level (2 IU/ml) and only a weak skin reaction to grass pollens, but the reason for her seasonal increase of Fc&R+ PBL is unknown. Perhaps the percentage of FcsR+ PBL is a sensitive indicator of even a small IgE response that cannot be detected by measuring serum IgE levels. Despite a significant change in the per-

Fc

RECEPTORS FOR

IgE

73

centages of Fed+ PBL, the total IgE serum level and the number of specific IgE anti-pollen antibodies measured by the radioallergosorbent test did not change appreciably in these patients; any changes that occurred did not correlate with the percentages of FcER+ PBL. Lymphocytes from four atopic donors were depleted of mIgM+/ mIgD+ B lymphocytes to investigate whether the patients’ FcsR+ PBL are B or T cells (Spiegelberg et al., 1979). Depleting PBL of B cells removed 90% or more of the FcER+lymphocytes, indicating that most of them are B cells. However, not all FcsR+ cells are B cells. Yodoi and Ishizaka (1979a) first reported that ragweed-sensitive atopic donors have a small but significant percentage of Fc&R+peripheral blood T cells (TEcells). To investigate the TE cells in our atopic patients, we reacted their isolated T cells with monoclonal antibodies to T cell subsets, and analyzed the IgE rosetting cells for positive immunofluorescence (Thompson et al., 1983). Less than 0.1% IgE rosetting T cells were detected in seven nonatopic donors, 4 of 7 atopic donors with mild atopic disease, and 6 of 7 donors with severe atopic dermatitis and highly elevated IgE levels, indicating that even atopic donors have few TE cells in the peripheral blood. The patients with severe atopic dermatitis not only had few TE cells, they also had significantly fewer FcyR+ T cells (Ty cells), a mean ? SD of 3.1 2.7% as compared to 10.5 4.1 for the nonatopic, and 7.2 ’. 3.7% for the mildly allergic donors. This observation has also been made by others (Canonica et al., 1979; Schuster et al., 1980). As shown in Table V, both TE and Ty cells from atopic donors that had more than 0.1% TE cells showed similar patterns of reactivity with monoclonal antibodies. These cells reacted with the pan-T cell monoclonal antibody Lyt-3 (Kamoun et al., 1981) that detects the sheep erythrocyte receptor on T cells, and with the OKT8 antibodies that react with the cytotoxic/suppressor T cell subset (Reinherz and Schlossmann, 1980), and OKM1, which reacts with killer and natural killer T lymphocytes (Lobo, 1981; Van de Griend et al., 1982). However, these TEand Ty cells did not react with the “pan-T cell” antibody OKT3, which reacts with most but not all T cells (Fox e t al., 1981), or OKT4 antibodies that detect helperhnducer T cells (Reinherz and Schlossman, 1980). As described previously, less than 0.1% TEcells were detected in nonatopic donors. Therefore, they could not be studied for their reactivity with monoclonal antibodies. However, the Ty cells of atopic and nonatopic donors reacted similarly with monoclonal antibodies to T cell subsets (Table V), suggesting that T cells expressing Fc receptors have the same phenotype in both donor groups. However, severely atopic donors appear to have

*

*

74

HANS L. SPIEGELBERG

PERCENTAGES OF IgE

TABLE V I$ ROSETTE-FORMING T CELLS REACTINGWITH MONOCLONALANTIBODIES

AND

IgE rosetting T cells (%) positive for

FceR+ (%) Donors Atopic 8 HS 9 JB 10 GN 15 BS Mean kSD

PBL

T

Lyt-3

OKT3

OKT4

OKT8

OKMl

/.LIB

4.0 3.4 4.0 6.7

0.8 0.6 1.3 0.8

81.8 90.9 81.3 100.0

0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0

21.4 61.5 21.4 28.6

35.7 68.8 29.4 38.5

0.0 0.0 0.0 0.0

4.5 k1.5

0.9 20.3

88.5 *8.9

0.0 kO.O

0.0 kO.O

33.2 219.2

43.0 217.6

0.0 20.0

IgG rosetting T cells (%) positive for

FcyR+ (%) Nonatopic

1 CB 3 JP 4 MG 5 RL

PBL

T

Lyt-3

OKT3

OKT4

OKT8

OKMl

p/8

15.0 17.5 16.0 12.5

6.0 10.4 7.0 7.1

70.3 97.4 94.1 96.8

0.0 3.9 0.0 4.8

0.0 0.0 0.0 0.0

28.3 23.7 11.8 38.8

84.6 75.7 78.3 78.6

5.1 0.0 0.0 0.0

Mean kSD

15.3 k2.1

7.6 21.9

84.6 k 13.2

2.2 k2.5

0.0 20.0

25.7 211.2

79.3 23.8

1.3 22.6

Atopic 8 HS 9 JB 10 GN 15 BS

12.8 18.7 18.0 10.5

4.8 12.0 12.8 3.0

80.9 76.0 96.0 78.9

0.0 2.0 4.0 0.0

0.0 0.0 0.0 0.0

23.1 22.2 3.3 8.8

68.9 94.1 71.4 95.5

0.0 0.0 0.0 0.0

15.0 *4.0

8.2 k5.0

83.0

1.5 kl.9

0.0 kO.0

14.4 29.9

82.5

k8.9

2 14.3

0.0 20.0

Mean &SD

a numerical deficiency of Fc&R+and FcyR+ peripheral blood T cells. Interestingly, the four atopic donors who had 0.6-1.3% T E cells showed these relatively high percentages only during the summer months. Whether T E cells in the peripheral blood of atopic donors show the same seasonal variations as has been shown for FcsR+ PBL is presently being investigated in our laboratory. VII. FcER+ Lymphocytes in Normal and Parasitically Infected Rats and Mice

Lymphocytes bearing F ~ E have R been demonstrated by rosette assays of blood samples from rats (Fritsche and Spiegelberg, 1978;

Fc

RECEPTORS FOR

IgE

75

Yodoi and Ishizaka, 1979a) and mice (Chen et ul., 1981; Vander-Mallie et ul., 1982). FcsR+ lymphocytes were also identified by measuring cell-bound IgE before and after acid treatment of mouse cells using a fluorescence-activated cell sorter (Katona et ul., 1983). The characteristics of rat and mouse FcsR+ lymphocytes closely resemble those of the human cells; they are mostly B cells. However, normal rats and mice have higher percentages of FcsR+ lymphocytes, for example, rats have 13% PBL compared to 1% in normal humans (Fritsche and Spiegelberg, 1978). Rat splenic lymphocytes are 14% FcsR+ (Table VI), and the several strains of mice tested including nude mice show 26-40% F ~ E R lymphocytes ' (Vander-Mallie et ul., 1982). Chen et al. (1981) found a much lower percentage-only 4% FcsR+ lymphocytes-in spleens of CAFl mice, a difference that may result from variation by strain or perhaps from the use of slightly undersized indicator erythrocytes. As in humans, FcsR on rat and mouse lymphocytes react specifically with IgE. IgE rosettes cannot be inhibited with Ig of other classes or heat-denatured IgE. They also cannot be inhibited with human IgE, but rat and mouse IgE can cross-inhibit. In contrast to human FcsR+ lymphocytes, rat and mouse FcsR+ lymphocytes form rosettes with rabbit IgG antibody-sensitized ox erythrocytes, indicating that these cells carry both FcsR and FcyR. Rat lymphocytes carry at least two distinct FcyR, one for IgGz, and one for IgG, (Spiegelberg, 1981). FceR' lyniphocytes do not react with normal IgG, (Fritsche and Spiegelberg, 1978); therefore, they may bear only FcyR for IgGl . However, investigation into the nature and specificity of the FcyR on FceR+ rat and mouse lymphocytes is not yet complete. The characterization of FcyR on these cells is a worthy endeavor because IgE has been shown to inhibit rosette formation on rat lymphocytes formed by rabbit IgG-coated erythrocytes (Yodoi and Ishizaka, 1979b). Preliminary experiments performed in our laboratory suggest that rat IgE inhibits IgGl but not IgGB, rosetting on rat lymphocytes. Yodoi and Ishizaka (197913) reported that the FcyR on rat lymphocytes are not Ig class specific and that the reaction of IgE with FcyR induces the formation of FcsR. We could not demonstrate cross-reaction between IgE and FcyR on human or murine lymphocytes. Therefore, it remains to be investigated whether the cross-reaction of IgE with FcyR is a general phenomenon that is important for induction of FcsR. The parasite Nippostrongylus brusiliensis induces a rise in IgE formation of infected rats (Jarrett and Bazin, 1974) and mice, which offers an experimental system for studying FcsR+ lymphocytes under conditions of increased IgE formation. Yodoi and Ishizaka (19794 showed

76

HANS L. SPIEGELBERG

first that numbers of FccR+ lymphocytes increase dramatically after infection of rats with N . brasiliensis, an observation that was confirmed in our laboratory (Spiegelberg, 1981). As Table VI illustrates, 1 month after infection of rats with N . brasiliensis, the percentages of Fc&R+lymphocytes increased in the spleens and particularly in the mesenteric lymph nodes with little change in the proportion of FcyR+ lymphocytes. Other conditions in which the IgE serum level increases are also associated with elevated percentages of Fc&R+cells. Rats bearing the IgE-producing myeloma IR162 (Bazin and Becker, 1976) and rats injected with purified IgE developed percentages of FccR+ lymphocytes in their mesenteric lymph nodes. The spleens of the rats bearing the IR162 tumor had marked infiltrations of FURtumor cells that isolated with lymphocytes. Therefore, they had low numbers of FccR+ cells in their spleens. As Table VI shows, injection of native but not reduced and alkylated IgE (which does not bind to FccR) induced an increase of FccR+ lymphocytes, indicating that IgE per se induces FcsR in normal rats. This finding corroborates in uitro studies of FccR induction on rat and mouse lymphocytes. Yodoi et al. (1979) showed for normal rat lymphocytes and Chen et al. (1981)for normal mouse lymphocytes that FceR+ cells appear after culturing in media containing 1-10 pg/ml of IgE. As described earlier, these findings contrast to those for human and monkey lymphocytes. Neither Ishizaka and Sandberg (1981) nor Spiegelberg et al. (1979) could inTABLE VI PERCENTAGES OF IgE AND IgG (RABBIT) ROSETTE-FORMING SPLEEN AND MESENTERICLYMPH NODE(MLN) LYMPHOCYTES FROM NORMAL RATS,N . BrUSilienSiS-INPECTED RATS, RATSBEARING THE I@-PRODUCING TUMOR IR-162, AND RATS INJECTED WITH PURIFIED IgE Rosetting cells (%o)" ~

FcER+

Normal (n = 12) N . brasiliensis-infected (n = 5) IgE IR162 tumor-bearing (n = 5) IgE IR162-injected (n = 3)b IRE IRl62 reclriced and alkylated injected (11 = 4)

~~

~~

~~~

FcyR+

Spleen

MLN

Spleen

MLN

13.9 f 2.4 21.5 +. 2.9

1.9 2 1.1 20.3 +. 7.8

34.3 f 10.6 27.8 -+ 5.1

22.2 5.9 12.4 2 6.6

8.7 f 2.1 20.2 2 3.7

10.3 f 3.7 11.0 f 6.0

52.6 f 7.8 44.2 f 12.8

27.4 f 5.8 22.0 +. 14.8

9.6 f 4.0

3.0

* 2.4

*

49.1 f 1.8

25.6 5 2.3 ~~

Mean f SD. Lymphocytes were examined 1 day after ip injection of 5 mg native or reduced and alkylated IgE IR162.

Fc

RECEPTORS FOR

IgE

77

duce FcsR+ lymphocytes by culturing human cells in media containing IgE. The FcsR+ lymphocytes induced by infection of rats and mice with N . brasiliensis are predominantly B cells, although a small fraction consists of TE cells. Yodoi and Ishizaka (19794 and Vander-Mallie et al. (1982) found approximately 3% TEin mesenteric lymph node lymphocytes of N . brasiliensis-infected rats and mice. Using dual parameter FACS analysis with anti-T cell and anti-IgE antibodies, Katona et al. (1983) detected less than 3% Thy-1.2+ IgE+ lymphocytes in N . brasiliensis-infected mice. Chen et al. (1981) reported that approximately 50% of Fc&+ cells induced in vitro on mouse lymphocytes are T cells. We isolated mesenteric lymph node lymphocytes from N . brasiliensis-infected rats and removed the B cells, leaving approximately 1% FcsR+ cells in the remaining population. When the T cells were reacted with the monoclonal antibodies W3/25, which characterize helper T cells (White et al., 1978), or 0x8, which detects cytotoxic and suppressor cells (Brideau et al., 1980), all IgE rosetting cells in the T cell preparations reacted with the antibody 0x8; no W3/25+ Fc&+ T cells were detected (Spiegelberg and Gilman, unpublished). In conclusion, rats and mice contain more FceR+ lymphocytes than humans, and purified IgE whether added in vitro to normal lymphocytes or injected into normal rats induces FceR+ cells. The majority of the FcsR+ lymphocytes are B cells; however, a small but significant percentage of OX8+ FceR+ T cells are detectable in N . brasiliensisinfected animals. VIII. Fc&R+ Monocytes in Nonatopic and Atopic Humans

Latex phagocytosing peripheral blood monocytes isolated b y adherence to plastic dishes were analyzed for IgE rosette formation (Melewicz and Spiegelberg, 1980). As one sees in Table VII, approximately 15% of peripheral blood monocytes and 8%alveolar M+ (Melewicz e t al., 1982b) from nonatopic donors form IgE rosettes. The IRE rosettes were specific for the native IgE Fc fragment; they could b e inhibited only with IgE or Fc IgE fragments but not with myeloma proteins of the other Ig classes, denatured IgE, or heterologous IgE. Rosette assays with monocytes must be performed strictly in the cold by keeping the cells on ice, otherwise large concentrations of IgE are necessary to inhibit the reaction (Melewicz and Spiegelberg, 1980). The major FcR on normal monocytes and M+ appears to be the FcyR, since 80% of monocytes and an average of 64% of the alveolar M+ formed IgG rosettes.

78

HANS L. SPIEGELBERG

TABLE VII IRE Roswrti FORMATION AND 51cr RELEASEFROM I ~ E - C O A I T D CHICKEN E R y r H R O C Y ' r w BY MoNocvTtis FROM NONArOPLc HEALTHY AND A T O P lC DONORS Cell type Normal Monocytes (n = 15) Alveolar M+ (n = 15) Atopic Mild atopic (n = 12) Severe atopic (n= 6) Severe atopic treated systemically with corticosteroids

Rosettes (%)"

"Cr release (%)

15.3 f 4.8 8.0 I2.6

2.6 f 2.2 n.t.6

18.6 1 3 . 5 56.0 t 39.3

2.9 t 2.9 16.3 _t 8.4

5.1 I4.8

1.0 I1.4

" Mean IJ

2 SD. n.t., Not tested.

The three categories of atopic patients, those with mild atopic disease, with severe atopic dermatitis, and with atopic dermatitis and severe asthma treated systemically with corticosteroids, were analyzed for IgE rosetting monocytes (Melewicz et al., 1981a) (Table VII). Patients with mild atopic disorders studied at a time when they did not have active disease had percentages of Fc&R+monocytes similar to those of the nonatopic control donors. In contrast, the patients with atopic dermatitis and highly elevated IgE serum levels had significantly more IgE rosetting monocytes than the nonatopic donors and patients with mild allergic disease. The patients with severe asthma who received systemic corticosteroid therapy showed the lowest percentages of FceR+ monocytes. Apparently, corticosteroids suppress FCERexpression not only on lymphocytes (Spiegelberg et al., 1979) but also on monocytes. To study whether FCER on monocytes mediate phagocytosis and lysis of IgE-coated target cells, mononuclear cells from healthy nonatopic donors (Melewicz and Spiegelberg, 1980) and atopic patients (Melewicz et al. 1981b) were incubated with 51Cr-labeled chicken erythrocytes that were sensitized with covalent conjugates of rabbit Fab' anti-chicken erythrocyte antibodies and a human IgE myeloma protein. Such sensitized target cells were lysed by monocytes but not by lymphocytes. As shown in Table VII, the mononuclear cells of patients with severe atopic dermatitis and elevated percentages of FmR+ monocytes induced significantly more 51Cr release from the IgE-coated chicken erythrocytes than mononuclear cells from the

Fc

RECEPTORS FOR

IgE

79

nonatopic donors or mildly atopic donors. Kinetic studies indicated that the chicken erythrocytes were first phagocytosed before being lysed. When the percentages of W r release were compared with amounts of FcsR+ monocytes in the mononuclear cell preparations, a good correlation (T = 0.83)was observed. In contrast, the extent of "Cr release did not correlate to the total number of latex phagocytosing cells in the preparations ( r = 0.46). These studies provide evidence that FcsR on monocytes promote phagocytosis and lysis of IgE-coated target cells and that IgE is an opsonin for monocytes. IX. FcER+ Rat and Mouse M4

Capron et al. (1975) reported the first experiments indicating a cytophilic interaction between IgE and M+. These authors showed that schistosomules adhere to rat M+ in the presence of immune sera containing IgE and that the killing of the schistosomules depends on IgE antibodies. Subsequent work by the same group of investigators (Joseph et al., 1978) demonstrated killing of schistosomules by human and baboon monocytes in the presence of IgE-containing immune sera. These sera apparently contained IgE-soluble parasite antigen complexes rather than monomeric IgE antibodies, and the IgE immune complexes induced the killing of the parasites. To characterize FCERon M+, we used the rosette assay for rat M+ (Boltz-Nitulescu and Spiegelberg, 1981; Boltz-Nitulescu et al., 1981) and mouse M+ (Boltz-Nitulescu et al., 1982). Approximately 80% of alveolar and 50% of proteose peptone-induced rat and mouse M+ formed IgE rosettes. In Table VIII, the IgE rosettes on rat M+ are IgE specific, since myeloma proteins of other classes and subclasses were not inhibitory; whereas rat and mouse IgE cross-inhibited, human IgE did not. Rat M+ were analyzed for rosette formation with fixed ox erythrocytes coated with rat myeloma proteins of all classes because of the reported cross-reaction of rat IgE with FcyR on rat lymphocytes (Yodoi and Ishizaka, 1979b) ad rat basophilic leukemia cells (Segal et al., 1981). Rat M+ formed rosettes with ox erythrocytes coated with IgGzB,IgGgb, and IgGl but not IgG,,, IgA, IgM, or IgD myeloma proteins. As shown in Table VIII, IgG2, and IgE rosettes were specific for IgGz, and IgE, respectively. In contrast, IgG1 and IgG21, rosettes were inhibited by IgG,, IgGzb, and IgE. They were also inhibited by heterologous IgG, whereas IgG2, rosettes were not. Presumably IgG1 and IgGzb react with the same receptor. Most likely, IgE reacts also with the FcR for IgGl and IgGzl,; however, the inhibition could also result from steric hindrance by IgE bound to its own receptors if they

80

HANS L. SPIEGELBERG

TABLE VIII I&%,, I#&,, I g c , , AND IgE ROSEXTE FORMAHON BY R m MYELOMA OF DIFFERENT CLASSES AND HETEROLOCOUS I g c ON RAT ALVEOLARM 4 PROTEINS

INHIBITION OF

Rosette inhibition (%)n

a

Inhibitor (1mg/ml)

IBCZ,

IgGzb

IgCi

IgE

Rat IgCz, (IR33) Rat IgC2b (IR863) Rat IgGl (IR595) Rat IgE (IR162) Rabbit IgC Human IgCl Mouse IgGl (MOPC2l) Mouse IgC, (UPC10)

66 17 2 8 19 20 3 4

8 83 79 71 68 65 nkb n.t.

9 81 78

7 5 8 83 4 4 n.t. n.t.

85 84 79 68 64

Mean of three experiments. n.t., Not tested.

are localized in close proximity to the Fcy112bR. Until more information is available on the structure of these receptors, the reason for the unidirectional cross-inhibition of IgGlIzb rosettes by IgE remains obscure. Mouse MC#Ialso formed IgGl, IgG,,, and IgG2b rosettes, but rat IgE failed to inhibit any of these reactions. Therefore, the inhibition of IgGlI2b rosettes by IgE seems to be unique for rat cells. To study whether mouse IgE can act as an opsonin for rat M+, BoltzNitulescu et al. (1982) incubated TNP-conjugated sheep erythrocytes sensitized with a mouse IgE anti-DNP monoclonal antibody (Liu et al., 1980). TNP-conjugated unsensitized cells were not phagocytosed after 30 minutes of incubation at 3TC,whereas both alveolar M+ and cultured M+ ingested the IgE-sensitized erythrocytes, indicating that IgE does act as an opsonin for mouse M+, X. FceR+ Eosinophils

Although evidence suggesting that eosinophils may bear F ~ E R had been reported by several investigators (Ishikawa et d.,1974; Hiibscher, 1975),this possibility was not generally accepted because FcgR are not easily demonstrable on normal eosinophils. However, Capron et aZ. (1981) provided evidence that eosinophils form specific rosettes with IgE-coated ox erythrocytes. Eosinophils from patients with eosinophilia and elevated IgE levels showed an average of 47% IgE rosetting cells. Apparently, eosinophils of nonatopic healthy donors do not express many F ~ E RTherefore, . most normal eosinophils do not

Fc

RECEPTORS FOR

IgE

81

form IgE rosettes, like normal lymphocytes and monocytes. However, conditions in which IgE formation increases seem to induce FcsR on eosinophils, as shown with the rosette assay. We described in Section I1 that IgE rosettes are inhibited by an antiserum to lymphocyte FcsR, indicating a close relationship between FcsR on eosinophils and lymphocytes (Capron et al., 1984). Furthermore, the antireceptor serum inhibited eosinophil-mediated ADCC for schistosomules, indicating that the FceR on eosinophils are directly involved in the killing of parasites (Capron et d.,1984). By employing the IgE rosette assay, Capron et al. (1981) also demonstrated FcsR on rat eosinophils. Normal rats had an average of 21%, whereas Schistosoma mansoni-infected rats had 69% FcsR+ eosinophils. This finding shows that IgE induces FcsR on rat eosinophils much as it does on human eosinophils. XI. Conclusions: Induction and Function of FceR on Lymphocytes and M+

Investigations performed in several laboratories over the past several years demonstrate that subpopulations of lymphocytes, monocytes, macrophages, and eosinophils carry IgE-specific cell membrane receptors that differ from the “classical” FceR on basophilic granulocytes and mast cells in two major aspects. First, they bind monomeric IgE with a relatively low affinity, and second, the cell populations expressing FcsR vary and increase under conditions of elevated IgE production. Basophils and mast cells bind monomeric and the IgE dissociates slowly IgE with a high affinity ( K ; , 10“W1), from the FcsR. In contrast, lymphocytes and MC#Ibind IgE with a K , lo7 M - l , and the IgE dissociates rapidly from these cells. This suggests that the two types of FcsR differ structurally and functionally. Indeed, the IgE-binding peptides isolated from lymphocytes and MC#I differ in molecular weight from those isolated from basophils. Moreover, antiserum to lymphocyte FceR failed to react with basophil FcsR, providing evidence for their structural difference. Whereas the function of FCERon MC#I is promotion of phagocytosis, killing of IgEcoated target cells, and release of substances that mediate inflammation, the function of FceR on lymphocytes is unresolved. However, all evidence obtained to date suggests that they are most likely involved in the regulation of IgE antibody formation. FCER are characteristic membrane receptors of all basophils and mast cells whether the host produces small or large quantities of IgE (Yodoi et al., 1979). In contrast, the percentages of lymphocytes, monocytes, and eosinophils that bear FcsR as shown by IgE rosette

-

-

82

HANS L. SPIEGELBERG

formation increase under conditions of heightened IgE formation, presumably because the cells switch from FcyR to FcsR production (Yodoi and Ishizaka, 1979b). The mechanism that induces FceR is not fully understood and appears to involve several factors. Nonatopic healthy donors having low IgE levels showed transient increases in numbers of FceR+ cells, presumably because of exposure to allergens during the grass pollen season, when they were tested. Similarly, the percentages of FcsR+ lymphocytes in grass pollen-sensitive atopic patients increased without measurable increases of total and specific IgE serum levels. Therefore, allergens alone may induce FcsR, perhaps by stimulating lymphokine and monokine production after interaction with MC#Jand T cells. Normal rats and mice have higher percentages of FceR+ lymphocytes and MC#Jthan humans, possibly from exposure to FceR-inducing antigens. In addition to these antigens, IgE also modulates FcsR expression. In rats and mice, monoclonal IgE induces both in vitro (Yodoi and Ishizaka, 1979b; Chen et al., 1981; Spiegelberg et al., 1983) and in vivo (Spiegelberg, 1981) an increase of FceR+ lymphocytes. Although human IgE myeloma proteins do not show this effect (Spiegelberg et d., 1979), a mixture of allergen and IgE induces FCERin vitro on lymphocytes of atopic but not nonatopic donors (Yodoi and Ishizaka, 1979b). Perhaps the allergen initiates the FceR production, and newly formed IgE subsequently modulates the number of FceR per cell. However, additional as yet undefined factors may also play a role in the control of FcsR expression, because cord blood lymphocytes, which are exposed neither to exogenous allergens nor to high IgE levels, contain a large percentage of Fc&+ cells. The possible function of FcR on B cells is controversial. From experiments demonstrating that IgG antibody-antigen complexes inhibit B cell proliferation in vitro, several investigators concluded that cocapping of cell membrane-bound Ig with FcR provides a B cell signal leading to inactivation of proliferation and differentiation of B cells (Sinclair, 1969; Kerbel and Davis, 1974; Ryan et al., 1975; Sidman and Unanue, 1976). However, subsequent investigations on B cell activation mechanisms suggest that FcR on B cells may be involved in induction of B cell proliferation. Addition of Fc fragments (Berman and Weigle, 1977; Berman et al., 1979; Morgan and Weigle, 1981) or monoclonal antibodies directed against FcR (Lamers et al., 1982) to B cell cultures stimulates B cell proliferation. This latter finding is compatible with the observation demonstrating that patients with atopic dermatitis and very high IgE levels also have high percentages of FcsR+ B cells (Spiegelberg et al., 1979). If FcsR on B cells

Fc

RECEPTORS FOR

IgE

83

down-regulated IgE formation, one would expect these patients to have low IgE serum levels, because they certainly do not have a numerical deficiency in FcsR+ B cells. In contrast, these patients appear to be deficient in FcsR+ T cells (Thompson et al., 1983). Therefore, FcER+T cells rather than B cells may be responsible for downregulating IgE production. Indeed, a fraction of human and rat TE cells reacted with monoclonal antibodies defining T suppressor cells (OKT8, OX8), whereas no TE cells reacted with monoclonal antibodies that define helper T cells (OKT4, W3/25). However, it will be necessary to perform functional studies to demonstrate the possible suppressive role of TEcells on IgE production. Mechanisms by which TE cells could suppress IgE formation are shown schematically in Fig. 2. Yodoi and Ishizaka (1980a,b) and Ishizaka and Sandberg (1981) reported that TEcells release an “IgE-binding factor” that suppresses IgE formation in vitro analogous to a previously reported IgG-binding factor that inhibits IgG synthesis in vitro (Gisler and Fridman, 1975; Joskowicz et al., 1980). Ishizaka and his colleagues have expanded these findings and characterized two types of IgE-binding factors in rats, one being glycosylated and enhancing IgE synthesis and one lacking carbohydrate and suppressing IgE formation in vitro (Yodoi and Ishizaka, 1980a,b; Suemura et ul., 1980; Yodoi et al., 1980; Hirashima et ul., 1980, 1981; Yodoi et al., 1982a; Uede et al., 1983). Since the source of the binding factors appears to include F ~ E R T cells, whether these IgE-binding factors are fragments of FcsR derived from TE cells, as illustrated in Fig. 2, remains to be shown. Furthermore, it is difficult to imagine how IgE- and IgG-binding factors secreted by T cells could reach the membrane IgE+ and membrane IgG+ B cells, because the binding factors would be adsorbed onto serum IgE and IgG. Therefore, one could propose that the function of FcR on TE and Ty cells may be recognition of membrane IgE+ and membrane IgG+ B cells, as illustrated at the bottom of Fig. 2. Once the TE cells made contact with membrane IgE+ B cells, they could release regulatory factors such as those described by Ishizaka and Kishimoto and their co-workers (Suemura et al., 1980; Kishimoto et al., 1978; Suemura et d.,1981) in close proximity to the target B cells. Although much research remains to test this hypothesis, the possibility that IgE-induced TE cell down-regulation of IgE production proceeds via F ~ E R in an “isotype-specific network” fashion is intriguing and worth further exploration, particularly since evidence has been obtained for a similar FcaR T cell-induced system (Hoover et al., 1981; Yodoi et al., 1982b). The outcome could speed the search for means to control IgE formation and allergic disease in humans.

84

HANS L. SPIEGELBERG

FCER ( + I T cell

- IgE Binding Factor

mlgEi+) B cell

T, cell - mlgE B cell Interaction

A FIG.2. Scheme depicting the hypothetical function of immunoregulatory TEcells. A fragment of FceR containing the IgE-binding site (IgE-binding factor, Suemura et al., 1981) may be released from TE cells, bind to mIgE of B cells, and regulate them. Alternatively, FcsR on T cells may allow TE cells to recognize and regulate mIgE+ B cells through cell-cell contact.

TE cells are most likely heterogeneous and may consist of at least two distinct functional subsets. As described earlier, OKT8+ TE may be suppressor cells. A large percentage of human TE cells react with the monoclonal antibody OKM1, which also reacts with Ty cells, and OKMl+ Ty cells have been reported to mediate ADCC and natural killer activity (Lobo, 1981; Van de Griend et al., 1982). Therefore, the OKM1+ TE subset may have natural killer and killer cell activity against IgE-sensitized target cells such as parasites. The function of FcsR on monocytes and M+ is much better understood than that on lymphocytes. FCERon M+ promote phagocytosis and killing of IgE-coated targets such as parasites (Capron et al., 1975, 1977). Aggregated IgE induces the release of mediators of inflammation from rat (Dessaint et al., 1979b) and human M+ (Joseph et al., 1980,1983). IgE immune complexes stimulate rat (Rankin et al., 1982) and mouse (Rouzer et al., 1982a,b) M+ to produce leukotriene C, a component of the slow-reacting substance of anaphylaxis. Whether FceR on M 4 are specifically involved in this leukotriene formation has not been proven. Preliminary experiments performed in our laboratory did not show a correlation between leukotriene formation and numbers of FcER+ monocytes. Therefore, it remains to be shown whether IgE-mediated leukotrienes released from FcER+alveolar M+ is a pathogenic mechanism in atopic asthma. However, it is clear that IgE antibodies interact through FCERwith M+ and that this interac-

Fc

RECEPTORS FOR

IgE

85

tion must play an important role in host defense mechanisms, particularly against IgE-inducing parasites.

ACKNOWLEDGMEN’I’S Publication No. 3037 IMM from the Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, California 92037. The work was supported by U.S. Public Health Services Gr;unts AI-10734 and A1-15:350 and Riomctlical Hcsr;irch Snpport Program (:rant RRO-5514. The author is indebted to his collaborators who contributed to this work, Mrs. Phyllis Minick for editing the manuscript, and Mrs. Margaret Stone for excellent secretarial help.

REFEHENCES Anderson, C. L., and Spiegelberg, H. L. (1981).J. lmmunol. 126, 2470. Bazin, H., and Becker, A. (1976). In “Molecular and Biological Aspects of the Acute Allergic Reaction” ( S . G . 0. Johansson, K. Stranberg, and B. Uvniis, eds.), p. 125. Plenum, New York. Berman, M. A., and Weigle, W. 0. (1977).J . E x p . Med. 146, 241. Berman, M., Spiegelberg, H. L., and Weigle, W. 0. (1979).J. lmmunol. 122, 89. Boltz-Nitulescu, G., and Spiegelberg, H. L. (1981).Cell. Immunol. 59, 106. Boltz-Nitulescu, G., Bazin, H., and Spiegelberg, H. L. (1981).J. E x p . Med. 154, 374. Boltz-Nitulescu, G., Plunimer, J. M., and Spiegelberg, H. L. (1982).J. lmmunol. 128, 2265. Bottcher, I., Ulrich, M., Hirayama, N., and Ovary, Z. (1980). l n t . Arch. Allergy Appl. lmmunol. 61, 248. Brideau, R. J., Carter, P. B., Mason, D. W., McMaster, W. R., and Williams, A. F. (1980). Eur. J. lmmunol. 10, 609. Canonica, G. W., Mingari, M. C., Melioli, G., Colombati, M., and Moretta, L. (1979).J . lmmunol. 123,2669. Capron, A., Dessaint, J. P., Capron, M., and Bazin, H. (1975).Nature (London)253,474. Capron, A., Dessaint, J. P., Joseph, M., Ronsseaux, R., Capron, M., and Bazin, H. (1977). Eur. J. lmmunol. 7, 315. Capron, M., Capron, A,, Dessaint, J. P., Torpier, G., Johansson, S. G. O., and Prin, L. (1981).J. lmmunol. 126, 2087. Capron, M., Spiegelberg, H. L.;Bennich, H., Dessaint, J. P., Prin, L., and Capron, A. (1984).J . lmmutlol. 132, 462. Chen, S . S., Bohn, J. W., Liu, F. T., and Katz, D. H. (1981).J.lmmunol. 127, 166. Conrad, D. H., Bazin, H., Sehon, A. H., and Froese, A. (1975).J.lmmunol. 114, 1688. Denibo, M., Goldstein, B., Sobotka, A. K., and Lichtenstein, L. M. (1978).J. lmmunol. 121, 354. Dessaint, J. P., Torpier, G., Capron, M., Bazin, H . , and Capron, A. (197%). Cell. lmmunol. 46, 12. Dessaint, J. P., Capron, A,, Joseph, M.,and Bazin, H. (1979b). Cell. lmmunol. 46, 24. Finbloom, D. S . , and Metzger, H. (1982).J . lmmunol. 129,2004. Finbloom, D. S., and Metzger, H. (1983).J . Immunol. 30, 1489. Fishkin, B. G., Orloff, N., Scaduto, L. E., Borucki, D., and Spiegelberg, H. L. (1972). Blood 39, 361.

86

HANS L. SPIECELBERC

Fox, R. I., Thompson, L.F., and Huddlestone, J. R. (1981).J . lmmunol. 126,2062. Fritsche, R., and Spiegelberg, H. L. (1978).J . lmmunol. 121, 471. Gisler, R., and Fridman, W. (1975).J . E x p . Med. 142, 507. Gonzalez-Molina, A., and Spiegelberg, H. L. (1976).J . lmmunol. 117, 1838. Gonzalez-Molina, A., and Spiegelberg, H. L. (1977).J . Clin. lnuest. 59,616. Hallberg, T., Gurner, B. W., and Coombs, R. R. A. (1973). Int. Arch. Allergy A p p l . lmmunol. 44, 500. Hellstrom, U.,and Spiegelberg, H. L. (1979). Scand. 1. lmmunol. 9,75. Hempstead, B. L., Parker, C. W., and Kulczycki, A. (1979).J . Zmmunol, 123,2283. Hirashima, M., Yodoi, J,, and Ishizaka, K. (1980).J . lmmunol. 125, 1442. Hirashima, M., Yodoi, J., and Ishizaka, K. (1981).J . lmmunol. 126,838. Hirata, A. A., and Brandriss, M. W. (1968).J . lmmunol. 100,641. Hoover, R. G., Gebel, H. M., Dieckgraefe, B. K., Hickman, S., Rebbe, N. F., Hirayama, N., Ovary, Z., and Lynch, €3. G. (1981). lmmunol. Reo. 56, 115. Hiibscher, T. (1975). J . lmmunol. 114, 1379. Ishikawa, T., Wicker, K., and Arbesman, C. E. (1974). lnt. Arch. Allergy Appl. lmmunol. 46,230. Ishizaka, K., and Ishizaka, T. (1968).J.Immunol. 101,68. Ishizaka, K., and Ishizaka, T. (1978). lmmunol. Reo. 41, 109. Ishizaka, K., and Sandberg, K. (1981).]. lmmunol. 126, 1692. Ishizaka, K., Ishizaka, T., and Hornbrook, M. M. (1966).J.lmmunol. 97, 55. Ishizaka, K., Ishizaka, T., and Lee, E. H. (1970). Immunochemistry 7,687. Ishizaka, T., and Ishizaka, K. (1974).J. Immunol. 112, 1078. Ishizaka, T., and Ishizaka, K. (1978).J . lmmunol. 120,800. Ishizaka, T., and Levy, D. A. (1980). In “Adv. Allergology and Clin. Immunol.” (A. Oehling, I. Glazer, E. Mathov, and C. Arbesman, eds.), p. 585. Pergamon, Oxford. Ishizaka, T., Ishizaka, K., Salmon, S., and Fudenberg, H. (1967).J . lmmunol. 99,82. Jarrett, E., and Bazin, H. (1974). Nature (London)251,613. Joseph, M., Capron, A+,Butterworth, A. E., Suttrock, R. F., and Houba, V. (1978). Clin. E x p . Immunol. 33,48. Joseph, M., Tonnel, A. B., Capron, A., and Voisin, C. (1980). Clin. E x p . lmmunol. 40, 416. Joseph, M., Tonnel, A. B., Torpier, G., Capron, A., Arnoux, B., and Beneveniste, J. (1983).J . Clin. Invest. 71,221. Joskowicz, M.,Rabourdin-Combe, C., Neauport-Sautes, C., and Fridman, W. H. (1978). J . Immunol. 121, 777. Kagey-Sobotka, A,, Dembo, M., Goldstein, B., Metzger, H., and Lichtenstein, L. M. (1981).J . lmmunol. 127,2285. ’ Kamoun, M., Martin, P., Hauser, J., Brown, M., Siadak, A., and Nowinski, R. (1981). J . E x p . Med. 153, 207. Katona, I. M., Urban, J. F., Scher, I., Kanellopoulos-Langerin, C., and Finkelman, F. D. (1983).J . lmmunol. 136,350. Kerbel, R. S., and Davis, A. J. S. (1974). Cell 3, 105. Kishimoto, T., Hirai, Y., Suemura, M., Nakamishi, K., and Yamamura, Y. (1978). J . lmmunol. 121,2106. Kulczycki, A., and Metzger, H. (1974).J. E x p . Med. 140,1676. Lamers, M. C., Heckford, S . E., and Dickler, H. B. (1982). Nature (London) 298, 178. Lawrence, D. A,, Weigle, W. O., and Spiegelberg, H. L. (1975).J.Clin. Inuest. 55,368. Liu, F. T., Bohn, J. W., Ferry, E. L., Yamamoto, H., Molinaro, C. A,, Sherman, L. A,, Klinman, N. R., and Katz, D. H. (1980).J . lrnmunol. 124,2728.

Fc

RECEPTORS FOR

IgE

87

Lobo, P. I. (1981).J . Clin. Invest. 68, 431. Meinke, G . C., Magro, A. M., Lawrence, D. A,, and Spiegelberg, H. L. (1978).J . Zmmunol. 121, 1321. Melewicz, F. M., and Spiegelberg, H. L. (1980).J . Zmmunol. 125, 1026. Melewicz, F. M., Zeiger, R. S., Mellon, M. H., O’Connor, R. D., and Spiegelberg, H. L. (1981a). J . Immunol. 126, 1592. Melewicz, F. M., Zeiger, R. S., Mellon, M. H., OConnor, R. D., and Spiegelberg, H. L. (198lb). Clin. E x p . Immunol. 43, 526. Melewicz, F. M., Plummer, J. M., and Spiegelberg, H. L. (1982a)J. Zmmunol. 129,563. Melewicz, F. M., Kline, L. E., Cohen, A. B., and Spiegelberg, H. L. (1982b). Clin. E x p . Immunol. 49,364. Metzger, H., Budman, D., and Lucky, P. (1976). Zrnmunochemistry 13, 417. Metzger, H., Goetze, A., Kanellopoulos, J., Holowka, D., and Fewtrell, C. (1982). Fed. Proc. Fed. Am. Soc. Exp. Biol. 41,8. Moretta, L,, Ferrarini, M., Durante, M. L., and Mingari, M. C. (1975). Eur.J. Zmmunol. 5, 565. Morgan, E. L., and Weigle, W. 0. (1981).J . E x p . Med. 154, 778. Perlmann, H., Perlmann, P., Pape, G. R., and Hallden, G. (1976). Scund. J . Zmmunol. 5 (Suppl. 5), 57. Rankin, J. A., Hitchcock, M., Merrill, W., Back, M. K., Brashler, J. R., and Ashkenase, P. W. (1982). Nature (London)297,329. Reinherz, E., and Schlossmann, S. (1980). Cell 19,821. Rouzer, C. A., Scott, W. A., Hamill, A. C., and Cohn, Z. A. (198O).J.Exp. Med. 152,1236. Rouzer, C. A., Scott, W. A., Hamill, A. L., and Cohn, Z. A. (1982a).J.E x p . Med. 155,720. Rouzer, C. A., Scott, W. A., Hamill, A. L., Liu, F-T., Katz, D. H., and Cohn, Z. A. (1982b). J . E x p . Med. 156, 1077. Rudolph, A. K., Burrows, P. D., and Wable, M.R. (1981). Eur. J. Immunol. 11, 527. Ryan, J. L., Arbeit, R. D., Dickler, H. B., and Henkart, P. A. (1975).J.E x p . Med. 142,814. Schuster, D. C., Bongiovanni, B. A., Pierson, D. L., Barbaro, J. F., Wong, D. T. O., and Levinson, A. I. (1980).J . Zmmunol. 124, 1662. Segal, D. M., Sharrow, S. O., Jones, J. F., and Siraganian, R. P. (1981).J.Immunol. 126, 138. Sidman, C. L., and Unanue, E. R. (1976).J . E x p . Med. 144,882. Sinclair, N. R. St. C. (1969).J . E x p . Med. 129, 1183. Spiegelberg, H. L. (1981). Zmmunol. Rev. 56, 199. Spiegelberg, H. L., and Dainer, P. M. (1979). Clin. Exp. Immunol. 35,286. Spiegelberg, H. L., and Melewicz, F. M. (1980). Clin. Zmmunol. Zmmunopathol. 15, 424. Spiegelberg, H. L., and Simon, R. A. (1981).J. Clin. Invest. 68,845. Spiegelberg, H. L., O’Connor, R. D., Simon, R. A., and Mathison, D. A. (1979).J.Clin. Invest. 64, 714. Spiegelberg, H. L., Boltz-Nitulescu, G., Plummer, J. M., and Melewicz, F. M. (1983). Fed. Proc. Fed. Am. Soc. E x p . Biol. 42, 124. Sterk, A. R., and Ishizaka, T. (1982).J . Zmmunol. 128, 838. Strausbach, P., Sulica, A,, and Givol, D. (1970). Nuture (London)227, 68. Suemura, M., Yodoi, J., Hirashima, M., and Ishizaka, K. (1980).J . Zmmunol. 125, 148. Suemura, M., Shiho, O., Deguelis, H., Yamamura, Y., Bottcher, I., and Kishimoto, T. (1981).J . Immunol. 127,465. Thompson, L. F., Mellon, M. H.. Zeiger, R. S., iintl Spiegelberg, H. L. (1983).J. Zmmunol. 131. 2772.

88

HANS L. SPIEGELBERG

Uede, T., Hirata, F., Hirashima, M., and Ishizaka, K. (1983).J . Immunol. 130, 878. Van de Griend, R. J., Ten Berge, I., Tanke, H. J.,Roos, D., Schellekens, P. T. A., Melief, C. J. M., Zeijlemaker, W. P., and Astaldi, A. (1982).J . Immunol. 128, 1979. Vander-Mallie, R., Ishizaka, T., and Ishizaka, K. (1982).J . Zmmunol. 128,2306. White, R. A. H., Mason, D. W., Williams, A. F., Galfre, G., and Milstein, C. (1978). J . Exp. Med. 148, 664. Woda, B. A., Yguerabide, J., and Feldman, J. D. (1981).J . Cell Biol. 90, 705. Yodoi, J., and Ishizaka, K. (1979a). J . Immunol. 122, 2577. Yodoi, J., and Ishizaka, K. (197913).J . Immunol. 123, 2004. Yodoi, J., and Ishizaka, K. (1980a). J , Immunol. 124, 934. Yodoi, J., and Ishizaka, K. (1980b). J . Immunol. 124, 1322. Yodoi, J., Ishizaka, T., and Ishizaka, K. (1979). J . Zmmunol. 123, 455. Yodoi, J., Hirashima, M., and Ishizaka, K. (1980). J . Immunol. 125, 1436. Yodoi, J., Hirashima, M., and Ishizaka, K. (1982a).J. Immunol. 128, 289. Yodoi, J., Adadis, M., and Masuda, T. (198213).J . Immunol. 128, 888.

ADVANCES IN IMMUNOLOGY, VOL. 35

The Murine Antitumor Immune Response and Its Therapeutic M anipulatio n ROBERT J. NORTH Trudeau Institute, Inc., Saronac Lake, New Yark

I. Introduction.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Analysis of Antitumor Immunity by Adoptive Immunization against Established Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Tumor-Induced T Cell-Mediated Immunosuppression as a Barrier to Adoptive Immunotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. The Specificity of Adoptive Immunotherapy and Its Suppression . . . . . . C. Cyclophosphamide Facilitates Adoptive Immunotherapy of Tumors by Eliminating Suppressor T Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. y-Irradiation-Facilitated Adoptive Immunotherapy . . . . . . . . . . . . . . . . . . 111. The Meaning of the Adoptive Immunization Assay . . . . . . . . . . . . . . A. The Expression of Adoptive Immunity against an Established T Requires Generation of Cytolytic T Cells .......................... B. Suppressor T Cells Inhibit Cytolytic Response in the Recipient. . . . . . . C. Adoptive Immunotherapy with T Cells Generated in Vitro . . . . . . . . . . . D. Adoptive Immunization of Irradiated Tumor-Bearing Recipients as a Special Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. The Meaning of the Local Passive Transfer Assay. . . . . . . . . . . . . . . . . . . IV. Analysis of Concomitant Immunity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Paradox of Passive Transfer of Immunity against an Established Tumor with T Cells from a Donor with an Established Tumor . . . . . . . . B. Kinetics and Decay of Concomitant Immunity as Measured by Adoptive Immunization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. The Decay of Concomitant Immunity Is Associated with the Generation of Suppressor T Cells.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Tumor Immunotherapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Immunotherapy by Immunopotentiation ........................... B. Immunotherapy by Immunofacilitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Discussion, . . . . . . . . . . ......................................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....................

89 92 93 97 100 105

108 115 117 118 120 122 122

127 131 133 135 138 142 153

1. Introduction

The purpose of this article is to discuss the immune response to the progressive growth of transplantable immunogenic tumors in mice. The discussion will be built mainly around the results of experiments performed in this laboratory over the past 3 years, as well as around results of studies still ongoing. This is not to promote these results over the important results of others. It is because the discussion will 89 Copyright 8 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.

ISBN 0-12-022435-6

90

ROBERT J. NORTH

concentrate on an analysis of concomitant antitumor immunity, an aspect of tumor immunology that has received relatively little attention. It will be argued that a knowledge of concomitant immunity is a prerequisite for the formulation of a hypothesis to explain why tumors with transplantation rejection antigens are not rejected by their immunocompetent hosts, and why such tumors are, for the most part, refractory to attempts to cause their regression by immunotherapy. Tumor immunology began with the discovery by Foley (1953) that chemically induced, transplantable tumors can be immunogenic. H e showed that removal of a tumor by ligation leaves the host with a mechanism of acquired immunity to growth of an implant of cells of that tumor, but not other tumors. This finding was soon confirmed by others who showed that immunization against an implant of tumor cells can be achieved by surgical excision of a tumor, or by repeated injection of nonreplicating X-irradiated tumor cells. Additional confirmation of tumor immunogenicity was followed by investigations of the type of immunity that tumors evoke. In keeping with what was being discovered about antiallograft immunity, it was shown that immunity to syngeneic tumors can be passively transferred from tumorimmunized mice to normal recipients with lymphoid cells, but not with serum. This period of the history of tumor immunology has been comprehensively reviewed by a number of authors (Old et al., 1962; Old and Boyse, 1964; Sjogren, 1965; Klein, 1966; Hellstrom and Hellstrom, 1969; Herberman, 1974). It served to establish, on the one hand, that chemically induced tumors can be immunogenic, and to pose, on the other, the central paradox of tumor immunology: the progressive growth of an immunogenic tumor in its immunocompetent host. It was followed by attempts to analyze antitumor immunity in vitro, and by attempts to cause the regression of tumors, many of unproven immunogenicity, by immunotherapy. The in vitro studies 1973; Plata et ul., 1973; Burton et d., 1975; left no doubt (Rouse et d., Burton and Warner, 1977; Wagner et aZ., 1980a) that cocultivation of nonreplicating tumor cells with lymphoid cells from tumor-immunized mice results in the generation of T cells cytolytic for target tumor cells in vitro. It was also shown that these T cells are capable of neutralizing the growth of an implant of tumor cells in vivo, according to a local passive transfer assay (Burton and Warner, 1977). In contrast, except for rare examples, the numerous attempts to cause the regression of established tumors by immunotherapy have been unsuccessful. The general conclusion, therefore, is that immunogenic or not, most established transplantable tumors are refractory to attempts to cause their regression by treatment with immunoadjuvants with a

MURINE ANTITUMOR IMMUNE RESPONSE

91

known capacity to augment immune responses in general. The failure of a host to reject its immunogenic tumor, even after intratumor injection of agents that would be expected to enhance weak immune responses, prompted attempts to explain the escape of tumors on the basis of tumor-mediated avoidance mechanisms (Kamo and Friedman, 1977). It was suggested, on the basis of experimental evidence (Old et ul., 1968), for example, that tumors may escape immune destruction by hiding their surface antigens from a postulated mechanism of antitumor immunity. Another suggested mechanism of escape was that tumors secrete antiinflammatory factors (Pike and Snyderman, 1976) that serve to prevent the migration of host cells from blood to a site of extravascular tumor growth. Indeed, the need to find an explanation of tumor escape became more compelling with the numerous in uitro demonstrations that neoplastic cells can be destroyed, not only by cytolytic T cells, but also by macrophages (Fink, 1976) and N K cells (Herberman and Holden, 1978). Most of the earlier explanations of tumor escape appear to have been put aside, however, with the demonstrations in several laboratories (Eggers and Wunderlich, 1975; Fujimoto et ul., 1976; Small and Trainin, 1976; Reinisch et al., 1977) that mice bearing established tumors can generate suppressor T cells. These examples of the tumorinduced production of suppressor T cells are based on a variety of assays. They serve collectively to suggest the possibility that a host fails to generate a mechanism of immunity against its immunogenic tumor because of its possession of T cells with proven ability to suppress immune responses in general. It can be postulated, however, that suppressor T cells are a product of an immune response. Therefore, in order to determine their functional role in a tumor-bearing host, one would need to investigate the possibility that they are generated to “down-regulate” a preceding antitumor immune response. That such an immune response is generated in tumor-bearing mice is evidenced by examples of the paradoxical expression of specific immunity to a tumor implant by a host bearing a progressive tumor. These examples of concomitant antitumor immunity have been reviewed b y Gorelik (1984), who discusses evidence which shows that this immunity is T cell mediated, and that its failure to develop in immunodepressed mice and its decay in immunocompetent mice are associated with the emergence of tumor metastases. However, failure of the literature to draw any unifying conclusion about the significance of concomitant immunity, and the absence of discussion of concomitant immunity in most general reviews on tumor immunology, means that concomitant immunity has received very little experimen-

92

ROBERT J. NORTH

tal analysis. It will be argued here that an analysis of the generation and decay of concomitant immunity is essential for an understanding of why immunogenic tumors grow progressively in their immunocompetent hosts, and why these tumors fail to regress in response to active or passive immunotherapy. II. Analysis of Antitumor Immunity by Adoptive Immunization against Established Tumors

A review of the literature reveals (Rosenberg and Terry, 1977) that while it has proved relatively easy to transfer immunity passively to growth of a tumor implant, particularly if tumor cells and immune lymphocytes are admixed at the same site, according to the Winn assay, it has proved extremely difficult to demonstrate that passively transferred, tumor-sensitized T cells are capable of causing the regression of an established growing tumor. This situation might be interpreted as being in line with the argument that, because tumor-associated antigens are “weak antigens,” the difficulty encountered in demonstrating adoptive immunotherapy of established tumors is based on the weakness of the immunity that is passively transferred. However, this argument need not apply to adoptive immunization experiments, because it is possible to compensate for the weak immunity of the donor by increasing the number of its T cells infused. Indeed, it was shown a number of years ago (Borberg et al., 1972) that regression of a relatively large immunogenic fibrosarcoma could be achieved by the passive transfer of spleen cells from hyperimmunized donors, provided large numbers of donor spleen cells were repeatedly infused. Nevertheless, this and another demonstration of successful immunotherapy of established tumors are the exception rather than the rule. We are left with the options, therefore, of either accepting the evidence as indicating that adoptive immunotherapy of established immunogenic tumors will be difficult to achieve because of the “ weakness” of antitumor immunity, or searching for another explanation for the refractoriness of an established tumor to an infusion of tumor-sensitized lymphocytes. Before discussing the subject of adoptive immunotherapy of established tumors, however, it is important to define what is meant here when a tumor is referred to as being established. For the purpose of this article, an established tumor is one that is in the process of growing progressively. It will be recalled that when tumor cells are implanted intradermally or subcutaneously, there follows a period of latency before a palpable tumor emerges. The length of the period of

MURINE ANTITUMOR IMMUNE RESPONSE

93

latency is inversely related to the number of tumor cells implanted but has little effect on the rate of growth of the tumor that eventually emerges. After it emerges the tumor grows progressively, so that a straight-line relationship is obtained when increases in tumor weight, diameter, or thickness are plotted against time. After a certain stage, however, its rate of growth decreases because of the inadequacy of the blood supply to the center of the tumor, and this is associated with the development of a central core of necrosis. The notion that implanted tumor cells do not form a tumor until the site of implantation becomes adequately vascularized under the influence of tumor angiogenesis factor (Folkman and Cotrau, 1976) is undoubtedly important for explaining the latency of tumor cells implanted in avascular sites, like the cornea of the eye. However, it is doubtful if this applies to the early stages of growth of tumors implanted in highly vascularized tissues such as the skin. Indeed, it has been a common finding in this laboratory that tumors begin growing within 24 hours of being implanted in the skin or in a footpad, provided the implant is large enough. It is important to point out that the curve representing the regression of an established tumor, which is obtained by plotting decreases in mean tumor diameter against time, does not necessarily represent the curve for the destruction of tumor cells. Indeed, the meaning of tumor regression in terms of death of tumor cells versus resorption of dead tumor tissue needs to be determined. It should be stated here, however, that the curves for immunologically mediated tumor regression to b e described in the sections that follow can give a false impression of the duration of expression of immunity. It takes very little experience to be able to predict from the appearance of the tumor within 1-2 days of the onset of regression that it will go on to regress completely. In fact, the generalized necrosis that is apparent within 1-2 days of the onset of regression leaves little doubt that the curve for regression after this time mostly represents resorption of dead tissue and a scab.

A. TUMOR-INDUCED T CELL-MEDIATED IMMUNOSUPPRESSION AS A BARRIER TO ADOPTIVE IMMUNOTHERAPY In searching for an immunological reason for why it is difficult to demonstrate the expression of passively transferred immunity against an established tumor, the subject of concomitant antitumor immunity must be considered. For it stands to reason that if immunogenic tumors evoke the generation of a state of concomitant immunity in their hosts, then any attempt to immunize adoptively against these tumors

94

ROBERT J. NORTH

would represent an attempt to superimpose a donor immune response on an already ongoing recipient immune response. If anything, this situation should be conducive to the expression of adoptive immunity against the tumor, if it were not for the fact that concomitant immunity undergoes decay after the tumor grows beyond a certain size. Moreover, the decay of concomitant immunity must be considered in light of evidence that has been accumulating to show (Naor, 1979) that the growth of immunogenic tumors can cause the host to generate suppressor T cells. The possibility exists, therefore, that the eclipse of a concomitant immune response is an active process mediated by suppressor T cells. If this were the case, it would serve to explain the refractoriness of an established tumor to the antitumor action of passively transferred tumor-sensitized T cells, because any mechanism in a tumor-bearing recipient that suppresses the recipient’s own immune response would be expected also to suppress an immune response passively transferred from a donor. On the basis of the reasoning that a mechanism of T cell-mediated immunosuppression might function to “down-regulate” concomitant immunity, it was postulated that it should be possible to immunize adoptively against an established tumor growing in a recipient, provided the recipient is treated in a way that prevents it from generating concomitant immunity. It was known already (North and Kirstein, 1977) that mice made T cell deficient by thymectomy and lethal y-irradiation, and protected with bone marrow cells (TXB mice), are incapable of generating concomitant immunity to their immunogenic tumors. Therefore, the experiments to test the postulate involved infusing TXB recipient mice bearing the chemically induced Meth A fibrosarcoma with spleen cells from donors that were immunized against this tumor. The donors were immunized by causing the regression of their established Meth A tumors by p a r e n t e d injection of bacterial endotoxin (Berendt e t al., 197th) or by intralesional injection of Corynebacterium parvum (Dye et al., 1981).Details of these immunization procedures will be discussed in a later section. It is enough to point out at this stage that endotoxin-induced and C. paruum-induced tumor regression are associated with the emergence of a long-lived state of specific immunity to growth of a tumor implant. Mice so immunized were employed as donors after the complete regression of their tumors. The experiments showed (Berendt and North, 1980) that, whereas intravenous infusion of one organ equivalent (1.5 x 10’) of immune spleen cells failed to cause the regression of established Meth A tumors growing in immunocompetent recipients, the same number of

MURINE ANTITUMOR IMMUNE RESPONSE

95

immune spleen cells caused the complete regression of the samesized tumors growing in TXB recipients. This result was obtained with tumors growing intradermally in the belly region or in a hind footpad. Moreover, because tumor regression did not commence until about 6 days after adoptive immunization, the tumors had time to grow to a relatively large size before adoptive immunity was expressed. It was shown, in addition, that the donor cells that transferred immunity were T cells, as evidenced by the elimination of their antitumor function by treatment with anti-Thy-1.2 antibody and complement. These results were interpreted as meaning, therefore, that progressive growth of an immunogenic tumor results in the generation of a tumor-induced, T cell-dependent mechanism that inhibits the antitumor function of passively transferred tumor-sensitized T cells. It was next predicted that if this T cell-dependent barrier to adoptive immunotherapy exists in tumor-bearing immunocompetent mice, it should be possible to demonstrate its presence by showing that it can be passively transferred to TXB recipients with spleen cells from tumor-bearing donors, but not with spleen cells from normal donors. Its successful passive transfer would be revealed by its ability to block tumor regression caused by passive transfer of immune T cells. This prediction proved correct, as evidenced by the finding (Berendt and North, 1980) that intravenous infusion of spleen cells from donors with established tumors inhibited the capacity of immune spleen cells infused several hours before to cause regression of established tumors in TXB recipients. The cells that passively transferred this suppression were T cells, as evidenced by the demonstration that treatment with anti-Thy-1.2 antibody and complement abolished their capacity to inhibit the expression of adoptive immunity. In contrast, infusion of spleen cells from normal donors had no suppressive effect. These results provided enough reason to hypothesize, therefore, that progressive growth of the Meth A fibrosarcoma evokes in its syngeneic host the generation of a mechanism of T cell-mediated immunosuppression. This is in keeping with the additional finding that concomitant immunity to this tumor undergoes progressive decay after a relatively early stage of tumor growth. The results suggest that failures in the past to cause the regression of established tumors by the passive transfer of tumor-sensitized T cells were caused by a lack of awareness of the presence in a tumor-bearing recipient of a tumor-induced mechanism of T cell-mediated immunosuppression. It is important to point out, moreover, that these results with the Meth A fibrosarcoma have been confirmed by workers in another laboratory (Bonventre et al., 1982) who employed athymic, BALB/c nude mice as immunoin-

96

ROBERT J. NORTH

competent tumor-bearing test recipients, and heterozygous littermates ad donors of immune cells and suppressor cells. It was important to determine next, however, whether the results obtained with the Meth A fibrosarcoma were peculiar to this tumor, or whether immunosuppression was likely to be a common consequence of the growth of most immunogenic tumors. Therefore, the same experiments were performed with the P815 mastocytoma, which is generally considered to possess weak immunogenicity. Moreover, unlike the Meth A fibrosarcoma, the P815 mastocytoma possesses the ability to metastasize to distant lymph nodes and to the liver and spleen. It provided an opportunity, therefore, to examine the effect of passively transferred, sensitized T cells on tumor metastases. The experimental results obtained with the P815 mastocytoma were essentially the same as those obtained with the Meth A fibrosarcoma. It was shown (Dye and North, 1981) that infusion of tumor-sensitized spleen cells from appropriately immunized donors caused the complete regression of P815 tumors growing in TXB recipients, but not in normal recipients. Again, the capacity of tumor-sensitized cells to cause tumor regression in TXB recipients was inhibited by infusion of spleen cells from donors with well-established tumors. The cells that passively transferred immunity, as well as those that inhibited the expression of this immunity, were destroyed by treatment with anti-Thy-1.2 antibody and complement. The results with the P815 mastocytoma, therefore, are consistent with the hypothesis that progressive growth of an immunogenic tumor results in the generation of a mechanism of T cell-mediated immunosuppression. Subsequent studies with another tumor, the P388 lymphoma (Dye and North, 1983), have given essentially the same results. As for employing the P815 tumor to study the feasibility of adoptively immunizing against tumor metastases, it was first necessary to show that tumor metastases are seeded before adoptive immunization is performed. Determining the stage of intradermal tumor growth at which metastases are seeded simply involved determining the survival time of mice that had their primary tumor removed at different stages of tumor growth. The results of this study revealed that immunocompetent mice could be cured if their primary tumor was excised before days 8-9 of growth. These mice lived beyond a 60-day period of observation. Excising the tumor after this time resulted in the growth of lymph node, spleen, and liver metastases and the death of all mice in 26 days. The results with TXB mice were different, in that they showed that metastases were already seeded in 80%of TXB mice by day 3 of tumor growth, and in 100% of them by day 6. It

MURINE ANTITUMOR IMMUNE RESPONSE

97

should be pointed out, in this connection, that the adoptive inimunization experiments with the P815 tumor just described showed that passive transfer of immune T cells on clay 4 of tumor growth cured 100% of TXB mice. Beca~isethe adoptive immunity was not expressed until about 6 days after passive transfer, it is almost certain that adoptive immunization destroyed metastases that already were seeded in the TXB recipients. Direct evidence for this was obtained b y determining whether TXB mice which had their tumors surgically removed on day 6 of tumor growth, and which were destined to die of nietastatic disease, could be rescued by an intravenous infusion of tumorsensitized T cells. It was found (Dye and North, 1981) that, whereas an infusion of normal spleen cells failed to prevent any of the TXB recipients from dying with a medium survival time of 19 days, an infusion of immune spleen cells enabled all recipients to live for at least the 60-day period of observation. Therefore, adoptive immunization was successful at destroying tumor metastases. The much earlier seeding of metastases in TXB mice is consistent with the interpretation of others (Gorelik, 1983) that concomitant immunity serves to retard the growth of metastases.

B. THE SPECIFICITY OF ADOPTIVE IMMUNOTHERAPY AND ITS SUPPRESSION There is a large body of evidence to show that immunity to growth of a tumor implant in immunized animals is specific. This has proved to be the case in studies that employed tumors from the same mouse strain, induced in the same tissue, with the same carcinogen-a finding that argues against the role of common virus-coded antigens in tumor immunogenicity. Exceptions to the general finding should not be surprising, however, because there are no rules against tumors sharing cross-reacting immunogens. Indeed, experiments designed to test the specificity of antitumor immunity that already is known to be mediated by sensitized T cells are really designed to test for crossreacting immunogens. It was anticipated on the basis of the bulk of the evidence that the chances were good that the expression of adoptive immunotherapy against any two immunogenic tumors chosen for study would prove to be specific. However, assuming that the suppression of this adoptive immunotherapy also would be specific might not be justified, on the grounds of a literature that shows (Naor, 1979) that suppression can be expressed nonspecifically by T cells in witro, and that macrophages also can suppress immune responses nonspecifically in uitro. It is essential, therefore, to provide formal evidence that T cell-mediated suppression of passively transferred antitumor immu-

98

ROBERT J. NORTH

nity is specific. Such evidence is needed in order to formulate a hypothesis to explain the mechanism of suppression. Specificity of suppression would mean that suppressor T cells and effector T cells are generated in response to the same antigenic determinants. This would suggest, in turn, that the induction and production of suppressor cells in response to an expanding population of effector T cells, under conditions of an increasing load of tumor antigen, might be explained in terms of idiotype-anti-idiotype interactions of the type postulated to explain T cell-mediated suppression of delayed sensitivity to certain haptens (Clamen et al., 1980; Greene et al., 1982). Therefore, experiments were performed first to determine whether the expression of passively transferred immunity against tumors growing in TXB recipients is specific, and second, to determine whether the suppression of this immunity by suppressor T cells from tumor-bearing donors also is specific. Specificity was tested with two DBAI2 tumors, the P815 mastocytoma and the P388 lymphoma. Donors of immune cells were immunized to one or the other of the tumors by causing regression of the tumors by intratumor therapy with C . parvum (Dye et al., 1981).Donors of suppressor cells were immunocompetent mice bearing 12- to 15-day P815 or P388 tumors intradermally. It was found (Fig. 1) that the expression of passively transferred immunity against an established tumor was specific, in that intravenous infusion of P815-sensitized splenic T cells caused regression of the P815 tumor but not the P388 tumor in TXB recipients. Conversely, P388-sensitized T cells caused regression of the P388 tumor but not the P815 mastocytoma. T cell-mediated suppression of this adoptive immunity also proved to be specific. It was demonstrated that intravenous infusion of spleen cells from donors bearing an established P815 tumor inhibited adoptive T cell-mediated regression of the P815 tumor, but not the P388 tumor i n TXB recipients. In the opposite direction, splenic T cells from P388 tumor bearers only inhibited adoptive T cell-mediated regression of the P388 tumor. Similar results were obtained with reciprocal passive transfer with the P815 tamor and L5178Y lymphoma (North et al., 1982). It should be mentioned, however, that this particular experimental result is not as easy to reproduce to completion as might be indicated by Fig. 1. The reason for this is that both the P815 and P388 tumors metastasize to other organs including the spleen after about day 10 of tumor growth in normal mice, and much earlier in TXB mice. Consequently, although the proportion of tumor cells to host cells is relatively small in the spleens of donors of suppressor T cells, the few tumor cells that are present are free to grow unrestrictedly when infused into a TXB tumor bearer that

99

MURINE ANTITUMOR IMMUNE RESPONSE A

0

10

20

30

40

0

10

20

30

40

DAYS OF TUMOR GROWTH

FIG.1. Evidence that the T cells from immune donors that cause tumor regression in TXB recipients, as well as the suppressor T cells from tumor-bearing donors that inhibit this tumor regression, are specific for the tumor that evokes their generation. A P815 tumor (A) was caused to undergo regression only by P8l5-sensitized T cells, but not the P388-sensitized T cells. Reciprocally, a P388 tumor (B) was caused to undergo regression only by P388-sensitized T cells. The T cell-mediated suppression of adoptive immunity similarly was specific in a reciprocal manner. Means of five mice per group.

has been adoptively immunized against the other tumor. This can result, in some experiments, in the recipients dying of systemic disease from the control tumor while they are still in the process of vigorously rejecting their experimental tumor. Even in those cases where this happens, however, it is obvious that the expression and suppression of adoptive imniuiiotherapy is specific. Other tumors are currently being examined in this laboratory for the purpose of avoiding the inconvenience of systemic disease. The idea that contaminating tumor cells are responsible for the observed suppression needs to be kept in mind but is highly unlikely for several reasons. First, suppression has been demonstrated with the Meth A fibrosarcoma, a tumor that is notoriously nonmetastatic. Second, additional experiments that were part of the foregoing specificity study (Dye and North, 1983) show that adoptive immunity against an established P815 tumor is not suppressed by intravenous infusion of 5 x lo5 replicating P815 tumor cells-a number that is capable of killing TXB mice in less than 2 weeks. Third, the suppressors of inimunity to the Meth A fibrosarconia and P815 mastocytoma are susceptible to treatment with anti-Thy-1.2 antibody and complement,

100

ROBERT J. NORTH

whereas the cells of these tumors are resistant to this treatment. In addition, subsequent studies have shown (Mills and North, unpublished) that the T cell suppressors of adoptive immunity are destroyed by treatment with anti-Ly-1 antibody and complement. Even so, the possible suppressive influence of contaminating tumor cells must always be kept in mind when examining new tumor models of adoptive immunity and its suppression. The possible influence of contaminating antigen also should be kept in mind in interpreting the results of in vivo models of suppression of delayed-type hypersensitivity (DTH) to haptens, where the donors of suppressor T cells are mice that have been infused up to a week earlier with a large number of haptencoupled syngeneic lymphocytes. The possibility that appreciable quantities of hapten are carried over with suppressor T cells needs to be investigated.

FACILITATES ADOPTIVEIMMUNOTHERAPY OF C. CYCLOPHOSPHAMIDE TUMORS BY ELIMINATING SUPPRESSOR T CELLS An impressive amount of published evidence exists to show that pretreating mice with the alkylating agent, cyclophosphamide, can enhance cell-mediated immune responses. This has been shown to be the case for the generation of delayed-type hypersensitivity to certain antigens (Goto et ul., 1981) and for the generation of T cells cytolytic for allogeneic cells (Rdlinghoff et al., 1977) and for SV40-induced syngeneic tumor cells (Glaser, 1979). The general finding has been that cyclophosphamide needs to be given just before giving the antigen. In some cases, moreover, it was demonstrated that the immunoaugmenting effect of cyclophosphamide could be cancelled by infusing the cyclophosphamide-treated animal with T cells from normal donors (Glaser, 1979). This type of result was interpreted as meaning that treatment with cyclophosphamide preferentially eliminates suppressor T cells that function to “down-regulate” cell-mediated immune responses. As far as using this type of pretreatment to show augmented immunity against a tumor implant is concerned, unpublished experiments with the immunogenic tumors employed in this laboratory have failed to show that pretreatment with cyclophosphamide has any detectable effect on the growth of tumor implants given 1-2 days later. Therefore, if cyclophosphamide pretreatment does cause an augmented concomitant immune response to implants of these tumors, the level of immunity generated is not sufficient to alter the rate of tumor growth. Therefore, this type of pretreatment study was not pursued in this laboratory to analyze antitumor immunity. Instead, studies were designed to determine whether cyclophos-

MURINE ANTITUMOR IMMUNE RESPONSE

101

phamide treatment of a tumor-bearing animal would facilitate the expression of antitumor immunity passively transferred with sensitized T cells. However, the rationale for treating a tumor-bearing recipient with cyclophosphamide to facilitate the expression of passively transferred immunity was not based on the belief that cyclophosphamide would eliminate suppressor T cells preferentially. On the contrary, it was based on the simple rationale that cyclophosphamide treatment, by suppressing the generation of concomitant immunity, would prevent the generation of suppressor T cells that “down-regulate” concomitant immunity, and thereby enable passively transferred immune T cells to express their antitumor function. It goes without saying that a model of adoptive immunotherapy that employs cyclophosphamidetreated tumor bearers is therapeutically more appealing than one that employs TXB tumor bearers. In designing the experiments to study the predicted facilitating effect of cyclophosphamide on the expression of adoptive immunity against established tumors, it was desirable to choose a tumor that is relatively resistant to the direct cytotoxic action of‘ the drug. Otherwise, the experiments merely would measure the ability of passively transferred T cells to prevent the regrowth of a tumor that already has been destroyed almost completely by the drug. Preliminary studies revealed that the Meth A fibrosarcoma growing either intradermally or in a hind footpad, is relatively resistant to a single 100 mg/kg dose of the drug, in that the tumor does not undergo regression but stops growing for several days before resuming its normal rate of growth. The basic experiment to test whether cyclophosphamide treatment would facilitate adoptive immunotherapy of this tumor consisted of treating mice bearing an established Meth A tumor with a 100 mg/kg dose of the drug, and infusing them 1 hour later with 1.5 X lo8 spleen cells from Meth A-immune donors. Because the drug is rapidly converted to the active form and has a half-life in mice of about 20 minutes, it has no effect on spleen cells infused 1 hour later. It was found (North, 1982) that combination therapy with cyclophosphamide and immune spleen cells caused the complete regression of tumors in all mice. In contrast, cyclophosphamide alone caused only a temporary halt in tumor growth, whereas immune spleen cells alone had no effect on tumor growth at all. This result shows, therefore, that cyclophosphamide facilitates the expression of adoptive immunity by eliminating a cyclophosphamide-sensitive mechanism from the tumor-bearing host that functions to inhibit the antitumor function of passively transferred immune spleen cells. If this is so, it should be possible to restore this suppressor mechanism

102

ROBERT J. NORTH

to the cyclophosphamide-treated tumor bearers by infusing them with spleen cells from donors with established tumors, but not with spleen cells from normal donors. This was found to be the case, as evidenced by experimental results which showed that regression of the Meth A fibrosarcoma that followed combination therapy with cyclophosphamide and intravenous infusion of immune spleen cells, was completely inhibited by intravenous infusion 1 hour later of 1.5 x lo8 spleen cells from mice with established Meth A tumors, but not by the same number of spleen cells from normal donors. In keeping with the results obtained with TXB mice, the immune spleen cells that caused tumor regression in cyclophosphamide-treated recipients, as well as the spleen cells from tumor-bearing donors which inhibited this tumor regression, were T cells, as evidenced by their susceptibility to treatment with anti-Thy-1.2 antibody and complement. Therefore, the evidence is consistent with the hypothesis that cyclophosphamide facilitates the expression of passively transferred immunity against an established tumor by eliminating a tumor-induced population of suppressor T cells. It was shown, in support of this hypothesis, that the splenic T cells that cause the suppression of adoptive antitumor immunity were destroyed by treating the tumor-bearing donors of suppressor cells with the same dose of cyclophosphamide that was employed to facilitate the expression of adoptive immunity in tumor-bearing recipients. An element of selectivity of the drug for suppressor T cells was suggested by the additional finding that, unlike suppressor T cells, the immune T cells that were employed routinely to transfer immunity passively proved totally resistant to a 100 mg/kg dose of cyclophosphamide. It is known, however, that these are memory T cells (see later) and that they are different from the effector T cells generated as part of the concomitant immune response. In fact, cyclophosphamide treatment of Meth A tumor bearers is immunosuppressive, because it results (North, unpublished) in a rapid loss of concomitant immunity to growth of a Meth A challenge implant. This is in keeping with results published from another laboratory (Glaser, 1979) which show that a 100 mg/kg dose of cyclophosphamide abridges an ongoing cytolytic T cell response. It is also in keeping with the results of additional experiments, to be discussed later, which show that a 100 mg/kg dose of cyclophosphamide destroys T cells cytolytic for the P815 mastocytoma. Even so, if the precursors of suppressor T cells are regenerated at a slower rate than the precursors of effector T cells after cyclophosphamide treatment, the possibility would exist that the therapeutic

MURINE ANTITUMOR IMMUNE RESPONSE

103

effect of this drug, in many cases, might be attributed to its ability to enhance antitumor immunity by delaying the suppression of concomitant immunity. Indeed, that host immunity contributes to tumor regression that follows treatment with cyclophosphamide, or other chemotherapeutic agents has been postulated repeatedly over a number of years. For example, an increased therapeutic effect of cyclophosphamide in terms of tumor re-emergence has been observed in animals preimmunized against tumors (Moore and Williams, 1973; Mathe et d.,1977; Chassoux et al., 1978), and in animals that receive immune cells (Fefer et al., 1976; Greenberg et al., 1980). A possible enhancing effect of cyclophosphamide on concomitant immunity is seen in convincing results published by Mokyr et al. (1982), which show that the therapeutic action of the drug against a syngeneic plasmacytoma requires that the tumor be relatively large, in that small tumors are refractory. The interesting additional finding that a small tumor could be made responsive to cyclophosphamide by the presence contralaterally of a larger tumor strongly suggests that the therapeutic action of the drug depends on an underlying concomitant immune response. The dose of cyclophosphamide required to facilitate adoptive immunotherapy was determined by infusing tumor bearers with 1.5 x loRimmune spleen cells 1 hour after giving them graded doses of the drug intravenously. It was found (Fig. 2) that the dose required to facilitate the expression of adoptive immunity was 100 mg/kg or more. A dose of 50 mg/kg or less was without effect. This is partly in keeping with the pretreatment dose required (Glaser, 1979) to give maximal augmentation of cytolytic T cell production in response to an implant of cells of an SV40-induced tumor in syngeneic mice. It is larger, however, than the dose required to eliminate suppressor T cells selectively in certain other models of T cell-mediated immunosuppression (Greene, 1980). It also is important to know the duration of the facilitating action of cyclophosphamide, because a long period of facilitation would enable tumor-sensitized T cells to be infused repeatedly and to function in an additive way against the tumor. The duration of facilitation was investigated by giving a 100 mg/kg dose of cyclophosphamide on day 4 of growth of a Meth A fibrosarcoma and determining how long after giving the drug an infusion of tumor-sensitized T cells would cause tumor regression. The results shown in Fig. 3 reveal that the facilitating action of the drug lasted for about 2 days. It should be realized, however, that because the tumor was larger at each progressive time

104

ROBERT J. NORTH

70 0

E E

0

60-

IMM.

+ 25mg/kg

IMM.

+ 50mgIkg

TUMOR 50-

U 0

I 40-

+3 $

30-

W

N

a 20-

0

4

8

12

16

20

DAYS

FIG. 2. Cyclophosphamide-facilitated expression of adoptive T cell-mediated regression of an established intrafootpad Meth A tumor. The expression of adoptive immunity was facilitated by giving 100 mg/kg of the drug, but not by giving 50 mg/kg or less. In this experiment, cyclophosphamide was given 1 hour before infusing immune spleen cells (IMM, 1.5 x loR)on day 4. Means of five mice per group.

interval tested, immune cells given at these times were faced with a larger tumor burden to reject. Even so, it is known from the results of other experiments (North, 1982) that the number of immune spleen cells employed was capable of regressing tumors larger than the day 8 tumors in this experiment, provided that cyclophosphamide was given 1 hour before the immune cells. It seems safe to conclude, therefore, that a 100 mg/kg dose of cyclophosphamide retards the generation of suppressor T cells in the tumor-bearing recipient for a period of 2 days, plus the amount of time that has to elapse before the passively transferred immune T cells cause the onset of tumor regression. The reason for the period of delay before tumor regression commences will be dealt with in a later section, where it will be argued that it represents the time needed for the passively transferred immune T cells to give rise to the generation of a state of active immunity in the recipient-an event that is highly sensitive itself to cyclophosphamide.

MURINE ANTITUMOR IMMUNE RESPONSE

70

-

60

-

105

E 50c

0 pe

40-

0

zi

3 I-

30-

LL

0 W

y

20-

u)

10

0

4

I

I

I

1

8

12

16

20

I 24

DAYS

FIG.3. Evidence that immunofacilitating effect of cyclophosphamide requires that the drug be given no more than 2 days before immune spleen cells. In this experiment, cyclophosphamide was given on day 4 of intrafootpad tumor growth, and immune spleen cells (1.5 x loH)infused 1 hour, or 2, 4, or 6 days later. Means of five mice per group.

D. ~-IRRADIATION-FACILITATED ADOPTIVEIMMUNOTHERAPY The results obtained with tumor-bearing TXB recipients, as well as those obtained with tumor bearers treated with cyclophosphamide, would suggest that any treatment that suppresses concomitant antitumor immunity and its T cell-mediated suppression should enable passively transferred tumor-sensitized T cells to cause tumor regression. If this prediction were true, it would follow that ionizing radiation, because of its immunosuppressive action, should also facilitate the expression of adoptive T cell-mediated immunity against an established tumor. This prediction was tested with four different immunogenic tumors: the P815 mastocytoma and P388 lymphoma (DBA/S), the SA1 sarcoma (A/J), and the Meth A fibrosarcoma (BALBlc). The experiments consisted of exposing mice bearing 4-day tumors to 500 rad of y-irradiation 1 hour before infusing them with spleen cells from immune donors. The donors were immunized in the standard way by

106

ROBERT J. NORTH

causing regression of their tumors with intralesional C. ~ U W therapy. It was found (North, 1984) with all four tumors (Fig. 4)that y-irradiation followed 1 hour later by intravenous infusion of 1.5 x lo8 immune spleen cells resulted, after about a 6-day delay, in complete regression of tumors in all mice. In contrast y-irradiation alone caused only a temporary reduction in the rate of tumor growth, whereas immune cells alone had even less effect. Because irradiation alone had only a marginal effect on tumor growth, it can be concluded that most of the tumor regression that occurred was immunologically mediated by the passively transferred, tumor-sensitized T cells. It is obvious, therefore, that y-irradiation eliminated from the tumor-bearing recipients a radiosensitive mechanism that functions to inhibit the antitumor action of passively transferred immune T cells. Evidence that this immunofacilitating action of y-irradiation depends on its ability to eliminate suppressor T cells was obtained by further analysis with the Meth A tumor. It was found (North, 1983) that tumor regression caused by combination therapy with y-irradiation and immune spleen cells could be prevented by intravenous infusion several hours later of A

B

70-

3

0 5 10 15 20 DAYS FIG.4. Evidence that 500 rad of y-radiation on day 4 of growth of an intrafootpad (A) Meth A fibrosarcoma or (B) SA1 sarcoma enabled 1.5 x 10*immune spleen cells infused 1 hour later to cause complete tumor regression. The same result was obtained with two other tumors (not shown). Means of five mice per group. 0

5

10

15

20

U

~

MURINE ANTITUMOR IMMUNE RESPONSE

107

spleen cells from donors with well-established Meth A tumors. Moreover, the cells from tumor-bearing donors which inhibited adoptive immunotherapy of the Meth A fibrosarcoma were T cells, as evidenced by their susceptibility to treatment with anti-Thy-1.2 antibody and complement. Just as important was the additional finding that these suppressor T cells were radiosensitive, in that they were eliminated from the spleens of tumor-bearing donors b y the same dose of‘ y-irradiation employed to facilitate adoptive immunotherapy in tumor-bearing recipients of immune T cells. It should be pointed out that these experiments with y-irradiation were not planned on the assuniption that y-irradiation would facilitate adoptive imniunotherapy by destroying suppressor T cells, or their precursors, selectively. However, additional findings, to he discussed later, indicate that this may well be the case. 111. The Meaning of the Adoptive ImmunizationAssay

Adoptive immunization has been one of the most powerful analytical tools employed in immunology. It is a physiological assay which, although reductionistic, is not as reductionistic as many of the in vitro assays on which a lot of modern immunology is based. It was responsible for distinguishing between cellular and humoral immunity, and for showing that lymphocytes are the central participants in immune responses. The assay is based on the simple rationale that, since acquired immunity to a given antigen is based on an acquired population of antigen-sensitized lymphocytes, it should be possible to harvest these lymphocytes and to use them to transfer immunity passively to a normal recipient. In practice, however, the assay is not as simple as this, because in most cases immunity is difficult to transfer passively unless the recipients are immunodepressed by X-irradiation, or by some other procedure that causes a depletion of lymphocytes. This radiosensitive barrier to adoptive immunization has been referred to as the “isogenic barrier,” and it has yet to be identified and explained. In the meantime, it will continue to be ignored by those who employ irradiation to demonstrate adoptive immunization successfully. The isogenic barrier to adoptive immunization is least problematic in those cases where the recipient is challenged with antigen either immediately or soon after adoptive immunization is performed. This is the case, for example, with most published studies of the passive transfer of delayed-type hypersensitivity and of T cell-mediated antibacterial immunity. The problem of the isogenic barrier might eventually need to be dealt with if a complete understanding of

108

ROBERT J. NORTH

the mechanisms of expression and suppression of adoptive antitumor immunity is to be gained. A. THEEXPRESSION OF ADOPTIVE IMMUNITY AGAINST AN ESTABLISHED TUMOH REQUIRES GENEHATION OF CYTOLYTIC T CELLS Aside from the isogenic barrier, another common unknown in many published studies of adoptive immunization is the type of donor immunity that is passively transferred. In most cases, moreover, an absence of information about the type of donor immunity is coupled with a disregard for the events that need to take place in the recipient itself before immunity can be expressed. As far as the type of donor immunity is concerned, it is surely realized that the generation of T cell-mediated immunity results sequentially in two different states of immunity: a short-lived state of active immunity and a state of immunological memory. Active immunity is characterized by the presence of effector T cells that disappear when the antigen that evokes their generation is eliminated. The host is then left with a state of immunological memory that enables it to regenerate effector T cells in an accelerated manner when it encounters antigen at a later time. Obviously, the outcome of an adoptive immunization experiment that involves an attempt to cause the rejection of a vascularized allograft or tumor syngraft will depend on whether the recipient is infused with active effector cells or memory cells. It stands to reason, for example, that if a donor with a state of immunological memory needs to regenerate a population of effector T cells itself, in order to express immunity against a second encounter with an allograft or a syngeneic tumor, then, so too would the recipient of the memory T cells from this donor need to generate a population of effector T cells. If so, adoptive immunity should not be expressed in recipients of memory T cells until after a delay. Indeed, an examination of the literature reveals that there is commonly a considerable period of delay, sometimes more than 2 weeks, after the passive transfer of immunity to allografts and to tumor syngrafts before the rejection of the target graft commences. It will be recalled, in this connection, that it was stated in the preceding sections that the expression of passively transferred antitumor immunity against established tumors growing in TXB or y-irradiated recipients does not begin until 6-8 days after tumor-sensitized T cells are infused intravenously. There may be a number of ways to explain this delay, but the most likely explanation, and the one worth testing, is that it represents the time needed for passively transferred helper or

MURINE ANTITUMOR IMMUNE RESPONSE

109

memory T cells to give rise to the production in the recipient of a large enough number of cytolytic effector T cells to destroy the recipient's tumor. Needless to say, it is essential to know if the recipient needs to generate an adoptive, secondary cytolytic T cell response to reject its tumor, in order to understand how suppressor T cells inhibit the expression of this adoptive immunity. Experiments were performed, therefore, to determine whether the expression of adoptive immunity against an established tumor growing in TXB recipients is preceded by a cytolytic T cell response. The P815 mastocytoma was employed in this study because of its suitability as a target for the in oitro 51Cr release assay. The recipients were TXB mice bearing 3-day footpad tumors. They were infused intravenously with 1.5 x loHspleen cells from donors immunized against the tumor by C. parvum therapy. The experiment involved following the production, in the node draining the tumor, of T cells capable of lysing P815 target cells in oitro according to a 6-hour "Cr release assay. It was found (Mills and North, 1983), in agreement with the results of previous experiments, that the passive transfer of P815-sensitized T cells resulted, after an 8-day delay, in the onset of progressive tumor regression in TXB recipients, but not in normal recipients. It was found, in addition, that the onset of tumor regression in TXB recipients was immediately preceded by a substantial production of cytolytic T cells in the draining lymph node (Fig. 5). According to the 51Crrelease assay, cytolytic T cell production did not begin until 4 days after passive transfer of sensitized T cells, and peaked 3 days later at the time of onset of tumor regression. The response then underwent rapid decay. It is important to realize, moreover, that a cytolytic T cell response also was generated in immunocompetent recipients of immune T cells, but it was of much lower magnitude, and this was in keeping with the failure of these recipients to cause the regression of their tumors. Furthermore, a similar lowmagnitude cytolytic response was generated in TXB recipients of normal spleen cells, whereas no cytolytic response at all was generated in TXB tumor-bearing controls. Taken as a whole, these results are consistent with the interpretation that the expression of passively transferred T cell-mediated immunity against established tumors growing in TXB recipients does not begin until an adequate number of cytolytic T cells is generated in the recipients. This does not represent proof that cytolytic T cells are the ultimate effectors of antitumor immunity, but the evidence strongly suggests that these cells are essential participants in the expression of immunity. This is particularly so, in view of the fact that the sensitized donor T cells infused intravenously had no cytolytic activity of their own in oitro, and possessed

110

ROBERT J. NORTH A

1

B

D A Y S 01 TUMOR G R O W T H FIG.5. Evidence that the onset of regression of the P815 mastocytoma in T cell-

deficient recipients of 1.5-2 x 1oX immune spleen cells from P815-immune donors is preceded by the production, in the recipients’ draining lymph node, of T cells cytolytic for P815 tumor cells in uitro.Failure of TXB recipients of normal spleen cells (A), and of normal recipients of immune spleen cells (B), to destroy their tumors was associated with a cytolytic T cell response of much lower magnitude.

MURINE ANTITUMOR IMMUNE RESPONSE

111

no immediate capacity to cause tumor regression in the TXB recipients. In contrast, the cytolytic T cells generated in the TXB recipients where Ly-2+ T cells that lysed the P815 in the 51Cr release assay in a specific manner. They were capable, in addition, of neutralizing the growth of P815 tumor cells according to the Winn assay in normal recipients. Even so, proof that these cytolytic T cells are the effectors of tumor regression has yet to be supplied. In the meantime, it is necessary to consider recently published evidence that has been interpreted as showing that cytolytic T cells are not the effectors of antiallograft and antitumor immunity. These published studies show that adoptive immunity to tumor allografts in TXB mice (Loveland et al., 1981), and adoptive immunity to skin allografts (Dallman et al., 1982) and virus-induced syngeneic tumors (Fernandez-Cruz et al., 1982) in TXB or irradiated recipient rats, can be passively transferred with memory or helper T cells, at the exclusion of cytolytic T cells and their precursors. For example, it was shown in the case of antiallograft immunity in mice that immunity can be passively transferred with T cells enriched for the Ly-1+2- phenotype by treatment with anti-Ly-2 antibody and complement. In the case of the experiments with rats, the passive transfer of immunity to skin allografts was denionstrated with W3/25-positive helper T cells at the exclusion of OX-8-positive cytolytic T cell precursors. In the case of the rat allograft study, moreover, infusion of highly purified OX-8positive cytolytic T cell precursors failed to cause rejections of the recipient’s graft. This latter finding is in keeping with the demonstration (Dallman et al., 1982) that helper T cells generated in vitro which could passively transfer immunity against a syngeneic virus-induced rat tumor growing in sublethally irradiated recipients were destroyed by treatment with anti-W3/25 antibody and complement. In all of the cases, therefore, immunity was passively transferred with T cells displaying the surface phenotype of helper T cells rather than cytolytic T cells. Needless to say, this type of evidence could be used to argue against cytolytic T cells as being the effectors of allograft and tumor rejection (Howard and Butcher, 1981). I n bet, it has been used as the basis for postulating (Loveland and McKenzie, 1982) that allografts and tumors are rejected, not by cytolytic T cells, but by a delayed-type hypersensitivity reaction mediated by Ly-l+ T cells in the graft. The postulate relies on the belief, however, that the surface Ly phenotypes of T cells are stable differentiation markers, and that TXB recipients are incapable of generating cytolytic T cells of their own. So far as the latter belief is concerned, there is ample evidence to the contrary. It has been demonstrated convincingly that TXB mice (Duprez et al.,

112

ROBERT J. NORTH

1982), as well as nude mice (Gillis et ul., 1979; Wagner et ul., 1980a), possess cytolytic T cell precursors that can differentiate into cytolytic effector cells in response to antigen in the presence of interleukin-2. Because Ly-l+ helper T cells secrete interleukin-2 on stimulation with specific antigen, the possibility that the passive transfer of these T cells into antigen-stimulated TXB recipients will cause the generation of cytolytic T cells from cytolytic T cell precursors of the recipient seems highly likely. It would seem essential, therefore, that before cytolytic T cells are discounted as the effectors of allograft and tumor rejection, the presence of these cells should be looked for in the TXB recipients at the time their grafts are in the process of being rejected. It is important to point out, in this connection, that in the case ofthose descriptions (Dallman et ul., 1982) of skin allograft rejection in TXB recipient rats infused with purified helper T cells, a surprisingly large number of T cells with the surface phenotype ofcytolytic T cells was present in the graft at the time of its rejection. Even in the absence of this information, there is ample reason to doubt a direct effector role for helper T cells in the models under discussion. For example, it is difficult to understand if donor helper T cells are either the mediators or effectors of graft rejection, why it takes up to 2 weeks for these T cells to cause graft rejection after they are infused into the TXB recipients. This very long delay surely means that the helper T cells have no capacity themselves to mediate graft rejection at the time they are passively transferred, and that they either need to acquire this function over a 2-week period, or they need to recruit recipient T cells into the response. The long delay before graft rejection commences is certainly not in keeping with the capacity of passively transferred Ly1+ helper T cells to immediately mediate a delay-type hypersensitivity reaction to specific antigen in the recipient. Therefore, until it is shown that the TXB mice and rats that are employed as recipients of purified helper T cells have no capacity to generate cytolytic T cells of their own at the time of graft rejection, it can be argued, because of no conclusive evidence to the contrary, that the expression of adoptive immunity against an established tumor in a TXB recipient infused with memory T cells requires that the recipient generate a secondary cyto!ytic T cell response of its own. This argument is supported, although not firmly, by the published demonstration (Frost et d., 1982) that the immunologically mediated regression of a sarcoma virus-induced tumor in syngeneic mice is associated with an influx into the tumor of large numbers of cytolytic T cells and their precursors. The argument is also reinforced by additional experiments performed with cyclophosphamide in this laboratory which were de-

MURINE ANTITUMOR IMMUNE RESPONSE

113

signed to determine whether treating TXB tumor-bearing mice with this drug after, instead of before they are adoptively immunized would prevent the adoptive immunity from being expressed. It will be recalled from experiments discussed in a preceding section (North, 1982)that cyclophosphamide has no effect on the memory T cells that are used routinely to transfer immunity passively. However, if these cells need to give rise to an adoptive secondary response in order to destroy the recipient's tumor, this response would be expected to be ablated by cyclophosphamide. It was found (Fig. 6), in agreement with the resistance of donor memory T cells to cyclophosphamide, that, whereas giving 100 mg/kg of the drug 1 hour before infusing immune T cells facilitated the expression of adoptive immunity, giving the drug 2 days after infusing immune T cells prevented the recipients from destroying their tumors. These results are consistent with the interpretation that, although the memory T cells that passively transfer immunity are cyclophosphamide resistant, they need to give rise to a population of cyclophosphamide-sensitivecells in the recipient before immunity is expressed. It is possible that further experiCYCLOPH. CONTROL

70

60 1

E E

-

-

50

0 K

0 40

I 3 kLL

0

30

w

N

;2 0 10

0

0

4

8

12

16

20

24

DAYS

FIG.6. Regression of an intrafootpad Meth A tumor in recipients given an intravenous infiision of iniinrine spleen cells on day 4 oftnmor growth was facilitated by giving cyclophosphamide 1 hour before immune spleen cells, but not by giving cyclophosphamide 2 days after immune cells, or later. Means of five mice per group.

114

ROBERT J. NORTH

ments will show that the cells in the recipient that are sensitive to cyclophosphamide are the cytolytic T cells that are generated to destroy the tumor. Indeed, additional evidence consistent with the view that the cytolytic T cells are involved in the rejection process comes from yet another study in this laboratory (Dye and North, 1983), which compared the ability of splenic T cells taken at the peak of a C . pauurnpotentiated anti-P815 tumor response (day lo), with the ability of spleen cells taken 20 days later (memory T cells) to cause regression of an established intradermal P815 tumor growing in TXB recipients. It was found (Fig. 7) that, whereas cells taken at peak response caused the onset of progressive tumor regression within 2 days of passive transfer, cells taken 20 days later did not cause tumor regression until after the standard 6- to 8-day delay. It is known (Mills et al., 1981) that the peak of the donors’ C . pauurn-potentiated response on day 10 is associated with the presence of maximal numbers of cytolytic T cells, and that cytolytic T cells are rapidly lost thereafter. The results cannot be explained in terms of there being more of a given type of T cells present in the donor on day 10 than on day 30, because substantially increasing the number of day 30 T cells infused made little difference to the delay before tumor regression commenced. Moreover, day 10

-&-.

30 DAY IMMUNE

DAYS OF TUMOR GROWTH

FIG.7. Regression of a 4-day intradermal P815 tumor in TXB recipients following infusion of active T cells (10 day) and memory T cells (30 day) from donors generating immunity to an immunizing admixture of lo6 P815 cells and 100 pg of C . pareurn (see text). Passive transfer of one spleen equivalent of active T cells caused the rapid onset of tumor regression, whereas passive transfer of one spleen equivalent of memory T cells did not cause tumor regression until after a 6-day delay. Means of five mice per group.

MURINE ANTITUMOR IMMUNE RESPONSE

115

cells were highly sensitive to treatment with vinblastine sulfate or with cyclophosphamide, whereas day 30 cells were highly resistant to treatments with both drugs. The different functional and physiological properties of “early” and “late” T cells generated in response to tumor allografts were recognized much earlier by others (Denham et al., 1970; Grant et ul., 1973).

T CELLSI N H ~ B I CYTOLYI.IC: ’~ RESPONSK IN THE RECIPIENT There is additional evidence to justify the assumption that the expression of adoptive immunity against an established tumor is mediated by cytolytic T cells. It comes from the results of experiments which tested the proposition that the ability of suppressor T cells from donors with established tumors to inhibit adoptive T cell-mediated tumor regression in TXB recipients depends on the ability of the suppressor cells to inhibit the production of cytolytic T cells in the recipients. The experiment involved determining whether the cytolytic T cell response that is generated in the draining node of tumorbearing TXB recipients of immune T cells is inhibited if the recipients are also infused with suppressor T cells from donors with established tumors. The results of these experiments were as predicted (Mills and North, 1983).They showed (Fig. 8) that the failure of tumors to undergo regression in adoptively immunized TXB recipients that received suppressor T cells was associated with a greatly diminished production of cytolytic T cells in the recipients’ draining lymph nodes. This evidence is consistent with the hypothesis that suppressor T cells generated in response to progressive tumor growth function to prevent the production of cytolytic effector T cells. Presumably, they achieve this either by inhibiting the function of helper T cells or by directly inhibiting the replication and maturation of effector T cells. This interpretation is supported by publications which show that T cells from syngeneic donors bearing an established P815 tumor (Takei et al., 1976, 1977)or a thymoma (Frost et ul., 1982) can inhibit a secondary cytolytic T cell response by sensitized spleen cells in vitro. The conclusion about the latter tumor was that suppressor T cells inhibit the generation, rather than the function, of cytolytic effector T cells. This is in keeping with the results obtained with models of in uiuo suppression in this laboratory, which show (Dye and North, 1981) that adoptive T cells-mediated regression of the P815 mastocytoma in TXB recipients is little affected if suppressor T cells are passively transferred at about the time of onset of tumor regression, that is, at the time that active immunity is being expressed. B.

SUPPHESSOR

116

ROBERT J. NORTH

CELLS TRANSFERRED IMMUNE

0-0

+

IMMUNE SUPPRESSORS 0-0 IMMUNE NORMAL

+

NONE

CELL TRANSFER

0

D A Y S O F TUMOR G R O W T H

FIG.8. Evidence that suppressor T cells from mice with a progressive P815 tumor inhibit adoptive immunotherapy of an established intrafootpad P815 tumor in TXB recipients by suppressing an adoptive cytolytic T cell response in the recipients. Tumor regression failed to occur in TXB recipients of immune spleen cells, if the recipients were also infused with suppressor T cells (A), and this was associated with a greatly reduced production of cytolytic T cells (B) in the draining lymph node. Infusion of normal T cells, instead of suppressor T cells, had no inhibitory effect on tumor regression or cytolytic T cell production.

MURINE ANTITUMOR IMMUNE RESPONSE

c. ADor.rrvE

117

IMMUNOTHEKAPY WIT11 T CELLS GmEKA'rEL) in Vitro

The generation in vitro of T cells cytolytic for allogeneic cells, hapten-coupled syngeneic cells, and immunogenic tumor cells by cocultivating T cells from primed mice with the appropriate mitomycin C-treated target cells has become a routine procedure. Moreover, the knowledge that T cell growth factor (interleukin-2) can be used to expand greatly populations of cloned or uncloned cytolytic T cells in vitro is of obvious therapeutic significance. It provides the means to generate a large population of the tumor bearer's own lymphocytes in vitro for intravenous infusion after debulking the tumor by surgery, irradiation, or chemotherapy. It is apparent, however, that attempts to use in vitro-generated cytolytic T cells to cause the rejection of allografts or tumors have given disappointing results. The general finding has been that, even though the cytolytic T cells employed are highly efficient at specifically lysing tumor target cells in oitro, their antitumor function in vivo can only be demonstrated by employing a Winn assay that involves admixing the cytolytic T cells with tumor cells and injecting them together into a subcutaneous site. Attempts to prevent the growth of a subcutaneous implant of tumor cells, let alone an established tumor, by infusing the cytolytic T cells intravenously for the most part have failed. Needless to say, this type of result is discouraging, because it indicates that cytolytic T cells generated in vitro do not possess the capacity to migrate from the intravascular compartment to an extravascular site of tumor growth. This functional deficiency not only would prevent cytolytic cells from attacking a primary tumor but also would prevent them from attacking extravascular tumor metastases. Although this type of finding might be interpreted by some as representing additional evidence that cytolytic T cells are not the effectors of T cell-mediated antitumor and antiallograft immunity, the more likely reason is that cytolytic T cells generated in vitro are not equipped to function normally in vivo. Indeed, experiments with radiolabeled in vitro-generated cytolytic T cells show that these cells show (Lotze et al., 1980; Dailey et al., 1982) abnormal distribution after intravenous infusion into a syngeneic recipient, in that an abnormally large proportion of them become trapped in the liver and lungs. It is possible, therefore, that in spite of the intactness of their cytolytic function, the physiology and surface structure of cytolytic T cells generated in vitro are abnormal enough to prevent them from circulating normally in blood and lymph of recipient animals, and of interacting normally with endothelial cells. It is even possible that they are seen

118

ROBERT J. NORTH

as effete by macrophages of the reticuloendothelial system. It should be pointed out, moreover, that because some of the attempts to determine whether cytolytic T cells can destroy tumors in vivo have employed immunocompetent recipients, the recipients’ isogenic barrier may have played some part in the failure of the T cells to express their function. Indeed, in those few cases where attempts to cause the regression of established tumors with in vitro-generated cytolytic T cells have been successful (Eberlin et al., 1982), the tumor-bearing recipients were sublethally irradiated. However, it will be argued below that the successful demonstration of adoptive immunotherapy of established tumors in irradiated recipients may give a significant overestimation of the capacity of passively transferred T cells to cause tumor regression.

D. ADOPTIVE IMMUNIZATION OF IRRADIATED TUMOR-BEARING RECIPIENTSAS A SPECIALCASE In general, whole-body exposure to ionizing radiation is immunosuppressive. Indeed, it was on the basis of the belief that the sublethal y-irradiation of tumor-bearing recipients would suppress the development of concomitant antitumor immunity, and consequently the suppressor T cells that “down-regulate” it, that it was predicted that yirradiation would enable passively transferred tumor-sensitized T cells to cause tumor regression. All of the results discussed in a preceding section were consistent with this interpretation. They showed with mice bearing 4-day tumors that, whereas y-irradiation alone or intravenous infusion of immune cells alone caused only a temporary reduction in the rate of tumor growth, combination therapy consisting of y-irradiation followed 1 hour later by infusion of immune spleen cells caused complete regression of tumors in all mice. However, the published results of others, as well as the results of ongoing experiments in this laboratory, suggest that the reason y-irradiation facilitates the antitumor function of passively transferred, tumor-sensitized T cells may be more complicated than originally proposed. Indeed, it is now apparent that by preventing or delaying the generation of suppressor T cells in a tumor-bearing animal, y-irradiation may serve not to depress but rather to augment a concomitant immune response that can function either synergistically, or in an additive way, with passively transferred T cells to cause tumor regression. Evidence which indicates that this is a likely possibility is supplied by the published impressive demonstration (Hellstrom et al., 1978) that sublethal irradiation of mice bearing an established immunogenic fibrosarcoma can result, after a delay, in complete regression of tumors in a significant

MURINE ANTITUMOR IMMUNE RESPONSE

119

proportion of the mice. Moreover, because it also was shown that this curative effect of irradiation could be prevented by infusing the mice with T cells from normal donors, it was proposed that regression was not caused by a direct effect of irradiation on the tumor. This was supported by the additional findings that irradiation was not therapeutic if given too early during tumor growth. The results were interpreted as meaning that irradiation preferentially eliminates suppressor T cells or their precursors, and thereby allows an antitumor immune response to develop to a high enough level to destroy the growing tumor. In other words, by selectively delaying negative regulation, whole-body irradiation may result in an augmentation of concomitant antitumor immunity. Needless to say, this type of result is not obtained with all immunogenic tumors, including most of the tumors studied in this laboratory. However, it has been a common observation with most of these tumors, but only if they had been implanted intradernially, that 500 rad of y-irradiation results, after a delay, in a pronounced although temporary reduction in the rate of tumor growth. Subsequent unpublished experiments in this laboratory with the SA1 sarcoma have revealed (Fig. 9) that this tumor growing intradermally can be caused to regress completely in all animals by giving 500 rad of y-irradiation after 5-6 days of tumor growth, but not earlier. Moreover, in agreement with the results of others (Hellstrom et ul., 1978), regression of this tumor did not begin until several days after irradiation, during which time the tumor continued to grow progressively to a large size. This delay suggests, again, that regression was not caused by a direct effect of irradiation on the tumor. In fact, additional experiments (Fig. 9) in this laboratory have shown that the pronounced irradiation-induced regression of the SA1 sarcoma and reduction in the rate of growth of some of the other tumors studied fails to occur if the tumors are growing in TXB mice. This irradiationinduced antitumor effect likely means, therefore, that irradiation serves to augment concomitant immunity. Additional evidence to support this interpretation will be discussed later. It is important to point out here, however, that if whole-body y-irradiation of a tumor bearer results in an augmentation of its concomitant antitumor immune response, any antitumor effect caused by infusing tumor-sensitized T cells into such an animal would need to be interpreted with caution. The tumor regression observed could well be the result of giving immune T cells to a host that has generated almost enough immune T cells of its own to cause tumor regression, depending on the immunogenicity of the tumor. In other words, the results may greatly overestimate the therapeutic capacity of the passively transferred immune T

120

ROBERT J. NORTH TXB 500R

I

E E

0 K

0

5 IA

0 W

!

fn

8ol 60

40i

20 -

I

0

1

0

4

1 8

I

I

12

16 DAYS

1

I

I

20

24

28

FIG.9. Spontaneous regression of an established, intradermal S A l sarcoma following exposure of the host, on day 6 of tumor growth, to 500 rad of whole-body y-irradiation. The tumors of three of the five mice were completely rejected. y-Irradiation failed to cause the regression of smaller tumors or of tumors growing in T celldeficient (TXB) mice. Means of five mice per group.

cells, regardless of whether they were generated in v i m or in uitro. This would be a particularly important possibility to consider with tumors that are immunogenic enough to be prone to spontaneous regression when implanted at certain sites in the host. It was demonstrated some years ago (Grant et al., 1973) that cytolytic effector cells generated in response to a growing tumor allograft are highly resistant to irradiation. If cytolytic T cells are the mediators of concomitant immunity to a tumor syngraft, these cells also would be expected to survive the doses of sublethal irradiation commonly employed to prepare a recipient for adoptive immunization. Evidence showing that cytolytic T cells are relatively radioresistant has been reviewed (Anderson and Warner, 1976).

E. THEMEANING OF THE LOCALPASSIVE TRANSFER ASSAY Before leaving the discussion of the meaning of adoptive immunization, it is perhaps worthwhile to mention the interpretation of results obtained with the Winn assay (Winn, 1960). In the first place, there is

MURINE ANTITUMOR IMMUNE RESPONSE

121

little doubt that this local adoptive immunization assay is popular because of the difficulty experienced by most researchers in attempting to transfer antitumor immunity systemically. The passive transfer of the immunity systemically is without question the most physiologically convincing method of assay, because it measures the capacity of the passively transferred lymphocytes to migrate from the vascular compartment to an extravascular site of antigen, a function that is essential for the expression of acquired immunity in most cases. Therefore, failure to transfer immunity systemically in the face of successful transfer by the Winn assay might indicate that the Winn assay is measuring the antitumor effects of cells that do not normally leave the circulation to express this function. It is more likely, however, that because it involves admixing lymphocytes and target cells and injecting them subcutaneously together, the Winn assay is simply a more sensitive assay. Even so, it is necessary to consider the types of sensitized T cells that are needed to destroy tumor target cells in this assay. There is certainly evidence that cytolytic T cells can indeed function efficiently in the Winn assay (Burton and Warner, 1977). Indeed, the original description of the assay by Winri (1960) indicates that he harvested donor lymphocytes at a time when the donor would have been making effector T cells against the allogeneic test tumor. In many of the cases, however, the assay is performed with donors, the immunization of which is not described in sufficient detail to know whether memory T cells or cytolytic T cells have been employed. It eventually may need to be explained, therefore, how T cells with no proven direct antitumor function can prevent the growth of a tumor implant in the Winn assay. It should be brought to mind, in this regard that Ly-l+ helper T cells are capable of locally transferring a DTH reaction and that a local DTH reaction is capable of nonspecifically destroying a tumor implant. It has been shown (Tuttle and North, 1975; Lagrange and Thickstun, 1979),for example, that the growth of a tumor implant can be neutralized by eliciting a DTH reaction to bacterial or other antigens at the site of tumor implantation. Thus, while cytolytic T cells could function in the Winn assay by specifically killing tumor cells, helper T cells or memory T cells might mediate tumor cell destruction by attracting nonspecific effector cells, such as macrophages, to kill tumor cells at the site of delayed inflammation. It should be realized, in this connection, that the antitumor fiinction of macrophages has been demonstrated (Feedman et al., 1980) with the Winn assay. Again, it has been demonstrated (Nakayama et al., 1978) that “late T cells” generated in vivo against an allograft can destroy syngeneic tumor cells in a Winn assay, but only if allogeneic stimula-

122

ROBERT J. NORTH

tor cells are present in the admixture that is injected subcutaneously. Considering the popularity of the Winn assay, there is a definite need to determine which immune functions this assay measures. IV. Analysis of Concomitant Immunity

The adoptive immunization experiments discussed in the preceding sections show that the progressive growth of a transplantable immunogenic tumor eventually results in the generation of a state of T cell-mediated immunosuppression. It was shown that this tumor-induced state of immunosuppression needs to be either eliminated or prevented from developing, in order for passively transferred, tumorsensitized T cells to express their antitumor function and cause tumor regression. It is apparent from these and other findings about the suppression of adoptive immunity that suppressor T cells are generated to “down-regulate’’ a preceding concomitant immune response before it generates enough effector T cells to destroy the tumor (North et al., 1982). It should be pointed out that all of the experiments performed to demonstrate tumor-induced immunosuppression of passively transferred antitumor immunity employed tumor-sensitized T cells from donors that were immunized by causing their tumors to regress by intralesional therapy with C. paruum. The reason for using immunized donors was based on the historical belief that, because the immunity generated against an immunogenic tumor is weak, demonstrating that it can be passively transferred from a tumor-bearing donor and expressed against a tumor implant, let alone an established tumor, in a syngeneic recipient would not be possible. Therefore, mice that served as donors of tumor-sensitized T cells were highly immunized by employing immunoadjuvants to augment the immune response to the immunizing tumor. Presumably, this is the reason why it is a common practice to give repeated injections of X-irradiated tumor cells to immunize donor mice. Indeed, there is little doubt from the experiments performed in this laboratory that adoptive immunization against established tumors requires adequately immunized donors. If the donors are not adequately immunized, the passive transfer of their T cells will cause either no regression or only partial regression of recipients’ tumors. A. THEPARADOX OF PASSIVE TRANSFER OF IMMUNITY AGAINST AN ESTABLISHED TUMOR WITH T CELLFROM A DONOR WITH AN ESTABLISHED TUMOR It goes without saying that the preceding discussion would tend to discourage any entertainment of the idea that adoptive immunity

MURINE ANTITUMOR IMMUNE RESPONSE

123

against an established tumor could be successfully demonstrated with the lymphocytes of a donor animal supporting the progressive growth of a lethal tumor. Yet, there is no real reason to believe that an animal generating concomitant immunity against a progressive tumor should not contain enough tumor-sensitized T cells to transfer immunity passively against a tumor implant, or to cause at least partial regression of a recipient tumor, provided the cells are harvested from the tumor bearer before the onset of suppression of concomitant immunity. In other words, the tumor-bearing donors that are used routinely in this laboratory as a source of suppressor T cells for inhibiting the expression of adoptive immunotherapy could also represent a source of immune T cells capable of adoptively immunizing against an established tumor. It is only possible to make this suggestion, however, because of the revelations about the type of recipient that needs to be employed to demonstrate adoptive immunity against an established syngeneic tumor. The recipient must be treated in a way that prevents it from generating suppressor T cells in response to its tumor. This paradoxical possibility of employing T cells from tumor-bearing donors to cause the regression of established tumors in immunocompetent recipients was investigated with the SA1 sarcoma and the P815 mastocytoma. It is known from experiments performed in this laboratory (Mills e t al., 1981) and elsewhere (Tuttle et al., 1983) that growth of the P815 mastocytoma evokes in its host the generation of a concomitant immune response that is evidenced by a low-magnitude cytolytic T cell response that peaks on days 9-10 of tumor growth and then decays. Concomitant immunity to the SA1, in contrast, has been measured only in terms of the ability of the tumor bearer to inhibit the growth of a challenge implant. According to this assay, immunity to the SA1 sarcoma is first detected between days 6 and 9 of tumor growth (North and Kirstein, 1977), although the decay of this immunity has not been documented. It was reasoned, on the basis of this information, that in the case of both tumors, a larger enough number of tumor-sensitized T cells should be present in the host on day 9 of tumor growth to transfer passively at least some level of immunity against tumors growing in recipients that have been y-irradiated to prevent them from generating suppressor T cells. To determine whether this is the case, prospective recipient mice bearing 4-day intradermal tumors were given 500 rad of y-irradiation, and infused 1 hour later with one organ equivalent (1.5-2 x lo8)of spleen cells from donors bearing 9-day intradermal tumors. The growth of the recipients’ tumors was followed against time. The type of result that routinely is being obtained is shown in Fig. 10, where it can be seen that spleen cells from donor mice destined to die of their progressive tu-

124

ROBERT J. NORTH

140

/

*CONTROL

20

0 0

4

a

I 12

16

20

24

DAYS

FIG. 10. The paradoxical demonstration that splenic T cells (1.5-2.0 x loR)from donor mice bearing an established intradermal SA1 sarcoma are capable, on passive transfer, of causing the regression of an SAl sarcoma growing in y-irradiated (500 rad) recipients. Donor spleen cells were infused intravenously 1 hour after exposing recipients bearing a 4-day tumor to y-irradiation. Means of five mice per group.

mors were capable, on passive transfer, of causing the regression of established tumors growing in y-irradiated recipients. It goes without saying that demonstrating that a mouse that is incapable of causing the immunologically mediated regression of its own progressive tumor, nevertheless possesses enough tumor-sensitized T cells in its spleen to cause the regression of a tumor growing progressively in a recipient mouse, is a very paradoxical situation. Indeed, the situation appears even more paradoxical when one considers the likelihood that additional sensitized T cells are present in the tumor bearer’s lymph nodes and in its circulation at large. However, the paradox is more apparent than real when one considers that the recipient tumor, although established, was smaller than the donor tumor at the time of passive transfer. A logical interpretation of the result, therefore, is that the donor would have been capable of causing the regression of its own tumor at the time of harvesting its spleen cells on day 9,

MURINE ANTITUMOR IMMUNE RESPONSE

125

if it had possessed the same-sized tumor as the recipient. In other words, too little immunity was generated by the donor too late. Moreover, available evidence suggests that this situation cannot be changed by implanting a smaller number of tumor cells with the intent of causing enough concomitant immunity to be generated before the tumor becomes too large. On the contrary, it is apparent from published results obtained with the SA1 sarcoma (North and Kirstein, 1977) that concomitant immunity is not generated until the tumor emerges and grows beyond a given size. Reducing the number of tumor cells implanted results in an increased period of latency before the tumor emerges, but this has no effect on the rate of growth of the tumor that emerges, or the need for a given tumor mass to trigger the induction of Concomitant immunity. Even so, an explanation of the paradox of concomitant immunity based on the tumors being too large to be rejected by the immunity that is eventually generated, does not explain why the immunity that is generated fails to cause either partial regression of the tumor or a temporary reduction in the rate of tumor growth. A possible explanation for this might be that the method for measuring tumor growth and regression is too crude to detect the consequences on tumor size of a 10-20% loss of tumor cells. As stated in an earlier section, the relationship, at any one time, between the extent of tumor regression and the number of tumor cells destroyed is not known. It is well to point out, however, that it has been repeatedly observed in this laboratory that immunogenic growing intradermally can display a reduced rate of growth for a period of time that corresponds to the time of generation of concomitant immunity. One cannot help wondering, therefore, whether a host generating concomitant immunity could reject its tumor if it possessed the capacity to direct all of its effector T cells into the tumor. Indeed, evidence for the existence of an untapped source of effector T cells in the circulation of tumor-bearing mice was revealed by the results of additional experiments that were designed to determine whether a tumor-bearing, concomitantly immune mouse can concentrate effector T cells in a peritoneal inflammatory exudate. The experiments were aimed at showing that induction of a peritoneal exudate by intraperitoneal injection of 5% neutral casein solution on day 8 or intradermal growth of the SA1 sarcoma or P815 mastocytoma would result in an influx over 48 hours of enough effector T cells into the peritoneal cavity to cause the regression of a 4-day intradermal tumor growing in a y-irradiated recipient. The rationale for using peritoneal exudate cells was based on the knowledge that inflammatory exudates are a favorite source of T cells for passive transfer experi-

126

ROBERT J. NORTH

ments, because sensitized T cells have a propensity for entering sites of inflammation (Koster et d.,1971; North and Spitalny, 1974). The results shown in Fig. 11 leave no doubt the two exudate equivalents (2 x lo7) of cells harvested from the inflamed peritoneal cavities of mice bearing a 10-day SA1 sarcoma were capable of passively transferring enough immunity to cause the complete regression of established SA1 tumors growing in y-irradiated recipients. This experimental result is obviously highly significant from the point of view of immunotherapy, because it shows that a host bearing an immunogenic tumor may have enough effector T cells in its lymphoid tissues and its circulation at large to destroy its tumor, provided a way could be

IMM. ANTI-THY

rLu

+

120

K

0

I 3

+

60

0

4

8

12

16

20

DAYS

FIG. 11. Peritoneal exudate cells (PEC) from mice bearing a 10-day established intradermal SA1 sarcoma were capable, on passive transfer, of causing the regression of an established SA1 sarcoma in y-irradiated recipients. Peritoneal exudates were induced by intraperitoneal injection of 5% neutral casein on day 8. The antitumor function of 2 x lo7exudate cells is compared with the antitumor function of (1.5-2.0 x loR) spleen cells (SPL) after treatment with complement alone (IMM PEC COMPL, IMM SPL COMPL), or after treatment with anti-Thy-1.2 antibody (ANTI-THY) and complement. Means of five mice per group.

MURINE ANTITUMOR IMMUNE RESPONSE

127

found to speed up the rate at which these cells enter the tumor. The proposition that a host’s sensitized T cells fail to cause the regression of its tumor because the T cells are prevented from entering the tumor by tunior-produced anti-inflammatory fktors seems highly unlikely. It is discounted by the knowledge that these same T cells, on passive transfer, can cause the regression of tumors growing in y-irradiated recipients. Indeed, all of the adoptive immunization experiments discussed in this article strongly argue against explanations of tumor escape based on the presence of anti-inflaminatory factors that inhibit the migration of T cells from the vascular compartment to the extravascular site of tumor growth.

B. KINETICSAND DECAY OF CONCOMITANT IMMUNITY AS MEASURED BY ADOPTIVEIMMUNIZATION It was demonstrated in the preceding section that a mouse bearing a progressive immunogenic tumor generates sensitized lymphocytes that are capable, on passive transfer, of causing the regression of the same tumor growing progressively in a y-irradiated recipient. This paradoxical finding provides an adoptive immunization assay, therefore, to follow the generation and decay of concomitant immunity to a progressively growing tumor. It was stated earlier that the generation and decay of concomitant immunity to the P815 mastocytoma already has been measured in terms of the generation and loss, in the draining lymph node, of T cells capable of lysing P815 cells in vitro, according to the 51Crrelease assay (Mills et al., 1981; Tuttle et al., 1983).However, in spite of the fact that some of those who work with cytolytic T cells believe they are working with the ultimate effectors of tumor and allograft rejection, there are others who believe (Loveland and McKenzie, 1982) that cytolytic T cells may not be involved in rejection mechanisms. There currently is a need, therefore, to measure the generation and loss of concomitant immunity in terms of the generation and loss of T cells capable of adoptively immunizing against an established tumor in a suitable recipient. With this in mind, experiments were performed to measure changes, against time of progressive tumor growth, in the capacity of one organ equivalent of spleen cells from the tumor-bearing host to cause the regression of a 4-day tumor growing in y-irradiated recipients. The passive transfer of spleen cells was performed at a single time with recipients that were all bearing a 4-day intradermal tumor. The donors were injected intradermally with lo6tumor cells 3 , 6 , 9 , 12,15, or 18 days before harvesting their spleen cells. The tumor-bearing recipients were given 500 rad of y-irradiation 1 hour before they received donor spleen cells.

128

ROBERT J. NORTH

The results obtained with the P815 mastocytoma are shown in Figs. 12 and 13. It can be seen that between days 6 and 9 of tumor growth, the spleens of tumor-bearing mice rapidly acquired splenic T cells capable of causing the regression of tumors growing in y-irradiated recipients. After day 9, however, there was a progressive decline in the ability of spleen cells to immunize adoptively against the recipient's tumor. Experiments with the SA1 sarcoma, Meth A fibrosarcoma, and P388 lymphoma have given results similar to those obtained with the P815 mastocytoma. Indeed, it is obvious from these results that successful adoptive immunotherapy with splenic T cells from tumorbearing donors requires that the donor spleen cells be harvested at the peak of the concomitant immune response. Harvesting spleen cells 3 days before or after peak immunity results in iiicomplete tumor re-

140-

120-

c.

E E

100-

r

@

80-

K

0

5I-

60-

LL

0 W

! cn

40-

20

0

4

8

12

16

20

24

28

DAYS FIG.12. Kinetics of generation, in donor mice bearing a progressive intradermal P815 tumor, of splenic T cells capable, on passive transfer, of causing the regression of an established P815 tumor growing in y-irradiated recipients. Donor spleen cells (1 organ equivalent) were infused 1 hour after giving 500 rad to recipients with a 4-day tumor. The numbers on the curves represent the days of tumor growth on which donor immune cells were harvested. Peak numbers of protective cells were present in the spleen about day 9 of tumor growth. Means of five mice per group.

MURINE ANTITUMOR IMMUNE RESPONSE

129

I A

0 B

100

>

5

I’ 8 0 z I-

z a

f

60

0 0 z 40 0 0

tj X

20

0

3

6

9

12

15

18

DAYS

FIG.13. Reprrsentation of kinetics of generation iund clccny of concomitnnt iminunity to intradennal growth of the P815 tumor, as calculated from the results in Fig. 12. (A) Growth of donors’ tumors. (B) Concomitant immunity is expressed as a resistance index which was obtained by subtracting mean tumor size in recipients of immune cells from mean tumor size in control recipients 17 days after passive transfer.

gression in the recipients. It is reassuring to know, in the case ofmice bearing the P815 mastocytoma, moreover, that the curve representing the generation and loss of splenic T cells capable of adoptively immunizing against this tumor in a y-irradiated recipient is almost identical to the curve for the generation 0f.T cells capable of lysing P815 cells

130

ROBERT J. NORTH

in uitro. The cytolytic T cell response, therefore, appears to be a legitimate measure of the generation and decay of Concomitant immunity. Obviously, it was a foregone conclusion that concomitant immunity to the P815 tumor would decay shortly after peaking on day 9 of tumor growth, because day 12-14 tumor bearers are used routinely as a source of suppressor T cells to suppress adoptive T cell-mediated regression of tumors in TXB test recipients. Indeed, because 12- t o 1 4 day tumor bearers can act as a source of splenic suppressor T cells, it might be considered puzzling that the spleens of these same mice can cause partial regression of 4-day tumors in y-irradiated recipients. However, this apparent contradiction is the result of employing yirradiated tumor-bearing recipients to assay for the presence of effector T cells in concomitantly immune mice on the one hand, and of using TXB tumor-bearing recipients to assay for suppressor T cells on the other. In fact, additional experiments have shown that the passive transfer of concomitant immunity is difficult to demonstrate with TXB tumor-bearing recipients. The reason for this was discussed in a preceding section, where it was argued that, unlike y-irradiated tumor bearers, TXB tumor bearers are incapable of generating their own concomitant immune response. It was pointed out that, because 500 rad of y-irradiation does not suppress an ongoing concomitant immune response, tumor regression in a y-irradiated tumor bearer infused with spleen cells from a concomitantly immune donor is a measure of the antitumor function of donor T cells plus the recipient’s own sensitized T cells. The assay gives a false impression, therefore, of the potential antitumor function of the T cells that mediate concomitant immunity. Nevertheless, the assay is a legitimate and powerful one to analyze concomitant immunity, provided it is realized that the recipient is contributing to the regression of its tumor. TXB tumorbearing recipients could be employed in these experiments, but complete regression of their tumors would not be achieved, unless the tumors were smaller, or the number of T cells passively transferred were substantially increased. In any case, there is no doubt that the results with y-irradiated recipients provide convincing evidence that a host does, indeed, generate sensitized T cells in response to the growth of its tumor, but that the generation of these T cells is abridged before enough of them are acquired to destroy the tumor. Evidence presented in earlier sections shows that the abridgement of concomitant immunity is an active process mediated by a population of suppressor T cells. The logical next step in this study, therefore, was to determine whether the progressive loss of T cells capable of passively

131

MURINE ANTITUMOR IMMUNE RESPONSE

transferring concomitant immunity is associated with the progressive acquisition of T cells capable of suppressing immunity.

C. THE:DECAY OF CONCOMITANT IMMUNITY Is ASSOCIATED wrm THE GENERATION OF SUPPRESSOR T CELLS The kinetics of generation of suppressor T cells in mice bearing a progressive P815 mastocytoma was determined (North, unpublished) by measuring changes against time in the capacity of one organ equivalent of spleen cells from the tumor bearers to inhibit the expression of adoptive T cell-mediated tumor regression in TXB recipients. This is the same assay described in a preceding section. It does not measure the suppression of passively transferred concomitant immunity, but suppression of immunity passively transferred from donors immunized to the tumor by causing tumor regression by intralesional C. pamum. It was shown earlier that this measures the ability of the infused suppressor T cells to prevent memory T cells from giving rise to the generation of cytolytic effector T cells in the TXB tumor-bearing recipient. The results of the experiment are shown in Figs. 14 and

-

50

7

40

E E

-

I

TUMOR C O N T R O L b

,DAY

1 2 SUPP. Y 1 5 SUPP.

0

D A Y 18

30

D A Y 10 SUPP.

I 3

+ LL

20

0 w

0 0

5

10

15

20 DAYS

25

30

35

40

FIG.14. Kinetics of generation of suppressor T cells in mice bearing an intrafootpad P815 mastocytoma (Fig. 12 is based on results obtained with an intradermal tumor). The numbers on the individual curves represent the days of tumor growth that splenic T cells were harvested. The results show changes against time in the ability of one organ equivalent of spleen cells from tumor-bearing donors to suppress tumor regression caused by one organ equivalent of immune spleen cells from immunized donors in T cell-deficient recipients. Suppressor T cells were progressively acquired after days 9-10 of tumor growth. Means of five mice per group.

132

ROBERT J. NORTH

15, where it can be seen that suppressor T cells were acquired progressively by the tumor-bearing donor from the time its concomitant immunity began to decay progressively as shown in the preceding section. In other words, progressive production of suppressor T cells occurred while T cells capable of passively transferring concomitant immunity and of lysing P815 cells in vitro were being lost from the spleen. The evidence is consistent with the hypothesis, therefore, that concomitant immunity is “down-regulated’’ by a population of suppressor T cells generated in response to a change in antigenic conditions caused by the tumors growing beyond a certain critical size. This

-8 c

0

20-

/

4

I-

g

/

O

10/

!

O

o/o

W

v,

O

l ’o

I

I

I

I

9

12

15

B

A 40

-

X W

8-

30-

z

e$

20-

W

a a

4

10-

v)

0

I 0

3

P

I

6

DAYS

FIG. 15. Kinetics of generation of suppressor T cells in response to progressive intrafootpad growth of the P815 mastocytoma. (A) Growth of the tumor. (B) Suppression is presented as a suppression index that was calculated from the results in Fig. 14, by subtracting the mean tumor thickness in TXB recipients of immune cells from mean tumor thickness in TXB recipients of immune cells plus suppressor cells on day 30.

MURINE ANTITUMOR IMMUNE RESPONSE

133

information surely would need to be considered in any attempt to cause tumor regression by immunotherapy. V. Tumor lmmunotherapy

The aim of tumor immunotherapy is to eradicate or slow the progress of a neoplastic disease by augmenting the host’s antitumor defense mechanisms. When the therapeutic modality is aimed at increasing the production of tumor-sensitized T cells or specific antibody, it is referred to as specific immunotherapy. Nonspecific immunotherapy, in contrast, is designed to enhance the nonspecific antitumor functions of macrophages, NK cells, and other cells which have been shown to be tumoricidal in vitro, regardless of whether or not the tumor cells possess tumor-associated transplantation antigens. Indeed, when one considers the number of antitumor defense mechanisms that have been demonstrated to function in vitro, one might have good reason to be optimistic about the therapeutic potential of immunotherapy. However, the literature reveals that, despite a substantial effort by many people over the past 10 years or more, the results of attempts to use immunotherapy to cause the regression of transplantable animal tumors, as well as human tumors, have been discouraging to say the least (Terry and Windhorst, 1978; Bartlett, 1979; Baldwin, 1982). It should be brought to mind, moreover, that the considerable difficulty of causing the immunologically mediated regression of transplantable immunogenic animal tumors is not appreciated by those who contend (Hewitt, 1979, 1982; Nossal, 1980) that chemically induced animal tumors are not suitable models of the spontaneous disease in humans. Presumably, the underlying basis for this contention is the erroneous belief that, so far as immunotherapy is concerned, the experimental tumor immunologist has an easier task than the clinical oncologist. There is not much evidence to support this contention, however, because the great majority of transplantable animal tumors are no more susceptible to immunotherapy than human tumors. Granted, there is substantial literature to show that an animal can be prophylactically immunized against the growth of tumor implants, but this represents nothing more than the classical way of showing that tumors can be immunogenic. Causing the immunologically mediated regression of the same tumors once they become established and begin growing progressively is an entirely different matter. In this case, immunotherapy for the most part has been shown to be without effect, unless the immunotherapy is given unrealistically, either at the same time, or soon after tumor cells are implanted. It is

134

ROBERT J. NORTH

obvious, therefore, that something happens at an early stage of growth of a transplantable immunogenic tumor that causes it to become refractory to treatment with agents that are commonly employed by immunologists to augment specific immune responses. Those few cases where an antitumor effect has been observed are the exceptions rather than the rule. One of the exceptions to the rule is the line 10 hepatocarcinoma, syngeneic in strain 2 guinea pigs, which can be caused to regress completely by intratumor injection of the bacillus Calmette-Guerin (BCG) strain of Mycobacterium bovis. This model of successful immunotherapy is based on the rationale that the irnmunoaugmenting action of adjuvants, such as BCG, requires that the adjuvant be injected at the same site as the antigen. It has been convincingly demonstrated (Zbar and Rapp, 1974) that the line 10 hepatocarcinoma growing intradermally can be caused to regress completely by intralesional injection of living BCG, or intralesional injection of the cell wall fraction of this organism suspended appropriately in oil. Even with this model, however, successful immunotherapy is limited to tumors less than about 1cm in diameter, in that tumors larger than this are for the most part completely refractory to immunotherapy. The evidence that BCG-induced regression of the line 10 tumor is immunologically mediated has been reviewed by some of those who have worked with this tumor (Bast et al., 1976). It consists of the demonstration that tumor regression is associated with the acquisition by the host of specific immunity to the growth of a tumor implant, and with lymphocytes that can passively transfer this immunity to normal recipients. It is possible to view BCG-induced regression of the line 10 tumor as a prototype model of immunotherapy which demonstrates the requirements and the limitations of specific immunotherapy. It shows that the immunoadjuvant must be injected into the tumor, and that the tumor must be below a certain critical size. This model provides further evidence, therefore, that tumors become refractory to immunotherapy after a certain stage of their growth. An obvious first step in analyzing this state of refractoriness would be to consider the possibility that it is caused by T cell-mediated suppression of concomitant antitumor immunity. Indeed, it is obvious from the discussion in the preceding sections that any rational attempt to treat an immunogenic tumor by immunotherapy would need to take into account the possible consequences of the concomitant antitumor immune response of the host. If the generation of concomitant immunity is a consequence of the growth of most immunogenic tumors, it follows that any attempt to treat an immuno-

MURINE ANTITUMOR IMMUNE RESPONSE

135

genic tumor by immunotherapy represents an attempt to augment an already ongoing antitumor immune response that may be undergoing negative regulation. The outcome of any immunotherapeutic treatment might be predicted, therefore, from a knowledge of whether the host is in the process of generating concomitant immunity, or losing this immunity under the regulatory influence of suppressor T cells. A successful therapeutic outcome would indicate that the therapeutic agent has served to augment concomitant immunity to a high enough level to cause partial or complete regression of the tumor. In contrast, failure of the agent to give an antitumor effect would indicate that the agent is incapable of overcoming T cell-mediated immunosuppression. Indeed, it would be logical to predict that the best time to give an immunotherapeutic agent would be liefore concomitant immunity undergoes negative regulation. However, I~eforeconsidering the design of immunotherapeutic procedures that might l x employed to take advantage of a knowledge ofthe timing ofthe generation and loss of T cell-mediated concomitant immnnity, it is important to point out that the convincing regression of experimental tumors has been achieved in two major ways: (1) by intralesional iiijection of immunoadjuvants such as BCG and C. jm-vziin, and (2) by intravenous injection of l~acterialendotoxin. It will I I ~argued in the sections that follow that, whereas intralesional injection of C. pcirvunz causes tumor regression by augmenting the generation of the tumor-sensitized effector T cells of concomitant immunity, parenteral injection of bacterial endotoxin causes tumor regression by fkilitating the expression of an already acquired mechanism of concomitant immunity. It will be suggested, in addition, that an immunogenic tumor becomes refractory to Imth of these inimunother~ipeuticagents immediately after the host’s concomitant immune responses hegin to decay.

A. IMMUNOTHERAPY BY IMMUNOPOTENTIATION With a view to obtaining direct evidence that the intratumor presence of an immunoadjuvant like C. parvurn or BCG can cause augmentation of the production of tumor-sensitized effector T cells, advantage was taken of a model of adjuvant-induced tumor regression which involves implanting tumor cells admixed with the adjuvant. This model has been favored by others (Baldwin and Pimm, 1978) to demonstrate the antitumor effect of intratumor immunoadjuvants, almost certainly because injection of the immunoadjuvant a day or so after the tumor cells are implanted gives a poor therapeutic response. The admixture model was employed in this laboratory for the same reason. Nevertheless, it is a useful model for the analytical purposes

136

ROBERT J. NORTH

intended, because injecting tumor cells admixed with the immunoadjuvant results in the growth of a tumor that does not begin to undergo regression for 10 days. Consequently, the model provides an opportunity to analyze the mechanisms of regression of a relatively large tumor mass, a mechanism that would need to function in any proper model of specific immunotherapy. Therefore, the model was employed with the aim of determining whether C. parvum-induced regression of the P815 mastocytoma is based on C. paruum-induced augmented production of T cells capable of passively transferring immunity to normal recipients, and of T cells capable of lysing P815 tumor cells in uitro. It was found, in general agreement with the results obtained by others (Dye et aZ., 1981),that when an admixture consisting of lo6P815 cells and 100 p g of C. paruum was injected in a hind footpad of mice, a tumor rapidly emerged, grew progressively for 9-10 days, and then underwent progressive and complete regression in about 90% of the animals. It was shown that the onset of tumor regression was associated with the acquisition by the host of specific immunity to growth of a tumor implant, and with T cells in the spleen capable of passively transferring immunity to a tumor implant systemically to normal recipients. This evidence, plus the finding that C. paruum failed to cause tumors to regress in TXB mice, was consistent with the hypothesis that tumor regression was caused by a C. paruumpotentiated T cell-mediated antitumor immune response. This does not represent proof of the hypothesis, however, because the possibility remained that the immunity that emerged was the result of the nonspecific destruction of the tumor by an immunological response directed against intratumor C. parvum. This seemed unlikely, however, in view of the results of an additional experiment designed to determine whether the mechanism responsible for regression of a C. paruum-treated tumor would have an effect on an untreated tumor of similar size growing in the contralateral footpad. It was found that the untreated tumor underwent regression in concert with the treated tumor. This could only mean that regression of the treated tumor was associated with a generation of a systemically distributed mechanism of specific antitumor immunity. Direct proof that intratumor C. paruum causes an increase in the generation of tumor-sensitized effector T cells was supplied by an experiment that measured the generation, in the node draining the treated tumor, of T cells capable of lysing P815 tumor targets in uitro according to a 6-hour W r release assay. The results showed (Mills et al., 1981) that the presence of C. paruum in the tumor resulted in a greatly augmented production of cytolytic T cells capable of specifi-

MURINE ANTITUMOR IMMUNE RESPONSE

137

cally lysing 51Cr-labeled P815 cells in vitro. Cytolytic T cells were not detected until between days 4 and 6 of tumor growth and reached peak production on day 10. Peak production was coincident, therefore, with the onset of progressive tumor regression. A cytolytic T cell response of much lower magnitude was generated against an untreated footpad tumor growing in control mice, although its kinetics were similar to those of the potentiated response. Both responses underwent progressive decay after peaking on day 10. It can be suggested on the basis of this finding, as well as those discussed in a preceding section, that the cytolytic T cell response and its decay in tumor-bearing control mice represents the generation and decay of concomitant immunity. In can be suggested, in turn, that C. parvum functions to augment this concomitant immune response to a level capable of causing tumor regression. The possibility that immunosuppression might be responsible for the inability of C. parvum treatment to cause the regression of tumors after they begin growing was investigated by testing the possibility that the growth of an untreated tumor growing on one side of the host can prevent the cytolytic T cell response to a C. parvum-tumor cell admixture injected in the opposite side. This experiment was based on the knowledge (Dye et al., 1981) that a therapeutic admixture of replicating or nonreplicating tumor cells and C. parvum has a pronounced therapeutic effect on a contralateral test tumor, but only if the therapeutic admixture is given at the time that the test tumor is implanted. A partial therapeutic effect is obtained against the test tumor if the admixture is given 2-4 days after tumor implantation, but no therapeutic effect is seen at all if the admixture is given 6 days after tumor implantation. It was predicted that failure of the admixture to give a therapeutic effect against a 6-day test tumor would be associated with the absence of a potentiated cytolytic T cell response to the admixture. This prediction proved to be correct. The experiment revealed (Mills et al., 1981) that injecting the therapeutic admixture on one side of the host at the same time as implanting the test tumor on the contralateral side, resulted in a cytolytic T cell response of high magnitude, not only in the node draining the therapeutic mixture, but also in the node draining the test tumor. This was associated with regression of the test tumor. In contrast, a cytolytic T cell response of very low magnitude was generated in response to an admixture given to a host bearing a 6-day test tumor, and this was associated with a failure of the host to cause the regression of its tumor. The results can be interpreted as indicating that an established tumor has a suppressive effect on the ability of C. parvum to potentiate an antitumor

138

ROBERT J. NORTH

immune response. However, direct evidence that the immunopotentiating action of C. parvurn is inhibited by tumor-induced suppressor T cells has yet to be obtained.

B. IMMUNOTHERAPY BY IMMUNOFACILITATION It was suggested in preceding sections which dealt with concomitant immunity that a tumor-bearing host generating peak concomitant immunity might possess enough effector T cells in its lymphoid tissues and its circulation at large to reject its tumor, provided a way could be found to direct these T cells to function in the tumor. An immunotherapeutic agent capable of ,achieving this end could be called an “immunofacilitator,” as opposed to an “immunopotentiator,” because it would function to cause the expression of an already acquired mechanism of immunity. The purpose of this section is to discuss evidence consistent with the hypothesis that parenteral injection of bacterial endotoxin can cause the regression of immunogenic tumors by this type of facilitating action. However, before discussing this evidence, it is well to point out that investigations of the antitumor actions of “bacterial toxins” go back to the turn of the century to Coley’s descriptions (1896) of the spontaneous regressions of large tumors in humans following natural attacks of erysipelas, or after deliberate injections of “mixed bacterial toxins.” An interesting review of the impressive case reports has been compiled by Nauts et al. (1953). Investigations of the effect of bacterial toxins on established animal tumors began in the 1920s, and a substantial literature accumulated over the following 50 years. The realization that the most active “toxins” are those present in the culture filtrates of gram-negative bacteria suggested that these bacteria produce a common active principal (Zahl et al., 1943), and this was shown (Shear and Turner, 1943) to be the lipopolysaccharide component (endotoxin) of the bacterial cell wall. A review of the literature up to 1943 (Shear, 1943) allowed several conclusions to be made about the antitumor action of endotoxin. These were that (1) autochthonous as well as transplantable tumors can be susceptible; (2) the antitumor action always is seen as an extensive hemorrhagic necrotic reaction in the core of the tumor; (3)necrosis is not followed, in most cases, by tumor regression; and (4) the tumor needs to be relatively large, in that small tumors appear refractory to endotoxin therapy. As far as mechanism of action is concerned, there have been attempts to explain the therapeutic effect of endotoxin in terms of the antitumor properties of a factor called tumor necrosis factor that is liberated into the circulation of animals shortly after endotoxin is ad-

MURINE ANTITUMOR IMMUNE RESPONSE

139

ministered. Experiments with mice have revealed (Carswell et uZ., 1975) that post-endotoxin serum can cause the necrosis of tumors in recipient mice, and the antitumor activity is associated with a seruni glycoprotein. Moreover, the partially purified glycoprotein has been shown to be toxic, to a larger or lesser degree, for some tumor cells in vitro. Hence, the reason for postulating that tumor necrosis factor is the effector of endotoxin-induced necrosis. It is important to realize, however, that the major source of tumor necrosis factor has not been the seruni of tunior-bearing mice hut the serum of mice that have had their macrophage systems highly activated by treatment with BCG or C. puruunz. It should be pointed out, in addition, that tumor necrosis factor has been studied for the most part in terms of its ability to cause core-confined iiecrosis rather than tumor regression. The distinction between necrosis and regression, and evidence that the former need not be of therapeutic significance, were supplied by results of a published study (Parr et al., 1973) that investigated the possibility that endotoxin-induced tumor regression is based on an immune response to the tumor. This study showed that the susceptibility to endotoxin of the tumors tested was related to their antigenicity, and that endotoxin failed to cause the regression of tumors growing in mice that had been immunodepressed by pre-exposure to whole-body X-irradiation, or by treatment with antilymphocyte serum. In contrast, immunosuppression had little effect on the ability of endotoxin to cause tumor necrosis. It is interesting that this particular study confirmed the general finding of others that small tumors tend to be refractory to endotoxin therapy. In fact, it was mainly on the basis of this last piece of information that experiments were performed in this laboratory to test the postulate that the therapeutic action of endotoxin is based on concomitant antitumor immunity. The experiments were performed with four different tumors. Two of these, the SA1 sarcoma and Meth A fibrosarcoma, were immunogenic. The other two, the BP3 fibrosarcoma and CaD2 mammary adenocarcinoma, were nonimmunogenic as determined by the classical procedure of testing for specific immunity to growth of a tumor implant 7-14 days after excising an established 7-day tumor. The experi1978a), in agreement with the results of ments showed (Berendt et d., others (Parr et al., 1973), that intravenous injection of endotoxin caused regression only of the immunogenic tumors, and that regression failed to occur if the tumors were growing in TXB mice or in mice that had been lethally irradiated before tumor implantation. Moreover, endotoxin-induced tumor regression was associated with the acquisition by the host of long-lasting immunity to growth of a

140

ROBERT J. NORTH

tumor implant, and with the acquisition of splenic T cells capable of passively transferring immunity to growth of a tumor implant. Evidence in favor of a role for concomitant immunity was obtained from experiments that measured concurrently the changes against time of tumor growth in the susceptibility of the tumor to endotoxin treatment, and immunity of the host to a tumor implant. It was demonstrated (Berendt et d.,197%) that the tumors did not respond to endotoxin therapy until the hosts developed concomitant immunity to an implant of cells of the tumor given at a distant site. Moreover, concomitant immunity to the Meth A fibrosarcoma underwent decay after about day 7 of tumor growth, and this decay of immunity was associated with a loss of susceptibility of the tumor to endotoxin. On the other hand, a more sustained state of Concomitant immunity generated against the SA1 sarcoma was associated with a longer period of susceptibility of this tumor to endotoxin therapy. This suggests, therefore, that the therapeutic effect of endotoxin is dependent on the generation of concomitant immunity and serves to explain why small tumors are refractory to endotoxin. Nevertheless, the evidence is circumstantial and does not prove that endotoxin serves to facilitate the antitumor function of an already acquired state of T cell-mediated concomitant immunity. Proof that this is the case would require the demonstration that a TXB mouse that is incapable of generating concomitant immunity and consequently of causing its tumor to regress under the influence of endotoxin, can have this capacity immediately conferred on it by an intravenous infusion of tumor-sensitized T cells from a tumor-bearing donor with concomitant immunity, that is, by T cells from a donor with the capacity to cause the regression of its own tumor in response to endotoxin therapy. Experiments of this type currently are being performed in this laboratory as part of an ongoing investigation of immunofacilitation. The basic design of the experiments was suggested by the convincing demonstration in a preceding section of the passive transfer of concomitant immunity against established tumors growing in y-irradiated tumor-bearing recipients. It was not possible to employ y-irradiated tumor bearers in these experiments, however, because an intravenous infusion of T cells from concomitantly immune donors causes tumors to undergo regression in these recipients without the need for endotoxin therapy. It will be recalled that it was argued that the expression of passively transferred concomitant immunity against a tumor growing in a y-irradiated recipient was dependent on the participation of the recipient’s own radioresistant concomitant immune response. However, the passive transfer of the same number of immune spleen

MURINE ANTITUMOR IMMUNE RESPONSE

141

cells from concomitantly immune donors causes only partial or no regression of tumors growing in TXB recipients that are incapable of generating concomitant immunity. This made it possible to use TXB recipients to determine whether an otherwise inadequate level of passively transferred concomitant immunity could be made to cause the regression of a tumor if the recipients also were treated with endotoxin. Experiments were first performed with the SA1 sarcoma, because it is highly susceptible to endotoxin therapy. The donors were mice bearing 9-day intradermal tumors. One organ equivalent (1.5 x lo8) of their spleen cells was infused into TXB recipients bearing 7day intradennal tumors. The results in Fig. 16 show that neither intravenous infusion of immune spleen cells alone on day 7 of tumor

I-"

*-*

CONTROL ANTI-THY - IMM.+ ENDO

f

B-B 0-0

17-1

0

0-0

,2;9'

ENDO IMM.+ ENDO COMP1.-IMM.+ENDO

4

'

iz

8

A

1.6

DAYS

FIG.16. Evidence that endotoxin-induced regression of the SA1 sarcoma is dependent on T cell-mediated concomitant immunity. Shown is the fate of intradermal SAl sarcomas growing in T cell-deficient (TXB) recipients given endotoxin alone on day 8, or spleen cells alone (1.5 x lo8)from 9-day tumor-bearing donors on day 7. The tumor did not undergo regression in response to intravenous injection of 50 pg of endotoxin unless the mice were infused 24 hours before with spleen cells from concomitantly immune donors. The spleen cells that primed the recipients' tumors for endotoxininduced regression were eliminated by treatment with anti-Thy-1.2 antibody and complement. Means of five mice per group.

142

ROBERT J. NORTH

growth, nor intravenous injection of endotoxin alone on day 8, had any appreciable effect on tumor growth. In contrast, combination therapy consisting of intravenous infusion of immune cells of day 7 and intravenous injection of endotoxin on day 8 caused rapid and complete regression of tumors in all animals. The immune cells from tumorbearing donors that primed the recipients’ tumors for endotoxin-induced regression were T cells, as evidenced by the failure to obtain a therapeutic effect with spleen cells that were treated with anti-Thy1.2 antibody and complement. This experiment leaves no doubt, therefore, that the therapeutic action of endotoxin is based on its capacity to facilitate the expression of T cell-mediated concomitant antitumor immunity. Indeed, additional experiments (North, to be published) show quite clearly that the curve for the acquisition and loss of susceptibility of the SA1 sarcoma to endotoxin therapy is exactly superimposable on the curve for the generation and decay of concomitant immunity, as measured by adoptive immunization. It is apparent, therefore, that endotoxin facilitates T cell-mediated tumor regression by creating conditions in the tumor that are conducive to the functioning of an already acquired population of tumor-sensitized T cells. However, the mechanism responsible for this facilitation remains unknown, as does the ultimate effector of tumor regression in this model. VI. Discussion

The major purpose of the foregoing discussion has been to present a framework of results which show that a host bearing a progressive immunogenic tumor responds immunologically to its tumor by generating a state of concomitant antitumor immunity. Because a host with an already established tumor represent the real situation that a therapist is confronted with, the possible possession by the host of a state of concomitant immunity would need to be considered in any attempt to cause regression of the tumor be immunotherapy. Indeed, it would seem reasonable to state that the rationale design of any immunotherapeutic modality would not be possible without a knowledge of whether or not the tumor is immunogenic, and whether the host is generating a concomitant antitumor immune response at the time that therapy is given. There is little doubt that the subject of concomitant immunity has been little studied of late, and that articles dealing with it since it was described b y Ehrlich (1906) have been neglected for the most part by tumor immunologists. It is apparent, from the absence of discussion,

MURINE ANTITUMOR IMMUNE RESPONSE

143

that concomitant immunity has been viewed as a novelty. Yet, there is nothing novel about a host generating an immune response to a tumor that displays tumor-associated rejection antigens. The only thing perhaps novel about the concomitant immune response is that it fails to reject the tumor that evokes its generation. But this represents nothing more than the central problem that tumor immunologists have been confronted with since the immunogenicity of tumors was first convincingly demonstrated in the 1950s. Moreover, it should be brought to mind that there is nothing peculiar about an immune response failing to fulfill its protective role. The many examples of chronic parasitic and infectious diseases attest to the fact that the possession of an immune system does not guarantee all individuals of a species freedom from colonization of their tissues by potential pathogens. Indeed, concomitant antitumor immunity may prove to be an extremely useful model to analyze situations where the immune response fails to fulfill a protective role. To an immunologist a tumor can be viewed as a replicating antigen which, like pathogenic microorganisms, has the theoretical potential to increase in quantity exponentially. It can be imagined in the case of a replicating agent with “weak antigens” that the host’s immune system cannot detect the agent until the agent replicates to supply the critical amount of antigen needed to trigger the induction of the immune response. Consequently, at a time when the immune response is being generated, it is confronted with a very large tumor mass that in a relatively short period of time is capable of supplying a superoptimal dose of antigen that resembles the doses of antigen required to induce a state of unresponsiveness as in other models of T cell-mediated immunosuppression. Otherwise, there is no reason to believe, at this time, that tumor-associated rejection antigens are any different from minor histocompatibility antigens that commonly are presented experimentally to an allogeneic recipient in the form of a skin or organ graft with no capacity to increase progressively in size. Both types of antigens are defined in terms of their ability to engender immunological rejection mechanisms, because both are extremely difficult to raise antibodies against (Klein, 1975). Indeed, it is doubtful that any antiserum or monoclonal antibodies yet have been raised against truly tumor-specific antigens (Hellstrom et al., 1982). Even if such antibodies are produced, it will remain to be determined whether the antigens they recognize are the same as those that function as immunogens in evoking T cell-mediated immune responses. Evidence that antigen recognition by T cells is based on surface recognition structures that may not be encoded by

144

ROBERT J. NORTH

variable-region antibody genes, and consequently that T cells may see certain foreign molecular configurations that are not seen by the antibody system has been discussed by Jensenius and Williams (1982). Reference to tumor-associated antigens as weak antigens is based, of course, on a comparison of these antigens with major histocompatibility antigens against which a strong immunological rejection mechanism are engendered. In fact, it is considered legitimate by some workers in the field to employ allogeneic tumors as models to analyze tumor immunology. There is nothing wrong with this view, except that having determined the type of rejection mechanism that needs to be generated in order to reject an allogeneic tumor, the central problems of tumor immunology remain unanswered. It would seem highly likely, however, that a scientific distinction between “strong” and “ weak” transplantation rejection antigens will be forthcoming. In the meantime, it would seem reasonable to propose that a concomitant immune response is generated against an allogeneic, as well as against a syngeneic tumor, but it is only against the former that enough cytolytic T cells are generated soon enough to cause tumor regression. In fact, the literature, as well as experienced in the laboratory, indicates that there is no difference between the growth of a tumor syngraft and a tumor allograft until about 7 days or so of tumor growth, after which the allograft, but not the syngraft, is vigorously rejected. The results presented in this article leave little doubt that a syngeneic host can, indeed, generate an immune response against its immunogenic tumor, but that the response is “down-regulated” by suppressor T cells before enough effector T cells are generated to destroy the tumor. This conclusion is based on the results of two main lines of investigation which involved the use of an adoptive immunization model illustrated diagrammatically in Fig. 17. This first involved demonstrating that tumor-bearing mice acquire a T cell-mediated mechanism of immunosuppression that prevents passively transferred, tumor-sensitized T cells from causing the regression of their tumors. It was shown that, whereas intravenous infusion of T cells from immunized donors failed to cause the regression of established tumors growing in immunocompetent recipients, the same immune T cells caused regression of the same-sized tumors growing in TXB recipients. The subsequent finding that passive transfer of T cells from immunocompetent tumor bearers inhibited the capacity of immune T cells to cause tumor regression in TXB recipients leaves little doubt that the progressive growth of an immunogenic tumor evokes the generation of suppressor T cells in an immunocompetent host. These

145

MURINE ANTITUMOR IMMUNE RESPONSE I M M U N I Z E D DONOR OF A C T I V E

T X B OR GAMMA-IRRADIATED

OR MEMORY T CELLS

T E S T RECIPIENT

( A N A L Y S I S OF I M M U N O G E N I C I T Y )

W I T H ESTABLISHED TUMOR

TUMOR-BEARING DONOR GENERATING C O N C O M I T A N T IMMUNITY

f

I

A N A L Y S I S OF GENERATION AND SUPPRESSION OF C O N C O M I T A N T IMMUNITY. DONOR OF E F F E C T O R AND SUPPRESSOR T C E L L S

1

I T H I S I S THE RESPONSE T H A T NEEDS TO BE M A N I P U L A T E D

ANALYSIS

FIG. 17. Diagrammatic representation of the basic adoptive immunization assays that are being employed to analyze the inimune response to progressive growth of an immunogenic tumor.

results have been confirmed in another laboratory with the Meth A fibrosarcoma (Bonventre et d.,1982). They allowed the prediction that treating a tumor-bearing host with any agent that is immunosuppressive should prevent the generation of suppressor T cells and thereby enable passively transferred, tumor-sensitized T cells to cause tumor regression. This proved to be the case for tumor-bearing recipients treated with cyclophosphamide or sublethal y-irradiation. The second line of investigation dealt with the generation and decay of concomitant antitumor immunity as determined by measuring changes against time in the capacity of splenic T cells from mice with a progressive tumor to transfer immunity passively against an established tumor growing in y-irradiated recipients. These experiments showed that mice bearing the P815 mastocytoma or SA1 sarcoma acquire T cells capable, on passive transfer, of causing tumor regression in y-irradiated recipients. These T cells were first generated on about day 6 or tumor growth, peaked on about day 9, and then were progressively lost. Moreover, in the case of the P815 mastocytoma, the kinetics of generation and loss of T cells capable of adoptively immunizing

146

ROBERT J. NORTH

against an established tumor in y-irradiated recipients were identical to the kinetics for the generation and loss of T cells cytolytic for P815 tumor cells in uitro. That cytolytic T cells are generated against this tumor and are then lost during progressive tumor growth has been reported by other workers (Takei et al., 1976, 1977; Tuttle et al., 1983). Subsequent experiments with the P815 mastocytoma revealed that the progressive loss of splenic T cells capable of passively transferring antitumor immunity against an established tumor occurred in concert with the progressive acquisition of T cells in the spleen capable of suppressing the expression of adoptive immunity against an established tumor growing in TXB recipients. This surely is strong evidence that concomitant immunity is progressively “down-regulated” by an increasing number of suppressor T cells. It was shown, however, that the interpretation of results obtained with the adoptive immunization assay requires an understanding of complexities of the assay. For example, it was argued that the expression of adoptive immunity against an established tumor in a TXB recipient requires the generation of a cytolytic T cell response in the recipient. Additional evidence indicates, moreover, that the ability of suppressor T cells from tumor bearers to inhibit the expression of adoptive immunity against tumors growing in TXB recipients is based on the capacity of suppressor T cells to prevent the generation of cytolytic T cells in these recipients. Again, failure of passively transferred tumor-sensitized T cells to cause tumor regression in immunocompetent mice likewise is associated with failure of the recipient to generate an adequate cytolytic T cell response. Attention also was directed at consideration of the complexity involved in employing yirradiated tumor bearers to show that concomitantly immune mice with progressive tumors contain T cells capable of causing tumor regression in recipients. It was suggested, on the basis of the available evidence, that because sublethal y-irradiation of a tumor bearer does not suppress concomitant immunity, the use of y-irradiated tumor bearers as recipients of immune T cells can give an overestimation of the functional capacity of donor T cells. The use of y-irradiated tumorbearing recipients provided a useful and legitimate assay, however, for measuring changes against time in the relative number of tumorsensitized T cells in a tumor-bearing donor. Results obtained with the assay leave little doubt that a tumor bearer first acquires and then loses T cells capable of passively transferring immunity against an established tumor in an irradiated recipient. Taken together, the results show that progressive growth of an immunogenic tumor results in the generation of a mechanism of T cell-mediated immunity that,

MURINE ANTITUMOR IMMUNE RESPONSE

147

after peaking on about day 9 of tumor growth, progressively undergoes decay under the negative regulatory influence of an increasing number of suppressor T cells. The “down-regulation” of concomitant immunity by suppressor T cells before it reaches sufficient magnitude to destroy the tumor seems the most plausible explanation for the escape of an immunogenic tumor in an immunocompetent host. The finding that suppressor T cells from mice bearing an established P815 tumor inhibit the generation of cytolytic T cells in TXB recipients infused with tumor-sensitized nieniory or helper T cells is in keeping with what one would expect the negative regulatory function of suppressor T cells to be. It essentially is in agreement with the published results of others (Takei et ul., 1976, 1977), which show that splenic T cells from mice bearing P815 tumors inhibit the production of T cells cytolytic for cells of this tumor in vitro. These published studies employed responder T cells from mice with small tumors and mitomycin C-treated P815 cells as stimulators. They also demonstrated that growth of the P815 tumor evokes the generation in its host of a low-magnitude cytolytic T cell response that decays as the tumor grows larger. Therefore, their responder T cells were harvested from mice generating concomitant immunity. Similar in vitro findings with another tumor were reported from another laboratory (Frost et al., 1982), except that in this case it was shown more clearly that suppressor cells suppress the generation rather than the function of cytolytic T cells. The properties of the suppressor T cells in our models of adoptive antitumor immunity currently are being investigated in detail. The findings reported in preceding sections that the suppressor cells are sensitive to y-irradiation and cyclophosphamide likens them to the suppressor T cells that function in other models of suppression. On the other hand, the demonstration that they are functionally eliminated by treatment with anti-Ly-1 monoclonal antibody and complement (Mills and North, study in progress), but not by anti-Ly-2 antibody and complement, make them different from the Ly-23+ suppressor T cells in certain other models. They are similar in this regard, however, to the T cells that suppress in vivo anti-Leishmania immunity (Liew et al., 1982) and in vitro immunity to minor H antigens (Macphail and Stutman, 1982). However, because the assay for suppression is an in vivo one, it is not possible to know whether Ly-l+ T cells are the ultimate suppressors. Even so, their Ly-l+ phenotype makes them different from the Ly-23+ suppressor T cells that function in what is perhaps the most quoted in vivo model of suppression of antitumor immunity: the suppression of immunity to an implant of the

148

ROBERT J. N O R T H

S1509a sarcoma in syngeneic A/J mice immunized against the tumor.

The series of published findings about T cell-mediated suppression of' immunity to this tumor has been the subject of review (Greene, 1980). The assay for suppression is based on the ability of an intravenous infusion of suppressor T cells from tumor-bearing donors to cause an increased rate of growth of a tumor implant in a preimmunized recipient, thus indicating that the recipient's immunity is suppressed. It was shown that T cells that cause the suppression are Ly-23+, I-J' T cells that are generated in thymus, draining node, spleen, and bone marrow within 24 hours or so of implanting tumor cells subcutaneously. This extraordinarily rapid, tumor-induced emergence of antigen-specific suppressor T cells is in agreement with the additional findings that treating normal A/J mice with very small volumes of anti-I-J serum (Greene et ul., 1977) or anti-I-J monoclonal antibody (Drebin et nl., 1983) immediately after implanting S1509a tumor cells causes a pronounced reduction in the rate of growth of the tumor that emerges 24 hours later. The obvious interpretation of these results is that the reduced rate of tumor growth caused by anti-I-J antibody is the result of the elimination of I-J+ suppressor T cells. However, these indications that the suppressor T cells in this model are generated and exert their suppressive effect during the first few days of tumor growth cannot be reconciled with the findings presented here, which show that suppressor T cells are not generated until an immunogenic tumor grows to a relatively large size. On the contrary, with all of the immunogenic tumors studied in this laboratory it takes at least 9 days of tumor growth before peak concomitant immunity is generated, and in keeping with their negative regulatory function, suppressor T cells are not generated until concomitant immunity begins to decay. Moreover, the generation of concomitant immunity would appear to be a common consequence of the progressive growth of immunogenic tumors in general (Nelson et uZ., 1979; Gorelik, 1983), but is generated only after the tumors are well established. This would lead one to believe that there is no immunity to suppress for the first 6 days or so of tumor growth. It should also be brought to mind in this connection that even a tumor allograft grows progressively for 7 days or more before antiallograft immunity is generated and expressed against it. Therefore, the ability of anti-I-J serum to cause a reduced rate of S 1509a tumor growth immediately after tumor implantation, and the demonstration of suppressor T cells in host lymphoid organs within 24 hours is probably important. It could be the reason why some tumors fail to show immunogenicity. Another important example of T cellmediated suppression of antitumor immunity is illustrated by the in-

MURINE ANTITUMOR IMMUNE RESPONSE

149

teresting findings of Kripke and co-workers (Kripke, 1976, 1977) about the immune response to ultraviolet light-induced skin tumors in mice. This niodel of T cell-mediated immunity and its suppression is relatively unique, and its importance would be best appreciated b y reading the original articles. According to the results presented in'this article, it is justifiable to present the immune response to a progressively growing immunogenic tumor diagrammatically as depicted in Fig. 18. The knowledge that a concomitant immune response is generated, peaks on day 9, and then progressively decays during the progressive production of suppressor T cells allows predictions to be made about the outcome of attempts to manipulate the immune response therapeutically at differ-

k

w

N VI 0 LL

1

c

4

x-

t z

I E

TIME+

Fit. 18. Diagriimniatic interpretation of the generation of concomit;int iniinunity and its T cell-mediated suppression. 'I'his (1iagr;ini predicts those times during the concomitant inimrine response at which iminrinotlierapy with inimr~nopoteiiti;ttion(intralesional HCG or C. purozcni), or iminanofacilitators (intravenous entlotoxin) is likely to givr a therapeutic result.

150

ROBERT J. NORTH

ent stages of tumor growth. For example, attempts to potentiate the antitumor immune response by intratumor injection of an immunoadjuvant like C. parvum would have to be made early enough so that potentiation would occur before the onset of immunosuppression at about days 9-10 of growth of the tumors that have been studied in this laboratory. The immunoadjuvant would need to be given well before days 9- 10, however, because potentiation of antitumor immunity almost certainly is based on the immune response to the adjuvant itself (Bast et al., 1976), which in the case of C. pamum takes 6-7 days (Tuttle and North, 1975).Therefore, in order for C. parvum to cause regression of the P815 mastocytoma, this immunoadjuvant would need to be injected into the tumor very shortly after the tumor is implanted. This is in agreement with experimental findings. However, in order for BCG to show an antitumor effect, it would need to be given before the tumor is implanted, because it takes the immune response to this immunoadjuvant about 12-14 days to develop (Miller et al., 1973). According to this line of reasoning, therefore, successful immunotherapy of immunogenic murine tumors with C. parvum or BCG should be a rare event, and this would appear to be supported by the general experience. Presumably, the guinea pig line 10 hepatocarcinoma can be successfully treated with BCG after it becomes palpable, because this tumor induces T cell-mediated suppression at a later stage of its growth. It is predicted, then, that causing the regression of immunogenic tumors by injecting them with immunoadjuvants will be difficult to demonstrate without taking steps to delay selectively the onset of immunosuppression. However, intravenous injection of bacterial endotoxin causes the regression of relatively large immunogenic murine tumors. According to the results discussed in a preceding section, moreover, there can be little doubt that endotoxin causes tumor regression by facilitating the expression of an already acquired state of concomitant immunity. However, the level of concomitant immunity has to be relatively high in order for endotoxin to cause tumor regression. Consequently, this therapeutic agent only functions when given at the time of peak concomitant immunity. It might also b e suggested from Fig. 18 that the immunological consequences of excising an immunogenic tumor will depend on whether excision is performed before or after the onset of tumorinduced immunosuppression. It will be recalled that showing that ligation or surgical removal of a tumor can result in immunity to growth of an implant of cells of the same tumor was the method originally employed to reveal the immunogenicity of syngeneic tumors

MURINE ANTITUMOR IMMUNE RESPONSE

151

(Foley, 1953). If it turns out to be true, as predicted here, that the growth of most immunogenic tumors results in the generation of a state of concomitant immunity that eventually decays under the regulatory influence of suppressor T cells, surgical removal of a tumor after the onset of decay of concomitant immunity may not result in the host showing immunity to a subsequent tumor implant. The subject of concomitant immunity and the effect on this immunity of excising the tumor has been reviewed by Gorelik (1983), who lists examples of where excision of some immunogenic tumors does not result in immunity to an implant. If it is found that removal of a tumor during “downregulation” of concomitant immunity fails to result in the emergence of antitumor immunity because the host remains immunosuppressed, then, the idea that specific immunotherapy of animal or human tumors is likely to be more successful after surgically debulking the tumors would need to be questioned. Indeed, it is not possible to leave this subject without mentioning experimental results obtained in this laboratory (Bursuker and North, 1983) with the Meth A fibrosarcoma. They show that, whereas surgical removal of the tumor before the onset of suppression of concomitant immunity results in the host retaining immunity to growth of a tumor implant for many weeks, surgical removal of the tumor after the onset of suppression is not followed by emergence of immunity to an implant for at least a 4-week period. On the contrary, for at least 4 weeks after the removal of a 15-day Meth A tumor the host retains T cells in its spleen which are capable of suppressing the capacity of immune T cells to cause the regression of Meth A tumors in TXB test recipients. As for the relevance of results obtained with chemically induced, transplantable murine tumors to the immunology of human tumors, nothing scientific can be said, because there is no generally accepted in vitro assay for measuring the immunogenicity of human tumors. In fact, if syngeneic strains of mice were not available to test for transplantation rejection antigens in vivo, we would be in no position to know whether the immunogenic murine tumors employed as models by experimental immunologists are immunogenic or not. Contrary to the belief of critics of the subject (Nossal, 1980; Hewitt, 1982), if it were not for transplantable tumors and an in vivo tumor rejection assay, there would be no reason to believe, even with highly immunogenic tumors, that there is any immunological restraint on the growth of these tumors by their hosts. The fact is that most immunogenic tumors grow at the same rate in immunodepressed animals as they do in immunocompetent animals. This is not to say that attention should not be drawn to the finding (Hewitt et al., 1976) that a number of

152

ROBERT J. NORTH

spontaneous tumors have been found to be nonimmunogenic when tested for immunogenicity by a certain assay. It would seem dangerously unwise, however, to state on the basis of such results that chemically induced immunogenic tumors are unrealistic models of neoplastic disease in humans. In the first place, the tests employed for immunogenicity were hardly exhaustive, and probably would not have revealed that immunogenicity of the weakly immunogenic P815 mastocytoma employed in this laboratory: a chemically induced tumor that is considered by some to be nonimmunogenic. In the second place, a convincing in vitro test for the immunogenicity or nonimmunogenicity of human tumors does not exist and is unlikely to be discovered without studying animal tumors of known immunogenicity according to in vivo rejection assays. Indeed, even the in vivo rejection assays are somewhat arbitrary, in that they rely on empirical attempts to immunize a syngeneic host against a tumor implant by repeated injections of X-irradiated tumor cells, or by ligating or excising a tumor, without any knowledge of the strength, duration, or mechanism of the immunity that these procedures are expected to engender. Therefore, the best test for immunogenicity may not yet exist, and the sensitivity of those tests that do exist is not known. It is even possible that testing for immunogenicity by repeatedly injecting nonreplicating tumor cells, with or without an immunoadjuvant, might be self-defeating. The fact is that practically nothing is known about the nature of tumor-associated rejection antigens or about the inductive events that need to be set in motion in order for these antigens to give rise to an antitumor immune response. The best was to answer the criticism that animal models of immunotherapy are not models of the human disease is to point out that, in spite of the discussion in the preceding pages, nobody in this laboratory would know how to cure a wild mouse of a palpable subcutaneous neoplasm that proved to be malignant at biopsy. There would be no generally acceptable way to test for immunogenicity in vivo or in vitro, and there would be no way of knowing whether the tumor already has induced immunosuppression. Even if it were assumed that the tumor is immunogenic, the fact that it is large enough to be noticed would indicate, on the basis of results obtained with transplantable immunogenic tumors, that it already has induced a state of immunosuppression that would prevent attempts to cure it by immunotherapy, even after surgical debulking. There is no doubt, therefore, as stated by some (Nossal, 1980), that tumor immunology is in a sad state. But so is the immunology of leprosy, tuberculosis, leishmaniasis, malaria, and other infectious and parasitic diseases that kill

MURINE ANTITUMOR IMMUNE RESPONSE

153

more humans than do cancers. The fact is that the exciting and elegant findings of modern immunology, which rightly have come from models that employ nonreplicating antigens, have had very little impact on the treatment of neoplastic, infectious, or even immunological diseases. Thus, the problems of tumor immunology are the problems of immunology as a whole, and these would be better dealt with by intelligent experimentation than by unwise critiques.

ACKNOWLEDGMENTS This work was supported by Grants CA-16642 and CA-27794 from the National Cancer Institute and a grant-in-aid from R. J. Reynolds Industries, Inc.

REFERENCES Anderson, R. E., and Warner, N. L. (1976). Adz;. Immunol. 24,215. Baldwin, R. W. (1982).Springer Semin. Immunopathol. 5 , 113. Baldwin, R. W., and Pimm, M. V. (1978). Adv. Cancer Res. 2 8 , 9 l . Bartlett, G. L. (1979). Semin. Oncol. 6, 515. Bast, R. C., Bast, B. S., and Rapp, H. J. (1976).Ann. N.Y. Acad. Sci. 277, 60. Berendt, M. J., and North, R. J. (1980).J . E x p . Med. 151, 69. Berendt, M. J., North, R. J., and Kirstein, D. P. (1978a).J. E x p . Med. 148, 1550. Berendt, M. J., North, R. J., and Kirstein, D. P. (1978b).J . E x p . Med. 148, 1560. Bonventre, P. F., Nickol, A. D., Ball, E. J., Michael, J. G., and Bubel, H. C. (1982).J. Reticuloendothel. SOC. 32, 25. Borberg, H., Oettgen, H. F., Chandy, K., and Beattie, E. (1972). Int. J. Cancer 10,539. Brunner, K. T., MacDonald, H. R., and Cerottini, J. C. (1981).J.E x p . Med. 154, 362. Bursuker, I., and North, R. J. (1984).J. E x p . Med. 159 (in press). Burton, R. C., and Warner, N. L. (1977). Cancer Immunol. Zmmunother. 2,91. Burton, R., Thompson, J., and Warner, N. L. (1975).J. Immunol. Methods 8, 133. Carswell, E. A,, Old, L. J., Kassel, R. L., Green, S., Fiore, N., and Williamson, B. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 3666. Chassoux, D. M., Botch, F. M., and MacLennan, I. C. M. (1978). Br.1. Cancer 38,211. Claman, H. N., Miller, S. D., Conlon, P. J., and Moorhead, J. W. (1980).Adz;. Immunol. 30, 121. Coley, W. B. (1896). Am. J. Med. Sci. 112,251. Dailey, M. O., Fathman, C. G., Butcher, E. C., Pillemer, E., and Weissman, I. (1982).J. Immunol. 128,2134. Dallman, M. J., Mason, D. W., and Webb, M. (1982). Eur. J . Zmmunol. 12, 511. Denham, S., Grant, C. K., Hall, J. G., and Alexander, P. (1970). Transplantation 9,366. Drebin, J. A., Waltenbaugh, C., Schatten, S., Benacerraf, B., and Greene, M. I. (1983).J. Immunol. 130, 506. Duprez, V., Hamilton, B., and Burakoff, S. J. (1982).J. E x p . Med. 156, 844. Dye, E. S., and North, R. J. (1981).J. E x p . Med. 154, 1033. Dye, E. S., and North, R. J. (1984).J. Leukocyte R i d . (in press). Dye, E. S., North, R. J., and Mills, C. D. (1981).J. E x p . Med. 154, 609. Eberlin, T. J., Rosenstein, M., and Rosenberg, S. A. (1982).J. E x p . Med. 156,385. Eggers, A. E., and Wunderlich, J. R. (1975).J. Immunol. 114, 1554. Ehrlich, P. (1906). In “Arbeiten aus dem koniglichen Institut fur experimentelle Therapie zu Frankfurt A.M.” (P. Ehrlich, ed.), pp. 77-103. Fisher, Jena.

154

ROBERT J. NORTH

Fefer, A., Einstein, A. B., and Cheever, M. A. (1976). Ann. Acad. Sci. Fenn. Ser. AZI Chem. 277,492. Fernandez-Cruz, E., Gilman, S. C., and Feldman, J. D. (1982).J. Zmmunol. 128, 1112. Fink, M. A., ed. (1976). “The Macrophage in Neoplasia.” Academic Press, New York. Foley, E. J. (1953). Cancer Res. 13, 835. Folkman, J,, and Cotran, R. (1976). Znt. Reu. E x p . Pathol. 16, 207. Fujimoto, S., Greene, M. I., and Sehon, A. H. (1976).J. Zmmunol. 116, 791. Freedman, V. H., Calvelli, T. A., Silagi, S., and Silverstein, S. C. (1980).J. E x p . Med. 152, 657. Frost, P., Prete, P., and Kerbel, R. (1982). Int. J. Cancer 30, 211. Gillis, S . , Union, N. A., Baker, P. E., and Smith, K. A. (1979).J. E x p . Med. 149, 1460. Glaser, M. (1979).J. E x p . Med. 149, 774. Gorelik, E. (1983). Ado. Cancer Res. 39, 71. Goto, M., Mitsuoka, A., Sugiyama, M., and Kitano, M. (1981).J. E x p . Med. 154, 204. Grant, C. K., Evans, R., and Alexander, P. (1973). Cell. Zmmunol. 8, 136. Greenberg, P. D., Cheever, M. A., and Fefer, A. (1980). Cancer Res. 40,4428. Greene, M. I. (1980). Contemp. Top. Zmmunobiol. 5, 91. Greene, M. I., Dorf, M. E., Pierrs; M., and Benacerraf, B. (1977). Proc. Natl. Acad. Sci. U.S.A.74, 5118. Greene, M. I., Nelles, M. J., Sy, M-S., and Nisonoff, A. (1982).Ado. Immunol. 32,253. Hellstrom, K. E., and Hellstrom, I. (1969). Ado. Cancer Res. 12, 167. Hellstrom, K. E., Hellstrom, I., Kant, J. A., and Tamerius, J. D. (1978)J. E x p . Med. 148, 799. Hellstrom, K. E., Hellstrom, I., and Brown, J. P. (1982). Springer Semin. Zmmunopathol. 5, 113. Herberman, R. B. (1974). Adu. Cancer Res. 19,207. Herberman, R. B., and Holden, H. T. (1978).Ado. Cancer Res. 27,305. Hewitt, H. B. (1979). Clin. Radiol. 30, 361. Hewitt, H. B. (1982).J.Biol. Resp. Mod$. 1, 107. Hewitt, H. B., Blake, E. R., and Walder, A. S. (1976). Br. J. Cancer 33, 241. Howard, J. C., and Butcher, G. W. (1981).Scand. J. Immunol. 14,687. Jensenius, J . C., and Williams, A. F. (1982). Nature (London) 300, 583. Kamo, I., and Friedman, H. (1977).Ado. Cancer Res. 25, 271. Klein, G. (1966). Annu. Reu. Microbiol. 20,223. Klein, J. (1975). “Biology of the Mouse Histocompatibility-2 Complex.” SpringerVerlag, Berlin and New York. Koster, F. T., McGregor, D. D., and Mackaness, G. B. (1971).J. E x p . Med. 133,400. Kripke, M. L. (1976).J. Natl. Cancer Znst. 53, 1333. Kripke, M. L. (1977). Cancer Res. 37, 1395. Lagrange, P. H., and Thickstun, P. M. (1979).J. Natl. Cancer Znst. 62, 429. Liew, F. Y., Hale, C., and Howard, J. G. (1982).J. Immunol. 128, 1917. Lotze, M. T., Line, B. R., Mathisen, D. J., and Rosenberg, S. A. (198O).J.Zmmunol. 125, 1487. Loveland, B. E., and McKenzie, I. F. C. (1982). Transplantation 33,407. Loveland, B. E., Hogarth, P. M., Ceredig, R. H., and McKenzie, I. F. C. (1981).J. E x p . Med. 153, 1044. Macphail, S., and Stutman, 0. (1982).J. E x p . Med. 156, 1398. Mathe, G., Halle-Pannenko, O., and Bouret, C. (1977). Cancer Immunol. Immunother. 2, 139.

MURINE ANTITUMOR IMMUNE RESPONSE

155

Miller, T. E., Mackaness, G. B., and Lagrange, P. H. (1973).J . Natl. Cancer Inst. 51, 1669. Mills, C. D., a n d North, R. J. (1983).J . E x ~ JMetl. . 157, 1448. Mills, C. D., North, R. J., and Dye, E. S. (1981).J . E x p . Med. 154, 621. Mokyr, M. B., Hengst, J. C. D., and Dray, S. (1982). Cancer Res. 42, 1974. Moore, M., and Williams, D. E. (1973). Int. J . Cancer 11, 358. Nakayama, N., Sendo, F., Miyake, T., and Fuyama, S. (1978).J . Immunol. 120,619. Naor, D. (1979). Ado. Cuncer Res. 29,45. Nauts, H. C., Fowler, G. A., and Bagatko, F. H. (1953).Actu Med. Scnnd. 145 (Suppl.), 103. Nelson, D. S., Nelson, M.,and Hopper, K. E. (1979).Ado. E x p . Med. Biol. 121B, 541. North, R. J. (1982).J . E x p . Med. 55, 1063. North, R. J. (1984). Cuncer I t n m u t i d . I m t n u t i ~ ~ t (/ i~n ~press). ~r. North, R. J., and Kirstein, D. P. (1977).J . E x p . Med. 145, 275. North, R. J., and Spitalny, G. L. (1974). Infect. Immun. 10, 849. North, R. J., Dye, E. S., Mills, C. D., and Chandler, J. P. (1982). Springer Semin. Immunopathol. 5, 193. Nossal, G. J. V. (1980).Immunol. Today 1, 5. Old, L. J., and Boyse, E. A. (1964).Annu. Reo. Med. 15, 166. Old, L. J., Boyse, E. A., Clark, D. A., and Carswell, E. (1962).Ann. N.Y. Acad. Sci. 101, 80. Old, L. J., Stockert, E., Boyse, E. A., and Kim, J. H. (1968).J.E x p . Med. 127, 523. Pam, I., Wheeler, E., and Alexander, P. (1973). Br. 1.Cancer 27,370. Pike, M. C., and Syndennan, R. (1976).J . lnimunol. 117, 1243. Plata, F., Gomard, E., Leclerc, J.-C., and Levy, J. P. (1973).J . Immunol. 111, 667. Reinisch, C. L., Andrew, S. L., and Schlossman, S. F. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 2989. Rollinghoff, M., Starzinski-Powitz, A., Pfizenmaier, K., and Wagner, H. (1977).J . E x p . Med. 145, 455. Rosenberg, S. A., and Terry, W. D. (1977). Ado. Cancer Res. 25,323. Rouse, B. T., Rollinghoff, M., and Warner, N. L. (1973). Eur. J . Immunol. 3, 218. Shear, M. J. (1943).J.Natl. Cancer Inst. 4,461. Shear, M. J., and Turner, F. C. (1943).J.Natl. Cancer Inst. 4,81. Sjogren, H. 0. (1965).Prog. E x p . Tumor Res. 6,289. Small, M., and Trainin, N. (1976).J . Immunol. 117, 292. Takei, F., Levy, J. G., and Kilburn, D. G. (1976).J . Immunol. 116,288. Takei, F., Levy, J. G., and Kilburn, D. G. (1977).J . lmmunol. 118, 412. Terry, W. D., and Windhorst, D., eds. (1978). “Immunotherapy of Cancer: Present Status of Trials in Man.” Raven, New York. Tuttle, R. L., and North, R. J. (1975).J . Natl. Cancer Inst. 55, 1403. Tuttle, R. L., Knick, V. C., Stopford, C. R., and Wolberg, G . (1983). Cancer Res. (in press). Wagner, H., Hardt, C., Heeg, M., Rollinghoff, M., and Pfizenmaier, K. (1980a).Nature (London)284,278. Wagner, H., Pfizenmaier, K., and Rollinghoff, M. (1980b).Ado. Cancer Res. 31, 78. Winn, H. J. (1960).Natl. Cancer Inst. Monogr. No. 2, 113. Zahl, P. A,, Hatner, S. H., and Cooper, F. S. (1943). Proc. Soc. E x p . Biol. Med. 54, 187. Zlxir, B., nnd Rapp, H. J. (1974).Cnnccr Res. 34, 1532.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY. VOL. 35

Immunologic Regulation of Fetal-Maternal Balance DAVID R. JACOBY,* LARS B. OLDING,t AND MICHAEL B. A. OLDSTONE' 'Department of Immunology, Scripps Clinic and Research Foundation, La Jolla, California, and +Departmentof Pathology, University of Goteborg, Goteborg, Sweden

I. Preface ......................................... 11. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Proposed Mechanisms for Maintenance of the A. Maternal Contributions to Fetal Maintenance during Pregnancy B. Placental Contributions to Fetal Maintenance. . . . . . . . . . . . . . . . . . . . . . C. Fetal Contributions to Prevent Allogeneic IV. Fetal Expression of Histocompatibility Antige Ontogeny and Development of Histocompatibility Antigens. . . . . . . . . . . . . V. Immunologic Basis of Lymphocyte Interactions between Mother and Fetus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Imniunocompetence of the Fetus B. Immunocompetence of the Mother. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Permeability of the Placenta to Lymphocytes VI. Human Maternal and Neonatal Lymphocyte Interactions. . . . . . . . . . . . . . . . A. Suppressor Activity of Lymphocytes from Human Newborns . . . . . . . . . B. Characterization of the Newborn Lymphocyte Subset Responsible for

............................................ on by Cord T Lymphocyte Subsets D. JMechanisms of Cord T Lymphocyte-Mediated Suppression . . . . . . . . . . E. Helper Activity of Cord Blood T Lymp ....... F. Regulatory T Lymphocytes in Newborns and Adults . . . . . . . . . . . . . . . . VII. Conclusions . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

157

164

170 172 172 177 179 180 180 184 188 192 196 197 199 202

I. Preface

How a fetus survives and by doing so escapes the expected immunologic assault of its mother is a filndamental yet largely unsolved question in biology. The answer to the question is important from a practical standpoint of securing successful pregnancy (i.e., avoiding unwanted spontaneous abortions). Further, the answer is important if we are to understand transplantation immunity and to be successful in its manipulation. Hypotheses to explain successful fetal-maternal balance fall into three general categories. The first presupposes mechanical barriers to segregate the maternal immunologic attack system from the fetus. Included is the absence of molecules (associated with the major histocompatibility complex, MHC) needed for cell-cell recognition and 157 Copyright 8 1984 by Academic Press. Inc. All rights of reproduction in any form reserved. ISBN 0-12-022435-6

158

DAVID R. JACOBY ET AL.

immune assault. The second view seeks evidence for active maternal factors involved in this process, and the third considers the fetus as actively interfering with a normal maternal graft rejection mechanism. All three hypotheses are reviewed in this article, primarily as they relate to humans. However, it is the third hypothetical pathway that we will discuss in the greatest depth. II. Overview

The fetus of an outbred mammalian species expresses paternal antigens from the earliest developmental stage, yet is not automatically rejected as an allograft. This apparent paradox in transplantation immunology is not related to the mother’s immunocompetence. Tissue transplants from paternal and fetal sources grafted to other maternal sites are rejected normally, and maternal sensitization to paternal histocompatibility antigens is clearly a common event during successful pregnancies. The mechanisms that contribute to fetal maintenance throughout gestation are important for survival and, further, preserve the selective advantages that heterozygosity conveys. Studies with mice (Warburton, 1968; Kirby, 1968; Beer et al., 1975) have shown that disparity between maternal and fetal histocompatibility antigens favors greater invasiveness of the trophoblast and larger size of the placenta. Also, women who share many histocompatibility loci with their husbands have a significantly higher rate of chronic spontaneous abortions than women who express different histocompatibility antigens (Beer et al., 1981). The antigenicity of fetal tissue has been known since the early 19OOs, when preeclampsia, toxemia of pregnancy, and rhesus (Rh) disease were shown to be associated with maternal sensitization to various determinants of fetal blood cells. However, it was not until 1954 that Medawar first addressed the subject of immunologic mechanisms whereby the fetus would be maintained and nourished during pregnancy despite genetic dissimilarity to the mother. H e hypothesized that (1)the uterus is an immunologically privileged site; (2) the invasive fetal tissues are antigenically immature; (3) the fetus is separated anatomically from the mother; and (4)pregnancy alters the maternal immunologic response. These ideas have provided the guidelines for contemporary research on maternal-fetal immunology. Although specific points have been refined or altered, three important concepts introduced by this work continue to provide the foundations for this field. First, several diverse mechanisms are responsible for maintenance of the fetus; second, both the maternal and fetal immune

REGULATION OF FETAL-MATERNAL

BALANCE

159

systems are involved; and third, the placental expression of histocompatibility antigens is important. Further clarification of the mechanisms involved in maintenance of the fetus will have direct clinical impact on treatment of chronic idiopathic abortions. Currently, 25% of spontaneous abortions in the first and second trimesters have no known etiology. Also, research in this field will continue to provide information on the ontogeny of immune function and neonatal susceptibility to disease. In-depth description of the various facets constituting maternalfetal immunobiology is beyond the scope of this article. Rather, the evidence and ideas thought to be important in maintenance of the fetus are reviewed and documented for the reader’s specific pursuit. We focus on the fetal immune system, specifically the ability of fetal and newborn lymphocytes to suppress maternal immune responses. This information is given in the context of the general maternal and fetal immunocompetence, the fetal expression of alloantigens, and immune sensitization of mothers to fetal antigens. 111. Proposed Mechanisms for Maintenance of the Fetus

A. MATERNALCONTRIBUTIONS TO FETAL MAINTENANCE DURING PREGNANCY The successful semiallogeneic transplant represented by the fetus has prompted investigation into the status of the maternal immune system during pregnancy, specifically whether foreign antigens are being presented during pregnancy, or the predicted immune response is being suppressed either systemically or locally at the site of the placenta.

1 . Humoral and Cellular Sensitization to Paternal Antigens

The pregnant mother produces a wide range of antibodies against fetal antigens (Winchester et al., 1975).As a consequence, certain fetal and neonatal diseases may result, particularly erythroblastosis fetalis due to maternal Rh incompatibility and neonatal thrombocytopenia due to incompatibility of platelet antigens. Sera from pregnant women commonly contain antibodies to HLA-A, -B, and -C series antigens (class I histocompatibility antigens). These antibodies have been detected in the sera of approximately 25% of women during their first pregnancy, and the incidence and titer of anti-HLA antibodies increase commensurate with further pregnancies (Van der Werf, 1971; Daughty and Gelsthorpe, 1974). Most maternal antibodies to HLA

160

DAVID R. JACOBY ET AL.

antigens are IgG and as such are capable of being transported across the placenta by an Fc receptor system (Brambell, 1970; Balfour and Jones, 1978).If the antibodies are directed toward an HLA specificity not manifest during that pregnancy, cord blood will contain maternal antibody. However, cord blood does not contain maternal anti-HLA when the specificity of the antibody is the same as that represented by the present pregnancy (Jeannet et al., 1977). This is likely a result of placenta] tissue that expresses HLA antigens. These incompatible cells bind the anti-HLA and thereby act as an immunoadsorbent removing maternal antibody and denying its entrance to fetal circulation. Sequestering of maternal antibody may also play a role in blocking the further sensitization by the mother to foreign antigens expressed by the fetus. For example, anti-HLA antibodies synthesized in the mother apparently cross the placenta and may react with fetal lymphocytes to mask effectively the paternal HLA determinants expressed by these cells (Dumble et al., 1977). Quantitative studies on the HLA antigens showed that the expression of maternal determinants on fetal lymphocytes was approximately the same as that on maternal cells, whereas determinants of paternal origin were decreased 30 to 90% compared with the amount on paternal cells. These antibodies have been termed blocking factors and are absent from the sera of some women with chronic spontaneous abortions of unknown etiology (Rocklin et al., 1976; Stimson et al., 1979; McIntyre and Faulk, 1983; Unander and Olding, 1983). These blocking factors have been demonstrated as IgG molecules with fetal specificity; however, whether they act by modulating the expression of these antigens or by blocking recognition by binding is unknown. However, raising of blocking factors using leukocyte transfusions has been employed to prevent chronic abortions (Taylor and Faulk, 1981; McIntyre and Faulk, 1983). The mixed lymphocyte reaction (MLR) is an in vitro correlate of allogeneic recognition and reactivity. Cocultures of allogeneic lymphocytes lead to blast formation and proliferation, primarily of T lymphocytes in response to alloantigens controlled by class I1 major histocompatibility (MHC) loci, and to a lesser extent, by a class I and some non-MHC loci. Sera from multiparous women can inhibit the MLR between maternal and paternal lymphocytes, as well as between maternal and fetal cells, and maternal and adult cells with different HLA specificities (Robert et al., 1973; Gatti et al., 1975). Similar results have been obtained in other in vitro systems. Lymphocytes from multiparous women make macrophage-inhibitory factor (MIF) in response to paternal antigens and placental extracts, whereas lympho-

REGULATION OF FETAL-MATERNAL

BALANCE

161

cytes from women during or after one or two pregnancies fail to make MIF in response to paternal alloantigens (Youtananukorn and Matangkasombut, 1972; Rocklin et ul., 1973). This suggests that only presensitized maternal lymphocytes can produce this factor. The secretion of MIF by presensitized maternal lymphocytes cultured with paternal cells can be blocked by autologous serum but not by sera from other multiparous women. Further, maternal cytotoxic activity toward semiallogeneic fetal trophoblast cells can be blocked if they are grown in maternal serum (Taylor and Hancock, 1975). The blocking factor present in maternal serum has been partially characterized and studied for its possible biologic function in pregnancy. It has been identified as IgG, based on an electrophoretic mobility similar to 7s molecules and removal of its activity after passage of maternal sera over an anti-IgG affinity column (Rocklin et al., 1976). These antibodies can be absorbed away from maternal sera by paternal cells but not by pooled human platelets, suggesting they are directed primarily to the HLA-D/DR antigens of fetal tissues (Rocklin et al., 1979).These antibodies are specifically cytotoxic to B cells, and their F(ab’)zfragments inhibit spontaneous rosette formation when treated with neuraminidase-treated sheep red blood cells (SRBC). Class I1 histocompatibility antigens are involved in antigen presentation in allogeneic lymphocyte responses. Therefore, the blocking of these receptors with antibody is conceptually a mechanism by which the fetus may not be rejected despite maternal sensitization to paternal antigens. An important role of these antibodies has been suggested by clinical studies of women who have recurrent idiopathic spontaneous abortions. Rocklin et al. (1976) found that, although the lymphocytes from chronic aborters were sensitized to paternal antigens as judged by their ability to secrete MIF, sera from these women did not contain a blocking factor that suppressed the MIF response to their mate’s lymphocytes. In contrast, sera from normal multiparous women could inhibit the M I F response to paternal lymphocytes. If a chronic aborter went on to have a successful pregnancy, the blocking factor could be demonstrated in her sera. This blocking activity has been detected in the sera of 80% of primagravid women between 16 and 24 weeks of gestation, but not in all women who have successful pregnancies (Stimson et al., 1979). However, adsorption from circulation by fetal tissue may decrease antibody titers to below detectable levels in these cases. Differentiated cytotoxic T lymphocytes (CTL) are rarely found in an active state in healthy individuals. Rather they are generated from

162

DAVID R. JACOBY ET AL.

precursors only after appropriate stimulation, as in response to virus infection or allogeneic cells. However, CTL found in maternal peripheral blood can specifically kill fetal cells and, to a lesser extent, paternal cells. Studies measuring specific cytotoxicity against semiallogeneic fetal lung cells demonstrated that 0 of 6 individual maternal lymphocyte samples were reactive at 8 to 14 weeks postgestation; 7 of 12 (58%)were reactive at 15 to 17 weeks, and 7 of 8 (88%) were reactive at 19 to 23 weeks (Timonen and Saksela, 1976). Autologous sera diminished the cytotoxicity of maternal lymphocytes for fetal cells in these assays, thus this activity may be blocked in uiuo. Studies in animals have shown both humoral and cellular responses to alloantigens during pregnancy. Most studies have been done in mice and rats because inbred strains are easily manipulated, and the structure of the placenta is similar to that in humans (hemochorial anatomical type). The antibody activities developed during pregnancy against paternal antigens are primarily hemagglutinating (Herzenberg and Gonzales, 1962), leukoagglutinating (Mishell et al., 1963),and cytotoxic (Carlson and Wegmann, 1978).The cell-mediated immune response to paternal H-2 antigens has been demonstrated in uivo by rejection of paternal skin grafts and in uitro in the MLR (Maroni and Parrott, 1973; Smith et al., 1978). In these animals, the maternal graft-versus-host reaction can occur across both H-2 and nonH-2 (minor histocompatibility) barriers (Baines et al., 1976). Lymphocytes sensitized during pregnancy are cytotoxic to embryonic tissue (Hamilton et al., 1976). Moreover, immune complexes are frequently found in the sera and basement membrane of renal glomeruli taken from pregnant mice (Tamerius et al., 1975). Two lines of evidence suggest that these complexes are the result of an immune response against neoantigens of pregnancy or of the fetus. First, immune complexes of this type are not found in pseudopregnant mice, and second, immunoadsorbent-coupled antibodies prepared against mouse embryo antigens remove these complexes. The circulating immune complexes are of the IgG isotype and able to block maternal cytotoxicity to tumor targets, and less dramatically to fetal tissue. In humans, immune complexes have been demonstrated in the sera of pregnant women (Masson et al., 1977), and these may be able to activate suppressor cells into secreting soluble suppressor factors (Moretta et al., 1977). However, other investigators have questioned the presence of immune complexes during a normal pregnancy (Theofilopoulos et al.,

1981). Several studies have shown that sera from parous women have autoanti-idiotypic antibodies that bind to autologous lymphoblasts sensi-

REGULATION OF FETAL-MATERNAL

BALANCE

163

tized in vitro against their husbands’ HLA-D/DR antigens (Suiu-Foca et al., 1983). Serologic determinants (idiotypes) on the antigen-binding receptors of B and T lymphocytes function as important antigenic stimuli that can induce the formation of anti-idiotypic antibodies. The strength of the reaction between idiotypic and anti-idiotypic receptors may be a major factor in regulating a specific immune response (Eichmann, 1978; Bona et al., 1981). Anti-idiotypic antibodies were found both in women who developed anti-HLA antibodies and in those who did not. This would suggest that the anti-idiotype antibodies could inhibit the anti-HLA reactivity during pregnancy by a network type of suppression. The anti-idiotypic antibodies present in sera bind to autologous allostimulated T lymphocytes and thereby inhibit their capacity to activate the autologous MLR.

2 . Maternal Hormones During pregnancy, production of hormones from several organs is enhanced. Experimental administration of some of these hormones, notably cortisone, results in lymphocytopenia, involution of lymphoid tissues, suppression of inflammation, and weakening of allograft reactivity (Bjornboe et al., 1951; Medawar and Sparrow, 1956; Cohen and Claman, 1971; Weston et al., 1973). However, human immune responses are relatively insensitive to corticosteroids administered in vivo (Long, 1957). Corticosteroids inhibit T lymphocyte proliferation in response to mitogens such as concanavalin A (Con A) and phytohemagglutinin (PHA) at concentrations comparable to physiologic levels. However, it remains unclear whether this in vitro sensitivity suggests a possible modulatory role in vivo. Indications to the contrary are that addition of 10 times the physiologic concentration of hydrocortisone is necessary to inhibit MLR between maternal lymphocytes and allogeneic cells (Schiff et al., 1975). Furthermore, the level of corticosteroids in maternal sera does not correlate with the ability of those sera to block the MLR, or maternal cytotoxicity to fetal and paternal tissue. Progesterone is produced primarily by the human placenta throughout pregnancy, during which much higher plasma concentrations are maintained than in normal nonpregnant females. Progesterone has been shown to suppress reactivity to skin allografts in experimental animals and inhibit the in vitro mitogenic response of lymphocytes (Ablin et al., 1974; Wyle and Kent, 1977). Physiologically, only 10% of progesterone is in a free, active form, the remainder being bound to transcortin and albumin. Pregnant women whose progesterone levels are very low, such as those hypophysectomized as protection against

164

DAVID R. JACOBY ET AL.

hormone-dependent malignancies, have borne healthy babies (Kaplan, 1961). Estrogens are synthesized from fetal and maternal precursors by the placenta, and are present in increasing levels as pregnancy proceeds. High doses of estradiol prolong skin grafts in mice and corneal grafts in sensitized rabbits (Hulka and Mohr, 1967; Waltman et al., 1971). However, there is no evidence that physiologic levels of estrogens are immunosuppressive; rather, they stimulate proliferation induced by Con A or PHA (Tomada et al., 1976). Pregnancy-associated a-macroglobulin is a high molecular weight glycoprotein, the level of which rises dramatically during pregnancy, then decreases to a comparatively low amount within 6 weeks postpartum (Stimson, 1976). This glycoprotein nonspecifically suppresses the MLR of maternal lymphocytes with paternal lymphocytes, PHAstimulated proliferation, and the production of MIF. However, the immunosuppressive properties of this glycoprotein cannot account for the total blocking activity of maternal sera. In summary, the general immunocompetence of a pregnant woman is at best only minimally affected by nonspecific factors in her circulation. Hence, the increased steroid levels in maternal blood afford only weak, if any, ancillary protection against rejection of the fetus as an allograft.

CONTRIBUTIONS TO FETAL MAINTENANCE B. PLACENTAL

1. The Uterus as a Privileged Site The placenta is an organ primarily of fetal origin that constitutes the interface between mother and fetus. Discerning the results of this intimate tissue contact is important to any understanding of maternalfetal immunology. A fully developed human placenta resembles a flat disk, approximately 15 cm in diameter (see Fig. 1).The outer fetal cell layers, or trophoblast, interdigitate and invade the endometrium of the uterus providing a large surface area for exchange between the two circulations. Early viable hypotheses to explain the maternalfetal balance during pregnancy postulated that (1) the invasive fetal tissues were antigen deficient; (2) the placenta provided an anatomic barrier separating maternal and fetal circulations; and (3)the maternal uterus constituted an immunologically privileged site, without effective lymphatic drainage, much like the anterior chamber of the eye or the hamster cheek pouch (Medawar, 1954). Current evidence indicates that the uterus is not an immunologically privileged site, because it is adequately drained by the pelvic

REGULATION OF FETAL-MATERNAL

BALANCE

165

Stratum spongiosum Limiting or boundary lavar Maternal vassals

Placental septum I

.Umbilical vain Marginal sinus

Umbilical corc

FIG.1. Anatomical structure of the fully developed human placenta. The maternal portion consists of the stratum spongiosum and placental septa, through which the uterine arteries and veins pass to and from the intervillous space. Fetal trophoblasts line the intervillous space. The fetal blood currents pass through the blood vessels of the placental villi, the walls of which separate the maternal and fetal circulations.

and para-aortic lymph nodes (Park, 1971). In experimental animals, transfer of allogeneic skin, leukocytes, or lymph node cells to one uterine horn stimulates ipsilateral hypertrophy of the para-aortic lymph nodes, sensitizes the host for accelerated rejection of a subsequent skin graft, and locally sensitizes the uterus for a delayed inflammatory reaction to a second inoculum of lymphoid cells (Beer and Billingham, 1974). These investigators further demonstrated that the uterine decidual response produced by pseudopregnancy or hormone injections prolongs allogeneic skin graft survival within the uterus, but the decidua alone were not sufficient to prevent rejection of the intrauterine grafts in presensitized hosts. Thus, the uterus does not seem to be protected from participation in immune reactions. 2. The Plucentu us an Immunologic Burrier Fetal arid maternal vascular systems are separate except where they interface at the placenta. Here, the barrier consists of the thin walls of the chorionic villi, and small amounts of fetal blood containing red cells, leukocytes, and lymphoid cells enter the maternal circulation as

166

DAVID R . JACOBY ET AL.

early as the fifteenth week of gestation (Adinolfi, 1975). In contrast to the easily observed passage of fetal cells into the maternal circulation, the transport of maternal cells across the placenta into the fetal circulation is uncommon. The placenta may act as an immunoadsorbent by selectively binding maternal antibodies with fetal specificity and preventing their entry into the fetal circulation. Maternal IgG not directed toward fetal antigens can cross the placenta and enter fetal circulation, and a degree of passive perinatal immunity results from this transferred antibody (Brambell, 1970; Gill, 1973). However, IgM, IgA, and IgE do not cross the placenta but can be eluted from trophoblast tissue (McCormick et al., 1971). Placental villous endothelium and villous macrophages have Fc receptors (Zuckerman and Douglas, 1975) that are able to bind circulating antigen-antibody complexes, thus potentially minimizing immune complex-mediated fetal damage. These immune complexes are reported to contain maternal antibodies to alloantigens in fetal heavy and light immunoglobulin chains (Faulk and Johnson, 1980). 3. Expression of Histocompatibility Antigens Placental tissues do express HLA antigens in humans and H-2 antigens in mice but not at the outermost cell layer. Extracts of mouse and rat placenta have been used to raise antisera that, when administered to animals of the same strain, cause placental damage and abortions (Koren, 1968; Beer et al., 1972). However, antisera to MHC class I, class 11, and &-microglobulin do not appear to react with the other trophoblast cell layer (Adinolfi et al., 1982). Faulk and co-workers (1978) have purified the trophoblast-specific minor histocompatibility antigens TA1 and TA2 from trophoblast cell cultures. Antibodies directed toward TAI blocks the MLR of maternal lymphocytes and allogeneic stimulator cells. Further, these antibodies show specificity for trophoblasts and are species specific (Faulk et al., 1979). Allogeneically stimulated lymphocytes express on their surface an antigen cross-reactive with TA1 (McIntyre and Faulk, 1979). Murine trophoblast cells of the placenta and peripheral lymphocytes have a highly sulfated sialomucoprotein on their surface. In theory, this substance renders the fetal tissues nonimmunogenic because of its intrinsic negative electrostatic charge, thereby repelling effector lymphocytes. Treatment of the trophoblast with neuraminidase to cleave off the sialic acid residues increased the immunogenicity and enabled ectoplacental cone trophoblast to induce transplanta-

REGULATION OF FETAL-MATERNAL

BALANCE

167

tion reactions in some experiments (Currie et al., 1968), although this was not confirmed (Simmons et al., 1971).

4 . Placental Hormones Hormones play little if any role in inhibiting the predicted allograft rejection of the fetus. However, certain hormones are found in greater concentrations at the placenta than in the general circulation. Because these agents are thought to modulate the immune response only at relatively high concentrations, they could in this way act specifically at the site of the placenta. Human chorionic gonadotropin (hCG) is a glycoprotein dimer synthesized in the Langerhans cells of the cytotrophoblast. The concentration of hCG rapidly rises to about 600 IU/g of tissue in the placenta during the second and third months of pregnancy. Thereafter, the hCG level declines until term (Loraine, 1961). Several studies have shown that hCG is an immunosuppressive agent. Thymic and splenic mass in rats was reduced after treatment with hCG, and in guinea pigs delayed hypersensitivity to the purified protein derivative of tuberculin (PPD) was diminished after injection of crude hCG (Han, 1975). Material extracted from urine of pregnant women, and containing hCG, inhibited material MLR with paternal cells and suppressed mitogen- or antigen-induced stimulation of maternal lymphocytes (Contractor and Davies, 1973; Adcock et al., 1973; Han, 1974). However, hCG suppression of PHA-stimulated lymphocyte proliferation was not strong (approximately 30%), and inhibition of the MLR required a dose range of 1000 to 10,000 IU/ml, above the physiologic concentration. Subsequently, isoelectric focusing demonstrated that the hCG preparations were heterogeneous and that suppressive activity in hCG preparations eluted more slowly from a molecular sieve Sephadex G-100 column than pure hCG (Caldwell et al., 1975). Thus, the suppressor activity associated with hCG may result from a lower molecular weight contaminant. Experiments with neuraminidase indicate that sialoglycoproteins found in the urine of pregnant women may be the inhibitory factor of these crude hCG preparations (Merz et al., 1976). Immunosuppressive properties have been described for human placental lactogen (hPL), a glycoprotein hormone first detectable in maternal plasma at 6 weeks of gestation then rising steadily in concentration throughout pregnancy. However, like hCG, hPL preparations with inhibitory activity are heterogeneous, and extensive purification of hPL removes this activity, even at doses much higher than physio-

168

DAVID R. JACOBY ET AL.

logic conditions. Hence, it is unlikely that either hCG or hPL significantly influences the immune reactivity of the mother. A number of other pregnancy-associated glycoproteins have been assigned immunosuppressive properties, but the evidence of in vivo effects is not firm. These include ovamucoid, &-glycoprotein (Gordon et al., 1977), and a-globulins such as pregnancy zone protein (PZ), pregnancy-associated a-macroglobulin, and pregnancy-associated plasma proteins A, B, C (PAPP-A, B, C) (von Schoultz and Stigbrand, 1973). The a-globulins are the best characterized of these hormones, appearing to immunosuppress T cell functions such as MLR, MIF production, rosetting with sheep red blood cells, and stimulation by mitogens and PPD. However, rigorous purity of these hormones has not yet been achieved. CONTRIBUTIONS TO PREVENT ALLOCENEICREJECTION C. FETAL

1 . Lymphocyte Interactions

The fetuses of humans and certain animals, such as mice, actively participate in their own nonrejection by potent suppression of maternal humoral and cellular reactivity (Weigle and Dixon, 1959; Olding and Oldstone, 1974; Mosier et al., 1977; Calkins and Stutman, 1978). Human T lymphocytes taken from the umbilical cord at term or from lymphoid tissue during pregnancy inhibit mitosis, PHA-induced proliferation, pokeweed mitogen (PWM)-stimulated immunoglobulin synthesis, and MLR of lymphocytes from mother and child. These phenomena are discussed in depth in Sections V1,A-D. 2 . a-Fetoprotein Human a-fetoprotein is a product of the embryonic yolk sac and fetal liver, and is one of the major serum components of the developing fetus (Alpert et al., 1971). Human a-fetoprotein production decreases in the late stages of pregnancy to be gradually replaced by serum albumin. Analysis of nucleic acid-coding regions has shown considerable homology between the a-fetoprotein and albumin genes (Beattie and Dugaiczyk, 1982). Murgita and Tomasi (1975) first reported that a-fetoprotein isolated from mouse amniotic fluid was a potent inhibitor of mitogen-induced proliferation and the MLR as well as primary and secondary humoral antibody responses. Preparations of a-fetoprotein consist of three polypeptide species of varying electrophoretic mobilities (Lester et al., 1976). A single species seems to contain all the immunosuppressive activity, but the biologic effect of contaminants is unknown. In mice, a-fetoprotein activates

REGULATION OF FETAL-MATERNAL

BALANCE

169

suppressor T lymphocytes, which may account for the inhibitory activity of neonatal lymphocytes because the normal physiologic decay of a-fetoprotein with increasing age largely parallels the loss of suppressor cell activity in the developing young (Murgita et al., 1981).

3. Prostaglandins Prostaglandins are a family of immunologically active products of arachidonic acid metabolism. Prostaglandins are present in fetal plasma, and concentrations increase with gestational age. These compounds are produced by placental tissues, fetal adrenal glands (Mitchell et al., 1982), and peripheral blood mononuclear cells (Johnsen, et aZ., 1982). In adults, a major source of prostaglandin biosynthesis is clearly monocytes, but controversy exists concerning the functional maturity of cord blood macrophages and their secretion of prostaglandins. Of their many biologic activities, the prostaglandins PGEl and PGEz are the major immunosuppressive species, inhibiting a broad range of immunologic functions. PGE added to cultures of T lymphocytes suppresses functions such as PHA- and Con A-induced proliferation (Goodwin et a/., 1978), colony formation in soft agar (Bockman and Rothschild, 1979), and proliferation and cytotoxicity of lymphocytes in MLR (Darrow and Tomar, 1980). PGE also suppresses unstimulated natural killer (NK) cell activity; however, these cells become resistant to PGEz when activated by interferon (Roder and Klein, 1979; Leung and Koren, 1982). Macrophage functions such as phagocytosis, locomotion, collagenase synthesis, plasminogen activator secretion, and inhibition of interferon-induced tumoricidal activity can be suppressed by PGE (Schultz et al., 1979). PGEz is synthesized by human fetal leukocytes (Johnson et al., 1983b) at concentrations known to be immunosuppressive for human adult lymphocytes in vitro (Johnsen et al., 1983a). Cord blood lymphocytes can suppress the mitogen-induced proliferation of maternal lymphocytes in cocultures. Addition of indomethacin or another prostaglandin synthetase inhibitor [5,8,11,14-eicosatetraynoic acid (ETYA)] abrogates the cord lymphocyte-mediated suppression of maternal proliferation as measured by I3H1thymidine (Tdr) uptake (Johnsen et al., 1982). However, the suppression of maternal cell division is restored when exogenous PGEz is added to cultures with cord lymphocytes at concentrations comparable to physiologic conditions (Johnsen et aZ., 1983a). The degradation and turnover of these molecules are very rapid, suggesting that they act locally at the placenta but probably do not remain active in the maternal vascular system.

170

DAVID R. JACOBY ET AL.

IV. Fetal Expression of Histocompatibility Antigens

ONTOCENCY AND DEVELOPMENT OF HISTOCOMPATIBILITY ANTIGENS 1. Preimplantation

Semen contains many antigenic components from the testes and accessory glands of the male reproductive tract as well as spermatozoa (Jones, 1976). MHC antigens are expressed on spermatozoa in both human and mouse as determined by indirect immunofluorescence and complement-mediated cytotoxicity assays. The ovum is surrounded by a mucoprotein coat, which allows specific interaction with spermatozoa and, after fertilization, becomes inpenetrable as a result of the release of active mucoids and glycoproteins (Yanagimachi, 1977). This mucoprotein coat, the zona pellucida, may protect the blastocyst during passage through the fallopian tube and uterus, although zona pellucida-free eggs can survive and implant in the uteruses of sensitized recipients (Kirby, 1969). In the mouse, immunoglobulins coat the blastocyst by the time of implantation and the zona pellucida is impermeable to them (Bernard et al., 1977; Sellens and Jenkinson, 1975). Murine blastocysts are capable of eliciting an immune response. Transfer of a blastocyst to ectopic sites in presensitized hosts leads to rejection, although the presence of the zona pellucida prolongs the time course of rejection (Kirby, 1968; Searle et al., 1976). Moreover, trophoblast cells from 4-day murine blastocysts inhibit the in vitro growth of macrophage monolayers (Fauve e t al., 1974). Minor histocompatibility antigens (non-H-2 determinants) have been demonstrated on the trophectoderm surface of the 4- to 8-cell morula and the early blastocyst by immunoperoxidase labeling, cytotoxicity, and immunofluroescence techniques (Searle et al., 1974; Heyner, 1973; Billington et al., 1977). Most investigators have not been able to identify major H-2 antigens on the outer trophectoderm of the blastocyst, but they are present on the inner cell mass (Searle et al., 1976; Webb et al., 1977).

2. Implantation As implantation nears, the zona pellucida is lost, and the outer trophectoderm’s non-H-2 antigenic determinants are no longer detectable. This occurs simultaneously with the first intimate tissue contact, and in conjunction with the development of polyploidy of other trophoblast cells, as well as the secretion of proteolytic enzymes and

REGULATION OF FETAL-MATERNAL

BALANCE

171

appearance of phagocytic properties associated with implantation (Searle et aZ., 1976). Trophoblast differentiation in vitro leads to loss of surface antigens, as it does in vivo, so this modulation of antigen expression probably is not dependent on the maternal environment. A variety of methods have been employed to determine whether the postimplantation trophoblast is immunogenic. However, many techniques have failed to separate trophoblast cells from other placental tissue, and contamination leads to unreliable results due to the expression of MHC antigens on nontrophoblast tissue. Youtananukorn et aZ. (1974) used a papain cleavage method of preparing soluble, surface histocompatibility antigens of the human placenta which were cultured with autologous lymphocytes. After the fifth month of gestation, lymphocytes from eight of eight primagravid women tested were sensitized to these antigens as judged by secretion of MIF. In the mouse, H-2 haplotypes of both maternal and paternal origin were demonstrated on the surfaces of trophoblast cells in suspensions of collagenase-treated placenta from 9 to 18 days after implantation, by using a sensitive radiometric technique (Chaterjee-Hasrouni and Lala, 1979). With this method, antigen density on 12- to 14-day trophoblast cells was comparable to that on F1 thyniocytes. Perfusion of the placenta via the uterine artery with radioiodinated antibodies specific to the paternal haplotype determined that significant levels of these antigens were expressed on the sinusoidal face of murine trophoblast cells in situ and in the labyrinthine region of the placenta (ChaterjeeHasrouni and Lala, 1982). However, trophoblast giant cells at the periphery of the placenta did not bind antibodies specific to paternal halotype, suggesting that they do not express detectable levels of H-2 antigens. Thus, in the mouse, class I MHC antigens are not detectable on outer trophoblast cells in contact with maternal uterine decidua, but the surfaces of trophoblast cells that lie in direct contact with maternal circulation do express these antigens. Murine trophoblast cells do not express detectable levels of MHC class I1 (Ia) antigens. Frozen sections of human placentas studied by immunohistologic techniques had HLA-A, -B, -C, and µglobulin reactive sites in mesenchymal and endothelial cells, but not on the surfaces or base1982). Similarly, ment membranes of trophoblast cells (Adinolfi et d., sections of human amniotic epithelia did not express detectable amounts of these antigens. Faulk and co-workers (1978) have used trophoblast cells to raise an antiserum to placental surface antigens. Two groups of these antigens exist: one group is also expressed on normal leukocytes (designated TAZ), and another group is expressed on certain human cell lines (designated TA1 ). Theoretically, anti-TA2

172

DAVID R. JACOBY ET AL.

antibodies produced in normal pregnancies block maternal lymphocyte alloreactivity. However, these antisera did not appear to have type specificity, because they reacted equally well with 25 different placental preparations. Studies in mice and humans suggest that histocompatibility antigens are expressed postimplantation by tissue that has direct contact with maternal blood but not by the cell layer invading the maternal uterine tissue. This distinction may account for sensitization of the maternal immune system to fetal antigens and subsequent production of blocking factors, and also for the unusual refractoriness of the outer trophoblast cell layer to lysis and other graft rejection mechanisms. There is general agreement that transplantation antigens are present on nontrophoblastic cells early in the human embryo’s development. Quantitative assays have revealed that expression of HLA antigens in several fetal tissues are very similar, relative to weight, to those of corresponding adult tissues (Dumble et ul., 1977). V. Immunologic Basis of Lymphocyte Interactions between Mother and Fetus A. IMMUNOCOMPETENCE OF THE FETUS

The intrinsic maturity of fetal lymphocyte functions is crucial in determining whether the fetus is actively involved in preventing its rejection or whether it is as immunologically inert tissue during pregnancy. The higher susceptibility of human newborns to life-threatening bacterial and viral infections has promoted many-faceted investigations of cellular and humoral immune competence of the fetus and neonate, and the development of these functions until maturity. The first immune-like reactions of the human fetus can be demonstrated with cells from the liver. At 7 to 10 weeks postgestation they are capable of strong proliferative responses when stimulated with allogeneic or xenogeneic cells in MLR, and also of binding antigen (Dwyer and Mackay, 1972; Carr et al., 1973; Stites et al., 1974; Asantila et al., 1974). However, both of these functions seem to be immunologically nonspecific. Thymic epithelium appears at approximately the sixth week of gestation and is populated with small lymphocytelike cells by the ninth to tenth week (Kay et al., 1962).T cells capable of forming spontaneous rosettes with sheep red blood cells are also observed after the ninth week of gestation (Asma et al., 1977). 1 . Responsiveness to T Lymphocyte Mitogens The ontogeny of T cell function has been mapped by fetal responsiveness to mitogens during pregnancy. The first PHA responses have

REGULATION OF FETAL-MATERNAL BALANCE

173

been observed with thymocytes taken at 10 weeks of gestation. Male fetal cells were screened by chromosome analysis to confirm that the dividing cells were not of maternal origin via transfer across the placenta. The onset of responsiveness to PHA by thymocytes appears at the developmental stage at which the thymus differentiates histologically into the cortex and medulla (August et al., 1971). PHA-responsive lymphocytes in the spleen and peripheral blood appear from approximately 14 weeks of gestation onward (Mumford et al., 1978). However, human fetal liver and bone marrow cells give only occasional proliferative responses to PHA. Comparison of mitogen responses in different fetal organs throughout gestation shows that the PHA response develops earlier and more strongly than the Con A response. The first cells to proliferate when stimulated with Con A are thymocytes at 13 to 14 weeks of gestation. In the spleen, cells are first responsive to Con A at approximately 18 weeks (Leino et d., 1980). This lag of Con A stimulation in comparison to PHA-induced proliferation by fetal lymphocytes is also demonstrable in animals, such as sheep (Leino, 1978), guinea pigs (Merikanto, 1979), and mice (Mosier, 1974). These findings support the concept that Con A and PHA stimulate at least partially different T cell subpopulations (Stobo and Paul, 1973). In humans, mitogenresponsive cells also appear first in the fetal thymus and thereafter in the spleen, suggesting that fetal T lymphocytes differentiate in the thymus.

2 . Cytotoxic Activity of Lymphocytes Cytotoxic activity toward maternal cells at the placenta is conceptually one way the fetus might effect immune suppression of a maternal graft response. However, fetal and neonatal cells have a limited or no capacity for specifically killing hapten-modified self, semiallogeneic (maternal), or allogeneic target cells (Granberg et al., 1979; Granberg and Hirvonou, 1980; Anderson et al., 1981; Olding et al., 1974; Olding and Oldstone, 1976). Irradiated, trinitrophenyl or fluorescein isothiocyanate-modified autologous cells incubated with presensitized cord bIood effectors were hyporesponsive or not detected in assay of cell-mediated lympholysis (CML). Additionally when neonatal lymphocytes are cultured with target lymphocytes from their respective mothers, no cytotoxicity was detectable ( W r release) after 6, 12, or 18 hours of incubation. Even at effector: target ratios as high as 100 : 1, CML was lacking, regardless of whether these studies were done in autologous serum (in the presence of blocking antibody) or with extensively washed cells in heat-inactivated fetal calf serum-containing

174

DAVID R. JACOBY ET AL.

culture media (in the absence of blocking antibody). Neither PHA nor soluble grown factors like interleukin 2 added to the cultures increased the specific CML. Thus, at birth, fetal cytotoxic lymphocytes apparently either are not generated or are inhibited by suppressor factors. 3. Activity of Nuturul Killer Cells NK cells are MHC nonrestricted lymphocytes that are spontaneously cytolytic for certain tumor cell lines as well as cells infected with several viruses or other infectious agents. It has been suggested that NK cells are not of a unique cell lineage but represent a maturational stage of either early monocytes or a T cell subset (Timonen and Saksela, 1977). Studies with mice have shown that NK activity is absent at birth, but develops soon thereafter, increases until animals are 6 to 8 weeks of age, and subsequently decreases (Kiessling et al., 1975; Dahl, 1980). The NK activity of human newborns is measurable but significantly lower than in adults. Cord peripheral blood mononuclear (PMN) cells tested for NK activity by lysis of a tumor cell line, K562, averaged 50% of adult activity at several effector :target ratios (Toivanen et al., 1981).Addition of leukocyte (Q-) interferon augments fetal NK activity to the same degree as adult cytotoxicity against tumor targets. Newborn leukocytes synthesize similar amounts of a-interferon as adult leukocytes (Bryson et al., 1979), suggesting that simple lack of interferon alone is not responsible for the decreased neonatal NK activity. Fetal lymphoid-like cells isolated from the liver as early as 9 to 11 weeks of gestation show NK activity at higher levels than cells taken from lymphoid tissue later in pregnancy, suggesting that a lymphoid precursor is demonstrating NK-like activity in the human fetus. 4 . Mounting Humoral Responses The activation of fetal B lymphocytes in vivo in response to antigens results mainly in IgM production, with the possibility of a very small contribution by fetus-derived IgG (Martensson and Fudenberg, 1965). Maternal IgG crosses the placenta during pregnancy, but at birth the individual has limited ability to synthesize IgG. B lymphocytes are first observed in the fetal liver at 10 to 11weeks of gestation, and express exclusively surface p+, until approximately the thirteenth week onward when the majority of B lymphocytes express p and y heavy chains. The frequency of B lymphocytes expressing different immunoglobulin heavy-chain isotypes increases as a function of age until adult proportions are reached around the fifteenth week of gestation. Virtually all of the fetal IgG+ and IgA+ cells

REGULATION OF FETAL-MATERNAL

BALANCE

175

coexpress IgM and IgD and therefore appear to have undergone intraclonal heavy-chain isotype switching (Lawton et ul., 1972; Hayward and Ezer, 1974; Gathings et al., 1977). B lymphocytes are found in highest frequencies in the blood (-30%), spleen (-35%), and lymph nodes (-13%) of fetuses, where the relative proportion of cells bearing each isotype is equal to or greater than that found in adult tissues (Gathings et ul., 1981). Thus, the isotype diversity at the B lymphocyte level appears to be generated early in gestation despite the relative paucity of stimulation from exogenous antigens. Full-term babies synthesize mainly IgM, but also IgG and IgA, and even premature babies are capable of vigorous antibody responses to a variety of antigens (Fink et ul., 1962; Uhr et ul., 1962). Nevertheless, human infants do not respond well to all antigens, and the nature of their antibody responses appears to be qualitatively different from that of adults. For example, young infants are low responders to a class of antigens, mainly polysaccharides, that are classified as thymus independent (Cowan et al., 1978). Moreover, in response to thymusdependent antigens, infants are slow to shift from IgM to IgG and IgA antibody production. The ontogeny of immunoglobulin production has been investigated in uitro using PWM, a T lymphocyte-dependent mitogen, or EpsteinBarr virsus (EBV), a T lymphocyte-independent mitogen that infects B lymphocytes and transforms them into autologously secreting plasma cells. In adults, polyclonal plasma cell responses of all immunoglobulin isotypes can be induced by both PWM and EBV. If cord blood lymphocytes are cultured in the presence of mitogenic concentrations of PWM, little or no immunoglobulin secretion is measurable after 7 to 21 days in culture (Andersson et d . , 1981; Jacoby and Oldstone, unpublished observations). However, cord blood lymphocytes have a greater rate of proliferation and incorporation of [3H]Tdr than adult lymphocytes in PWM-stimulated cultures. From birth onward, IgM and IgG secretion by lymphocytes in PWM-treated cultures increases until adult levels are reached when children are approximately 1 year old. EBV induces DNA synthesis and the formation of IgM plaque-forming cells (PFC; those B cells expressing surface p chain) of cord blood lymphocytes comparable to that obtained from lymphocytes of adults, indicating that the capacity for IgM formation is a fully mature function of the neonatal B lymphocyte. The kinetics of an IgM response to EBV-cultured cord blood lymphocytes is similar to that of adults, with maximum formation of IgM PFC on the seventh day of culture, tailing away by the tenth to fourteenth

176

DAVID R. JACOBY ET AL.

days. In contrast, cord B cells cannot form IgG PFC or IgA PFC in response to added EBV (Bird and Britton, 1979; Andersson et ul., 1980). In these experiments, depletion of T cells from the EBV cultures did not alter the immunoglobulin profile or response. Therefore, in vitro studies indicate the immaturity of B lymphocyte function in the human newborn in terms of IgG and IgA responses. EBV stimulates an IgM response comparable to that of adults; however, if cord blood lymphocytes are cultured in the presence of a T lymphocyte-dependent mitogen, limited immunoglobulin responses of all isotypes occur. This suggests that the humoral response of newborns is further inhibited by a lack or immaturity of helper T lymphocytes, or by active suppression of the B lymphocyte response.

5. MonocytelMacrophage Functions The ability of neonatal macrophages to process antigen, perform opsonization and phagocytosis, suppress PFC and immunoglobulin formation in cultures, and modulate immune responses by soluble factors such as prostaglandins has not been clearly defined. Argyris (1968) tested newborn mice and found that the antibody response to sheep red blood cells (SRBC) increased when this antigen was given with adult peritoneal exudate cells high in resident macrophage content. Transfer of adult macrophages to newborn rats protected the animals from lethal doses of Listeria monocytogenes and augmented the antibody response to keyhole limpet hemocyanin (Blaese and Lawrence, 1977). Most studies of macrophages from human newborns have dealt with nonspecific effector functions such as the ability to kill bacteria. Although this capacity is normal in newborns, these cells’ response to macrophage chemotactic factors is reduced (Kretschmer et al., 1976; Klein et al., 1976). Monocytes’ phagocytic dysfunction may result from the lack of opsonic factors rather than their inherent immaturity in the newborn. In adults, macrophage activation is dependent on a series of lymphokine signals, such as macrophage-activating factor (MAF) and y-interferon (Pace et al., 1983). However, newborn lymphocytes do not secrete y-interferon (Bryson et al., 1979), suggesting that macrophages are less active in neonatal immune responses. Current evidence suggests that fetal macrophages insignificantly suppress immunoglobulin synthesis in mitogen-stimulated cultures. Cord blood mononuclear cells profoundly suppressed immunoglobulin production in such cultures, and this suppression was fully reconstituted after removing monocytes by glass adherence, magnetic depletion of iron-ingesting cells, or passage over a G-10 or nylon wool column (Olding and Oldstone, 1976). Positive selection of cord blood

REGULATION OF FETAL-MATERNAL

BALANCE

177

macrophages by glass adherence and culturing with PWM and adult peripheral blood lymphocytes had no effect on normal mitogendriven PFC formation by lymphocytes from adults (Durandy et al., 1982). Others have shown that selection with the monoclonal antibody OKM 1did exert significant suppression in these cultures (Rodriguez et al., 1981). However, OKMl selects for monocytes, but also reacts with “null” cells and a subset of Ty lymphocytes (with Fc receptors for IgG) (Fox et al., 1981).In adults, mononuclear cells and especially macrophages are the main source of arachidonic acid metabolites, potent soluble modulators of several aspects of immune responsiveness. Human fetal circulation has significantly greater amounts of prostaglandins than that of adults, but these compounds appear to be synthesized by the placenta and fetal adrenal glands. PGEZ constitutes the main component of fetal prostaglandins, and this potent suppressive agent modulates the expression of Ia antigens in mice. Concentrations of PGEz found in fetal circulation decreased the expression of Ia on macrophages to below detectable levels in vitro (Snyder et al., 1982). This may account for the inability of mice to process certain antigens, and by extension, the human newborn’s inability to respond to certain polysaccharide antigens.

B. IMMUNOCOMPETENCE OF THE MOTHER The general ability of a mother to respond immunologically to antigenic challenge is an important issue in determining the mechanisms involved in the maintenance of the fetus. Medawar (1954) postulated that the state of pregnancy affected the normal immune response, possibly via hormonal influence, and that an impaired response to expressed paternal antigens contributed to survival of the fetus. Lymphoid tissues undergo temporary changes during pregnancy, including involution of the thymus and decreased size and cellular content of germinal centers of the para-aortic lymph nodes (Nelson et a1., 1973). Enumeration of lymphocytes during pregnancy indicated no differences in the overall circulating lymphocyte count between pregnant and nonpregnant individuals. However, in humans and other animals, the proportion of T to B lymphocytes decreased during the first trimester of pregnancy. This ratio gradually increased to normal levels after approximately 20 weeks of gestation (Strelkauskas et

al., 1975). Monoclonal antibodies have been developed that are specific for T lymphocytes and their functional subpopulations (Reinherz and Schlossman, 1980). In mothers tested at term with the monoclonal antibody OKT3, levels of circulating T lymphocytes closely resem-

178

DAVID R . JACOBY ET AL.

bled those of matched health adults (69.0 vs 71.5%, respectively, were OKT3+). Maternal blood has a higher proportion of OKT4+ (helper) lymphocytes at term (48.5 vs 41.0% OKT4+, respectively), whereas the levels of maternal OKT8+ (suppressor/cytotoxic) lymphocytes were normal (31.0 vs 29.5% OKT8+, respectively) (Jacoby and Oldstone, 1983). Lymphocytes of pregnant individuals synthesize immunoglobulins reactive to a variety of antigens. Production of specific antibodies to environmental antigens is normal, thus indicating that most helper T and B lymphocyte functions are unimpaired during pregnancy (Valquist et al., 1950). There are some reports of decreased levels of plasma IgG during the last trimester, although these may be due to placental transfer, rather than a defect in IgG synthesis (Maroulis et

al., 1971). The use of mitogens such as PHA to test maternal T cell reactivity has not given clear results. When specific cell-mediated immune reactions in MLR were evaluated, most reports described a normal or slightly depressed reactivity to fetal lymphocytes in one-way MLR. These experiments may be complicated by the presence of blocking antibodies in maternal sera capable of inhibiting MLR. Tuberculin skin reactivity is either normal or slightly depressed (approximately 25% of the cases) in the last trimester of pregnancy (Montgomery et al., 1968).Graft rejection can also be slightly impaired during pregnancy, as primary skin grafts undergo a longer time course until rejection. Presensitized mothers reject skin grafts within the normal time span (Peer, 1958). The level of complement, the main enhancer of antibody effect during graft rejection, has been evaluated during pregnancy. Amounts of total complement (CH50) and the third component (C3) progressively increase above normal levels during pregnancy (Kitzmiller et

al., 1973). In summary, the ability of the mother to respond to foreign antigens is normal during pregnancy. The maternal immune system remains functionally intact with observed changes playing minor or insignificant roles. Although pregnancy-related hormones are suppressive in uitro, the lack of significant suppression of general maternal immunocompetence indicates that physiologic concentrations of these molecules have no such effect on maternal immune responses. Thesefore, maintenance of the fetus is likely restricted at the placenta, not by systemic immunosuppression or reduced general immunocompetence.

REGULATION OF FETAL-MATERNAL

BALANCE

179

C. PERMEABILITY OF THE PLACENTA TO LYMPHOCYTES The placenta appears to be an incomplete anatomic separation of fetal and chorionic villi, and maternal circulation in the intervillous spaces can be permeable to leukocytes. Transfer of fetal lymphocytes into the maternal circulation occurs during a normal pregnancy. This has been documented by screening for the Y chromosome of mitogen-stimulated, dividing cells of mothers carrying a male fetus. Walknowska et al. (1969) demonstrated that the peripheral blood from 19 of 21 pregnant women contained lymphocytes from males. The incidence of fetal lymphocytes in the circulation was calculated to be 0.14 to 0.8%(1 fetal : 125 to 715 maternal lymphocytes). These observations have been confirmed using fluorescence-activated cell sorter (FACS) analysis. Fetal cells were segregated from large numbers of maternal peripheral blood lymphocytes by staining the peripheral lymphocytes with an antisera specific to paternal HLA antigens. In such studies, fetal cells occurred in maternal circulation as early as 14 to 15 weeks of gestation (Herzenberg et

al., 1979). The passage of maternal lymphocytes into the fetal circulation is less certain. After screening of male fetal blood for the 46XX karyotype, only 0.01%of mitogen-stimulated dividing cells of maternal origin were discovered (3 of 25,880 metaphase cells). However, screening for sex chromosomes of metaphase cells in mitogen-stimulated cultures is a poor test of this exchange, since fetal lymphocytes suppress division of maternal lymphocytes by 88 to 100%of their normal mitotic index (Olding and Oldstone, 1974). Although there is as yet no direct documentation of significant maternal lymphocyte transfer to the fetal circulation during normal pregnancy, there is evidence for transfer in disease states. Maternal T and B lymphocytes, identified by HLA typing, are present in the circulation of infants with severe combined immunodeficiency (Geha and Reinherz, 1983). Further, some human perinatal deaths are believed caused by graft-versus-host reactions caused by spontaneous transplacental passage of maternal lymphocytes into fetal circulation (Kadowaki et al., 1965). Beer and Billingham (1973) provided evidence for this concept in a series of elegant experiments in which the bone marrow cells of female inbred rats were destroyed by cyclophosphamide and then reconstituted with bone marrow from F1 hybrids possessing heterologous paternal antigens. These female chimeras were then mated with syngeneic males. More that 50% of the offspring suffered postnatal runt disease,

180

DAVID R. JACOBY ET AL.

which suggested that the chimeric cells of the mother had been transferred across the placenta, where they instigated graft-versus-host reactions and runt disease. VI. Human Maternal and Neonatal Lymphocyte Interactions

A. SUPPRESSOR ACTIVITYOF LYMPHOCYTES FROM HUMAN NEWBORNS Suppressor cells function to down-regulate normal humoral and cellular immune response mechanisms. Conclusions as to their prevalence and potency in the peripheral blood have largely relied on the development of antisera, particularly monoclonal antibodies, that recognize functionally distinct subpopulations of lymphocytes. The monoclonal antibodies OKT8 and Leu2 bind approximately 30% of adult peripheral T cells, and enriched populations of these cells, from adults, suppress immunoglobulin synthesis and mitogen-stimulated proliferation. Their role in vivo presumably relates to the homeostasis and regulation of immune responses, and they are rarely in an active state in healthy adults. However, some diseases such as multiple sclerosis and certain autoimmune disorders are characterized by active, profoundly suppressive lymphocytes in the peripheral blood (Strelkauskas et al., 1978; Huddlestone and Oldstone, 1979). Suppressor cells act, in most cases, directly on helper T lymphocytes, cells that produce the lymphokines required for B and T lymphocyte differentiation (Reinherz and Schlossman, 1980).

1 . Abrogation of Cell Division of Lymphocytes from the Mother Lymphocytes purified from the cord blood of human newborns have the ability to act as suppressor cells by several criteria. If one mixes peripheral blood lymphocytes from male fetuses in equal proportion to purified lymphocytes from their mothers and cultures them for 5 to 6 days without further stimulation, the mitotic index for cells having the 46XX karyotype is significantly reduced (Olding et al., 1974; Olding et al., 1977). In mitogen-unstimulated lymphocyte cultures, 90 to 100% of cells in metaphase are from the newborns (Table I). If the cord blood lymphocytes are cultured with those from nonpregnant women or from any unrelated adult, the same degree of mitotic index suppression is observed. Similar results were obtained from cultures in which mitogen is added. The prevalence of dividing lymphocytes of fetal origin in MLR results primarily from a 13-fold suppression in the mitotic index of lymphocytes from adults. Yet, the independent

REGULATION OF FETAL-MATERNAL

181

BALANCE

TABLE I LYMPHOCYTES FROM NEWBORNABROGATEMATERNAL LYMPHOCYTE CELLDIVISION IN MIXEDLYMPHOCYTE CULTURES Source of lymphocytes" Male

Female

PHA-P

Newborn Newborn Newborn Newborn Newborn Newborn Newborn Adult Adult Newborn Expected result

Mother Mother Unrelated mother Unrelated mother Unrelated mother Nonpregnant adult Nonpregnant adult Nonpregnant adult Nonpregnant adult Newborn 1: 1 Mixture

+' Od + + + 0 + + +

+ or0 + or0

Percentage of metaphase cells having Y chromosomeb 99 98 90 93 90 92 88 54 53 52 50

Equal numbers of lymphocytes (1-2 x lo6 cells) cultured together for 3 days.

' Cells treated with colchicine and stained with quinacrine dihydrochloride to mark Y chromosome. At least 50 cells in metaphase were examined from each culture. See Olding et al. (1974) for experimental details. Cultured in the presence of 50 pg/ml of PHA-P. Cultured in the absence of PHA-P.

growth of cord blood or maternal lymphocytes (cultured separately) is roughly equivalent. This suppression is not a property of the Y chromosome per se, since lymphocytes from adult males cannot inhibit division of lymphocytes from adult females. Despite suppression of maternal cell division, lymphocytes from newborns continue to proliferate at a slightly increased rate (twofold or less), suggesting that they are resistant to the inhibitory action they are responsible for. Furthermore, when lymphocytes from newborn males are mixed with lymphocytes from newborn females, lymphocytes from one newborn do not abrogate mitosis of another's lymphocytes (Olding and Oldstone, 1976). This suggests that neonatal lymphocytes may not express the receptor through which the inhibition takes place.

2. Suppression of Immunoglobulin Synthesis of Lymphocytes from the Mother Graded amounts of lymphocytes from newborns added to mitogendriven lymphocytes from adults synthesizing immunoglobulin, will progressively and strongly inhibit production of all immunoglobulin isotypes. This has been confirmed by a number of studies involving

182

DAVID R. JACOBY ET AL.

such mitogens as PWM, Con A, PHA, EBV, and water-soluble mitogen extracted from Nocardia opaca (NWSM) (Andersson et al., 1981; Miyawaki et al., 1981; Oldstone et al., 1977; Hayward and Lawton, 1977). An experiment that illustrates this phenomenon is the addition of cord blood lymphocytes to maternal peripheral blood lymphocytes (PBL) synthesizing IgG in a PWM-stimulated culture (Fig. 2). After 8 days in culture, suppression is assayed by the ability to inhibit the secretion of IgG into the supernatant in comparison to control cultures of maternal lymphocytes grown alone. The addition of relatively small numbers of cord blood lymphocytes suppresses IgG secretion of maternal cells, and the inhibition becomes greater as the ratio of newborn : maternal cells increases. Simply adding newborn’s lymphocytes to these cultures suppresses approximately 60% of the IgG synthesis, relative to the control cultures. Adding peripheral blood lymphocytes from adults to such cultures containing PWM has the expected effect of substantially boosting IgG synthesis presumably

0.1

1

I

1

1.o 2.0 Ratio of Added Adult and Cord Lymphocytes to Maternal PBLs

.5

FIG. 2. IgG synthesis by PWM-stimulated maternal PBL cultured with graded or autologous lymphocytes (i.e., from the amounts of cord blood lymphocytes (0-0) mother, 0--0)PBL . from the mother alone synthesize 1360 ng/ml of IgG. Addition of increasing amounts of cord blood lymphocytes results in suppression of IgG synthesis by lymphocytes from the mother. Addition of autologous lymphocytes (mother’s) increases IgG synthesis.

REGULATION OF FETAL-MATERNAL

183

BALANCE

from the addition of competent helper T cells and B cells (Jacoby and Oldstone, 1983).

3. Ontogeny and Persistence of Suppressor Activity Lymphoid-mediated suppression of adult lymphocyte proliferation and immunoglobulin production is an early immune function of the fetus. Cells isolated from the liver of an 8-week-old fetus that were of similar density and morphologically indistinguishable from lymphoid cells strongly suppressed the proliferation of adults' PHA-stimulated lymphocytes. This was also true for fetal lymphocyte samples obtained by the fourteenth week and later in gestation (Table 11; Unander and Olding, 1981). The appearance of suppressor activity in the human fetus coincided closely with the first emergence of other cellular immune reactions in vitro, for example, response to alloantigens in MLR and proliferation in response to PHA. Fetal suppression of PHAinduced adult cell division is strong throughout pregnancy, and progressively declines from birth to approximately 1 year of age (Table 11). These studies correlate very closely with those of Miyawaki et al., (1979), in which lymphocytes from children of various ages were added to adult peripheral blood lymphocytes and assayed for suppresTABLE I1 ONTOCENY AND POSTNATAL PERSISTENCE OF SUPPRESSOR ACTIVITY Fetal age" (weeks)

Infant age6

Source of lymphocytes

Percentage suppression of female lymphocytes"

6 weeks 3 months 8 months 11 months 1 year 4 months 1 year 5 months

PBL Liver PBL PBL Spleen Thymus PBL PBL PBL PBL PBL PBL

100 96 79 91 100 100 95 100 93 80 25 9

15 16 20

Weeks of gestation. Age after birth. Lymphocytes from male fetuses or newborns were cultured with lymphocytes from adult women in 1: 1ratio. Suppression determined by 1 - (mitotic index of female cells cultured alone/mitotic index of female cells cocultured with male cells) x 100. See Unander and Olding (1981) for experimental details. (1

184

DAVID R. JACOBY ET AL.

sion by the ability to inhibit differentiation of adult B lymphocytes (PFC), in PWM-stimulated cultures. Cord blood lymphocytes suppressed more than 80% of adult PFC formation, and suppression of PFC formation was still measurable in cultures containing lymphocytes from l-year-old, but not from 2-year-old children.

B. CHARACTERIZATION OF THE NEWBORNLYMPHOCYTE SUBSET RESPONSIBLE FOR SUPPRESSION

The attempts to characterize fetal and neonatal cell types that exert suppressor effects on the maternal immune response have utilized various fractionation techniques, such as rosetting, panning, and antiimmunoglobulin affinity chromatography (Rodriguez et al., 1981; Reinherz et al., 1982a; Chess et al., 1974). The resulting purified cell populations from cord blood are then added to adult lymphocytes in mitogen-stimulated cultures. Suppression is judged by inhibition of proliferation, B cell differentiation, and immunoglobulin synthesis. 1 . Thymus-Derived Lymphocytes

T lymphocytes isolated from cord blood, added to an equal number of maternal peripheral blood lymphocytes in PHA-stimulated cultures and incubated for 3 days, exhibited the same degree of suppression of maternal cell division as the addition of unfractionated neonatal lymphocytes (Table 111; Olding and Oldstone, 1976). From 12 such cultures examined, an average of 95% of the metaphase cells were from the male newborn. Reconstitution of cord B lymphocytes and monocytes to purified cord T lymphocytes in these cultures did not increase the observed inhibition. Moreover, purified populations of B lymphocytes and monocytes completely failed to suppress maternal lymphocyte division. Likewise, proliferating nonlymphoid cells from the fetus and newborn, such as amnion cells, spleen fibroblasts, and bone marrow fibroblasts, had no effect on the mitotic index of maternal lymphocytes (Olding and Oldstone, 1976). The addition of fetal or cord T blood lymphocytes to adult peripheral blood lymphocytes capable of synthesizing immunoglobulin in PWM-treated cultures profoundly inhibited PFC formation and IgG synthesis. The degree of suppression was largely dose related until ratios of cord T cells : adult lymphocytes begin to exceed 2 : 1, after which the addition of cord blood T lymphocytes had less effect (Yachie et al., 1981; Jacoby and Oldstone, 1983). Maximal suppression was approximately 85 to 100% of the normal immunoglobulin-synthesizing ability of adult lymphocytes (Fig. 3A). Purification of T lymphocytes from cord blood leads to a significant enrichment of suppressor activity, in comparison to that by unfractionated cord lymphocytes,

TABLE 111 CORDBLOODT LYMPHOCYTES INHIBITTHE PROLIFERATION OF LYMPHOCYTES FROM THE

MOTHER

Mitogen Cord blood cells"

PHA-P"

PWM

T T T T + B + M T + B + M B B M M

+ + + + +

0 0 0 0 0

+ +

+

98 96 90 95 95 33 59 7 0 50

+ +

0 0

Expected result?

Percentage of metaphase cells having Y chromosomec

0 0

or

+

T lymphocytes (T), B lymphocytes (B), and macrophages (M) from cord blood were isolated into homogeneous populations and cultured in equal numbers (2 x lo6 cells) with maternal PBL for 3 days. See Olding and Oldstone (1976) for experimental details. PHA and PWM added to cultures at a concentration of 50 pg/ml. Cells treated with colchicine and stained with quinacrine dihydrochloride to mark the Y chromosome. At least 50 cells in metaphase were examined from each culture. 1 : 1 Mixture of either lymphocytes from male and female babies or lymphocytes from male and female adults. Experimental tests gave mean 53 2 8%.

*

30001A B

2500

0.5 1

2

4

0.5 1

2

4

Ratio Of Cord Blood T Cells Added To 1 x 10' Maternal PELS In a PWM Stimulated Coculture

FIG.3. Effect of fractionated or unfractionated cord T lymphocytes on IgG synthesis. (A) Graded amounts of cord blood T lymphocytes (A-A) suppress maternal IgG synthesis, but after 1200 rad of y-irradiation (H-H),cord T lymphocytes are unable to suppress the IgG of maternal PBL cultured alone or with PWM (0-0). (B) OKT8- cord T lymphocytes (0-0) suppress IgG synthesis to the same degree as unfractionated cord T lymphocytes. OKT4- lymphocytes marginally effect the IgG synthesis of maternal PBL (0-0).

186

DAVID R. JACOBY ET AL.

indicating that this cell type is the major or exclusive mediator of suppressed immunoglobulin synthesis of adult lymphocytes in cocultures. Irradiation (1200 rad) of cord blood lymphocytes or purified T lymphocytes eliminates their ability to suppress adult proliferation in MLR, and immunoglobulin synthesis in PWM cultures. The dose response to varying amounts of y-irradiation is similar to that known to abrogate adult suppressor activity, although cord blood suppressor cells appear slightly more resistant to radiation. However, adult T lymphocytes continue to provide soluble factors required for B lymphocyte development and immunoglobulin synthesis even after receiving approximately 2000 rad of irradiation (Waldmann and Broder, 1982). Neonatal and adult T lymphocyte-mediated suppression is fully abrogated by 1200 rad of y-irradiation, illustrating the relative radiosensitivity of this activity, in comparison to the ability of adult T lymphocytes to provide helper functions.

2. Monocytes Cells of the monocyte and macrophage lineage play critical roles both as accessory cells required for the presentation of antigens in an immunogenic form and as negative regulators that suppress immune reaction. In adults, the capacity of large numbers of monocytes to suppress B lymphocyte activation was demonstrated by depleting peripheral blood mononuclear cells of monocytes, then showing augmented polyclonal activator-stimulated immunoglobulin synthesis (Laughter and Tworney, 1977). Reports conflict as to the functional maturity of cord blood monocytes and their ability to suppress adult immunoglobulin synthesis in PWM-stimulated cultures. Glass-adherent cells do not suppress adult lymphocytes division in PHA-stimulated MLR (Olding and Oldstone, 1976). Furthermore, neonatal macrophages added to PWM-stimulated adult lymphocyte cultures do not inhibit normal immunoglobulin synthesis (Durandy et al., 1982). Other evidence suggests that the inhibitory activities of cord and fetal peripheral blood mononuclear cells is not caused by or related to the presence of macrophages. First, depletion of cord blood monocytes by glass adherence, Sephadex G-10 chromatography, nylon wool columns, or magnetic separation of iron-ingesting cells did not affect the degree of suppression observed in mitogen-driven MLR, nor immunoglobulin synthesis by adult lymphocytes (Olding and Oldstone, 1976; Jacoby and Oldstone, unpublished data). Second, purification of cord blood T lymphocytes enriched this suppressor activity over that observed with unfractionated peripheral mononuclear cells (Jacoby

REGULATION OF FETAL-MATERNAL

BALANCE

187

and Oldstone, 1983). Third, monocyte-mediated suppression in PWM-treated cultures is unaffected by radiation doses of u p to 8000 rads (Gmelig-Myeling and Waldmann, 1981). In contrast, cord blood mononuclear cells subjected to relatively mild doses of radiation (1200 rad) completely lost their suppressive capacity in culture systems (Jacoby and Oldstone, 1983). 3. Enumeration of Cord T Lymphocyte Subsets The development of monoclonal antibodies that characterize subpopulations of T lymphocytes as functionally distinct in the adult has led to valuable comparisons between the prevalence of these markers in adult and cord blood. The monoclonal antibody OKT3 recognizes nearly all (>go%) of adult peripheral T lymphocytes; OKT4 reacts with a subset that provides T cell help in functional assays and comprises 50 to 60% of adult peripheral T cells; OKT8 marks a nonoverlapping population of adult peripheral T lymphocytes that are suppressor/cytotoxic in functional tests (Reinherz and Schlossman, 1980). Unfractionated peripheral blood lymphocytes from mothers at term, their infants, and unrelated adults have been analyzed for their distribution of these T cell phenotypes by indirect immunofluorescence and cytofluorographic analysis (Jacoby and Oldstone, 1983). Levels of OKT3+ T cells were similar in 26 pairs of maternal-cord blood lymphocyte samples and those from 33 unrelated adults. Of note, maternal-newborn samples differed significantly in the circulating levels of OKT8+ T lymphocytes. Of the 26 pairs investigated, 21 newborns had lower OKT8+ levels than their mothers. The lymphocytes from newborns were, on average, 20.5% OKT8+ compared to a 29.0% OKT8+ value obtained from their mothers. In contrast, OKT4+ levels did not differ significantly between newborns and their mothers; however, these values exceeded the proportion of OKT4+ lymphocytes in the peripheral blood of unrelated, nonpregnant adults. These studies corroborate those of Yachie et al. (1981)and Hayward and Kurnick (1981), who described significantly higher levels of OKT4+ lymphocytes and lower levels of OKT8+ cells in cord blood compared to peripheral T cells in unrelated adult samples. The OKT4+ : OKT8+ ratio is elevated in newborns relative to their mothers and unrelated adults, and gradually declines to adult levels as the infant reaches approximately 1.5 years of age. The decrease in this ratio parallels the progressive loss of suppressor activity exerted by newborns’ lymphocytes in uitro. The majority of resting peripheral T lymphocytes from adults do not express class I1 histocompatibility antigens (HLA-DRIMT loci), but may express these antigens after stimulation with mitogens, alloanti-

188

DAVID R. JACOBY ET AL.

gens, or soluble antigens (Indiveri et al., 1980; Winchester and Kunkel, 1980; DeWolf et al., 1979). Monoclonal antibodies specific to this antigen framework (OKIal) demonstrate that only 3% of unstimulated adult T lymphocytes express DR antigens by immunofluorescence. This level increases to nearly 50% when T lymphocytes are cultured in the presence of PWM (Miyawaki et al., 1982). The expression of DR is largely restricted to the OKT4+ subset, because the reciprocal OKT8+ subset neither proliferates nor expresses these antigens after PWM stimulation. Cord T lymphocytes do not express class I1 histocompatibility antigens either in an unstimulated state, or after culture with PWM for 3 to 21 days (Miyawaki et al., 1982; Jacoby and Oldstone, unpublished results). Stimulated T lymphocytes from newborns at varying ages from birth onward begin to express DR antigens, reaching adult levels at approximately 2 to 3 years of age. Yachie et al. (1982) have shown in adults that OKT4+DR+lymphocytes are critical in activating OKT8+ suppressor cells in PWM-stimulated cultures. Although cord blood has a greater proportion of OKT4+ T cells, it appears to lack the OKT4+DR+ subset. C. ASSAYS OF SUPPRESSION BY CORDT LYMPHOCYTE SUBSETS 1 . Separation by Fc Receptors Human T lymphocytes in the peripheral blood have been shown to have receptors for the Fc fragment of IgG (Ty) or IgM (Tp) (Moretta et al., 1977). Adult Tp cells when added to B lymphocytes in PWMtreated cultures are capable of inducing differentiation and immunoglobulin synthesis. Ty cells added to these cultures suppressed immunoglobulin synthesis, even if PHA, Con A, or PWM was the mitogen (Moretta et al., 1977; Pichler and Broder, 1981). Oldstone et al. (1977), using depletion and reconstitution procedures, demonstrated that cord blood lymphocytes bearing the Ty phenotype were responsible for the suppression of maternal immunoglobulin synthesis. Using a 5- to 7-day PWM culture system, they showed in all eight of the mother-baby pairs studied that purified Ty cells caused the same increment of suppression as did the total cord blood lymphocyte population. Further, addition of Tp cells and macrophages did not enhance the degree of suppression. With a different assay, Hayward and Lydyard (1978)noted that Tp cells suppressed adult lymphocyte immunoglobulin synthesis. This apparent discrepancy of Ty and T p suppressors was resolved by experiments of Durandy et al. (1979), who demonstrated that both Ty and Tp cells suppressed adult B cell differentiation in PWM-stimulated cultures. Tp lymphocytes exerted

REGULATION OF FETAL-MATERNAL

BALANCE

189

transient suppression on proliferation and PFC formation that ceased after 18 hours of in nitro cultivation. In contrast, Ty cells from newborns inhibited adult PFC formation by 85% in similar cultures, and the activity persisted during incubation for 5 to 7 days. Furthermore, cord blood Ty but not Tp cells inhibited the division of maternal lymphocytes in PHA-stimulated MLR. Hence, in the newborn, both Ty and Tp subsets define suppressor populations. The Tp suppressor cell is activated early and its action is transient. Suppressor activity of Ty cells occurs later and is longer lasting. In the adult, Ty cells can be induced to become suppressor cells only after the in vitro fixation of immune complexes (Moretta et al., 1977). Ty lymphocytes in cord blood are already in an active state, suggesting that immune complexes in maternal circulation (Masson et al., 1977) may activate fetal suppressor cells in vivo. However, treatment of neonatal Ty lymphocytes with trypsin or pronase to remove bound immune complexes did not abolish suppression of adult lymphocyte proliferation. Also, incubation of adult T lymphocytes in maternal or newborn sera did not induce suppressor activity, suggesting that cord blood suppressor lymphocytes may be primed by other mechanisms. 2. Fractionation b y Monoclonal Antibodies Cord blood T lymphocytes specifically depleted by monoclonal antibody and complement of cells expressing the OKT4 antigen (OKT4-, i.e., enriched for OKT8+ cells) are unable to exert a suppressor effect on IgG synthesis by maternal lymphocytes in PWM-treated cultures (Jacoby and Oldstone, 1983) (Fig. 3B). This remains true after incubation for 3, 5, 8, or 12 days and when the ratio of added cord blood T lymphocytes :maternal lymphocytes varies from 0.1 : 1 to 4 : 1. To the contrary, when cord blood lymphocytes are treated with OKT8 antibody and complement (OKT8-, enriched for OKT4+ lymphocytes) and cultured under similar conditions, potent suppression of IgG synthesis follows. This phenomenon is significant at low ratios of added neonatal T lymphocytes : maternal lymphocytes and increases as greater numbers of newborns’ T cells are added. The maximal degree of suppression observed is an average of 86% at the highest dose of added cord OKT8- cells (Table IV). These results and experiments by Yachie et al. (1981) indicate that cord OKT8- lymphocytes suppress B cell differentiation in mitogen-stimulated cultures. This outcome contrasts with those in functional assays of negatively selected adult lymphocytes in nitro. Adult OKT4- lymphocytes suppress proliferation, B cell differentiation, and immunoglobulin synthesis. Conversely, adult OKT8- lymphocytes provide soluble factors required for immunoglo-

TABLE IV OKT8- LYMPHOCYTES FROM NEWBORNS SUPPRESS IgG SYNTHESIS OF MATERNAL PBL Ratio of cord T cells :maternal PBL" in culture Cord blood T cells added to culture

0.5

1

2

(% suppression)b

(% suppression)

(% suppression)

4 (% suppression)

Unfractionated Irradiated (1500 rad) OKT4OKT8Negative control (media alone)

67.0 & 12.W 1.0 ? 9.0 18.0 2 10.0 61.5 f 7.0 -2.0 2 15.0

72.0 f 10.5 -5.0 t 10.0 13.0 f 10.0 74.0 5 5.5 1.0 2 14.0

78.0 t 10.0 -15.0 2 10.0 11.0 2 11.0 81.0 f 7.0 5.0 2 10.0

84.5 2 6.0 -19.0 2 11.5 13.0 f 11.0 86.0 t 5.0 -4.0 f 12.0

a Graded amounts of cord T cells added to 5 x lo5 maternal PBL in the presence of 15 pl/ml PWM. Cultured in U-bottom 96-well plates for 8 days in IgG secreted into the supernatant measured by human IgG ELISA. See Jacoby and Oldstone (1983)for experimental details. Suppression index (%) is calculated as 1 - (immunoglobulin synthesis of maternal PBL + cord T lymphocytes/immunoglobulin synthesis of maternal PBL) x 100. Negative indices indicate an elevated level of IgG synthesis. Mean 2 SD of four experiments.

REGULATION OF FETAL-MATERNAL

191

BALANCE

bulin synthesis when added to purified B lymphocytes, and boost the immunoglobulin synthesis of peripheral blood lymphocytes when added to PWM-treated cultures. Similar results with cord blood lymphocyte subsets were obtained from positive selection experiments. The functional profiles of newborn and maternal T cell subsets selected in this fashion differ substantially. OKT4+ lymphocytes and to a lesser degree OKT3" lymphocytes (total T) from mothers or unrelated adults, when added to autologous PBL, mediate a boost in IgG synthesis, indicating a predominate helper activity (Table V). In contrast, several ratios of OKT3+ and/or OKT4+ lymphocytes from newborns suppress the IgG synthesis by lymphocytes from their mothers. This suppression is comparable to that in the negatively selected OKT8- subset. Positively selected maternal OKT8+ cells suppressed autologous IgG suppression when added to cultures, confirming that maternal T lymphocyte subsets are functionally similar to those in nonpregnant adults tested in vitro. However, cord OKT8+ lymphocytes do not suppress IgG synthesis of maternal lymphocytes in uitro. The IgG secretion levels do not significantly differ when maternal lymphocytes are cul-

TABLE V T CELLSUBSETSFROM MOTHERS AND THEIRNEWBORNS HAVEDIFFERENT

FUNCTIONAL PROPERTIES Source of positively selected T cells

Ratio of'T cells : maternal PBL" in culture

0.5 Cord (baby)

Autologous (mother)

OKT3' OKT4' OKT8+

OKT3+

-

OKT4' OKT8+ -

Negative control (media alone)

ng/ml

1975 t 275" 1070 f 310 2165 5 370 1380 2 265 1245 f 180 1480 2 175 1650 f 275

2 p value

<0,05'

ng/ml

2130 ? 325 715 f 190 2390 f 485 820 ? 235 1030 f 120 1535 2 215 1740 f 170

p value

<0.005

Graded amounts of purified T cells added to 5 x lo5maternal PBL in the presence of 15 pl/ml of PWM. Cultured in U-bottom 96-well plates for 8 days and secreted IgG measured by ELISA. Mean 2 SD of three experiments. Probability that cord and maternal T cells are functionally indistinct.

192

DAVID R. JACOBY ET AL.

tured alone or with OKT8+ lymphocytes from cord blood, suggesting functional immaturity or inactivity of the OKT8+ T cells from newborns.

D. MECHANISMSOF CORDT LYMPHOCYTE-MEDIATED SUPPRESSION To determine whether the suppression exerted by cord blood lymphocytes results from an inducer/suppressor phenomenon in which newborn lymphocytes activate maternal suppressor cells in culture, or whether cord blood lymphocytes are direct effectors of inhibited maternal IgG synthesis, cord blood T lymphocytes were added to maternal lymphocytes specifically depleted of the suppressor subset (Jacoby and Oldstone, 1983) (Fig. 4). This was accomplished by two methods: first, treatment of maternal lymphocytes with OKT8 antibody and complement and, second, separation of maternal T and B lymphocytes and y-irradiation of the T lymphocytes with 1200 rad.

l00-J RATIO OF CORD T CELLS TO FRACTIONATED MATERNAL PELS IN A PWM STIMULATED COCULTURE

FIG.4. Suppression index (recorded as percentage suppression) of cocultures with cord T lymphocytes added to fractionated maternal PBL. Control cultures illustrate cord T lymphocyte suppression of unfractionated maternal PBL (0-0). Treatment of maternal PBL with OKT8 antibody and complement (W-W) or irradiation of T lymphocytes with 1200 rad (A-A) had no effect of the degree of suppression. Irradiation of added cord T lymphocytes (0-0) abrogated suppression and slightly stimulated maternal IgG synthesis. See Jacoby and Oldstone (1983) for experimental details.

REGULATION OF FETAL-MATERNAL

BALANCE

193

Cord blood lymphocytes were then added to the cultures in which maternal T and B lymphocytes were reconstituted. No significant difference in the degree of suppression resulted when graded amounts of cord blood T lymphocytes were added to unfractionated maternal lymphocytes or maternal lymphocytes treated with OKT8 antibody and complement. Although the suppression indices were similar, the maximal synthesis of IgG varied among these cultures. For example, 5 X lo5 maternal lymphocytes secreted an average of 1385 ng/ml IgG into culture supernatants, whereas 5 x lo5 OKT8- maternal lymphocytes produced 1890 ng/ml IgG, suggesting that depletion of the maternal suppressor cell subset resulted in increased immunoglobulin synthesis. Complement-mediated lysis of OKT8+ cells may not fully deplete maternal T cells of lymphocytes capable of becoming suppressors in PWM-stimulated cultures; however, when T lymphocytes from cord blood were added to cultures in which maternal T lymphocytes had been irradiated, no change was observed in the former’s ability to inhibit IgG synthesis. Therefore, cord blood T lymphocytes do not act by inducing suppressor lymphocytes among mothers’ cells in these cultures but rather act directly on the immunoglobulin-synthesizing mechanism. There is evidence that cord T lymphocyte-mediated suppression may block immunoglobulin synthesis at the plasma cell level. Blaese et al. (1982) have demonstrated that mitogen-stimulated neonatal T lymphocytes can inhibit the autologous secretion of immunoglobulin secretion by EBV-transformed B cells, whereas adult mitogen-induced suppressor cells are unable to suppress this immunoglobulin synthesis. Miyawaki et al. (1981)observed that using NWSM (Nocardia water-soluble mitogen), a T-independent B cell mitogen, to stimulate adult B cells, addition of cord T lymphocytes inhibits the IgG and IgA PFC formation in cocultures but does not affect the ability of adult B lymphocytes to express the p chain. Thus, it appears that suppression of adult B cell differentiation by OKT4+ newborn lymphocytes occurs when B cells express p chains, but before they become committed to expressing other heavychain molecules. Whether neonatal T lymphocytes suppress other interactions involved in immunoglobulin synthesis is unknown, but they can inhibit other T lymphocyte functions such as proliferation in the MLR.

1 . Soluble Suppressor Factors

The down-regulation of B cell responses by antigen-specific as well as nonspecific suppressor T lymphocytes is mediated by soluble suppressor factors in experiments with mice both in vitro and in uivo.

.

194

DAVID R. JACOBY ET AL.

Several investigators analyzing suppressor cells in humans have found soluble factors that mediate suppression. Saxon and Stevens (1978) reported that cell-free supernatants of PWM-stimulated suppressor T lymphocytes inhibited the immunoglobulin production by adult lymphocytes. Uytdehaag et al. (1979) demonstrated that suppression of the antibody response by human lymphocytes in vitro was mediated by a soluble factor released by T lymphocytes after activation with high doses of antigens for 24 hours. Moretta et al. (1977) provided evidence that Ty cells, induced to become suppressive by binding immune complexes, inhibited B lymphocyte differentiation by secreting a soluble suppressor factor. Marbrook-Diener culture systems have been used to investigate whether suppression mediated by cord blood T lymphocytes requires cell-to-cell contact or if soluble factors released by these lymphocytes elicit the same effect. The culture system consisted of two glass chambers, the inner chamber separated from the outer by a cell-impermeable membrane, such as a Nucleopore (0.4 pm), Millipore (0.22 pm), or dialysis membrane. Lymphocytes from the baby rested on the membrane of the inner chamber, and lymphocytes from the mother or an unrelated adult were placed in the outer chamber. These were cultured in the presence of PHA for 60 to 72 hours. After culture, the lymphocytes were transferred to microtiter plates and pulsed with [3H]Tdr. Cord blood T lymphocytes released a dialyzable product(s) capable of suppressing mitogen-induced proliferation of the maternal lymphocytes (Olding et al., 1977; Nagaoki et al., 1980). The suppression was not restricted to cells from the natural mother, but also affected lymphocytes from unrelated pregnant and nonpregnant adults. When adult cells were cultured in both chambers, no suppressor factor was released. Additionally, lymphocytes from one baby did not interfere with the proliferation of lymphocytes from another baby. The discrepancy noted between the degree of maternal lymphocyte suppression in MLR (90 to 100% of cell division), PWM-driven B cell differentiation (85 to loo%), and suppression of proliferation in the double chamber (40 to 71%) may have resulted from short-range activity of the suppressor factor released by cord blood lymphocytes. Also, these experiments do not rule out the possibility that both dialyzable and nondialyzable substances mediate the potent suppression in mixed cultures. Miyawaki and colleagues (1981) incubated PWMstimulated cord T cell supernatants with adult lymphocytes and observed that cell-free supernatants only suppressed adult B cell differentiation to the same degree as soluble, dialyzable factors in double-chamber experiments (45%). The differences of suppression

REGULATION OF FETAL-MATERNAL

BALANCE

195

in cocultures and in double-chamber experiments could arise from the instability of dialyzable suppressor factors and/or facilitation of suppression by cell-to-cell contact.

2. Characterization of Soluble Suppressor Factors Experiments with several sizes of Millipore and Auricon membranes in the Marbrook-Diener culture system indicated that the suppressive soluble factor secreted by cord blood T lymphocytes has a molecular weight of less than 500 and is extremely labile (Johnsen and Olding, unpublished). The possibility that these substances may be prostaglandins was investigated by exposing maternal peripheral mononuclear cells either to the dialyzable products of newborn lymphocytes stimulated with mitogen in a double-chamber system, or to newborn lymphocytes and their products in coculture. The effects of indomethacin (2.8 or 28 pM concentration), which inhibits the cyclooxygenase pathway of arachidonic acid metabolism, or of 5,8,11,14eicosatetraynoic acid (ETYA, 3.3 or 33 p M ) , which blocks both the cyclooxygenase and lipoxygenase pathways, were then assayed. Prostaglandin synthetase inhibitors completely abrogated the suppression exerted by neonatal lymphocytes and led to stimulation of maternal cell proliferation. Cells from one baby slightly stimulated those from another baby, and the addition of indomethacin enhanced this activity. Indomethacin per se had no stimulatory effect on the proliferation of maternal or newborn lymphocytes. Durandy et al. (1982)later demonstrated that indomethacin and anti-PGEz antibodies also blocked the suppression of adult B lymphocyte differentiation induced by cord blood lymphocytes. Thus, prostaglandins are the probable effector molecules in the suppression exerted by lymphocytes from newborns. Comparing the effects of several prostaglandins (A, E l , Ez, F I N ,F2a, 12) on Con A- or PHA-stimulated lymphocytes, Goodwin et al. (1978) and Johnsen et al. (1983a, b) found that PGEl and PGEz were the major immunosuppressive species. Further, PGEz added exogenously to cultures in which the endogenous synthesis had been blocked by indomethacin had differential effects on lymphocytes from newborns M supand adults. That is, PGEz at a concentration of 1.4 X pressed the proliferation of adult lymphocytes in contrast to the 1.4 x M or greater required to suppress of cord blood lymphocytes. Thus, lymphocytes from adults appear to be 100 times more sensitive to the immunosuppressive effects of PGE2 than lymphocytes from newborns. This suggests a potential mechanism by which cord blood lymphocytes suppress adult lymphocytes but not themselves. Binding

196

DAVID R. JACOBY ET AL,

assays of labeled PGE indicated that each maternal lymphocyte bound an average of 1800 molecules by high-affinity receptors. The binding of PGEz by cord blood lymphocytes was too small to allow an accurate calculation but was 30% or less of the binding by maternal lymphocytes (Johnsen and Olding, 1983). Thus, the difference in sensitivity between neonatal and maternal lymphocytes to prostaglandin E may be partly due to less high-affinity receptors present on cord blood lymphocytes.

E. HELPERACTIVITYOF CORDBLOODT LYMPHOCYTES The requirement for helper T lymphocytes in B cell differentiation and IgG synthesis in PWM-stimulated cultures was exploited to clarify whether T lymphocytes from newborns could provide the factors necessary for immunoglobulin production (Jacoby and Oldstone, 1983). Graded amounts of T lymphocytes from newborns were added to rigorously purified maternal B lymphocytes and cultured for 8 days in the presence of PWM (Fig. 5A). Unfractionated cord T blood lymphocytes added to these cultures elicited no IgG response from maternal B lymphocytes. In contrast, control cultures containing autologous maternal T lymphocytes added back to maternal B lymphocytes elicited a strong response. Cord blood T lymphocytes subjected to 1500

25007

-E

-

B

A

1

2000-

I

m

.- 1500Z I 5 1000-

2 -

% 500-

-

I

I

I

0:5

i

i

4

Ratio Of T Cells Added To Maternal B Cells (5x 10') In a PWM Stimulated Coculture

FIG.5. IgG synthesis of purified maternal B lymphocytes cultured with graded amounts of T lymphocytes. (A) Maternal T lymphocytes (0-0) elicit strong IgG response, but cord T lymphocytes (A-A) do not. Irradiated cord T lymphocytes ).-.( induce a significant B lymphocyte response. (B) OKT4+, irradiated cord T lymphocytes (0-0)give responses similar to those by irradiated cord T lymphocytes (H-H). OKT8+, irradiated T lymphocytes (0-0) provide no response.

REGULATION OF FETAL-MATERNAL

BALANCE

197

rad of y-radiation and subsequently added to the culture, raised the IgG synthesis of maternal lymphocytes substantially above the background of B lymphocytes cultured alone. Although experimentally reproducible and significant, the maximal response was limited to less than 40% of IgG produced by the positive control. Specific depletion of the cord blood OKT4 subset (OKT4-) or the OKT8 subset (OKT8-) by treatment with monoclonal antibody and complement did not enable the remaining T lymphocytes fo induce the synthesis of IgG from maternal B lymphocytes. Other experiments used panning for positive selection of purified cord blood OKT4+ or OKT8+ subsets, which were then exposed to 1500 rad of y-radiation and cultured with maternal B lymphocytes to determine whether radioresistant subpopulations induced IgG synthesis (Fig. 5B). Irradiated cord OKT4+ lymphocytes provided a significant degree of helper T lymphocyte factors as indicated by levels of IgG synthesis similar to those of cultures with irradiated, unfractionated cord blood T lymphocytes. Irradiated OKT8+ lymphocytes did not stimulate IgG synthesis in these cultures. Further evidence of helper activity in cord blood lymphocytes has been found in several studies. Not only was inhibition of maternal lymphocyte activity abrogated by irradiation of added cord T blood lymphocytes, but an increase in B cell differentiation and IgG synthesis were also noted (Fig. 3A, Table IV). Additionally, when indomethacin blocked the suppression of maternal PBL proliferation mediated by neonatal lymphocytes, the rate of maternal cell proliferation increased, even though indomethacin alone had no effect on maternal cells. Helper activity by cord blood T lymphocytes had not been described previously, yet premature and full-term babies are capable of vigorous antibody responses to certain antigens. This helper activity probably was not detected because it is normally obscured by the potent suppression exerted by lymphocytes from fetuses and newborns.

F. REGULATORY T LYMPHOCYTES IN NEWBORNS AND ADULTS The cord blood OKT4+T8- T lymphocyte subset has two distinct regulatory subpopulations: a potent radiosensitive suppressor population and a radioresistant helper population as demonstrated in PWMstimulated cocultures with lymphocytes from adults (Table VI). OKT8+ cells from the newborn appear to be functionally immature because of their inability to provide radioresistant helper activity, suppress maternal IgG synthesis, or exhibit significant cytotoxic responses to allogeneic cells, semiallogeneic cells, or hapten-modified syngeneic cells.

198

DAVID R. JACOBY ET AL.

TABLE VI FUNCTIONAL PROPERTIES OF T LYMPHOCYTE SUBSETS FROM MOTHERSAND NEWBORNS ~

~ ~ _ _ _ _ _ _ _ _ _ _

Newborn

Mother

Effect of 1200 rad y-irradiation

OKT4+T8OKT4+T8-

OKT4+T8OKT4+T8OKT4-T8+

Resistant Sensitive Resistant Sensitive Sensitive

-

Function

Activity

Helper" Suppressorb Helper Suppressor Suppressor

Minor Major Major Minor Major

Helper activity determined by culturing T lymphocyte subset with purified B lymphocytes in the presence of PWM for 8 days. IgG secreted into the culture supernatant measured by ELISA. Suppressor activity determined by adding T lymphocyte subsets to adult peripheral blood lymphocytes in PWM-stimulated cultures. Suppression was judged as the inhibition of IgG synthesis compared to that by the lymphocytes cultured alone.

Although OKT4+ populations in newborns and adults have different functional properties, recent reports of heterogeneity in the adult OKT4+ subset complement that found in neonatal regulatory cells. That is, not all adult OKT4+ lymphocytes stimulate B lymphocyte differentiation and immunoglobulin synthesis in mitogen-treated cultures. Studies with the monoclonal antibodies 5/9, OKT17, and TQ1 have differentiated functionally discrete populations within the OKT4+ subset (Corte et al., 1982; Thomas et al., 1982; Reinherz et al., 198213). For example, the monoclonal antibody 5/9 reacted with 15% of peripheral T cells that in functional studies provided soluble factors required for B cell differentiation and proliferated in response to soluble antigens or allogeneic challenge. This subset comprised only onethird of OKT4+ cells, and the remaining OKT4+5/9- cells demonstrated none of these properties, suggesting that they are not helper cells by these criteria. Evidence for suppressor cell activity within the adult OKT4+ subset comes from two separate studies. Juvenile rheumatoid arthritics (JRA) have naturally occurring anti-T cell autoantibodies that react predominantly with OKT4+ lymphocytes. Depletion of JRA+ T cells enhanced PWM-stimulated immunoglobulin production, despite a concomitant increase in OKT8+ cells (Morimoto et al., 1981).Further, Thomas et al. (1981)have shown that irradiated OKT4+ lymphocytes from normal healthy adults induced a greater B cell response than nonirradiated OKT4+ lymphocytes. The addition of OKT4+ cells to cultures of irradiated T lymphocytes and B lymphocytes suppressed the im-

REGULATION OF FETAL-MATERNAL BALANCE

.

5040

+

2 Z E

-

Radiosensitive Suppressor Cells

30-

=

20-

5

10-

0

199

I

Radioresistant Helper Cells

c 0

d

U 0

0-

Cord

Adult

T4+IT8+ 2.5 1.4 Diagram illustrating the phenotype populations and functional properties of FIG.6. peripheral blood lymphocytes from human newborns and adults. Cord blood has higher proportions of OKT4' lymphocytes and lower proportions of OKT8+ lymphocytes than adults. The OKT4+ subset from newborns is characterized by predominate radiosensitive (1200 rad of y-irradiation) suppressor activity (m). In contrast, the OKT4+ subset from adults is largely radioresistant helper lymphocytes (0).

munoglobulin response. The authors concluded that OKT4+ T lymphocytes from normal adults can differentiate into radiosensitive suppressor cells that act independently of the OKT8+ subset. This correlates with experiments in which active OKT4+ cells in the cord blood directly inhibited immunoglobulin synthesis and did not induce mature (adult) OKT8+ cells to suppress. The suppressor activity was diminished in adults compared to neonates, apparently obscured by an active helper cell population. Therefore, OKT4+ T lymphocyte populations from both adults and newborns are heterogeneous, containing both helper and suppressor cell subpopulations, and the overall functional differences between the two groups may be quantitative, rather than qualitative (Fig. 6). VII. Conclusions

Numerous hypotheses have been formulated to explain the pregnant mother's maintenance of a fetus despite its similarity to a nor-

200

DAVID R. JACOBY ET AL.

mally rejected semiallogeneic graft. This phenomenon is important in biology, both for survival and for accommodating variation in the gene pool of higher organisms. Conceptually, the great selective pressure that perpetrates this exception to normal transplantation immunology would suggest that fetal maintenance involves multifaceted and diverse mechanisms. This is supported by the multitude of experiments denoting different ways the fetus can be maintained during pregnancy. Among these mechanisms is strong evidence that the fetus plays an active role in its own protection via suppressor lymphocytes that are capable of inhibiting maternal immune responses (Fig. 5 ) . A fetus does not escape immune detection by its mother during pregnancy. In fact, mothers undergo cellular and humoral sensitization to foreign antigens expressed by their fetuses during pregnancy. This has several implications concerning the maintenance of the fetus. First, lack of expression or blocking of foreign histocompatibility antigens at the outermost cell layer of the trophoblast does not prohibit the sensitization of the mother to these determinants. This layer’s relative resistence to cell-mediated lympholysis suggests that it is a protective barrier, not a preventive one. Second, systemic immunosuppressive agents such as steroid hormones, glycoprotein hormones, and pregnancy-specific proteins are not present in concentrations great enough to modulate the action of primed maternal lymphocytes. Therefore, the mechanisms primarily responsible for fetal maintenance likely act locally at the placenta, and inhibit the activity of sensitized maternal lymphocytes. Fetal suppressor cells are capable of inhibiting immune responses by primed lymphocytes from adults in uitro. Lymphocytes from cord blood potently suppress lectin-stimulated proliferation, B cell differentiation, and immunoglobulin synthesis of lymphocytes from adults. These lymphocytes from the newborn also suppress the autologous secretion of immunoglobulin by EBV-transformed B cells. Fetal suppressor cells are in an active state, because they inhibit the division of their mother’s lymphocytes without the addition of mitogen. Furthermore, they inhibit maternal immunoglobulin synthesis directly and do not require an adult suppressor cell intermediate. We have characterized the phenotype and activity of cord blood lymphocytes. Cord cells that mediate suppression appear to be exclusively T lymphocytes that bear the OKT4+T8- phenotype, determined by both positive and negative selection techniques. Maximal suppression exerted by cord blood lymphocytes is 90 to 100% of proliferation normal for maternal lymphocytes in MLR, and 85 to 100%of IgG synthesized by maternal lymphocytes in PWM-stimulated cocul-

REGULATION OF FETAL-MATERNAL

BALANCE

201

tures. Inhibition of proliferation and IgG synthesis in this situation is mediated by soluble suppressor factors, as demonstrated by the ability of supernatants from cord blood lymphocytes to suppress proliferation and B lymphocyte differentiation in mitogen-treated cultures of maternal or adult lymphocytes. Further, cocultures in which newborn and maternal lymphocytes are separated by a cell-impermeable membrane still show suppression by cord blood lymphocytes. Here dialyzable substances diffuse across the membrane to inhibit cell division of lymphocytes from mothers or unrelated adults. Indomethacin and anti-PGE antibodies can block cord T lymphocyte-mediated suppression, implicating PGE as one of the soluble mediators. This substance is a potent immunosuppressive agent when added to maternal lymphocytes exogenously; however, cord blood lymphocytes express fewer receptors for PGE and appear to be approximately 100 times less sensitive to immunosuppression. Thus, cord blood lymphocytes suppress proliferation and differentiation of adult lymphocytes, but the suppressive factor does not inhibit their own ability to respond to antigens. PGE is labile, indicating that it has a limited range of activity, probably confined to the placental region (Fig. 7). Suppression mediated by cord blood lymphocytes is fully abrogated by 1200 rad of y-irradiation. Remaining radioresistant cells provide helper activity to purified mature B cells in PWM-stimulated cultures, and these lymphocytes map to the OKT4+T8- subset. This activity is relatively weak compared with that of adult helper T lymphocytes. The neonatal OKT4+T8- subset is functionally heterogeneous, containing a potent radiosensitive suppressor activity and a weak radioresistant helper activity. At birth, lymphocytes from newborns strongly suppress functions of adult lymphocytes in vitro, and this activity

FIG.7. Schematic diagram depicting the effect of lymphocytes from a fetus (a) in actively controlling potentially aggressive lymphocytes froin the mother (b).T lymphocytes from the fetus are shown adjacent to the thin chorionic villi of the placenta where they achieve their suppressive effect, in uiuo. They may either cross the placenta and inhibit primed lymphocytes from the mother by cell-cell interactions and/or release of soluble factors, thought to be prostaglandins E, from the fetal circulation. These suppressor lymphocytes from the fetus are OKT4+T8- and are sensitive to 1200 rad of irradiation.

202

DAVID R. JACOBY ET AL.

gradually declines until children are approximately 1.5 years of age, but remains detectable in the OKT4+ subset of normal adults. Helper activity is a weak function at birth, but increases until adult levels are reached when children are 2 to 3 years old. The development of helper activity in vitro correlates with expression of the OKT4+DR+ phenotype of stimulated lymphocytes, suggesting that the presentation of class I1 MHC antigens is involved in the development of the mature immune response. Early in pregnancy, fetal lymphocytes develop into cells that respond to mitogens, soluble antigens, and allogeneic or xenogeneic challenge. Fetal suppressor cells appear as early as the eighth week of gestation. The period of pregnancy excludes most antigenic stimuli for the fetus, but a greater threat is the mother’s potential rejection of antigenically different fetal tissues. Strong suppressor activity is one way that fetal lymphocytes may develop to respond to this challenge. After birth, the infant’s immune system must develop in order to respond to increased antigenic challenge. The fetus is capable of responding to certain antigens, but is unusually susceptible to infection until adult helper activity develops, B cells mature to synthesize all immunoglobulin isotypes, and cytotoxic T lymphocytes can be generated.

ACKNOWLEDGMENTS This is Publication No. 3001-IMM from the Department of Immunology, Scripps Clinic and Research Foundation, La Jolla, California. These studies have been possible only with the assistance and collaboration of Dr. S. Pruyn, Dr. J. Gambone, Mrs. M. Moreno, and the nursing staff at the U.S. Navy Regional Medical Center, San Diego, California. The authors wish to thank Phyllis Minick, Ana Garcia, Lisa A. Flores, and Sally Booz for their help in manuscript preparation. This research was supported by U.S.P.H.S. Grants AI-07007 and NS-12428.

REFERENCES Ablin, R. M., Bruns, G. R., and Guinan, P. (1974).J. lmmunol. 113, 705. Adcock, E. W., 111, Teasdale, E., August, C. S., Cox, S., Meschia, G., Battaglia, F. C., and Naughton, M. A. (1973). Science 181, 845. Adinolfi, M. (1975). In “Irnmunobiology of the Trophoblast.” Cambridge Univ. Press, London and New York. Adinolfi, M.,Akle, C. A., McColl, I., Fensorn, A. H., Tansley, L., Connolly, P., Hsi, B.-L., Faulk, W. P., Travers, P., and Bodmer, W. F. (1982). Nature (London) 298, 325. Alpert, E., Drysdale, J. W., Schur, P. H., and Isselbacher, K. J. (1971). Fed. Proc. Fed. Am. SOC. E x p . Biol. 30, 246. Andersson, U., Bird, A. G., and Britton, S. (1980). Eur. J. lmmunol. 10,888. Andersson, U., Bird, A. G., Britton, S., and Palacios, R. (1981). lmmunol. Reu. 57, 5 . Argyris, B. F. (1968). J. E x p . Med. 128, 459.

REGULATION OF FETAL-MATERNAL BALANCE

203

Asantila, T., Vahala, J., and Toivanen, P. (1974). Zmmunogenetics 1,272. Asma, G. E. M., Pichler, W., Schnit, H. R. E., Knapp, W., and Humans, W. (1977). Clin. E x p . Immunol. 29, 898. August, C. S., Berkel, I., Driscoll, S., and Merler, E. (1971). Pediatr. Res. 5, 539. Baines, M. G., Speers, E. A., Pross, H., and Millar, K. G. (1976). Zmmunology 31, 363. Balfour, A. H., and Jones, E. A. (1978).Znt. Arch. Allergy Appl. Zmmunol. 56,435. Beattie, W. G., and Dugaiczyk, A. (1982). Gene 20, 412. Beer, A. E., and Billingham, R. E. (1973). Science 179,240. Beer, A. E., and Billingham, R. E. (1974).J. Reprod. F e d . 21, 59. Beer, A. E., Billingham, R. E., and Yang, S. L. (1972).J.E x p . Med. 135, 1177. Beer, A. E., Billingham, R. E., and Yang, S. L. (1975). Biol. Reprod. 12, 176. Beer, A. E., Quebbeman, J. F., Ayers, J. W. T., and Haines, R. F. (1981).A m . ] . Obstet. Gynecol. 141, 987. Bernard, O., Ripoche, M. A., and Bennett, D. (1977).J. E x p . Med. 145, 58. Billington, W. D., Jenkinson, J., Searle, R. F., and Sellens, M. H. (1977). Transplant. Proc. 9, 1371. Bird, A. G . , and Britton, S. (1979). Zmtnunol. Reo. 45,41. Bjornboe, M., Fischell, E. E., and Stoerk, H. C. (1951).J. E x p . Med. 93,37. Blaese, R. M., and Lawrence, E. C. (1977).In “Development of Host Defenses.” Raven, New York. Blaese, R. M., Pike, S., and Tostato, G. (1982). Clin. Zmmunol. Zmmunopathol. 23, 162. Bockman, R. S., and Rothschild, M. (1979).J. Clin. Znoest. 62, 753. Bona, C., Heber-Katz, E., and Paul, W. E. (1981).J. E x p . Med. 153, 951. Brambell, F. W. R. (1970). In “Frontiers of Biology,” Vol. 18. North-Holland Publ., Amsterdam. Bryson, Y. J., Winter, H. S., Card, S. E., Fischer, T. J., and Stiehm, E. R. (1979). Cell. Zmmunol. 55, 191. Caldwell, J. L., Stites, D. P., and Fudenberg, H. H. (1975).J.Zmmunol. 115, 1249. Calkins, C. E., and Stutman, 0. (1978).J. E x p . Med. 147, 87. Carlson, G. A,, and Wegmann, T. G. (1978). Transplant. Proc. 10,403. Carr, M. C., Stites, D. P., and Fudenberg, H. H. (1973). Nature (London) 241, 279. Chaterjee-Hasrouni, S., and Lala, P. K. (1979).J . E x p . Med. 149, 1238. Chatejee-Hasrouni, S., and Lala, P. K. (1982).J. E x p . Med. 153, 1679. Chess, L., MacDermott, R. P., and Schlossman, S. F. (1974).J.Zmmunol. 113, 1113. Cohen, J., and Claman, H. N. (1971).J. E x p . Med. 133, 1026. Contractor, S. F., and Davies, H. (1973). Nature (London) 243,284. Corte, G . , Mingari, M. C., Moretta, A., Daminani, G., Moretta, L., and Bargellesi, A. (1982).J.Zmmunol. 128, 16. Cowan, M. J., Ammann, A. J., Wara, D. W., Howie, V. M., Schultz, L., Doyle, N., and Kaplan, M. (1978). Pediatrics 62, 721. Currie, G. A,, Van Doorninck, W., and Bagshawe, K. 0. (1968). Nature (London) 219, 191. Dahl, C. A. (1980).J. Zmmunol. 125, 1924. Darrow, T. L., and Tomar, R. H. (1980).Cell. Zmmunol. 56, 172. Daughty, R. W., and Gelsthorpe, K. (1974). Tissue Antigens 4, 291. DeWolf, W. C., Schlossman, S. F., and Yunis, E. J. (1979).J. Zmmunol. 122, 1780. Dumble, L. J., Tait, B. D., Whittingham, S., and Ashton, P. W. (1977). Zmmunogenetics 5, 345. Durandy, A., Fischer, A., and Griscelli, G. (1979).J . Zmmunol. 123,2644. Durandy, A., Fischer, A., and Griscelli, G. (1982).J . Zmmunol. 128, 525.

204

DAVID R. JACOBY ET AL.

Dwyer, J. M., and Mackay, I. R. (1972). Zmmunology 23,871. Eichmann, K. (1978). Ado. lmmunol. 26, 195. Faulk, W. P., and Johnson, P. M. (1980). Ado. Clin. Zmmunol. 2. Faulk, W. P., Temple, A., Lovins, R. E., and Smith, N. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 1947. Faulk, W. P., Yeager, C., McIntyre, J. A., and Veda, M. (1979).Proc. R. SOC. London Ser. B 206,163. Fauve, R. M., Hevim, B., and Jacob, H. (1974). Proc. Natl. Acad. Sci. U.S.A.71,4052. Fink, C. W., Miller, W. E., Dorward, B., and LoSpalluto, J. (1962).J. Clin. Znuest. 41, 1422. Fox, R. I., Thompson, L. F., and Huddlestone, J. R. (1981).J. lmmunol. 126,2062. Gathings, W. E., Lawton, A. R., and Cooper, M. D. (1977). Eur. J. lmmunol. 7, 804. Gathings, W. E., Kubagawa, H., and Cooper, M. D. (1981). lmmunol. Reo. 57, 100. Gatti, R. A., Svedmyr, E. A. J., Leibold, W., and Wigzell, H. (1975). Cell. Zmmunol. 15, 432. Geha, R. S., and Reinherz, E. L. (1983).J. lmmunol. 130,2493. Gill, T. J., 111 (1973). Lancet I, 133. Gmelig-Meyling, F., and Waldmann, T. A. (1981).J. lmmunol. 126, 529. Goodwin, J. S., Messner, R. P., and Peake, G. T. (1978).J. Clin. lnoest. 62, 753. Gordon, Y. B., Grudzinskas, J. F., Lewis, J. D., Jeffrey, D., and Letchworth, A. T. (1977). Br. J. Obstet. Cynecol. 84, 642. Granberg, C., and Hirvonen, T. (1980). Cell. lmmunol. 51, 13. Granberg, C., Hirvonen, T., and Toivanen, P. (1979).J. Zmmunol. 123,2563. Hamilton, M. S., Hellstrom, I., and van Belle, G. (1976). Transplantation 25, 263. Han, T. (1974). Clin. E x p . lmmunol. 18,529. Han, T. (1975). Immunology 29, 509. Hayward, A. R., and Ezer, G. (1974). Clin. E x p . Zmmunol. 23,279. Hayward, A. R., and Kurnick, J. T. (1981).J. lmmunol. 126, 50. Hayward, A. R., and Lawton, A. R. (1977).J.Zmmunol. 119, 1213. Hayward, A. R., and Lydyard, P. M. (1978). Clin. E x p . lmmunol. 34,374. Herzenberg, L. A., and Gonzales, B. (1962).Proc. Natl. Acad. Sci. U.S.A.48, 570. Herzenberg, L. A., Bianchi, D. W., Schroder, J., Cann, H. M., and Iverson, G. M. (1979). Proc. Natl. Acad. Sci. U.S.A.76, 1453. Heyner, S. (1973). Transplantation 16,675. Huddlestone, J. R., and Olstone, M. B. A. (1979).J. lmmunol. 123, 1615. Hulka, J. F., and Mohr, K. (1967). Am. 1.Obstet. Gynecol. 87,407. Indiveri, F. B. S., Wilson, C., Russo, C., Quaranta, V., Pellegrino, M. A., and Ferrone, S. (1980).J. Zmmunol. 125, 2673. Jacoby, D. R., and Oldstone, M. B. A. (1983).J . Znimtrnol. 131, 1765. Jeannet, M., Werner, C., Ramirez, E., Vassalli, P., and Faulk, W. P. (1977). Transplant. Proc. 9, 1417. Johnsen, S.-A., and Olding, L. B. (1983). Scand. J. lmmunol. 17,389. Johnsen, S.-A., Olding, L. B., Westberg, N. G., and Wilhelmsson, L. (1982). Clin. Zmmunol. lmmunopathol. 23,606. Johnsen, S.-A,, Olofsson, A., Green, K., and Olding, L. B. (1983a). Am. J. Reprod. lmmunol. 4, 45. Johnsen, S.-A., Olding, L. B., and Green, K. (1983b). lmmunol. Lett. 6,213. Jones, W. R. (1976). In “Immunology of Human Reproduction” (J. S. Scott and W. R. Jones, eds.). Academic Press, New York. Kadowaki, J. I., Zuelzer, W. E., and Brough, A. J. (1965). Lancet I, 1152.

REGULATION OF FETAL-MATERNAL

BALANCE

205

Kaplan, N. M. (1961).J.Clin. Endocrinol. 21, 1139. Kay, H. E. M., Playfair, J. H., Wolfendale, M., and Hopper, P. K. (1962). Nature (London) 196,238. Kiessling, R., Klein, E., Pross, H., and Wigzell, H. (1975). Eur. J . Zmmunol. 5, 117. Kirby, D. R. S. (1968). Transplantation 6, 1005. Kirby, D. R. S. (1969). Fertil. Steril. 20, 933. Kitzmiller, J. L., Stoneburger, L., Yelenosky, P. F., and Lucas, W. E. (1973). Am. J . Obstet. Gynecol. 198, 805. Klein, R. B., Rich, K. C . , Biberstein, M., and Stiehm, E. R. (1976). Clin. Res. 24, 108A. Koren, Z. (1968).Am. J . Obstet. Gynecol. 102, 340. Kretschmer, R. R., Stewardson, P. B., Papierniak, C. K., and Gotoff, S. P. (1976). J . Zmmunol. 117, 1303. Laughter, A. H., and J. J. Tworney. (1977).J . Zmmunol. 119, 1973. Lawton, A. R., Self, K. S., Royal, S. A., and Cooper, M. D. (1972). Clin. Zmmunol. Zmmunoputhol. 1, 84. Ledbetter, J. A., Evans, R. L., Lipinski, M., Cunningham-Rundles, C., Good, R. A., and Herzenberg, L. A. (1981).J . E x p . Med. 153, 310. Leino, A. (1978). Clin. Zmmunol. Zmmunopathol. 17,547. Leino, A., Hirvonen, T., and Soppi, T. (1980). Clin. Zmmunol. Zmmunopathol. 17,547. Lester, E. P., Miller, J. B., and Yachnin, S. (1976).Proc. Natl. Acad. Sci. U.S.A.73,4645. Leung, K. H., and Koren, H. S. (1982).J.Zmmunol. 129, 1742. Long, D. A. (1957). Znt. Arch. Allergy 10, 5. Loraine, J. A. (1961). In “Hormones in Blood” (C. H. Gray and V. H. T. James, eds.). Academic Press, New York. Maroni, E. S., and Parrot, D. M. V. (1973). Clin. E x p . Zmmunol. 13, 253. Maroulis, G. B., Buckley, R. H., and Younger, J. B. (1971).Am. J . Obstet. Gynecol. 109, 971. Martensson, L., and Fudenberg, H. H. (1965).J . Zmmunol. 94, 514. Masson, P. L., Delire, M., and Cambiaso, C. L. (1977). Nature (London) 266, 542. McCormick, J. N., Faulk, W. P., Fox, H., and Fudenberg, H. H. (1971).J . E x p . Med. 133, 1. McIntyre, J. A., and Faulk, W. P. (1979). Trunsplant. Proc. 11, 1892. McIntyre, J. A., and Faulk, W. P. (1983). Am. J . Reprod. Zmmunol. 4, 165. Medawar, P. B. (1954).Symp. Soc. E x p . Biol. 7, 320. Medawar, P. B., and Sparrow, E. M. (1956).J . Endocrinol. 14, 240. Merikanto, J. (1979).Immunology 38, 677. Merz, W. E., Schmidt, W., Hilgenfeldt, U., Scliackert, K., and Lenhard, V. (1976).J . Immunol. 152, 286. Mishell, R. I., Herzenberg, L. A., and Herzenberg, L. A. (1963).J . Zmmunol. 90, 628. Mitchell, M. D., Carr, B. R., Mason, J. I., and Simpson, E. R. (1982). Proc. Natl. Acud. Sci. U.S.A. 79, 7547. Miyawaki, T., Seki, H., Kubo, M., and Taniguchi, N. (1979).J . Zmmunol. 123, 1092. Miyawaki, T., Moriya, N., Nagaoki, T., and Taniguchi, N. (1981). Zmmunol. Rev. 57, 61. Miyawaki, T., Yachie, A., Nagaoki, T., Mukai, M., Yokoi, T., Uwadana, N., and Taniguchi, N. (1982).J . Zmmunol. 128, 11. Montgomery, W. P., Young, R. C., Aller, M. D., and Horden, K. A. (1968).Am.J . Obstet. Gynecol. 100, 829. Moretta, L., Webb, S. R., Grossi, C. E., Lydyard, P., and Copper, M. D. (1977).J . E x p . Med. 146, 184.

206

DAVID R. JACOBY ET AL.

Morimoto, C., Reinherz, E. L., Borel, Y., Mantzouranis, E., Steinberg, A. D., and Schlossman, S. F. (1981).J. Clin. lnuest. 67, 753. Mosier, D. E., (1974).J. Zmmunol. 112, 305. Mosier, D. E., Mathieson, B. J,, and Campbell, P. S. (1977).J. E x p . Med. 146, 59. Mumford, D. M., Sung, J. S., Wallis, J. O., and Kaufman, R. H. (1978). Pediatr. Res. 12, 171. Murgita, R. A., and Tomasi, T. B. (1975).J. E x p . Med. 141,269. Murgita, R. A,, Hooper, D. E., Stegagno, M., Delovitch, T. L., and Wigzell, H. (1981). Eur. J . Immunol. 11,957. Nagaoki, T., Moriya, N., Miyawaki, T., Seki, H., Yokoi, T., Okuda, N., and Tanaguchi, N. (1980).J. lmmunol. 125, 1563. Nelson, J. H., Lo, T., Hall, J. E., Krown, S., Nelson, J. M., and Fox, C. W. (1973).Am. J. Obstet. Cynecol. 117,689. Olding, L. B., and Oldstone, M. B. A. (1974). Nature (London) 249,161. Olding, L. B., and Oldstone, M. B. A. (1976).J. Zmmunol. 116, 682. Olding, L. B., Benirschke, K., and Oldstone, M. B. A. (1974). Clin. Zmmunol. Zmmunopathol. 3, 79. Olding, L. B., Murgita, R. A., and Wigzell, H. (1977).J. Zmmunol. 119, 1109. Oldstone, M. B. A,, Tishon, A., and Moretta, L. (1977). Nature (London) 269, 333. Ortaldo, J. R., MacDermott, R. P., Bonnard, G. D., Kind, P. D., and Herberman, R. B. (1979). Cell. lmmunol. 48, 356. Pace, J. L., Russel, S. W., Schreiber, R. D., Altman, A., and Katz, D. H. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 3782. Park, W. V. (1971).In “Choriocarcinoma: A Study of its Pathology.” Heineman, London. Peer, L. A. (1958).Ann. N.Y. Acad. Sci. 73, 854. Pichler, W. J., and Broder, S. (1981). Zmmunol. Reu. 56, 163. Reinherz, E. L., and Schlossman, S. F. (1980). N . Engl. J. Med. 303, 370. Reinherz, E. L., Penta, A. C., Hussey, R. E., and Schlossman, S. F. (1982a). Clin. Immunol. Immunopathol. 21,257. Reinherz, E. L., Morimoto, C., Fitzgerald, K. A., Hussey, R. E., Daley, J. F., and Schlossman, S. F. (1982b).J . Zmmunol. 128, 463. Robert, M., Betuel, H., and Revillard, J. P. (1973). Tissue Antigens 3, 39. Rocklin, R. E., Zuckerman, J. E., Alpert, E., and David, J. R. (1973). Nature (London) 241, 130. Rocklin, R. E., Kitzmiller, J. L., Carpenter, C. B., Garovoy, M. R., and David, J. R. (1976).N . Engl. J. Med. 295, 1209. Rocklin, R. E., Kitzmiller, J. L., and &ye, M. D. (1979). Annu. Reu. Med. 30,375. Roder, J. C., and Klein, M. (1979).J. Immunol. 123, 2785. Rodriguez, M. A., Bankhurst, A. D., Ceuppens, J. L., and Williams, R. C. (1981)J. Clin. Znuest. 68, 1577. Sato, T., Fuse, A., and Kuwata, T. (1979). Cell. lmmunol. 45, 458. Saxon, A., and Stevens, R. H. (1978). Clin. lmmunol. Immunopathol. 10,427. Schiff, R. I., Mercier, D., and Buckley, R. H. (1975). Cell. lmmunol. 20, 69. Schultz, R. M., Stoychkov, J. N., Pavlidis, N., Chirigos, M. A,, and Olkowski, Z. L. (1979).J. Reticuloendothel. SOC.26, 93. Searle, R. F., Johnson, M. H., and Billington, W. D. (1974). Transplantation 6, 1005. Searle, R. F., Sellens, M. H., Elson, J., Jenkinson, E. J., and Billington, W. D. (1976).J. E x p . Med. 143, 348. Sellens, M. H., and Jenkinson, E. J. (1975). Nature (London) 255,412.

REGULATION O F FETAL-MATERNAL

BALANCE

207

Simmons, R. L., Lipschultz, M. L., Rios, A., and Ray, P. K. (1971).Nature (London)231, 111. Smith, J. A., Burton, R. C., Barg, M., and Mitchell, G. F. (1978).Transplantation 25,216. Snyder, D. S., Beller, D. I., and Unanue, E. R. (1982). Nature (London) 299, 163. Stimson, W. H. (1976).Clin. E x p . Zmmunol. 25, 199. Stimson, W. H., Strachan, A. F., and Shepherd, A. (1979). t3r.J.Obstet. Gynecol. 86,41. Stites, D. P., Carr, M. C., and Fudenberg, H. H. (1974). Cell. Zmmunol. 11, 257. Stobo, J. D., and Paul, W. E. (1973).J . Immunol. 110, 362. Strelkauskas, A. J., Wilsin, B. S., Dray, S., and Dodsen, M. (1975).Nature (London)258, 331. Strelkauskas, A. J., Callery, R. T., McDowell, J., Borel, Y., and Schlossman, S. F. (1978). Proc. Natl. Acad. Sci. U.S.A.75, 5150. Suiu-Foca, N., Reed, E., Rohowsky, C., Kung, P. C., and King, D. W. (1983).Proc. Natl. Acad. Sci. U.S.A.80, 830. Tamerius, J., Hellstrom, I., and Hellstrom, K. E. (1975). Znt. J. Cancer 16, 456. Taylor, P. V., and Faulk, W. P. (1981). Lancet 11, 68. Taylor, P. V., and Hancock, K. W. (1975). lmmunology 28,973. Theofilopoulos, A. N., Gleicher, N., Pereira, A. B., and Dixon, F. J. (1981). Am. J. Reprod. Zmmunol. 1, 92. Thomas, Y., Rogozinski, L., Irgoyen, 0. H., Friedman, S. M., Kung, P. C., Goldstein, G., and Chess, L. (1981).J. E x p . Med. 154,459. Thomas, Y., Rogozinski, L., Irgoyen, 0. H., Shen, H. H., Talle, M. A., Goldstein, G., and Chess, L. (1982).J. Zmmunol. 128, 1386. Timonen, T., and Saksela, E. (1976). Clin. E x p . Zmmunol. 23,462. Timonen, T., and Saksela, E. (1977). Cell. Zmmunol. 33,340. Toivanen, P., Uksila, J., Leino, A., Lassila, O., Hirvonen, T., and Ruuskanen, 0. (1981). Zmmunol. Reu. 57, 89. Tomada, Y., Fuma, M., Miwa, T., Saiki, N., and Ishizulki, N. (1976). Gynecol. Znuest. 7, 280. Tosato, G., Magrath, I. T., Koski, I. R., Dooley, N. J., and Blaese, R. M. (1980).J. Clin. Znuest. 66, 383. Uhr, J. W., Dancis, J., Franklin, E. C., Finkelstein, M. S., and Lewis, E. W. (1962).J. Clin. Invest. 41, 1509. Unander, A. M., and Olding, L. B. (1981).J. Zmmunol. 127, 1182. Unander, A. M., and Olding, L. B. (1983).A m . ] . Reprod. Zmmunol. 4, 171. Uytehaag, F., Heijnen, C. J., Pot, K. H., and Ballieux, R. E. (1979).J.Zmmunol. 123,646. Valquist, B., Lagecrantz, R., and Nordbring, F. (1950). Lancet 11, 851. Van der Werf, A. J. M. (1971). Lancet I, 595. von Schoultz, B., and Stigbrand, T. (1973).Acta Obstet. Gynecol. Scand. 52,51. Warburton, F. E. (1968). Heredity 23, 151. Waldmann, T. A., and Broder, S. (1982).Ado. Zmmunol. 32, 1. Walknowska, J., Conte, F. A., and Grumbath, M. M. (1969). Lancet I, 1119. Waltman, S. R., Burde, R. M., and Berrios, J. (1971).Transplantation 11, 194. Webb, C. G., Gall, W. E., and Edelman, G. M.(1977).J. E x p . Med. 146, 923. Weigle, W. O., and Dixon, F. J. (1959).J.E x p . Med. 110, 139. Weston, W. L., Claman, H. N., and Krueger, G. G. (1973).J. Zmmutiol. 110, 880. Winchester, R. J., and Kunkel, H. G. (1980).Adu. Zmmunol. 28, 221. Winchester, R. J., Fu, S. M., Wernet, P., Kunkel, H. G . , Dupont, B., and Jersild, C. (1975).J . E x p . Med. 141, 924.

208

DAVID R. JACOBY ET AL.

Wyle, F. A,, and Kent, J. R. (1977). Clin. E x p . Immunol. 27, 407. Yachie, A., Miyawki, T., Yokoi, T., Nagaoki, T., and Taniguchi, N. (1981).J . Immunol. 127, 1314. Yachie, A., Miyawaki, T., Yokoi, T., Nagaoki, T., and Tanaguchi, N. (1982).J.Immunol. 129, 103. Yanagimachi, R. (1977). In “Immunology o f Gametes (M. Edidin and M. H. Johnson, eds.). Cambridge Univ. Press, London and New York. Youtananukorn,V., and Matangkasombut, P. (1972). Clin. E x p . Immunol. 11, 549. Youtananukorn,V., Matangkasombut, P., and Osathanondh, V. (1974). Clin. E x p . Immunol. 16,593. Zuckennan, S. H., and Douglas, S. D . (1975). Nature (London)255,412.

ADVANCES IN IMMUNOLOGY. VOL. 35

The Influence of Histamine on Immune and Inflammatory Responses DENNIS J. BEER,’ S T N E N M. MATLOFF,t AND ROSS E. ROCKLlNt ‘Pulmonary Medicine Section, Evans Memorial Department of Clinical Research, Boston University School of Medicine, Boston, Massachusetts, and +Allergy Division, Department of Medicine, New England Medical Center, Tufts University School of Medicine, Boston, Massachusetts

I. Histamine as an Autacoid.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Introduction. . . . . . .

A. Basophils.. . . . . B. Eosinophils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

C. Expression of Lymphocyte Receptors D. Monocyte/Macrophage Function.. . . .

209

............

215

.........

259

References.. . . . . . . . . . . . . . . .

1. Histamine as an Autacoid

A. INTRODUCTION Histamine is a vasoactive amine that is produced during an inflammatory response. Until recently, it was only considered in terms of its being an inflammatory mediator producing symptoms of allergic disease. Over the past 10 years, it has become increasingly apparent that histamine and other mediators such as prostaglandins and @mimetic catecholamines-termed “autacoids” by Melmon et al. (1981)-may be part of the physiologic process that regulates both immune and inflammatory events. These inflammatory mediators are endogenously generated during immune responses, and it is quite possible that concentrations of histamine achieved during these reactions are sufficient to allow them to modify the functions of cells involved in humoral and cell-mediated immunity. Enough data have now been accumulated to indicate that these mediators may directly affect a 209 Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-022435-6

210

DENNIS J. BEER ET AL.

number of stages in the immune response and also act as feedback modifiers connecting the early and late phases of these reactions. It should also be noted that receptors for vasoactive hormones such as histamine are not present in a random fashion on all lymphoid cells but develop concomitantly with the commitment of a clone of immunocompetent cells to express cell-mediated and humoral immune responses. In this review, the role of histamine will be explored in terms of its ability to modulate inflammation and immune responses. While it has been shown that histamine can induce an elevation in intracellular levels of cyclic 3’,5’-adenosine monophosphate or cyclic AMP (see later), it is not clear that its mechanism of action is entirely through this second messenger, or that all of its effects will be via a negative signal. That is, while some of the effects of histamine will be shown to be inhibitory to certain leukocyte and lymphocyte functions, histamine may also exert a positive influence on some cells, as will be discussed. B. PHARMACOLOGY OF HISTAMINE Histamine, 2-(4-imidazolyl)ethylamine,i s a low-molecular-weight (MW 111) biogenic amine that is present in a preformed state in almost all mammalian species (Fig. l).Histamine is formed by decarboxylation of the amino acid histidine, a reaction catalyzed by the enzyme L-histidine decarboxylase. It is metabolized via two major pathways (Beaven, 1976). In humans, histamine is primarily methylated to l-methylhistamine by the enzyme histamine methyltransferase. This product is converted to l-methyl-imidazole-4-acetic acid by the enzyme monoamine oxidase. In the other pathway, which is also present in humans, histamine is oxidized by diamine oxidase (histaminase) to imidazole-4-acetic acid, much of which is conjugated with ribose and is excreted as the riboside. Diamine oxidase is contained within polymorphonuclear neutrophilic leukocytes, while histamine methyltransferase is present in monocytes and tissue fluids. The presence of histamine methyltransferase in monocytes might explain the observation that relatively high concentrations of histamine are required to alter lymphocyte function in uitro, since histamine may be degraded during the culture period. Although a pool of non-mast cell-associated histamine may exist in certain tissues like the stomach (Kahlson and Rosengren, 1968), the majority of histamine in humans is found in the granules of circulating basophils and tissue mast cells. These granules also contain large amounts of heparin and proteolytic enzymes. Histamine is released from its intracellular stores by physical and chemical agents, antigen-

211

INFLUENCE OF HISTAMINE

AGONISTS Histamine

2-Methyl histamine

4-Methyl histamine

CHz-CHz-NH2

““XN “3 H

cnz-cHz-Nnz U

N

Dimaprit

ANTAGONISTS H 3 C O 0 c r ;

Pyrilamine

CHI

N-CHZ-CHZ-N CH3

Chlorpheniramine

cn

h C H Z - S - C H z -CHzNH-C-NH-CH3

E

Metiamide HN-N

Cimetidine

“’

rn

CHz-S-CHp-CH2NH-~-NH-cH,

NCZN

HN-N

FIG.1. Structures of some of the compounds acting on histamine receptors.

antibody reaction, and a variety of drugs. The highest concentrations of histamine are found in lung, skin, and gastric mucosa. In normal individuals plasma histamine levels are generally less than 1 pg/nil ( M ) . In conditions such as cholinergic urticaria and mastocytosis, associated with widespread flushing and itching reactions, plasma histamine levels may reach 3-5 pg/ml (Kaplan et al., 1978). In uiuo,

212

DENNIS J. BEER ET AL.

tissue levels of histamine may transiently reach as high as M (Adams and Lichtenstein, 1979). Histamine contracts some smooth muscles such as those of the bronchi and gut, but relaxes others, including those of the microvasculature. It is also a very potent stimulus to gastric acid production and elicits various other exocrine secretions. Some of these effects, such as bronchoconstriction and gut contraction, are mediated by one type of histamine receptor, the histamine type 1 (H1) receptor (Ash and Schild, 1966), which is blocked by pyrilamine and other such classical antihistamines (Fig. 1).Other effects, most notably gastric acid secretion and enhanced cardiac chronotropy, are completely refractory to such antagonists but are susceptible to inhibition by a newly developed group of antagonists, including burimamide, metiamide, impromidine, and cimetidine. These observations defined the presence of histamine type 2 (H2) receptors (Black et al., 1972). Finally, a number of histamine-induced responses such as vascular dilatation are mediated by both H1 and H2 receptor stimulation, and can only be abrogated when a combination of H1 and H2 blockers are present (Powell and Brody, 1976; Smith et al., 1980). The two classes of histamine receptors can also be defined by differential responses to various histamine-like agonists (Fig. 1).Thus, 2methylhistamine and 2-pyridylethylamine (2-PEA) preferentially elicit responses mediated via H1 receptors, whereas 4-methylhistamine and dimaprit have a correspondingly preferential effect mediated through Hz receptors (Black et al., 1972). As will be discussed, the availability of specific histamine receptor agonists and antagonists has greatly enriched the understanding of the physiology of histamine on immune responses. C. HISTAMINE RECEPTOR-MEDIATED CHANGES IN CYCLIC NUCLEOTIDES Histamine has been shown to induce a rise in intracellular levels of cyclic AMP via stimulation of H2 receptors in a variety of tissues including epidermis, neuronal tissue, and T lymphocytes (Aoyagi et al., 1981; Daly, 1975). Furthermore, elevation of intracellular cyclic AMP following the addition of histamine to cultures of human keratinocytes in vitro is associated with reduced mitosis by these cells. This latter effect can be reproduced by the addition of 4-methylhistamine (H2 agonist) and reversed by cimetidine (H2 antagonist) (Flaxman and Harper, 1975; Aoyagi et al., 1981). In several other tissues, such as mouse fibroblasts and lymphocytes, elevations in intracellular cyclic AMP have also been shown to inhibit cellular prolif-

INFLUENCE OF HISTAMINE

213

eration. In addition, histamine receptor stimulation may also induce an elevation in intracellular cyclic 3',5'-guanosine monophosphate (GMP) levels in these tissues. Elevation in cyclic GMP levels induces a stimulation of DNA synthesis in these cells (Friedman, 1976; Seifert and Rudland, 1974; Hadden et al., 1972). Histamine was shown to generate optimal amounts of intracellular cyclic AMP at concentrations M , whereas maximum levels of cyclic GMP production were of induced at M (Aoyagi et al., 1981). Maximal inhibition of cell M and was actually less when a concentration growth occurred at M histamine was employed. The authors concluded that histaof mine modulated cell growth through the generation of cyclic nucleotides. Histamine-stimulated increases in intracellular cyclic AMP resulted in inhibition of cell growth, while histamine-mediated increases in cyclic GMP partially antagonized this inhibition. In contrast to stimulation of the Hz receptor, there is evidence to suggest that stimulation of the H1 receptor is coupled to activation of guanylate cyclase and the formation of cyclic GMP. For example, mouse neuroblastoma cells and superior cervical ganglion tissue developed increased levels of intracellular cyclic GMP following histamine stimulation. This response was apparently mediated through the H I receptor, since H1 agonists simulated the histamine effect and H1 antagonists blocked this effect (Richelson, 1978a,b; Study and Greengard, 1978). Histamine induces a rise in intracellular cyclic AMP in murine T lymphocytes (Roszkowski et al., 1977). The state of cellular maturation of these cells determines their response to histamine. Thus, thymocytes generate little cyclic AMP in response to histamine, while spleen cells accumulate greater amounts of intracellular cyclic AMP.

D. METHODSFOR DETECTING HISTAMINE RECEPTORS ON LEUKOCYTES Various methods have been employed in an attempt to detect histamine receptors on leukocytes. One of the first techniques described involved the insolubilization of histamine by conjugation with a protein carrier. Histamine was conjugated to either of two types of carriers: either bovine or rabbit serum albumin, or random copolymers of DL-alanine and L-tyrosine. The conjugate was then linked to Sepharose agarose beads (H-RSA-S). The latter conjugate was then incubated with cells and their adherence to the histamine-protein carrier assessed (Melmon et al., 1972; Weinstein et al., 1973). Using this method, it was demonstrated that the majority of H-RSA-S beads were bound by leukocytes. This binding appeared to be specific for

214

DENNIS J. BEER ET AL.

histamine, since the binding of cells to H-RSA-S beads was prevented or reversed by high concentrations of histamine antagonists and soluble histamine itself, but not by catecholamines or their antagonists. The specific nature of the binding reaction was also documented by an experiment in which the protein and polymer carriers attached to Sepharose (RSA-S) in the absence of histamine resulted in less binding of the cells. Similarly, when histamine was bound directly to Sepharose, without the intervening protein or polymer carrier arm, specific binding did not occur. One of the advantages of this method is that the insolubilization of histamine prevented histamine from entering the cell interior, and therefore permitted the study of histamine-mediated cell membrane events on intact cell functions. H-RSA-S columns have been used to define the distribution of histamine receptor-bearing cells in a mixed leukocyte population (Weinstein et al., 1973). These investigators demonstrated that the majority of leukocytes had specific receptors for histamine. Thus, the presence of histamine receptors was detected on 20-40% human blood lymphocytes, 35-55% tonsil lymphocytes, and 10-55% thymic lymphocytes (Ballet and Merler, 1976). Subsequent studies have detected histamine receptor-bearing cells by means of rosette formation with histamine-coated erythrocytes (Kedar and Bonavida, 1974; Saxon et al., 1977). Mouse leukocytes taken from various lymphoid tissues formed rosettes with sheep erythrocytes that had been coated with histamine-rabbit serum albumin (HRSA) conjugates. The percentage of rosettes in lymphoid tissues of adult mice ranged from 8% in the thymus to 70% for peritoneal exudate cells; spleen, bone marrow, blood, and lymph node cells were intermediate in the numbers of rosetting cells (20-40%). HistamineRSA and unbound histamine specifically inhibited rosette formation, but RSA itself did not. The manner of presentation of histamine to the cell surface was an important factor in determining its affinity for the histamine-binding site. It was shown that monovalent histamine exhibited very low binding affinity compared with that of histamine conjugated to RSA, which was about 500 times more effective than free histamine in inhibiting rosette formation. These latter findings are in agreement with those obtained by Weinstein et ul. (1973), who also could not demonstrate the binding of cells to histamine directly linked to Sepharose beads. Another technique for the identification and analysis of histamine receptors on lymphocytes utilizes histamine conjugated to fluoresceinated human albumin (Osband et al., 1980). The cells are incubated with this conjugate, and binding is analyzed using a flow cytom-

INFLUENCE OF HISTAMINE

215

eter. The percentages of histamine receptor (H1and Hz)-bearing cells are enumerated by performing the incubation in the presence or absence of H1 or Hz antagonists. It was found that the histamine-fluoresceinated human albumin conjugate (his-FHA) bound to approximately 45% of T cells, and that this binding was apparently specific, since FHA alone did not bind to T cells, and soluble histamine inhibited his-FHA binding. The ratio of H1 : Hz receptors on normal T cells is approximately 1.5: 1 by this technique. The determination of the numbers of total histamine and specific H1 and Hz receptors per cell has been accomplished using a radioligandbinding method (Osband et al., 1981).This method involves the binding of [3H]histamine to formaldehyde-fixed lymphocytes, in the presence of unlabeled histamine (nonspecific binding), or the H1 or Hz receptor antagonists, diphenhydramine or cimetidine, respectively. The percentages of cells bearing specific histamine H1 and Hz receptors were estimated. However, the specificity of binding of ["Hlhistamine to T cells by this method has been questioned, since the affinity of the radioligand to the putative receptor was of the order of M . It thus becomes difficult to measure ligand-receptor bindto ing accurately with such low affinity of interaction. This observation may also explain in part why in some instances rather high M) concentrations of histamine are required to achieve in vitro biologic effects. Histamine H1 and Hz receptors on thymic lymphocytes have been solubilized, separated, and partially characterized (Osband and McCaffrey, 1979). The H1 and Hz receptors were reported to have molecular weights of approximately 50,000 and 40,000, respectively, based on gel filtration with Sephadex G-75. Furthermore, chromatography of these solubilized receptors on ion exchange columns effected a separation of the H1 and H2 receptors. [3H]Histamine association to the putative receptors appeared to exhibit specificity as evidenced by the observation that H1 antagonists (but not Hz antagonists) blocked binding to the 50,000 MW material, and Hz antagonists (but not H1 antagonists) blocked binding to the 40,000 MW material. 11. Histamine Modulation of Polymorphonuclear Inflammatory Cells

A. BASOPHILS That histamine might exert an influence on inflammatory events was first shown by Bourne et al. (1971), who observed that the exogenous addition of histamine to cultures of human basophils inhibited

216

DENNIS J. BEER ET AL.

the subsequent antigen-stimulated release of histamine (Table I). The concentration of histamine causing 50% inhibition of release was to lo-' M . The inhibitory effect of histamine on mediator release was localized to the activation phase and not to the release phase. Maximum suppression (80-90%) of histamine release occurred when histamine was added simultaneously with antigen but was reduced to 2040% when added as little as 2 minutes after the addition of antigen. These results were interpreted to mean that histamine interferes with an intracellular event occurring during activation that is not dependent on extracellular calcium or magnesium. The nature of the histamine receptor involved in this reaction was TABLE I HISTAMINE MODULATIONOF POLYMORPHONUCLEAR INFLAMMATORY CELLFUNCTIONS

Cell type Basophil (human)

Lung mast cell (human) Skin mast cell (human) Eosinophil (human)

Neutrophils (human)

Assay Histamine release Chemotactic response (C5a) Cyclic AMP Histidine uptake Histamine release Histamine release Chemotactic response (C5a) Chemotactic response (C5a) Chemotaxis (histamine) Cyclic AMP Complement receptors (C3b and C4) Chemokinesis Chemokinesis (f-Met-Leu-Phe, casein) Chemotaxis (f-Met-Leu-Phe, casein) Lysosomal enzyme release (activated serum) Cyclic AMP Adherence (C5a) Membrane potential (f-Met-Leu-Phe) Superoxide anion production (f-Met-Leu-Phe) Peroxide formation (f-Met-Leu-Phe)

Effect of histamine

Receptor specificity

HI

He

+ + + -

+ + + + -

-

+ +

+ + + + + +

INFLUENCE OF HISTAMINE

217

investigated by Lichtenstein and Gillespie (1973). They found that the inhibition of histamine release by exogenously added histamine was antagonized by the Hz antagonist burimamide ( M ) but not by HI antagonists such as chlorpheniramine, pyrilamine, and promethaM).The binding characteristics of these reactions zine (up to M concentrations of antagonists, occurred maximally with to and they therefore appeared to act through receptor blockade and not through a mechanism which suggested an alteration in histamine metabolism. Furthermore, Hz receptor agonists such as 4-methylhistamine and dimaprit (but not HI agonists such as 2-methylhistamine) mimicked the suppressive effects of histamine on histamine release (Lichtenstein and Gillespie, 1975). It is known that agents which raise intracellular levels of cyclic AMP inhibit histamine release (Lichtenstein, 1975). Changes in intracellular levels of cyclic AMP were measured in studies in which histamine was employed (Lichtenstein and Gillespie, 1973). These workers showed that the addition of histamine to cultures of basophils was associated with a significant rise in intracellular levels of cyclic AMP (Table I). They also showed a direct correlation between the inhibition of mediator release caused by exogenously added histamine and a corresponding rise in intracellular levels of cyclic AMP. Further, the rise in intracellular levels of cyclic AMP was shown to be mediated through the Hz receptor, since burimamide M ) , but no chlorpheniramine or pyrilamine, were capable of blocking it. The ability of histamine to inhibit basophil and mast cell secretion may vary depending on the releasing stimuli and the cell source. For example, basophil secretion induced by antigen was inhibitable by histamine, whereas that induced by the calcium ionophore A23187 was not suppressed by histamine or other cyclic AMP-elevating agents (Lichtenstein, 1975). Moreover, histamine inhibited secretion by human basophils and skin mast cells (Ting et al., 1980)but did not alter secretion by human lung mast cells challenged by antigen (Kaliner, 1978). In addition to affecting the secretion of mediators by basophils, histamine has also been shown to influence the migration of these cells (Table I). Experiments by Lett-Brown and Leonard (1977)demonstrated that histamine inhibited the chemotactic response of basophils to C5a but not to a lymphokine-derived chemotactic factor or the bacterial peptide, N-formylmethionylleucylphenylalanine (f-MetLeu-Phe). The ICs0 of this reaction occurred at M histamine. It should be noted that histamine itself did not influence the motility of basophils. In contrast, the macrophage chemotactic response was

218

DENNIS J. BEER ET AL.

measured in parallel in the latter study, and the investigators found that histamine did not inhibit their directed migration to C5a, the lymphokine-derived chemotactic factor, or f-Met-Leu-Phe. With regard to the nature of the histamine receptor involved, the investigators found that metiamide (Hz antagonist) blocked the histamine-mediated inhibition of C5a-directed basophil chemotaxis but required four orders of magnitude more antagonist than agonist to accomplish this. Pyrilamine, an H1 antagonist, did not block the histamine-induced inhibition of chemotaxis, although it did inhibit the chemotactic response by itself. Since metiamide has greater binding affinity for the Hz receptor than histamine, one wonders why it should require four orders of magnitude more of this antagonist to block the histamine effect. Therefore, the specificity of the response should be questioned. In any event, it appears that histamine may modulate the accumulation of basophils being attracted by complement-derived products, but that it permits the accumulation of cells during lymphocyte-mediated reactions. While the evidence accumulated thus far regarding inhibition of histamine secretion and cell movement by exogenously added histamine implies involvement of the Hz receptor, other studies suggest that stimulation of the H1 receptor may lead to changes in histamine metabolism (Table I). Studies by Stewart and Kay (1980) seem to M ) of histamine induce a 35% indicate that high concentrations ( enhancement of [ 14C]histidine uptake by basophils. Histamine analogs were not effective in causing this increase in [14C]histidine upM ) partially inhibited this increase. take, and H1 antagonists There was no effect by H2 antagonists. These studies seem to suggest that increased concentrations of histamine may promote further synthesis of the hormone.

B. EOSINOPHILS Histamine effects on eosinophil function include the regulation of cell migration, the appearance of complement receptors, and an elevation in intracellular cyclic nucleotide levels (Table I). Histamine appears to be directly chemotactic for eosinophils (Clark et al., 1975). Further, the chemotactic effect of histamine on eosinophils can be blocked only by the combination of H1 plus Hz antagonists and not by either alone, leading investigators to speculate that a third type of histamine receptor might be involved (Clark et al., 1977; Turnbull and Kay, 1976; Bryant and Kay, 1977). In addition to its direct effects on eosinophil migration, histamine has also been reported to inhibit the directed movement of eosinophils in response to endotoxin-acti-

INFLUENCE OF HISTAMINE

219

vated serum or kallikrein, at concentrations greater than M . ConM histamine increased the eosinophil chemocentrations below tactic response to endotoxin-activated serum. Facilitation of directed movement appeared to be mediated via H1 receptors, since only H1 antagonists, but not H2 antagonists, blocked the effect. In contrast, the chemotactic response to C5a was blocked by H2 antagonists but not H1 antagonists. Histamine also induced a rise in intracellular levels of cyclic AMP that could be blocked by H2 antagonists, but not H1 antagonists (Clark et al., 1977). The expression of certain membrane structures present on human eosinophils that could be influenced by histamine was studied by Kay (1979) and Anwar and Kay (1977, 1978). These investigators demonstrated that the exogenous addition of histamine increased the appearance of C3b and C4 receptors on these cells. Further, the H1 agonist, 2-aminoethylthiazole, increased C3b rosetting in a manner similar to histamine. In contrast, the H2 agonists, 4-methylhistamine and dimaprit, had no effect on the expression of C3b receptors. Similarly, enhancement of C3b receptors by histamine could be blocked by HI antagonists such as chlorpheniramine and mepyramine at concentraM but not by similar concentrations of burimamide or tions of 5 x metiamide. Increased expression of C3b receptors resulted in the facilitation of eosinophil-dependent killing of certain parasites such as schistosomula and trichinella (Anwar and Kay, 1980). The increased killing induced by histamine was again mediated through the H I receptor, since H1 antagonists, but not H2 antagonists, abrogated this effect. Since parasitism is associated with high levels of IgE, part of which is directed against the parasite, sensitization of mast cells with IgE may lead to antigen-induced degranulation and subsequent histamine release. It has been suggested that the released histamine recruits eosinophils to the site of the reaction and potentiates their ability to destroy the parasite. C. NEUTROPHILS

A considerable body of evidence has accumulated to indicate that histamine is capable of altering a number of functions of neutrophils including enzyme release, chemotaxis and cheniokinesis, adherence, superoxide anion production, and hydrogen peroxide formation (Table I). However, as will be noted, some of these reports present conflicting information. Selective release of inflammatory mediators from leukocyte lysosomes has been shown to be reduced by compounds which increase intracellular cyclic AMP levels and augmented by agents which in-

220

DENNIS J. BEER ET AL.

crease intracellular cyclic GMP (Weissmann et al., 1971; Ignarro and Colombo, 1973). Histamine, in a dose-dependent fashion ( 10-8-10-5 M ) , as well as other adenylate cyclase-active agents such as isoproterenol, PGEl , and cholera enterotoxin, all inhibited the secretion of lysosomal enzymes from cytochalasin B-treated human neutrophils in response to zymosan-activated serum (Zurier et al., 1974; Busse and Sosman, 1976).The H2 antihistamine, metiamide, prevented the histamine inhibition of lysosomal enzyme release as well as the parallel increase in intracellular cyclic AMP levels of granulocytes. Chlorpheniramine, an HI antihistamine, did not prevent the histamine-induced inhibition of granulocyte lysosomal enzyme release (Busse and Sosman, 1976). Subsequent observations (Busse et al., 1980) with the use of the specific H2 agonist dimaprit, confirmed the earlier findings that inhibition of lysosomal P-glucuronidase release was mediated by the Hz receptor. Furthermore, the H1 agonist 2-pyridylethylamine, had no effect on zymosan-stimulated lysosomal enzyme release. Interestingly, the granulocyte response to histamine and dimaprit was maximal at M , and at this agonist concentration, inhibition of lysosomal enzyme release was equal for both agonists. However, at lower concentrations ( 10-7-10-5 M ) , inhibition of enzyme release was greater with histamine, indicating that histamine was 5.5 times more potent than dimaprit. The histamine- and dimaprit-induced inhibition of lysosomal enzyme release was equivalently blocked by metiamide. Histamine has been reported to be a more potent stimulus to intracellular cyclic AMP accumulation and a more potent inhibitor of lysosmal enzyme release in neutrophils obtained from normal individuals than from those obtained from asthmatic subjects or individuals with active atopic eczema (Busse and Sosman, 1977; Busse and Lantis, 1979). It was shown that zymosan-activated particles caused an equivalent release of P-glucuronidase from neutrophils obtained from asthmatics, patients with active atopic eczema, and normal subjects. M , inhibited Histamine, in concentrations ranging from to lysosomal enzyme release in all three populations, but to a lesser extent in neutrophils obtained from asthmatics and subjects with active eczema. In neutrophils obtained from the individuals with asthma or active eczema, the expected histamine-induced increase in levels of intracellular cyclic AMP was reduced in parallel with the reduction in inhibition of enzyme release. Following incubation of neutrophils from normal subjects with certain viruses, the capacity of histamine to inhibit enzyme release was diminished (Busse et al., 1979). In addition, neutrophils which were

INFLUENCE OF HISTAMINE

22 1

poorly responsive to histamine also exhibited diminished responsiveness to isoproterenol and PGEI, suggesting that if an abnormality existed in virally infected neutrophils or neutrophils from asthmatic subjects, it was not due solely to decreased numbers of histamine receptors, but to a postreceptor defect. Of further interest, pretreatment of neutrophils from normal subjects with the P blocker, propranolol, did not alter the response to histamine. Therefore, the mechanism contributing to a decreased histamine response of neutrophils from asthmatics could not be explained by P-adrenergic blockade. In contrast to these findings, Marone et al. (1980) and Clark et al. (1977) have not been able to demonstrate either significant inhibition of neutrophil P-glucuronidase release, or significant increases in neutrophil cyclic AMP levels (except in the presence of potent phosphodiesterase inhibitors) induced by histamine. The discrepancies between these results and those which do define functional neutrophil H2 receptors have not been satisfactorily resolved. Histaminase is a non-granule-associated neutrophil enzyme responsible for histamine catabolism in humans. Herman et d.(1979) have reported that neutrophil histaminase release was not modulated by histamine. Interestingly, imidazole acetic acid, the product of the enzymatic action of histaminase on histamine, was shown to be a potent inhibitor of histaminase release induced by particle-bound C3b. Other histaminase release reactions were not affected by imidazole acetic acid, not did they have an effect on the secretion of P-glucuronidase. Histamine has also been reported to have complex effects on neutrophil motility. Anderson et al. (1977) observed that histamine, at M, consistently enhanced the to 5 x concentrations of 1 x chemokinetic response of neutrophils to both endotoxin-activated serum and casein, while the chemotactic response of neutrophils to these agents was inhibited by these concentrations of histamine. That is to say, histamine, in the presence of a chemoattractant, enhanced random neutrophil migration but diminished neutrophil migration that was dependent on a concentration gradient. The inhibition of chemotaxis could be abolished by pretreatment with metiamide, an H2 receptor antagonist, but not by diphenhydramine, an HI receptor antagonist. The histamine effects on neutrophil motility were associated with a parallel increase in intracellular levels of cyclic AMP, whereas cyclic GMP levels were unaffected. Other adenylate cyclaseactive agents such as isoproterenol and PGEl produced effects on neutrophil motility similar to those of histamine. A subsequent study by Seligman et al. (1983) analyzed the effects of

222

DENNIS J. BEER ET AL.

histamine on neutrophil chemotaxis and chemokinesis induced by fMet-Leu-Phe and confirmed the earlier observations relating to histamine modulation of neutrophil migration. Chemotaxis stimulated by fMet-Leu-Phe was inhibited by histamine ( M ) , and in contrast, chemokinesis induced by this agent was enhanced by histamine. In these studies, histamine alone had no effect on random nonstimulated neutrophil migration. In related experiments, it was shown that both H1 and H2 agonists enhanced f-Met-Leu-Phe-stimulated chemokinesis and inhibited chemotaxis, while only the Hz antagonists cimetidine reversed the effects of histamine and its other analogs. The H1 antagonist pyrilamine stimulated chemokinesis and inhibited chemotaxis at high concentrations. The simultaneous addition of both HI and Hz agonists, each at less than saturating concentrations, produced additive but not synergistic inhibition of chemotaxis, which suggested that a single receptor for histamine mediated both stimulation of chemokinesis and inhibition of chemotaxis. Taken together, these studies raise questions as to the specific histamine receptor subtype(s) involved in modulation of neutrophil motility. In addition to its effects on f-Met-Leu-Phe-induced neutrophil lysosoma1 enzyme secretion and neutrophil motility, histamine and both HI and H2 agonists, in a dose-dependent fashion, were also shown to inhibit the f-Met-Leu-Phe-stimulated changes in membrane potential, superoxide anion production, and hydrogen peroxide formation without blocking the binding or internalization of [3H]f-Met-Leu-Phe (Seligman et al., 1983). These effects could be blocked by cimetidine. Histamine was unable to influence these same neutrophil functions induced by phorbol myristic acid or the calcium ionophore A23187. Taken collectively, these data suggested that a single site with specificity for both H1 and H2 analog structures modulated the various fMet-Leu-Phe-stimulated functions studied. As suggested by Seligman et al. (1983), a single receptor for histamine could enhance chemokinesis while inhibiting chemotaxis by any of at least several mechanisms. Histamine could have multiple independent effects on steps early in neutrophil activation. Alternatively, if histamine selectively inhibits the detection of a gradient of chemotactic factor without inhibiting the stimulation of motility by a chemoattractant, then enhanced chemokinesis and inhibited chemotaxis could result. In addition, the inhibition by histamine of superoxide anion or hydrogen peroxide production could disrupt regulation of the chemoattractant gradient by the myeloperoxidase/halide/peroxide system (Clark et al., 1980) with subsequent differential effects on chemotaxis and chemokinesis.

INFLUENCE OF HISTAMINE

223

111. Histamine Modulation of Immune Effector Cells

A. T-LYMPHOCYTE FUNCTIONS 1 . Cell-Mediated Cytotoxicity a. Cytotoxic T Lymphocytes (CTL).Lymphocytes from specifically immunized animals become cytolytic for cells bearing antigenic determinants against which the animals have been sensitized. The precise mechanism of the interaction between cytotoxic lymphocytes and target cells is not fully understood. In an early study, Henney and Lichtenstein (1971) found that isoproterenol and theophylline, which produce an increase in the cyclic AMP content of a variety of cell types, inhibited the cytolytic activity of splenic lymphocytes obtained from C57BW6 mice immunized with DBN2 mastocytoma cells. The first demonstrated effect of histamine modulation of lymphocyte function was that of inhibition of T lymphocyte-mediated cytotoxicity (Table 11).In the animal model employed, within 7 days following immunization of mice (C57BL/6) with histoincompatible tumor cells (DBN2 mastocytoma cells), cytotoxic T lymphocytes (CTL) were detected in various lymphoid organs including the spleen. These CTL were fully differentiated effector cells, specific for the immunizing antigen, and were able to kill their target cells in 60 minutes or less (Henney et al., 1972). The activity of CTL from the spleen was reversibly inhibited by agents that increase intracellular cyclic AMP levels, including histamine (10-5-10-3 M). The amount of inhibition of the CTL response correlated with the magnitude of the rise in intracellular cyclic AMP levels. Unlike other adenylate cyclase-stimulating agents that were studied, the inhibition caused by histamine did not exceed 50% even at very high concentrations ( M).Additionally, the inhibition caused by a particular concentration of histamine, while highly reproducible in replicate analysis involving the same effector cell population, varied greatly between different cell populations. These data suggest that the number of effector cells bearing functional histamine receptor sites varies widely, and maximally approaches only 40-50% of the total cytolytically active cells. Subsequent observations (Plaut et al., 1975) suggested that the number of functional histamine receptors on effector T lymphocytes was related to the stage of immunocompetence of the population. That is, the number of effector cytotoxic cells bearing histamine receptors, detected by their susceptibility to histamine inhibition, appeared to increase with time after immunization.

224

DENNIS J. BEER ET AL.

TABLE I1 HISTAMINE MODULATIONOF LYMPHOCYTE FUNCTIONS

Cell type I n vivo primed cytotoxic

T lymphocyte (mouse) In vitro primed cytotoxic T lymphocyte (mouse) I n vitro induction of cytotoxic T lymphocyte (mouse) Natural killer cell (mouse) Lymphokine-producing cells (guinea pig, human) Proliferating T lymphocyte (mouse, guinea pig, human) Suppressor T lymphocyte (mouse, guinea pig, human)

Contrasuppressor cell (mouse) T Lymphocyte (human)

Thymic cells (mouse) T Lymphocytes (human)

T Helper cells (human) B Lymphocytes (mouse)

B Lymphocytes (human)

Assay

Effect of histamine

Receptor specificity

HI

H2

+ +

.1 t

Cytolytic activity Cyclic AMP Cytolytic activity

0

Induction of cytotoxicity

.1

+

Cytolytic activity

.1

?

MIF, LIF, interferon

3.

Inhibition of 13H]TdR uptake Inhibition of blastogenesis, lymphokine production, and antibody production Generation of hapten self-cytotoxic T cells Induction of chemoattractant lymphokines LCF LyM 1F35~ LYMIF~sK Thy-1 antigen expression E Rosette formation RFcy Expression I d D r antigen expression IgG response (SRBC) Antibody production Cyclic AMP Pokeweed-induced IgG production IgG, IgM, IgA PFC

.1

+ +

t

+ -

t

+

t

-

t t t

+ +(?)

+(?)

-

+ +

? ? ? +(?)

? ? ? +(?)

-

+

Histamine-induced inhibition of CTL has been demonstrated to be mediated via Hz receptors, as evidenced by the observation that the Hz receptor antagonists burimamide, metiamide, cimetidine, and ICI 125,211 reversed it, but not HI receptor antagonists such as diphenhy-

INFLUENCE OF HISTAMINE

225

dramine and pyrilamine (Plaut et al., 1973, 1975; Plaut and Roszkowski, 1979; Plaut and Lichtenstein, 1982). All of the HZreceptor antagonists behaved as competitive inhibitors, as suggested by antagonist-induced shifts, in parallel, to the right of agonist log doseresponse curves, and confirmed by Schild plots with slopes not significantly different from 1. In other words, several orders of magnitude more histamine were required to obtain the same levels of inhibition in the presence of the antagonist. Interestingly, the antagonist activity M )in this assay was significantly higher than of metiamide ( K B = that noted for experiments ( K B = 3 x M ) involving histamine inhibition of P-glucuronidase release from neutrophils (Busse et al., 1980).This difference may represent a variation within leukocyte subpopulations. In addition, selective Hz receptor agonists such as 4methylhistamine and dimaprit mimicked the effects of histamine by inhibiting cytotoxicity, while the HI receptor agonist 2-methylhistamine did so only at high concentrations, compatible with its weak Hz receptor agonist activity. These data suggest that the effects of the antagonists were via blockade of the Hz receptor as opposed to other possible mechanisms such as changes in the metabolism of histamine. In addition to coincubation with Hz receptor antagonists, the sensitivity of CTL to inhibition by histamine could be altered by several other experimental manipulations. Functional Hz receptors on CTL were lost within 3 hours of culture at 37°C (Plaut and Lichtenstein, 1982). Interestingly, cultured CTL also rapidly lost their sensitivity to other adenylate cyclase-active agents. The mechanisms by which these changes in cyclic AMP responsiveness occurred were not defined. Hz receptors in splenic CTL could be specifically desensitized within 1 to 3 hours following in uivo (intraperitoneal) injection of histamine, but the sensitivity of splenic CTL to histamine was restored after 24 hours (Plaut and Roszkowski, 1979). Cytotoxic T lymphocytes can be generated in vitro, as well as in vivo, when lymphoid cells from nonimniune mice are incubated with alloantigen for 4-6 days (Cerottini and Brunner, 1974). This immune response requires interaction of antigen-processing cells (macrophages) and helper T lymphocytes with cytotoxic T lymphocyte precursors, Both proliferation and differentiation are required for optimal cytotoxic responses. Memory (secondary) cytotoxic responses result from short-term (24 hours) in vitro restimulation with alloantigen (MacDonald et aZ., 1975). Shearer et aZ. (1977) examined the effects of fractionating spleen cells on insolubilized histamine columns before in uitro sensitization to alloantigens. Spleen cells from normal BALB/c mice were cultured

226

DENNIS J. BEER ET AL.

in vitro with irradiated C57BL/6 stimulating cells. Five days later CTL activity of the effector cells was assessed. Before the BALB/c responding lymphocytes were sensitized, they were fractionated by passing the spleen cells over H-RSA-S or RSA-S control columns. Fractionation of cells over the H-RSA-S column depleted or significantly reduced the cytotoxic potential of the nonretained cells. All cytotoxic potential was recovered when the cells that adhered to the H-RSA-S were eluted from the column. In contrast, no effect on responsiveness was detected after cells had been fractionated over the control column. The loss of response potential by the cells that did not adhere to H-RSA-S could not be accounted for by removal of macrophages, nor by the concentration of cells with suppressor activity in the effluent. These cell fractionation studies suggested that the progenitors or amplifiers of CTL possess histamine receptors. These studies did not explore the question whether or not the histamine receptor-bearing cells had their function influenced by adenylate cyclase-active agents. The results presented from the study just mentioned indicate that a cell-mediated cytotoxic response to alloantigens could be influenced by passing mouse spleen cells over H-RSA-S columns before in vitro sensitization. However, Plaut et al. (1975) found that some cytotoxic effector cells capable of responding to histamine with an elevation in intracellular cyclic AMP were not preferentially bound to insolubilized histamine column. This discrepancy may be explained by an inconsistency in the binding of cells with histamine receptors to insolubilized conjugates of H-RSA. One cannot assume that because chromatography via insolubilized conjugates of H-RSA can potentially separate cells on the basis of their complement of receptors for histamine, that all cells with such receptors will be separated by this procedure (Melmon et al., 1972). There may be other determinants on the cell membrane besides the receptor itself that are critical to eventual binding of the cells to the insolubilized hormone. Because CTL can be generated in the mouse not only after immunization in vivo but also after in vitro culture as previously discussed, Plaut (1979) was able to analyze the role of cyclic AMP in modulating CTL generated under both culture conditions. CTL were generated in the spleen by immunizing C57BL/6 mice intraperitoneally with allogeneic P815 cells, or by culturing C57BL/6 spleens cells in vitro with mitomycin C-treated P815 cells. Agents that increase intracellular cyclic AMP levels, including histamine ( 10-5-10-4 M ) , prostaglandin Ez, dibutyryl cyclic AMP, and theophylline, markedly inhibited the activity of in vivo-generated CTL. In contrast, these agents had

INFLUENCE OF HISTAMINE

227

little inhibitory effect on the activity of in uitro-generated CTL. This difference was found not to be a function of lymphocyte :target cell ratio, time of incubation, dead cell content of effectors, or phosphodiesterase levels. Inhibition of cytotoxicity caused by these agents in mixtures of in uiuo- and in uitro-generated CTL populations negated the possibility of an in uiuo cyclic AMP-active suppressor cell which was not present in vitro and thereby accounting for the discrepancy in intracellular cyclic AMP inhibition of CTL. When in viuo-generated CTL were restimulated in uitro with antigen within 24 hours, the inhibition by cyclic AMP-active agents of the resultant CTL was no greater than inhibition of in uitro-generated primary CTL. Thus, some factor(s) peculiar to the in uitro culture of cells appeared to have been responsible for the low or absent cyclic AMP responsiveness of in uitrogenerated CTL. Presumably, the in oitro culture experiment was either deficient in some component that was present in viuo, or contained an element that induced a loss of responsiveness. Subsequent studies (Plaut et al., 1980) suggested that different CTL populations have distinct patterns of responsiveness to histamine and other cyclic AMP-active agents. Further, two or more mechanisms, including changes in (1) receptor number and/or receptor-adenylate cyclase coupling, and in (2) post-cyclic AMP biochemical events, may contribute to these patterns. Another study (Schwartz et al., 1980) investigated the ability of histamine to inhibit the primary in vitro induction of murine cytotoxic T cell responses. Exposure of mouse (C57BU6) spleen cells to histamine ( 10-6-10-3 M ) inhibited the induction of cytotoxic T lymphocytes specific for either allogeneic or trinitrophenol (TNP)-modified syngeneic (TNP-self-specific) targets. The use of specific histamine agonists implicated H2 receptor (but not HI receptor) specificity in the mediation of these effects early in the induction of cytotoxic cells. In agreement with the findings of Plaut (1979),the addition of histamine to previously in uitro-primed cytotoxic lymphocytes failed to inhibit their cytotoxic effector function. Kinetic studies demonstrated that this difference might be due to a loss of functional histamine receptors after the initiation of the in uitro mixed leukocyte culture. The mechanism(s) of histamine-induced suppression of the induction of primary in uitro murine cytotoxic responses was not precisely defined. It is possible that this suppression was mediated through histamine activation of regulatory T cells. b. Natural KilEer Cells. Natural killer (NK) lymphocytes possess an important but poorly understood capacity to destroy tumor cells both

228

DENNIS J. BEER ET AL.

in vivo and in vitro (Kiessling and Wigzell, 1979). Some results have suggested that the lytic mechanism of NK cells was distinct from that of other effector cell types. A point mutation on chromosome 14 in a particular mouse strain caused a complete impairment in the postrecognition lytic pathway of NK cells with no effect on cytolysis mediated by T cells or macrophages (Roder and Duwe, 1979). Despite these observations, Roder and Klein (1979)found that dibutyryl cyclic AMP and the cyclic AMP-elevating agents, prostaglandin El, theophylline, and histamine ( 10-7-10-3 M), markedly suppressed mouse splenic NK cytolytic function (Table 11). These results were similar to the earlier findings concerning alloantigen-induced CTL. The inhibition by histamine was rapidly induced, dose dependent, and persisted in the presence of the drug. The receptor specificity of this histamine effect was not defined. A subsequent report (Ruiz-Arguelles et al., 1982) investigated the effect of cimetidine on human NK cell activity against K562 myeloid cells. A bimodal inhibitory dose-related effect was found employing peripheral blood mononuclear cells as effectors. When an enriched population of large granular lymphocytes served as effector cells, there was a dose-dependent inhibition of NK activity by cimetidine (10-'0-10-3 M), ranging from 20 to 50% inhibition of cytolysis. The mechanism(s) for cimetidine inhibition of human NK activity are not clear.

2. Lymphocyte Prolijeration a. Animal Studies. The ability of lymphocytes to undergo increased DNA synthesis (measured by incorporation of [3H]thymidine)and mitosis in response to a specific or nonspecific stimulus has been termed lymphocyte proliferation and reflects an expansion of immunoreactive clones of cells. Lymphocyte proliferation per se does not represent an immune response, but the sequence of events is probably analogous to that occurring during an immune response. The action of histamine on lymphocyte proliferation has been studied in a number of species (Table 11).In the mouse, high concentrations of histamine (10-5-10-4 M) added directly to the lymphocyte cultures inhibited Con A-induced proliferation (Plaut et al., 1975; Plaut and Roszkowski, 1979). The effects of histamine were only seen when suboptimal concentrations were utilized. The suppression of lymphocyte proliferation was found to be mediated through the Hz receptor as evidenced by a reversal of the histamine effect by adding the Hz receptor antagonist burimamide to the cultures. It is of interest that in contrast to the effects of histamine on Con A-induced prolifera-

INFLUENCE OF HISTAMINE

229

tion, histamine did not inhibit lipopolysaccharide (LPS)-induced proliferation. The latter response is though to be specific for B cells, whereas the above Con A mitogenic response represents primarily a T cell response. Histamine (10-7-10-3 M ) has also been reported to inhibit (in a dose-dependent fashion) the murine lymphocyte proliferative response to allogeneic stimuli elicited in a mixed lymphocyte culture (MLC) (Schwartz et al., 1977). Serotonin, another vasoactive amine, did not inhibit MLC-induced lymphocyte proliferation. The in vitro blastogenic response of guinea pig lymphocytes to specific antigen has been shown to be inhibited by histamine ( 1 W M) (Rocklin, 1976; Beets and Dale, 1979). In the former study, burimamide, but not chlorpheniramine, effectively restored the proliferative response to control values thereby reversing the histamine effect. This suggested that histamine inhibition was mediated through activation of an H2 receptor. b. Human Studies. Recent studies in normal subjects indicate that histamine also modulates the proliferative response of human lymphocytes. Concentrations of histamine as low as lop7M consistently inhibited the proliferative response of human lymphocytes induced by suboptimal concentrations of T cell mitogens such as Con A (Goodwin et al., 1979; Plaut and Berman, 1978; Martinez et al., 1979). This effect was apparently mediated by H2 receptors. However, histamine had only minimal inhibitory effects on proliferation induced by optimal or supraoptimal mitogen concentrations. Ballet and Merler (1976) reported that high concentrations of histamine inhibited the proliferative response to alloantigens induced by MLC. The influence of histamine (10-s-10-3 M) on the proliferative response of T cells in autologous and allogeneic mixed lymphocyte cultures (MLC) has also been studied (Damle and Gupta, 1981). Pretreatment of responder T cells, but not of stimulator non T-cells with histamine for 24 hours, resulted in a dose-dependent diminished proliferative response in both the autologous and allogeneic MLC. A minimum of 4 hours of incubation of responder T cells with histamine ( M ) was required to inhibit autologous and allogeneic MLC. Suppression of the MLC increased as the time of incubation with histamine approached 24 hours. Histamine-induced inhibition of T cell proliferation in an MLC (autologous or allogeneic) was prevented by the presence of an equimolar concentration of cimetidine but not of chlorpheniramine. Preincubation of responder T cells for 24 hours with 4-methylhistamine and dimaprit inhibited T cell proliferation in autologous and allogeneic MLC. The degree of inhibition seen with these H2 receptor agonists was comparable to that observed with his-

230

DENNIS J. BEER ET AL.

tamine. In contrast, preincubation of responder T cells with 2-methylhistamine and 2-pyridylethylamine failed to inhibit the proliferative response of T cells in either an autologous or an allogeneic MLC. T cells pretreated with histamine ( 10-5-10-4 M ) , followed by mitomycin C treatment, when cocultured with fresh autologous T cells, suppressed their proliferative response in both the autologous and allogeneic MLC. In addition, supernatants obtained from T cells cultured with histamine M ) for 24 hours also suppressed both autologous and allogeneic MLC responses by fresh T cells. Taken collectively, these observations suggest that inhibition of MLC-induced proliferation by histamine was mediated through activation of suppressor T cells via their H2 receptor with subsequent elaboration of a soluble suppressor factor(s). Histamine-induced suppressor cell activity will be discussed in more detail in Section III,A,4. Several reports have suggested that lymphocytes from different individuals vary in their responsiveness to histamine-induced suppression of blastogenesis. Con A-stimulated cultures of peripheral blood mononuclear cells from subjects with the HLA-B12 haplotype were significantly less inhibited than control cells by the addition of M histamine to the cultures (Staszak et al., 1980). If a suboptimally mitogenic concentration of PHA was employed to stimulate lymphocyte proliferation, histamine inhibition of blastogenesis was equivalent for HLA-Bl2 individuals and controls. Since Con A and PHA stimulate different subpopulations of human peripheral blood T cells (Engleman et al., 1979), it may be that these subpopulations differ in their sensitivity to histamine-induced suppression. Lymphocytes from HLA-BI2 individuals were also less inhibitable by PGE2 (3 x loT93x M ) . This suggests that the defect in histamine responsiveness may reside at a point beyond the receptor and not with the receptor itself. The results from other studies concerning individual responsiveness to histamine-induced suppression of lymphocyte proliferation are contradictory with each other and difficult to interpret. Strannegard and Strannegard (1977, 1979), employing different experimental conditions than other investigators, such as 5-day cultures of whole blood, found complex histamine dose-response curves including some proliferation-enhancing effects at lo-' M concentration. They reported that histamine and other adenylate cyclase-active agents inhibited proliferation of lymphocytes from atopic individuals (children), but not from normals. Another report (Brostoff et d.,1980) of increased sensitivity to histamine by lymphocytes from atopic donors is also difficult to interpret, since some histamine dose-response

INFLUENCE OF HISTAMINE

231

curves were biphasic, and Hz receptor antagonists potentiated histamine-induced suppression. Moreover, at least two studies (Martinez et al., 1979; Beer et al., 1982a) reported that, compared to lymphocytes from nonatopic individuals, the proliferation of lymphocytes from atopic donors were less inhibited by histamine. Finally, another report (Wang and Zweiman, 1978) did not find any difference in the histamine-induced suppression of lymphocyte blastogenesis between normal and atopic individuals. There are several reports concerning the capacity of cimetidine to modulate in vitro human lymphocyte proliferative responses to mitogens and antigens (Gifford and Schmidtke, 1979; Gifford et al., 1980; Festen et al., 1981).Neither the consistency of these findings nor their biologic significance has been established. It is possible that cimetidine is either blocking the inhibitory effects of histamine endogenously released from basophils contaminating the human peripheral blood lymphocyte preparation, or that cimetidine may have an as yet undefined immunostimulatory effect.

3 . L y mpho kine-Producing Cel 1s a. Macrophage Migration-Inhibiting Factor (MZF). T lymphocytes, when activated in vitro either by specific antigens to which they are sensitized or by nonspecific mitogens, produce a wide array of soluble factors having effects on a number of inflammatory cells. These nonantibody substances are collectively termed lymphokines and are involved in various aspects of the expression of cell-mediated immune reactions. Histamine has been shown to diminish partially the expression of delayed-type hypersensitivity (DTH) reactions in the skin and to alter profoundly certain lymphocyte functions in vitro (Rocklin, 1976). Concentrations of up to M histamine administered intradermally alone with the eliciting antigen maximally reduced (by 40-60%) the DTH skin reaction in guinea pigs immunized with orthochlorobenzoyl-bovine y-globulin (OCB-BGG) in complete Freund’s adjuvant. This suppression of DTH skin reactivity was completely reversed by burimamide (Hz receptor antagonist) but only partially corrected by chlorpheniramine (HI receptor antagonist). To explore the effects of histamine on cellular immune reactions further, a series of in vitro experiments were carried out on the guinea pig cells (T lymphocytes and macrophages) intimately involved. In a dose-dependent fashion, histamine ( 10-5-10-3 M ) suppressed the antigen-induced production of macrophage migration-inhibitory factor (MIF) (Fig. 2). Histamine inhibition of MIF production required that

232

DENNIS J. BEER ET AL.

$

0 -10I

Control

I 10-3

Histamine Histamine Histamine Histamine 10-4

+ t t

Burimomide Diphenhydramine Chlorpheniramine

10-5

10-6

10-7

CONCENTRAT/oN fM/

FIG.2. Effect of histamine on MIF production by OCB-BGG-immune guinea pig lymph node lymphocytes. Suppression of MIF by histamine is concentration related. Hz receptor antagonists restore M I F production to normal.

histamine be added within 1 hour of culture of sensitized lymphocytes with antigen; thereafter, the addition of histamine did not significantly alter the production of MIF. This histamine effect was reversible. In contrast to the results obtained with skin reactions, only H2 receptor antagonists such as burimamide and metiamide blocked the suppressive effects of histamine on lymphocyte MIF production (Table 11).Chlorpheniramine and diphenhydramine (HI receptor antagonists) failed to prevent histamine's action of MIF production. Although agents that raise intracellular levels of cyclic AMP have been demonstrated to inhibit guinea pig MIF production (Pick, 1974), it was not clear that histamine-induced inhibition of MIF production was mediated via this mechanism. In fact, observations by Rocklin et al. (1978) indicated that histamine inhibited MIF production indirectly, by activating suppressor cells (see later), which, in turn, acted on MIF-secreting cells. It was observed that the removal of a subpopulation of cells by passage over H-RSA-S columns rendered the remaining nonadherent population unresponsive to histamine, that is, lymphocytes produced MIF in response to the sensitizing antigen in the presence of soluble histamine (Fig. 3). Other studies by Rocklin (1977) demonstrated the presence of a nondialyzable macromolecule (23,000-40,000 daltons) elaborated by guinea pig lymphocytes (and not macrophages) stimulated with histamine that was capable of suppressing in vitro M I F production via a noncytotoxic mechanism. The

233

INFLUENCE OF HISTAMINE

IA

IB

40

30 20

to 0 1 0 -

Control

-

tu3~

I

~o-~M

Histarnine-

I

Control

-

~ o - ~ M IO-~M Histamine-

FIG.3. Immune lymph node cells (LNC) were chromatographed on H-RSA and RSA columns and either tested immediately (A) for their sensitivity to histamine or allowed to “recover” for 18 hours (B) before exposing them to histamine. Cells allowed to recover for 18 hours still produced M I F in the presence of histamine, indicating that histamine receptor-bearing lymphocytes were retained on the H-RSA column and did not pass through in a desensitized state.

biologic activity of this factor was not diminished after incubation with phosphodiesterase, suggesting that this histamine-induced factor was not cyclic AMP. b. Other Lymphokines. Elaboration of other lymphokines such as human leukocyte migration-inhibitory factor (LIF) (Riga1 et ul., 1979) and mouse immune interferon (Weinstein and Melmon, 1976) have also been demonstrated to be decreased by the addition of histamine in uitro. The former observation appears to be secondary to a soluble mediator with a MW of 10,000-50,000 that was elaborated by human T cells stimulated with histamine. Of related interest is the study by Reichman et al. (1979) showing that phytohemagglutinin and PPDinduced L I F production was significantly decreased from human mononuclear cells depleted of cells adherent to insolubilized histamine when compared to that elaborated by unfractionated cells. These results, in conjunction with the previously mentioned studies, suggested that not only can cells possessing histamine receptors suppress the production of L I F but also that there is a subpopulation of lymphocytes involved in LIF production which possibly expresses histamine receptors.

234

DENNIS J. BEER ET AL.

Clinical observations by Jorizzo et al. (1980) on four family members with chronic mucocutaneous candidiasis treated with cimetidine support the concept that histamine inhibits lymphokine production and thereby has antiinflammatory properties. During treatment with cimetidine, all four patients who initially lacked DTH skin responsiveness to candida antigen and whose lymphocytes did not produce LIF in response to candida in vitro showed improvement in these parameters of cellular immunity, The most likely interpretation of these results is that endogenous histamine blocks lymphocyte function in this subset of patients, either because the concentration of histamine is unusually high, or because their cells are very sensitive to its effects. Cimetidine may reverse the anergy by competing with endogenously generated histamine and thereby reduce suppressor tone. Of course, a nonspecific immunostimulatory effect of cimetidine in the latter study cannot be excluded. It should be noted, however, that the effect of cimetidine in the latter study was irreversible, that is, washed lymphocytes from treated patients made LIF, suggesting that its mechanism of action might not simply be due to Hz receptor blockade. In contrast to these inhibitory effects on lymphokine elaboration, two reports suggest that histamine enhances production of lymphocyte-derived factors. In the first report, histamine had no effect alone, but acted synergistically with 5-hydroxytryptamine to induce normal blood lymphocytes to generate a chemotactic factor(s) for monocytes M ) was shown (Foon et al., 1976). I n a second study, histamine to induce guinea pig lymphocytes to secrete a low molecular weight factor that interfered with the activity of complement-derived chemotactic factors for guinea pig and human eosinophils. The receptor specificity of these enhancing effects was not defined, although the latter was reported to be blocked by both HI and Hz antagonists (Kownatzki et al., 1977).

4 . Suppressor Cell Function a. Animal Studies. Immune responses are similar to other complex biologic processes, such as blood clotting and complement activation, in that they are modulated by a series of positive and negative regulatory factors (Cantor and Gershon, 1979). A major negative regulatory influence is provided by suppressor T cells. Suppressor lymphocytes can be activated in vitro by various stimuli including specific antigens, mitogens, antigen-antibody complexes, and histamine (see later). This activation process involves the participation of macrophages, inducer T lymphocytes, and suppressor cell precursors. Ma-

INFLUENCE OF HISTAMINE

235

ture suppressor T cells are of a lineage distinct from helper, MIFproducing, and cytotoxic T lymphocytes, and are themselves a heterogeneous population. Signals from cell to cell in immunoregulatory loops may be mediated by soluble suppressor factors. Indeed, addition of soluble suppressor factors to some in vitro systems such as antibody formation and mixed lymphocyte reactions mimics the effects of activated suppressor cells. The assessment of drug effects on suppressor cell function is complicated by the fact that the assay consists of both an activation phase and an effector stage (interaction of the activated suppressor lymphocytes with the target system). Evaluation of suppressor function is dependent on the model employed, and thus any apparent inconsistent effects of histamine may reflect the sensitivity of the various suppressor systems or the nature of cellular interactions involved in a given immune response. On the one hand, and in keeping with the concept of a negative modulatory role on lymphocytes of agents that activate adenylate cyM ) and other adenylate cyclase-active agents clase, histamine ( were shown to block the induction of antigen-induced suppressor cells (Mozes et al., 1974). Reduced immune response potential of SJL/J mouse spleen cells to a synthetic multichain polypeptide antigen was observed when the spleen cells were incubated in vitro with the antigen prior to in vitro challenge of the cells with the same antigen in syngeneic irradiated recipients (Mozes et al., 1974). Incubation of the cell-antigen mixture in the presence of histamine, prostaglandins, or cholera enterotoxin was associated with an increased accumulation of intracellular cyclic AMP in the spleen cell suspension, and abrogated the antigen-induced loss of responsiveness. AddiM ) to the tion of the H1 receptor antagonist diphenhydramine cell-antigen-histamine mixture abolished the effects of histamine in both the inhibition of suppression and increased intracellular cyclic AMP levels, suggesting that abrogation of antigen-specific suppressor cell activity by histamine was mediated via the HI receptor. The finding of H1 receptor-mediated elevation of intracellular cyclic AMP was unusual in that most investigators have found that agonist stimulation of lymphocyte H2 receptors, but not HI receptors, characteristically activates adenylate cyclase. M ) has also been shown to inhibit Con A-induced Histamine ( suppressor T cell activation (Schwartz et al., 1981). In a murine system, in vitro exposure of splenocytes to Con A resulted in the induction of T cells (Thy-1.2+, Lyt-1-23+), which suppressed the subsequent in vitro generation of trinitrophenol (TNP)-modified

236

DENNIS J. BEER ET AL.

self-specific cytotoxic T lymphocytes. Such suppressor populations did not contain detectable cytotoxic cells or cytotoxic precursors specific for TNP-modified syngeneic target cells. Histamine inhibited both the DNA-synthetic response and generation of suppressor T lymphocytes induced by Con A. When the H1 agonist, 2-pyridylethylamine (10-5-10-4 M ) was added to the Con A cultures, suppressors were induced which on transfer to the assay system resulted in strong inhibition of the generation of cytotoxic T cells. However, addition of the H2 agonist, dimaprit (5 x M ) , resulted in a decrease in suppressor induction equivalent to that seen by the addition of M histamine. This suggested that the ability of histamine to inhibit the generation of suppressor T cells was mediated through an Hz receptor. Since this study employed unpurified cell populations and because any given immune response involves multiple intercting subsets of immunoregulatory cells, the mechanism of histamine inhibition of Con A-induced suppressor activity could not be defined. In this regard, another class of immunoregulatory T cells have been defined which exert positive regulation in that they oppose suppressor cell activity. They have been termed “contrasuppressor cells,” and their activity has been referred to as “contrasuppression” (Gershon et al., 1981; Yamauchi et al., 1981). Subsequent investigations have suggested that interferon (Knop et al., 1982) and histamine (Siegal et al., 1982) might serve to activate contrasuppressor cells. Spleen cells from BGAFl mice treated in vivo with the toleragen trinitorbenzenesulfonic acid (TNBS) were employed as potential regulators of in vitro generation of primary antitrinitrophenyl self-cytotoxic T lymphocytes. This model system was employed because these cells themselves would not develop into killer cells. Under the conditions used, however, these spleen cells exerted no major regulatory effects. In contrast, if these cells were preincubated with histamine ( M ) for 3060 minutes, suppressor activity was induced, but this occurred inconsistently and with nonstoichiometric results. It then became apparent through the selective use of synthetic histamine H1 and H2 agonists that histamine may activate both suppressor and contrasuppressor cell subsets. When an HI receptor agonist (2-pyridylethylamine or 2-PEA) was used, no suppressor activity was observed in cells from TNBS-treated mice, whereas the Hz receptor agonist dimaprit tended to activate suppressor cells. However, when the regulatory cells were pretreated with both of the two synthetic agonists, the suppressive capacity of dimaprit was eliminated. Thus, it appeared that 2-PEA was not without effect in this system, but was exerting an effect that nullified suppressor cell induction by dimaprit.

INFLUENCE OF HISTAMINE

237

A mixing experiment was performed in order to determine whether cells made tolerant in vivo and then preincubated with 2-PEA could block the suppressive effects of regulatory cells that had been preincubated with dimaprit. A separate aliquot of regulatory cells was preincubated with 2-PEA and then added to the regulatory cells that had been preincubated only with dimaprit. This resulted in a significant reduction in the amount of suppression. The latter experiment suggested that histamine may have opposing actions that tend to obscure suppression. This bidirectional potential was also shown by treatment of 2-PEA-induced contrasuppressor cells (Z-J+) with complement and anti-Z-J antibody (Gershon et al., 1981; Yamauchi et al., 1981). The results of this treatment revealed a high level of suppressor cell activity that was not expressed until the opposing contrasuppressor cells were removed. Taken collectively, these data demonstrated that histamine can have more than one immunoregulatory effect and, under certain conditions, these two effects can cancel one another out. The two opposing effects induced by histamine can occur by its actions on separate cells and are not necessarily a cancellation of one cellular signal by another on the same cell. Results from other experiments also indicate that histamine provides a positive signal for activation of suppressor cells. T cells activated in this manner have been shown to suppress lymphocyte proliferation (Rocklin, 1976; Rocklin et al., 1978,1979; Thomas et al., 1981), lymphokine production (Rocklin, 1976, 1977), and immunoglobulin production (Lima and Rocklin, 1981; Garovoy et al., 1983). The mechanisms and cellular interactions by which histamine leads to suppression of certain lymphocyte functions have been analyzed. Rocklin (1977) observed that lymph node cells from nonimmune or immune strain 2 guinea pigs elaborated a nondialyzable factor (histamine-induced suppressor factor of HSF) into the culture supernatant M histamine. HSF, when cowhen incubated with lops to cultured with sensitized lymphocytes, suppressed their production of MIF and proliferative responses to antigen. HSF was made by lymphocytes but not by macrophages or B cells (Rocklin et al., 1979). The production of HSF by lymph node cells stimulated by histamine was concentration dependent. That is, as the concentration of histamine was decreased, less factor was made as evidenced by less suppression of MIF production (Fig. 4).Its production could be blocked by an Hz receptor antagonist (burimamide) but not by an HI receptor antagonist (chlorpheniramine). Furthermore, the inhibitory effect of H S F was reversible, as lymphocytes washed free of the factor after 24 hours and recultured with fresh medium and antigen were able to produce MIF.

238

DENNIS J. BEER ET AL.

H, Agonist (2-methylhistamine)

** H2Agonist (4-methylhistamine)

$

60-

h 40 -

s2i

2 8 $

20-

O-

s? -20

-

FIG.4. Production of HSF is preferentially induced by HZagonists. HI (Z-methylhistamine) and Hz (4-methylhistamine) agonists in varying concentrations ( 10-3-10-7 M ) were incubated with LNC for 24 hours as per histamine. Supernatants from these cells were analyzed for HSF activity using suppression of M I F production as before. MIF production in the presence of control supernatants varied from 44 to 56% inhibition of migration.

Molecular sieve gel filtration over Sephadex G-100 indicated that guinea pig HSF had an approximate molecular weight of 23,000 to 40,000. A subsequent report (Rocklin et al., 1978)showed that passage of immune lymph node cells over H-RSA-S columns rendered the nonadherent cells insensitive to histamine, that is, these cells made MIF and proliferated in the presence of histamine to the same extent as unfractionated cells. Cells that passed through control columns (RSA or histidine methyl ester-RSA columns) retained their sensitivity to histamine. That cells bearing histamine receptors were actually removed by H-RSA-S columns and were not present in an unresponsive state to histamine was suggested by experiments in which cells that were not adherent to columns were allowed to recover for 18 hours before being stimulated to make MIF (Fig. 3). In these experiments, H-RSA-S-passed cells did not regain their sensitivity to histamine. If these cells were present but unresponsive, they should have regained their responsiveness to histamine within that period of time.

INFLUENCE OF HISTAMINE

239

Moreover, lymphocytes chromatographed on H-RSA-S columns did not make HSF but the retained cells did. In contrast to the potent activity of histamine-stimulated supernatants, other agents that increase intracellular cyclic AMP levels, such as dibutyryl cyclic AMP itself, PGEl, isoproterenol, and cholera toxin, and other mediators of immediate hypersensitivity including seroM ) , slow-reacting substance of anaphylaxis (100 unitdml), tonin ( and eosinophilic chemotactic factor (tetrapeptide, M ) , did not generate HSF-like activity (Rocklin et al., 1979). The biochemical mechanism(s) by which histamine activates guinea pig suppressor cells had not been defined in the studies just mentioned. If, as has been demonstrated, only histamine and no other drugs that increase intracellular cyclic AMP leads to the elaboration of suppressor factor, then either the T cells which produce HSF do not possess functional receptors for the panoply of drugs that raise cyclic AMP, or the effects on HSF are not related to its accumulation but are related to an as yet undefined mechanism of Hz receptor-mediated histamine action. The fact that exogenously added dibutyryl cyclic AMP did not stimulate HSF production suggests that the mechanism by which histamine induced HSF production was not a result of an accumulation of cyclic AMP. Histamine-induced elevation of cyclic AMP levels in suppressor cells has not been demonstrated. In this regard, studies relating to histamine inhibition of cytotoxic T lymphocyte activity have shown this effect to be independent of and unassociated with histamine-induced suppressor factor production or histamine-activated suppressor cells (Plaut and Lichtenstein, 1982). The effect of histamine on the activity of cytotoxic lymphocytes was apparently by a direct effect on Hz receptors present on the cytotoxic cells with consequent activation of adenylate cyclase, and not via an “HSF.” The negative effect of supernatants derived from mixtures of immune lymphocytes, target cells, and histamine was due entirely to histamine itself, as the inhibitory effect was abolished following treatment of the supernatant with histaminase. Thus, although histamine can activate adenylate cyclase and thereby modulate certain lymphocyte functions without activating suppressor cells, histamine also appears to activate suppressor T lymphocytes and thereby influence particular lymphocyte functions without necessarily involving changes in intracellular cyclic AMP. A murine model of histamine-induced suppressor cell activity has been described (Suzuki and Huchet, 1981, 1982). Histamine-induced inhibition of lymphocyte responses to mitogens in mice was caused by an activation of suppressor T cells bearing Hz receptors and Ly2-3+

240

DENNIS J. BEER ET AL.

membrane antigen. Spleen cells preincubated for 24 hours with histamine and then washed were capable of inhibiting the response of syngeneic spleen cells of PHA in a coculture system. Such suppressive activity was abrogated after treatment with anti-Ly2.3 antiserum plus complement. Histamine ( M)-induced inhibition was related to the elaboration of a suppressor factor with an apparent molecular weight of 45,000 to 68,000 as determined by Sephadex G-100 chromatography. This factor's inhibitory activity was removed by passage over a rabbit anti-mouse Ig-Sepharose column; its effect was blocked by a prostaglandin synthetase inhibitor, suggesting that this activity was in part secondarily mediated by prostaglandin synthesis (see later). b. Human Studies. Several studies have demonstrated that histamine also activates human suppressor cells (Rocklin et al., 1980; Rocklin and Haberek-Davidson, 1981; Lima and Rocklin, 1981; Thomas et al., 1981; Beer et al., 1982b,c; Garovoy et al., 1983). Highly purified populations of human T cells, but not B cells, produced HSF in response to varying concentrations of histamine ( 10-9-10-4 M). The histamine dose-response curves for the generation of HSF were variable from one individual to another. In some instances, high concentrations of histamine (10-5-10-4 M) were required to produce HSF, whereas in other instances, low concentrations ( 10-8-10-6 M) generated peak HSF responses. Ongoing DNA metabolism was not required for the release of HSF, as mitomycin C treatment did not diminish HSF production but d e novo protein synthesis was essential. The T cells that synthesized HSF expressed H2 receptors and were retained on H-RSA-S (but not RSA-S) affinity columns; cells with histamine receptors have been found within the OKT8-positive population (Damle and Gupta, 1981; Lima and Rocklin, 1981) and comprised of approximately 50% of the population of T cells that possess Fc receptors €or IgG (Ty) but were not found in the population of T cells that possess Fc receptors for IgM (Tp). Moreover, cells that were not retained by histamine affinity columns had a reduced capacity (compared to unfractionated or control column-passed cells) to develop into suppressor cells following stimulation by Con A or specific antigen (Rocklin and Habarek-Davidson, 1981). Human HSF has been characterized by enzyme treatment, sensitivity to reduction, and alkylation, by molecular sieve chromatography and polyacrylamide gel electrophoresis (Rocklin et al., 1983a). HSF was found to have a wide pH stability (pH 3-10), sensitivity to temperatures greater than 80°C, and to have the properties of a glycoprotein by virtue of its sensitivity to chymotrypsin, trypsin, sodium periodate, and neuraminidase. HSF did not appear to have a serine

INFLUENCE OF HISTAMINE

24 1

group(s) in its active site, since its biologic activity remained intact following treatment with an irreversible serine esterase inhibitor. Furthermore, HSF did not appear to have inter- or intramolecular disulfide bridges, because treatment with denaturing and/or reducing agents, followed by alkylation, did not significantly alter its suppressor activity. Molecular sieve chromatography employing Sephadex G100 revealed an apparent molecular weight for HSF of 25,000 to 40,000 (Table 111). Electrophoresis of HSF in polyacrylamide gels at pH 8.7 under nonreducing conditions revealed two regions of activity, one migrating with albumin and the other anodal to albumin (acidic glycoproteins). Furthermore, gel filtration of supernatants generated by stimulating human mononuclear cells with histamine, dimaprit (but not 2-PEA), Con A, or specific antigen, resulted in similar elution profiles with regard to inhibition of lymphocyte proliferation. That is, 25,000 to 40,000 MW fractions of supernatants generated by each ligand suppressed lymphocyte proliferation to a similar degree. These findings provide indirect evidence that T lymphocytes, triggered by antigen-specific or nonspecific stimuli, elaborate suppressor molecules capable of modulating T cell function that share certain similarities. The effect of various metabolic inhibitors and agents which increase intracellular cyclic nucleotide levels on the generation of suppressor cells was investigated in the coculture system described previously (Rocklin and Haberek-Davidson, 1983). Human blood TABLE 111 PHYSIOCHEMICAL AND FUNCTIONAL CHARACTERISTICS OF HISTAMINE-INDUCED HUMAN LYMPHOKINES Receptor specificity

Physiochemical characteristics

Biologic activities

Histamine-induced suppressor factor (HSF)

H2

MW 25,000-40,000; b'0 species of acidic glycoproteins

Lymphocyte chemoattractant factor (LCF) Lymphocyte migration-inhibitory factor 35K (LyMIF& Lymphocyte migration-inhibitory factor 75K (LyMIF&

HZ

MW 56,000; cationic sialoprotein with a p l of 9.0 MW 300,000-40,000; cationic sialoprotein with a p l of 8.5 MW 70,000-80,000; cationic protein with a p l of 7.5

Suppression of lymphocyte proliferation, lymphokine production, and antibody production Selective chemokinetic activity for T lymphocytes Noncytotoxic inhibitor of T lymphocyte migration Noncytotoxic inhibitor of T lymphocyte migration

Factors

HI

HI

242

DENNIS J. BEER ET AL.

mononuclear cells (MNC) were stimulated with histamine ( 10-5-10-3 M)for 24 hours in the absence or presence of other agents, washed, cocultured with lectin-stimulated autologous MNC, and suppression of [3H]thymidine uptake measured, The generation of histamine-induced suppressor cells was shown to require active cellular metabolism, since inhibitors of transcription (1pg/ml actinomycin D), translation (5 pglml puromycin and 1 pglml cycloheximide), oxidative phosphorylation M sodium azide), and glycolysis (lop2M 2deoxyglucose) significantly reduced their activity. Inhibitors of cytoskeletal function such as cytochalasin B (5 pg/ml) and colchicine ( 3 x M ) also markedly reduced suppressor activity. There was no apparent requirement for DNA synthesis in the generation of suppressor cells, since treatment of MNC with mitomycin C (50 pg/ml) had no effect on suppression. Incubation of MNC with cyclic AMP-elevating M papaverine) for 2 M dibutyryl cyclic AMP and 2 x agents ( hours resulted in augmented suppressor activity, while longer incubations (24 hours) resulted in reduced function. Incubation (2-24 hours) M dibutyryl cyclic of MNC with cyclic GMP-elevating agents M imidazole) had no effect on suppressor cell activity. GMP and These agents may interfere with critical lymphocyte-macrophage interactions (lymphokine/monokine synthesis and/or secretion) that are required for intact suppressor responses. Employing a coculture assay, the requirement of the macrophage/ monocyte in the activation and expression of suppressor cells in response to histamine has been explored. Histamine activation of human suppressor T cells required the presence of monocytes (HBbert et al., 1980; Beer et al., 1982b,c). Nonadherent T lymphocytes containing less than 2% monocytes were unable to be activated by histamine to express suppressor activity. If, during the generation phase, the nonadherent T cell population was reconstituted by the readdition of 5-10% autologous monocytes in the form of glass-adherent cells, subsequent exposure to histamine resulted in suppressor activity. The requirement for intact monocytes in the activation process could be bypassed by suspending the nonadherent T lymphocyte population in supernatants derived from allogeneic monocytes stimulated with heat-killed Staphylococcus albus. These crude supernatants contained leukocyte pyrogen (LP) and lymphocyte-activating factor (LAF) activities. Sequential purification of the crude monocyte supernatants using gel filtration, immunoadsorption, and isoelectric focusing demonstrated that only those fractions containing both LP and LAF were capable of reconstituting nonadherent T cell histamine-induced suppressor activity. LP/LAF (interleukin-1) by itself

243

INFLUENCE O F HISTAMINE

did not activate T suppressor cells nonspecifically, but instead behaved as an accessory factor that was required for the generation of histamine-induced suppressor T cells (Fig. 5). Additionally, further investigation (Staszak and Goodwin, 1980; Beer et al., 1982b; Rocklin et al., 1983a) has demonstrated the requirement of monocytes in the effector stage of histamine-induced suppressor activity. The expected histamine-induced suppression of PHAstimulated lymphocyte blastogenesis was abrogated with either removal of nylon wool-adherent cells, or addition of indomethacin or compound RO 20-5720 (whose only known action is reversible inhibition of prostaglandin synthetase). These effects could be explained by a requirement for monocytes or monocyte products for suppressor cell activation and/or production of prostaglandins during the effector stage. A subsequent study (Beer et al., 1982b), employing a coculture assay for histamine-induced suppression, permitted delineation of the indomethacin effect between the generation phase (suppressor population) and/or the effector phase (indicator population). This study demonstrated that removal of monocytes from both the suppressor and indicator populations resulted in a marked reduction in the expression of histamine-induced suppressor activity despite the addition of human monocyte-derived interleukin-1 to both phases of the assay. Thus, unlike the generation phase of suppressor cell activation, where a monocyte-derived factor could partially replace or augment the function of intact monocytes, the effector phase appeared to require the presence of monocytes or a factor not present in the dialyzed, partially purified monocyte supernatants for full expression of histamine-induced suppressor activity.

/

LEUKOCYTE PYROGEN

PROSTAGLANDINS

HISTAMINE OF

OMIF ?

FIG. 5. A proposed model of the in uiuo monocyte-lymphocyte interactions required for the expression of histamine-induced suppression. Histamine, in the presence of monocytes or their secretory products, activates a subpopulation of T lymphocytes to elaborate soluble substances capable of stimulating mononuclear phagocytes to synthesize prostaglandins that subsequently suppress mitogen-stimulated lymphocyte proliferation.

244

DENNIS J. BEER ET AL.

The addition of indomethacin (1 pg/ml) at the time of coculture (effector phase) abrogated histamine-induced suppression when monocytes were present in the indicator phase population (Beer et al., 1982b; Rocklin et al., 1983a). In contrast, indomethacin had no effect on the activation of suppressor cells. In related experiments (Rocklin et al., 1983a), addition of exogenous PGE2 (10-8-10-7 M ) to mononuclear cell cultures reconstituted histamine-induced suppressor activity in the presence of indomethacin. Furthermore, supernatants from histamine- or dimaprit-stimulated mononuclear cells induced monocytes, but not lymphocytes, to increase their production of prostaglandins E2 and Fza and thromboxane B2. Suppressor supernatants were chromatographed by gel filtration on Sephadex G-100. The elution profiles for the factor(s) inducing suppression of lymphocyte proliferation (25,000-40,000 daltons) and augmenting PGE2 production (25,000 daltons) overlapped but were not identical. Collectively, these data suggest that HSF-mediated inhibition of lymphocyte proliferation may occur in part through the augmented production of prostaglandins and/or thromboxane Bz by human monocytes (Fig. 5 ) . 5. Lymphocyte Motility Delayed-type hypersensitivity reactions initiated by activated lymphocytes are characterized by a mononuclear cell infiltrate consisting of macrophages and lymphocytes, the majority of which are not specifically sensitized for the antigen eliciting the reaction (McCluskey et al., 1963). Also frequently noted in these reactions is the presence of basophils (Askenase et al., 1980) whose histamine content can be released through the action of a recently described lymphokine, histamine-releasing factor (Thueson et al., 1979). The accumulation of mononuclear cells is presumed to occur by antigen-induced release of chemoattractant and migration-inhibitory lymphokines. Ward et al. (1971) demonstrated chemoattractant activity for rat lymphocytes from antigen-stimulated guinea pig lymphocytes. Subsequent studies (Cohen et al., 1973; Ward et al., 1977) with extracts of guinea pig skin undergoing delayed-type hypersensitivity reactions and with mitogen-stimulated rat mononuclear cells suggested the presence of two lymphocyte chemoattractant factors. Lymphokines with lymphocyte migration-inhibitory activity have been described from antigen-stimulated murine lymphocytes (Kowalczyk et al., 1981; Zembala et al., 1981) and avian T and B cells (Subba Rao and Glick, 1977; Martin and Glick, 1982). These observations, taken together with the findings that sensitized murine lymphocytes could be chemotactically attracted to

INFLUENCE O F HISTAMINE

245

antigen (Wilkenson et al., 1977),suggested that antigen might attract the first sensitized lymphocytes to a site of inflammation, but the accumulation of subsequent nonsensitized effector T lymphocytes is a result, at least in part, of chemoattractant and migration-inhibitory lymphokines. Several investigations (Center et al., 1983; Berman et al., 1983) have identified and characterized three histamine (lo-* M)-induced human lymphokines which influence lymphocyte motility. Utilizing a modified Boyden chamber assay to assess lymphocyte migration, Center et al. (1983) identified chemoattractant activity for human blood and rat splenic T lymphocytes in histamine-induced human mononuclear cell supernatants. No neutrophil or monocyte chemoattractant activity was present. Sephadex G-100 molecular sieve chromatography of histamine-induced supernatants showed that the lymphotactic activity was eluted at a region corresponding to an apparent molecular weight corresponding to 56,000 (Table 111).This activity was cationic as determined by its elution pattern from a Sephadex QAE anion exchange matrix with a single isoelectric point of 9.0 determined by isoelectric focusing in sucrose. Its biologic activity was predominantly chemokinetic in nature, stable to heating at 56°C for 30 minutes, but sensitive to the effects of neuraminidase and trypsin. This lymphocyte chemoattractant factor (LCF) did not possess suppressor activity attributable to HSF. The elaboration of histamine-induced LCF was blocked by the addition of cimetidine to the culture, but not by coincubation with diphenhydramine, suggesting that LCF was the product of Hz receptor-bearing lymphocytes. In the course of purifying the LCF from histamine-stimulated human mononuclear cells, Sephadex G-100 molecular sieve chromatography also demonstrated the presence of two simultaneously generated noncytotoxic inhibitors of T lymphocyte migration. These two lymphocyte migration-inhibitory factors (LyMIFs) were distinct and were neutral to cationic proteins with apparent molecular weights of 70,000 to 80,000 (LYMIF~SK)and 30,000 to 40,000 (LyMIF3m). LyMIF3SKwas characterized as a single molecular species glycoprotein with an isoelectric point of 8.5, whose biologic activity was susceptible to neuraminidase treatment and heat inactivation (56°C for 30 minutes) but not destroyed by trypsin (Table 111). LyMIF75K had an isoelectric point of 7.5, and its biologic activity was susceptible to degradation by trypsin but was not sensitive to neuraminidase treatment or heating. These physiochemical characteristics suggested that these lymphokines were distinct from one another and different from

246

DENNIS J. BEER ET AL.

previously described migration-inhibitory lymphokines. Whether or not these lymphocyte migration-inhibitory lymphokines possess suppressor activity attributable to HSF remains to be determined. Sephadex G-100 gel filtration of supernatants generated from mononuclear cells incubated with histamine plus cimetidine resulted in selective production of LyM1F35~and L ~ M I F T ~no K LCF ; was produced, In contrast, cells incubated with histamine and diphenhydrawere detected. mine produced LCF, while no L Y M I For ~ LYMIF~~K ~~ Cells incubated with either histamine antagonist alone did not generate LCF or LyMIF activity. Collectively, these data suggested that ~ K products of HI receptor-bearboth L Y M I F ~and ~ K the L Y M I F ~were ing lymphocytes. Previous studies (Center and Cruikshank, 1982; Cruikshank and Center, 1982; McFadden et al., 1983; Berman et al., 1983) had described Con A-induced human chemoattractant and migration-inhibitory lymphokines which are physiochemically and functionally identical to those induced by histamine. These investigators assessed the ability of a lymphocyte population depleted of histamine receptorbearing cells to generate these lymphokines in response to histamine or Con A. Cells nonadherent to a histamine affinity matrix (histamineEACA-Sepharose) were subsequently stimulated with histamine or Con A and did not produce LCF or either of the LyMIFs, while supernatants generated in parallel from histamine or Con A stimulation of unfractionated or control column (EACA-Sepharose)-passed lymphocytes possessed all three lymphokines (Table IV). Interestingly, cells nonadherent to a histamine affinity matrix produced significant monocyte chemoattractant activity following activation with Con A (but not histamine). These data, taken together with the histamine antagonist-blocking studies, suggested that all LCF production in response to either histamine or the polyclonal lymphocyte activator Con A could be attributed to HS receptor-bearing lymphocytes and that all LyMIF production, regardless of stimulus, could be attributed to HI receptor-bearing lymphocytes.

B. B-LYMPHOCYTE FUNCTION 1 . Suppressor Cell Mechanisms a. Animal Studies. Indirect evidence that histamine may modulate

B cell immunoglobulin production began with the observations of Shearer et al. (1972) concerning the regulation of mouse antibody response by cells expressing histamine receptors. Spleen cells from immunized and nonimmunized mice were chromatographed over ei-

247

INFLUENCE OF HISTAMINE

REPRESENTATION OF

THE

TABLE IV MIGRATORYRESPONSESOF T LYMPHOCYTES TO PARTIALLY

HISTAMINEAND CONCANAVALIN A-INDUCED SUPERNATANTS FROM HUMAN LYMPHOCYTES FRACTIONATED ON HISTAMINE AFFINITY MATRICES

PURIFIED

Lyniphokine activities Stimulant Unfractionated M) Histamine ( Con A (10 pg/ml)b H-EACA-Sepharose" Histamine Con Ab EACA-Sepharosed Histamine Con Ah

L C F ~ ~ K " LyMIF&'

+ +

+ +

LYMIF~~K"

+ +

-

-

-

-

-

-

+ +

+ +

+ +

Data are expressed as presence (+) or absence (-) of expected lymphokine activities in fractions of appropriate molecular weights. Crude Con A-induced (but not histamine-induced) supernatants from unfractionated lymphocytes and from lymphocytes nonadherent to either a control or histamine affinity column all possessed significant monocyte chemoattractant activity. Histamine affinity matrix: histamine-e-aminocaproicacid-Sepharose. d Control column: e-aminocaproic acid-Sepharose.

ther H-RSA-S or RSA-S columns. The immune response potentials of 5 x lo6 cells excluded from the two columns were compared with each other, and with an equal number of unfractionated cells. The cell suspensions were mixed with sheep erythrocytes (SRBC)and injected into irradiated, syngeneic recipients (adoptive cell transfer). The direct and indirect anti-sheep erythrocyte plaque-forming cell (PFC) responses generated by the cell population passed over the H-RSA-S column were significantly greater than the responses resulting from the inocula of unfiltered cells or cells passed over control columns. These results suggested the existence of a cell population expressing surface receptors for histamine, which functioned to regulate antibody responses. The fact that H-RSA-S fractionation of spleen cells from nonimmunized donors also generated elevated PFC responses indicated that histamine receptor-bearing cells possessing regulatory function were present in the splenic compartment before active immunization. A subsequent study employing the same animal model of adoptive cell transfer (Shearer et al., 1974) substantiated these initial observations and, in addition, compared the kinetics of anti-sheep erythrocyte

248

DENNIS J. BEER ET AL.

direct and indirect PFC responses in recipients injected with unfiltered spleen cells, to the kinetics in host mice injected with cells filtered over H-RSA-S columns. Although the kinetics of the responses were similar, five times more PFC were generated by the HRSA-S-filtered spleen cells than by the unfiltered cells at the peak of the IgM response. Filtration of cells over H-RSA-S columns also resulted in an increase in the total number of nucleated cells detected in the repopulated host spleens. Similar enhancement of the humoral immune response potential of' mouse spleen cells was also obtained when H-RSA-A column-fractionated thymocytes were injected together with unfractionated bone marrow cells. In contrast, no enhancing effect was observed when fractionated marrow cells were mixed with unseparated thymocytes. Taken together, these data are compatible with the hypothesis that the H-RSA-S column filtration alters humoral immunity by removal of a T cell expressing histamine receptors which functions in a regulatory manner by direct cell-to-cell contact and/or by release of some soluble cytokine with suppressive properties. Further evidence for a histamine receptor-bearing T suppressor cell regulating humoral responses in mice was provided by an analysis of specific antigen-induced immune incompetence of murine splenocytes (Segal et al., 1981). Antigen-specific immunosuppression to SRBC in BALB/c mice was induced by repetitive injection of high doses of soluble components derived from hypotonic lysates of SRBC. This form of unresponsiveness induced both IgM- and IgG-specific suppression. Suppression of the T cell-dependent IgG response, but not of the IgM response, was stable and transferable in an adoptive cell transfer system. The transference of IgG unresponsiveness could be terminated by the removal of a cell population that adhered to a HRSA-S column. The derepression obtained by the depletion of the cells bound by H-RSA-S was not only found in the spleen but also was reflected peripherally, as illustrated by the higher IgG-hemagglutinating titers in the serum of the normal irradiated grafted recipients. Reconstitution of nonadherent splenocytes fractionated over HRSA-S columns with adherent cells eluted with cold buffer resulted in immunosuppression when injected into irradiated syngeneic mice. The immunosuppressive capacity of cells adherent to a H-RSA-S column was nullified by anti-8 antibody-induced cell lysis, suggesting that the suppressor cell was a T cell. The identity of the suppressor cell as a T lymphocyte was further strengthened by the demonstration that these cells exhibited phenotypic markers characteristic of T but

INFLUENCE OF HISTAMINE

249

not B lymphocytes. The generation of these suppressor T cells seemed to depend on exposure to antigen. This was illustrated by the ability of the H-RSA-S-adherent cells obtained from animals exposed to high doses of the antigen to suppress the anti-SRBC response when cocultured with cells obtained from normal mice. In contrast, cells from normal mice which were identically processed were unable to suppress such a response. Finally, the target for the suppressor cells appeared to be helper T cells, rather than B cells, as evidenced by the observation that the suppressor cells had no influence on the IgM response but only on the T-dependent IgG response. The enhancement of antibody elaboration obtained when murine splenocytes were chromatographed over a histamine affinity matrix before being mixed with antigen and transferred to syngeneic recipients differed in two respects from the findings reported for histamine column fractionation of potential cytotoxic responder cells (Shearer et al., 1977). First, the nonadherent cells were the most reactive for antisheep erythrocyte plaque responses, whereas the cytotoxic responses of nonadherent cells were the least reactive. Second, evidence that an adherent population of regulator cells that regulated antibody production was obtained (Bourne et al., 1974), whereas no indication of such a regulator cell was observed in any fraction of cells that participated in the induction of cell-mediated cytotoxicity. Collectively, these findings suggest, but in no way prove, that the precursors of the cytotoxic cells that were retained by the H-RSA-S columns may serve as regulators for humoral antibody production. It must be pointed out, however, that the model for antibody production utilized in vivo antigen (SRBC) challenge to induce an immune response, whereas the model of cytotoxicity utilized in vitro alloantigen sensitization to induce an immune response. In order to compare effects of separate cell populations on humoral and cellular parameters, a single antigen should be employed and a common in vitro or in vivo model should be adopted for generating the response. b. Human Studies. The modulatory role of histamine was investigated in an in vitro model of immunoglobulin production by human blood mononuclear cells (Lima and Rocklin, 1981). The addition of histamine ( 10-6-10-4 M ) in vitro had little effect on the spontaneous production of IgG (measured on day 7 by double-antibody radioimmunoassay) but significantly suppressed pokeweed mitogen-induced IgG production in a dose-dependent fashion. That the inhibitory effects of histamine on immunoglobulin production was not merely a toxic effect was suggested by kinetic experiments in which it was

250

DENNIS J. BEER ET AL.

found that histamine-induced inhibition was seen if the amine was added within 72 hours after the addition of the mitogen. Thereafter, adding histamine to the cultures had little effect on IgG production. An autologous coculture technique permitted the partial characterization of a histamine-responsive suppressor cell operative in the system just described (Lima and Rocklin, 1981). This suppressor cell was found to be an E rosette-positive cell (T cell) that was sensitive to irradiation (2000 rad). In addition, the histamine-activatable suppressor cell possessed a surface antigen identified by a monoclonal antibody (OKT8) directed against a T cell subset (Fig. 6). The nature of the functional histamine receptor present on the suppressor cell responsible for inhibiting IgG production was determined by exposing T-enriched cells to selective HI and H2 agonists and antagonists for 24 hours and then coculturing them with mitogenstimulated B cells (Lima and Rocklin, 1981). The HI antagonist, diM ) where it may have agophenhydramine, at a concentration ( nistic effects, inhibited IgG production somewhat more effectively than histamine. The combination of diphenhydramine ( M ) and

0

NONE PWM PWM + IO*M HISTAMINE P W M +5ppConA

BUFFER

CONTROLSERUM

OKT4-

OKTB-

Cell Treatment

FIG.6. Treatment of histamine- and Con A-induced suppressor cells with monoclonal antibodies directed against T cell subsets. T-Enriched cells were incubated in buffer alone, buffer + complement, OKT4 antiserum + complement, or OKT8 antiserum + complement, washed, and then incubated with or without histamine or Con A for 24 hours. After incubation, the remaining cells (OKT4- or OKT8-) were cocultured with autologous E rosette-negative cells that were then stimulated by pokeweed mitogen (PWM).

INFLUENCE OF HISTAMINE

25 1

histamine M ) resulted in the same degree of suppression as diphenhydramine alone, suggesting that there was a limit to the amount of suppression that could be achieved. In contrast, the Hz antagonist, cimetidine ( M ) , had no effect itself on in vitro IgG production, and the combination of equimolar concentrations ( lop4 M ) of histamine and cimetidine resulted in the same degree of suppression as histamine alone. Taken together, these data suggested that T cells bearing functional HI receptors may suppress in vitro IgG production by B cells. Finally, preincubation of cells with 2-pyridylethylamine and dimaprit, specific H I and Hz agonists, respectively, resulted in an equivalent suppression of mitogen-induced IgG production. This implied that activation of H2 receptor-bearing T cells may also suppress IgG production. A study by Hkbert et al. (1981) measured the effects of histamineinduced suppressor activity on pokeweed mitogen-generated in vitro IgG synthesis as measured by a modified reverse hemolytic plaque (PFC) assay. Addition of autologous or allogeneic histamine-induced suppressor T cells, if added in appropriate numbers at the end of a 7 day culture, caused a significant suppression of'the IgG response. In contrast, addition of suppressor cells at the beginning of the culture did not suppress the synthesis of IgG. These observations suggest that the histamine-induced suppressor cells act not by blocking the proliferation of B cells or B cell precursors but rather by inhibiting the synthesis or release of IgG by mature B cells. The latter results demonstrating that histamine-induced suppressor activity affects only the late phase of antibody production appears contradictory to the study of Lima and Rocklin (1981), which showed an effect of histamine-induced suppressor activity in the early phases of antibody production. These discrepancies may be related more to differences in experimental design than to actual disparate data. A subsequent study (Garovoy et al., 1983),to be described, further clarifies the kinetics of histamine-induced inhibition of humoral responses. Employing a model of polyclonal B cell activation induced during a primary mixed lymphocyte culture (MLC), the effect of a T cellderived histamine-induced suppressor activity (HISA) on the humoral immune response was examined (Garovoy et al., 1983).The number of plaque-forming cells generated during MLC was measured by a protein A plaque assay. HISA was produced by incubating lymphoM histamine. The addition of cytes from normal subjects with HISA on day 0 to MLC-induced plaques reduced the mean number of IgG, IgM, and IgA PFC by 60 to 80%. Unconcentrated HISA superna-

252

DENNIS J. BEER ET AL.

tants were active at a titer of 1/1000 and suppressed IgG, IgM, and IgA isotypes equally. The mechanism of suppressive action of histamineinduced supernatants was investigated by employing a coculture technique in which T cells were initially activated to become helper T cells by allogeneic lymphocytes in a unidirectional MLC (phase I) and then cocultured with autologous unprimed B lymphocytes (phase 11). If HISA was present only during the helper T cell generation phase or only at the time of coculture, the subsequent total plaqueforming B cell response was inhibited by up to 80% (Fig. 7). To study further the effect of HISA on helper T cell activation, the expression of Ia and DR antigens on activated T cells was monitored. I d D R antigens are normally detected on 50 to 60% of activated T cells generated by day 6 during an MLC. HISA reproducibly reduced by half the presence of detectable IdDR antigens. Whether the HISA inhibition of I d D R antigen expression in MLC-activated T lymphocytes was restricted in its action to the OKT4+, Ia+ T helper cell, or whether other T cell subsets might also be affected, was not determined. One mechanism by which activated T lymphocytes promote a B cell response is through the generation of specific and nonspecific helper factors (Armending and Katz, 1974; Friedman et al., 1980).The effect of HISA on B cell-helper factor interaction was investigated by cul-

0 EXP I E4 EXP2 'b t

'a CL

loo-

-

~

E

X

P

~

80-

-

6040-

Bg bp

-

20-

0-

-

INFLUENCE OF HISTAMINE

253

turing the supernatant generated after a 48-hour MLC with allogeneic B lymphocytes (Garovoy et al., 1983). Increasing concentrations of HISA led to significant inhibition of the response to preformed helper factors. Therefore, HISA appears to interfere directly with B cell function as well. Interestingly, to achieve a 50-70% reduction in B lymphocyte response to MLC-generated helper factor required at least half-strength HISA. A similar reduction in primary MLC-generated PFC responses was achieved with significantly lower concentrations (1/100 dilution) of HISA. These data suggest that in the sequence required for antibody production, the phase of T cell activation (phase I) may be more sensitive to the inhibitory effects of HISA than the subsequent phase of B cell activation (phase 11). In summary, the results just presented suggest that a soluble histamine-induced lymphocyte product inhibits MLC-induced polyclonal B cell activation by interfering with the generation and effector function of T helper cells as well as the B cell response to preformed helper factors. In addition, HISA inhibits the expression of Ia/DR antigens on MLC-activated T cells. The former, together with the latter observations, may serve to limit T-B cooperation and help explain the inhibitory effect of HISA on immunoglobulin production.

2. Non-Suppressor Cell Mechanisms Although the studies thus far described in this section suggest a role for histamine receptor-bearing suppressor cells in the modulation of a humoral immune response, one cannot exclude other mechanisms of inhibition of antibody synthesis by histamine. Melmon et (11. (1974a) analyzed the correlation between humoral stimulation of intracellular cyclic AMP accumulation and hormonal inhibition of plaque formation by niurine splenic leukocytes. In experiments with three mouse strains, histamine, P-adrenergic catecholamines, and prostaglandins of the E series were all found to inhibit plaque formation in a dose-dependent fashion that correlated with the activity of the same agents to stimulate accumulation of cyclic AMP in splenic leukocytes. Dibutyryl cyclic AMP also inhibited plaque formation in all three strains. Theophylline, an inhibitor of phosphodiesterase, an enzyme responsible for the degradation of cyclic AMP, potentiated the effects of histamine, isoproterenol, and the prostaglandins in both intracellular cyclic AMP accumulation and plaque formation. Cholera enterotoxin, a potent activator of adenylate cyclase in many tissues (Pierce et ul., 1971), caused a delayed increase in leukocyte cyclic AMP and a delayed inhibition of plaque formation. Both responses were blocked by a specific antitoxin to cholera. An

254

DENNIS J. BEER ET AL.

attempt to block specifically the effect of histamine in BALB/BL splenocytes was unsuccessful, because each of four antihistamines with HI or Hz receptor specificity (pyribenzamine, pyrilamine, burimamide, and antazolidine) significantly inhibited plaque formation when used alone but did not alter the cyclic AMP content of the splenocytes tested. Therefore, the specific histamine receptor (HI vs H2)involved in the foregoing sequence could not be determined, and, in addition, inhibition of plaque formation could occur by mechanisms unrelated to intracellular cyclic AMP accumulation. The data pr,esented in these elegant experiments did not distinguish between the possible mechanisms of inhibition of plaque formation. The latter could involve inhibition of release or secretion of antibody, or changes in intracellular rates of production or degradation of antibody. Employing the same in vitro model of antibody release, namely the formation of hemolytic plaques by splenic leukocytes from mice immunized with SRBC, these same researchers (Melmon et al., 1974b) provided stronger evidence that antibody-producing cells themselves bear functional histamine receptors. Their results indicated that a substantial portion of the splenic cells that produce antibodies to SRBC can be subtracted from the total spleen leukocyte population by chromatography over H-RSA-S columns. It could not be determined whether all cells that produce antibody had receptors for histamine, nor was it determined whether the ability to separate antibody-producing cells on the basis of their histamine receptors was dependent on the antigen utilized for immunization. These studies, in conjunction with the prior experiments employing the adoptive cell transfer model, suggest that receptors for histamine do not develop on or are not expressed on the precursors of antibodyforming cells. In the adoptive cell transfer experiments, spleen cells from either primed or unprimed donor mice passed over a H-RSA-S column produced three to four times as many recipient PFC responses as did transfer of equal numbers of unfractionated cells or cells excluded from a RSA-S column. If the precursor cells of antibody formation possessed receptors for the amine, one would have expected them to be retained within columns of insolubilized histamine. Thus, the transfers performed with cells excluded from H-RSA-S columns should have resulted in fewer PFC responses than controls. The fact that the opposite result occurred (i.e., enhancement of the PFC response) suggests, therefore, that receptors for histamine appear some time after the cell has become committed to the production of antibody.

INFLUENCE OF HISTAMINE

255

The in vioo effects of histamine injection on the immune reactivity to trinitrophenylated (TNP) bovine y-globulin (BGG) was studied in LAFl mice using plaque-forming cell (PFC) responses and their avidity distributions (Szewczuk et al., 1981). Splenic anti-TNP PFC responses of mice treated with histamine (5 x lop6 M ) intravenously were significantly reduced in number and restricted in heterogeneity, and were characterized by a preferential loss of high-avidity IgG PFCs. The reduced PFC response in histamine-treated mice was dose and time dependent. The possibility that the observed reduction in PFC responses in histamine-treated mice was mediated by suppressor cell activity was evaluated by use of cell transfer studies. When normal mice were injected with splenocytes from histamine-treated mice, there was no reduction in response of the subsequent PFC following antigen challenge. This suggested that histamine-induced suppressor activity was not the mechanism for the alterations in PFC responses. Interestingly, only histamine-treated mice produced a significantly high percentage of anti-idiotype-inhibitable, hapten-augmentable IgG PFCs, suggesting the presence of auto-anti-idiotypic activity. Furthermore, immune sera taken from histamine-treated mice caused an inhibition of anti-TNP PFC in tiitro by spleen cells from TNP-BGG immune mice. This PFC-inhibiting factor in immune sera of histamine-treated mice was an antibody of the IgG1 and IgG2, class, lacked anti-TNP antibody activity, but reacted with anti-TNP antibody of LAFl origin. Passive hemagglutination studies on these sera demonstrated the presence of anti-[anti-TNP F(ab’)2-IgG]antibodies. Taken collectively, these data suggested that histamine, in combination with antigen, induced an auto-anti-idiotypic antibody which, in turn, deregulated the anti-TNP PFC response to TNP-BGG in tiitio. A study by Badger et al. (1982) suggested that histamine inhibition of antibody responses in vivo was not caused by induction of suppressor cells but rather involved interference with the induction phase of the response to T-dependent antigens through the release of glucocorticoids. They reported that high doses of histamine injected into BALB/c mice 24 hours after sensitization with Con A resulted in almost complete abrogation of antibody synthesis to sheep red blood cells (SRBC) injected 2 hours later. This phenomenon occurred with nonimmunosuppressive doses of Con A and was strain specific (but not related to specific H-2 background). It did not occur in response to the T-independent antigen polyvinylpyrolidone (PVP) or if histamine was administered after the antigen. The antibody response to PVP does not require the presence of helper T cells or macrophages (An-

256

DENNIS J. BEER ET AL.

derson and Blomgren, 1971; Wong and Herscewitz, 1979). The lack of effect of histamine on antibody formation in response to PVP therefore suggested that this effect was not due to activation of T suppressor cells. I n addition, if suppressor cells were being generated by histamine, they should be evident in the experiment where histamine was injected after the SRBC. Further evidence for a lack of involvement by suppressor cells was provided by cocultivation of treated spleen cells with normal syngeneic cells in vitro. The latter failed to demonstrate a difference between the number of suppressor cells generated in the Con A- and histamine-treated animals and that by the Con A-treated controls. Studies to determine the histamine receptor interaction involved in the foregoing phenomenon supported the conclusion that this was an H I receptor-mediated interaction, since 2-methylhistamine7an H I receptor agonist, mimicked the effect of histamine, whereas dimaprit, an Hz receptor agonist, had no effect. Since histamine stimulation of HI receptors results in the release of adrenocorticotrophic hormone (ACTH), Badger et al. examined the effects of ACTH and corticosterone, the predominant glucocorticosteroid synthesized in mice (Spackman and Riley, 1979). In this system, they found that both drugs could mimic the effects of histamine. In addition, they were successful in reversing the histamine-induced phenomenon by pretreatment of the mice with op’-DDD, a drug that ablates adrenocortical cells. Although the exact target cell affected by the released steroid hormone was not determined, these authors hypothesized that glucocorticosteroid may interfere with the antibody response to SRBC either by a direct cytotoxic effect on helper T cells or by inhibition of helper factor production. Alternatively, the depressed antibody response could have resulted from interference with an antigen-processing step by macrophages or the release of a monokine required as a signal for the T cell release of T cell growth factor. In order to defined precisely the target cell involved in the Con A-corticosterone inhibition of antibody synthesis, an in vitro model of this in vivo phenomenon will have to be established. OF LYMPHOCYTE RECEPTORS C. EXPRESSION

1. Animal Studies Singh and Owen (1976) investigated the maturation of fetal mouse thymic cells by studying the effects of various agents known to increase intracellular cyclic AMP levels on the expression of T cell alloantigens. Histamine (2 X M ) was found to augment the ex-

INFLUENCE OF HISTAMINE

257

pression of Thy-1 antigen on cells as detected by dye exclusion cytotoxicity tests. This effect was apparently mediated through an Hz receptor, since it was inhibited by metiamide (H2antagonist), but not by mepyramine maleate (HI antagonist). The observed increase in the proportion of thymic Thy-l-positive cells may have reflected either initiation of de novo synthesis of the T cell alloantigen or an increase in the basal rate of synthesis of this antigen.

2. Human T Cells In addition to possessing cell surface determinants which identify T cells as being unique within the lymphoid system, these cells also express a variety of membrane receptors. The ability to bind to sheep erythrocytes (E) is often used as a means to distinguish human T lymphocytes from other imniunocompetent cells. Rosette formation results because these cells bind to certain monosaccharides present on the SRBC membrane. While some cyclic AMP-active agents such as P-adrenergic catecholamines have been shown to inhibit T lymphocyte-E rosetting (Chisari and Edington, 1974), the reported effects of histamine have been inconsistent (Verhaegan et al., 1977; DeCock et al., 1977, 1978; Scheinberg et al., 1978; Ambanelli et al., 1979; Hall et al., 1982). On the one hand, Verhaegen et al. (1977) found that histaM ) had no effect on the capacity of T lymphomine (3 x 10-7-3 x cytes to form E rosettes in healthy subjects, but significantly inhibited E rosette formation in atopic individuals. Levamisole, an agent capable of decreasing intracellular cyclic AMP levels (Hadden et al., 1975), restored the histamine-mediated suppression of T lymphocyte-E rosetting activity of patients with allergies. The inhibition of E rosette formation by histamine could be abrogated by pretreatment of the cells with both an HI and Hz receptor blocker. In contradistinction to the foregoing findings, Hall et al. (1982) demonstrated that histamine, in a dose-dependent fashion (5 x 10-'-5 x M ) , significantly and reproducibly inhibited E rosette formation of T cells from normal subjects, but had little effect on E rosette formation in atopic subjects. If SRBC underwent neuraniinidase treatment prior to incubation with lymphocytes, no inhibition of E rosette formation was seen in the normal subjects. In this study, high concentration (5 x lo-' M ) of histamine, 2-pyridylethylamine (HI agonist), and 4-methylhistamine (Hz agonist) were shown to inhibit E rosette formation of lymphocytes from normal individuals. Interestingly, the H1 agonist but not the Hz agonist produced similar results in atopic and normal subjects. There is no simple explanation for the discrepancies between this and the previous study.

258

DENNIS J. BEER ET AL.

Incubation of human blood T lymphocytes at 37°C with the H2 agonist impromidine ( 10-7-10-4 M ) has been shown to increase the percentage of cells bearing receptors for the Fc portion of IgG (RFcy) (Birch and Polmar, 1981).Maximal RFcy enhancement was observed M impromidine. The effect of impromidine after 60 minutes with on RFcy expression appeared to be mediated by an H2 receptor. Cimetidine blocked the impromidine-induced increase in the number of RFcy-positive cells, and 2-pyridylethylamine (HI agonist) decreased the number of RFcy-bearing lymphocytes. The impromidine induction of RFcy expression was inhibited by puromycin, azide, or incubation at 4"C,indicating that this was an active metabolic process requiring oxidative phosphorylation and protein synthesis. In this study, T cells were further fractionated according to the theophylline sensitivity of their E rosette receptors (Limatibul et al., 1978). Impromidine increased RFcy-bearing cells in the fractions containing T lymphocytes with theophylline-resistant (TR) E rosette receptors and not those with theophylline-sensitive (Ts ) receptors. These data suggest that there is a distinct T lymphocyte subpopulation bearing a theophylline-resistant E rosette receptor, as well as an H2 receptor, whose activation induces the expression of Fcy receptors. In the foregoing study (Birch and Polmar, 1981),it was also denionstrated that TR cells (enriched for RFcp cells), induced to express RFcy by impromidine, lost their ability to function as helper cells as evidenced by their failure to facilitate B cell differentiation. This study did not determine whether Fcy receptors were induced in the TR population bearing Fcp receptors or on TR cells displaying no Fc receptors initially. In subsequent experiments, these investigators examined the relationship between histamine modulation of RFcy expression on T lymphocyte subsets and modification of their immunoregulatory function (Birch and Polmar, 1982). Impromidine treatment of TR cells not only resulted in the loss of their ability to facilitate B lymphocyte differentiation, but also induced the generation of suppressor activity. I; was also shown that impromidine caused a reduction in the number of OKT4+ cells and a doubling of the number of OKT8+ cells, as well as increasing the proportion of TR cells expressing readily detectable surface &-microglobulin. Thus, impromidine appeared to modify both the antigenic and functional phenotype of these subsets of T lymphocytes (Birch et d., 1982). The suppressor cells that were induced in the TR cell fraction were distinct from the suppressor cells enriched for RFcy in the Ts fraction. Ts suppressor cells were characterized by the presence of theophyl-

INFLUENCE OF HISTAMINE

259

TABLE V EFFECTOF IMPROMIDINE(Hi AGONIST) ON THE PHENOTYPIC AND FUNCTIONAL OF T CELLSUBSETS EXPRESSIONS 1. Increased the numbers of FcRy-bearing cells and decreased the proportion of FcRpbearing cells

2. Increased the number of OKT8' cells and decreased the number of OKT4' cells 3. Decreased helper function within the FcRp population, which was also associated with increased suppressor activity within that population 4. Increased expression of FcRy-bearing cells within the theophylline-resistent E rosette receptor-positive population but no change in FcRy-bearing cells within the theophylline-sensitive E rosette receptor-positive population 5. Induced the development of radiosensitive (1000 rad) suppressor cells having spontaneous activity within the population characterized by the presence of theophylline-sensitive E rosette receptors, unstable FcRy, and which required DNA synthesis to express their function 6. Induced the development of radioresistant (1000 rad) suppressor cells having spontaneous activity within the population characterized by the presence of theophylline-resistant E rosette receptors, stable FcRy, and which did not require DNA synthesis to express their function

line-sensitive E rosette receptors, relatively unstable RFcy (labile in culture at 37°C for 24 hours), and having spontaneous suppressor activity (Table V). They also required DNA synthesis to express their suppressor activity and were radiosensitive (1000 rad). In contrast, T R impromidine-induced suppressor cells had theophylline-stable E rosette receptors and RFcy, did not require DNA synthesis to express their function, and were radioresistant. Moreover, treatment of Ts cells with an HI agonist produced both a loss of RFcy and a loss of suppressor activity. Whether the Fcy receptors were an integral part of suppressor cell function or only a cell membrane marker which was associated with suppressor activity was not determined. It is possible, although no proven, that impromidine-induced T R suppressor cells may be similar to the cells which produce HSF (discussed earlier).

FUNCTION D. MONOCYTE/MACROPHAGE 1 . Zmmune and Znflammutory Processes The appearance of functional histamine receptors on phagocytic mononuclear cells is not well documented, although a few direct effects of histamine on monocytelmacrophage function have been observed (Table VI). Unlike other cyclic AMP-active agents, histamine does not modulate the action of the preformed lymphokine, macrophage migration-inhibitory factor (MIF), on guinea pig macrophages (Rocklin, 1976; Rocklin et al., 1978). Moreover, histamine does not

260

DENNIS J . BEER ET AL.

TABLE VI HISTAMINE MODULATION OF MONOCYTE/MACROPHACE FUNCTION

Cell type

Assay

Macrophage (guinea pig)

Response to MIF Antigen binding Antigen presentation Lymphocyte-M$ rosette formation Cyclic AMP Respiratory burst Expression of C3b receptors Enhanced chemotaxis Production of C2 Secretion of C2, C4, C3, factor B, and pIH-globulin Synthesis of C5

Monocyte (human)

Macrophage (mouse)

Effect of histamine

Receptor specificity

HI

H2

+

-

.1

-

+

.1 .1

?

?

0 0 0 0 0

t

0 0

interfere with antigen binding to macrophages or the presentation of macrophage-bound antigen to lymphocytes (Rocklin et al., 1978),does not inhibit lymphocyte-macrophage interactions (rosetting) that lead to T cell activation (Rocklin et d.,1978), does not augment the appearance of monocyte C3b receptors (Kay, 1979), and is not directly chemotactic for monocytes, nor does it interfere with the chemotactic response of normal monocytes to C5a (Lett-Brown and Leonard, 1977). Histamine does not activate macrophage adenylate cyclase (Remold-O’Connell, and Remold, 1974). The fact that histamine does not activate macrophages or monocytes and does not stimulate increased levels of cyclic AMP in macrophages as well as not directly affecting a number of macrophage functions argues strongly against the presence of histamine receptors on macrophages. Alternatively, if such receptors are present, they may not be functionally coupled to adenylate cyclase. Certain functional responses of monocytes/macrophages to histamine have been reported. Histamine coupled (via rabbit serum albumin) to zymosan particles has been reported to induce a respiratory burst in guinea pig alveolar macrophages; this effect was blocked by H1 receptor antagonists (Diaz e t al., 1979a,b). Because there is some controversy concerning the specificity of receptor binding to histamine-protein conjugates, the significance of the latter observation is uncertain.

INFLUENCE OF HISTAMINE

261

2. Complement Studies Histamine has also been shown to influence the production of comM ) proplement proteins by human monocytes. Histamine (lO-‘-lO-” duced a rapid and irreversible dose-dependent noncytotoxic inhibition in the production of the second complement component (C2) by human monocytes (Lappin and Whaley, 1980). The specificity of this response was shown by the failure of histidine or histamine metabolites to inhibit C2 production. This effect was Hz receptor dependent as evidenced by the fact that cimetidine (H, receptor antagonist) prevented the histamine-induced response, whereas chlorpheniramine (Hl receptor antagonist) had no effect. Dimaprit and 4-methylhistamine, Hz receptor agonists, simulated the effect of histamine, whereas the H1 receptor agonist 2-(2-aminoethylthiazole) was ineffective. Thus, the effect of histamine on C2 production by monocytes was mediated through the H2 receptor. In further studies (Lappin et al., 1980), histamine ( 10-6-10-4 M ) was shown to inhibit the secretion of newly synthesized C2, C4, C3, factor B, and PIH-globulin by human monocytes in culture. However, these investigators detected increased amounts of intracellular, radiolabeled, acid-precipitable protein that was antibody specific for the complement components, C2, C4, C3, factor B, and PIH-globulin, within monocytes cultured with histamine. They concluded on the basis of their findings that histamine treatment of monocytes resulted in impaired secretion of complement proteins, and that their increased intracellular accumulation might exert a negative feedback effect on further protein synthesis. Employing mouse peritoneal macrophages, Ooi (1982) has shown that histamine (10-6-10-3 M ) stimulation of the Hz receptor exerted an inhibitory effect on the synthesis of C5 measured by decreased levels of hemolytic and antigenic C5 protein synthesized both intracellularly and extracellularly. There was a direct demonstration that the amount of precursor (pro-C5) protein synthesized was diminished when macrophages were incubated with histamine, and on the basis of the parallelism between the amount of suppressive activity exerted by histamine on antigenic (pro-C5) and on functional C5 protein synthesized, it appeared that the major mechanism by which histamine exerts its negative influence on C5 synthesis was by a reduction in the amount of precursor protein synthesized. The finding of decreased intracellular pro-C5 protein during incubation with histamine contrasts with the findings of Lappin et al. (1980), who detected increased amounts of intracellular precipitable complement proteins, whose secretion was impaired by histamine. The reasons for the discrepancies between

262

DENNIS J. BEER ET AL.

these two laboratories may relate to differences in cell type, animal species, and type of complement protein studied. To date, there is no information on the regulation of protein synthesis by histamine that assists in clarifying the role of this substance in modulating complement biosynthesis by the monocyte/macrophage. The ability of histamine to modulate complement production may result in negative feedback regulation and prevent further cleavage of C3 and C5 both by limiting production of C5 and by restricting the formation of C3 (C42) and C5 (C423b) convertases. The latter effects of histamine may have important implications for complement-mediated phenomena occurring during IgE and antigen-antibody complex reactions. IV. Conclusion

The foregoing review has provided much evidence in support of the concept that autacoids such as histamine must be placed into a broader perspective involving both the initiation and modulation of inflammatory and immune processes. The fact that histamine can influence the immune process at different stages, and that it can influence different subpopulations of cells at concentrations which probably exist in vivo during physiologic and pathologic events, indicates that it can and should be seriously considered as a significant modulator of inflammatory and immune processes. The interaction of antigen with specifically sensitized mast cells or basophils may provide an initial source of histamine. Antigen stimulation of sensitized T effector cells will result in the elaboration of lymphokines, among which is a histamine-releasing factor, which may augment the amount of histamine provided during an immune reaction. Further, the histamine-releasing factor may provide a source of histamine in the absence of IgE-mediated responses. Once histamine is made available, it is capable of interacting in the local milieu with lymphocytes/macrophages or other cell types present at the site of the reaction and modulating immune and inflammatory events. Some of its pro-inflammatory effects include stimulating T effector cells to produce chemoattractant and migration-inhibitory lymphokines, thus recruiting other immunocompetent lymphocytes to the reaction site and retaining them there. A histamine gradient will also result in the attraction of eosinophils to the reaction site and enhance the expression of C3b receptors on their surface. The latter will facilitate the destruction of parasites. Histamine may also activate contrasuppressor cells (in addition to suppressor cells), and, depending on the ratio of contrasuppressor : suppressor cells present in the local milieu, the immuno-

INFLUENCE OF HISTAMINE

263

logic response will be augmented (reduced suppressor activity) or reduced (enhanced suppressor activity). Some of the anti-inflammatory events mediated by histamine include the activation of suppression cells following their interaction with macrophages and/or their products (IL-1), which leads to the production of HSF. The latter augments the production of prostaglandins by macrophages/monocytes. These arachidonic acid metabolites may have inhibitory effects on T effector cells and reduce their proliferation and lymphokine production, thereby dampening cell-mediated immune reactions. Histamine may modulate the function of cytotoxic T cells and natural killer cells directly, thereby reducing their ability to mediate damage to their target cells. Histamine inhibition of complement synthesis by macrophages may serve to limit the severity of the inflammatory response. Furthermore, histamine may feed back negatively on the mast cell or basophil to reduce further mediator release as well as affecting their migration. Similar changes in the function of eosinophils and neutrophils may be mediated by histamine and may help to determine the nature of the ensuing inflammatory reaction. The production of various specific antibodies as well as a number of isotypes (IgG, IgM, IgA) is regulated to some degree by an effect of histamine directly on B cells or indirectly through inhibition of T helper cell function via a suppressor lymphokine. Much of the work presented in this review has been the result of in vitro studies. Much more experimentation is required under in vivo circumstances to substantiate further the exact role of histamine as a physiologic modifier of immune and inflammatory processes. This work has, however, identified certain areas where there is a potential to modify immune and inflammatory responses by antagonists to autacoids such as histamine.

ACKNOWLEDGMENTS This work was supported by USPHS Grants AI-16362 and HL-00898

REFERENCES Adams, M., and Lichtenstein, L. M. (1979).J. Zmmunol. 122, 555. Ambanbelli, V., Ferraccioli, G. F., Maganelli, P., and Vanona, G. L. (1979).Agents Actions 9, 474. Anderson, B., and Blomgren, H. (1971). Cell. lmmunol. 2, 411. Anderson, R., Glover, A., and Rabson, A. R. (1977).J.Zmmunol. 118, 1690. Anwar, A. R. E., and Kay, A. B. (1977).Nature (London)269,522. Anwar, A. R. E., and Kay, A. B. (1978).J. lmmunol. 121,1245. Anwar, A. R. E., and Kay, A. B. (1980).Clin. E x p . lmmunol. 42, 196. Aoyagi, T., Adachi, K., Halprin, K., Levine, V., and Woodgard, C. (1981).J. Invest. Dermatol. 76, 24.

264

DENNIS J. BEER ET AL.

Armending, D., and Katz, D. H. (1974).J. E x p . Med. 140, 19. Ash, A. S. F., and Schild, H. 0. (1966). Br. J. Pharmacol. 27,427. Askenase, P. W., Bursztajn, S., Gershon, M. D., and Gershon, R. K. (1980).J.E x p . Med. 152, 1358. Badger, A. M., Griswold, D. E., DiMartino, M. J., and Poste, G. (1982).J.Zmmunol. 129, 1017. Ballet, J. J., and Merler, E. (1976). Cell. Zmmunol. 24,250. Beaven, M. A. (1976).N. Engl. J. Med. 294,320. Beer, D. J,, Osband, M. E., McCaffrey, R.P., Soter, N. A., and Rocklin, R. E. (1982a).N . Eng1.J. Med. 306,454. Beer, D. J., Rosenwasser, L. J.. Dinarello, C. A., and Rocklin, R. E. (1982b). Cell. Zmmunol. 69, 101. Beer, D. J., Dinarello, C. A., Rosenwasser, L. J., and Rocklin, R. E. ( 1 9 8 2 ~ J. ) . Clin. Invest. 70, 393. Beets, J. L., and Dale, M. J. (1979). Br. J . Pharmacol. 66,365. Berman, J. S., Cruikshank, W. W., McFadden, R. G., Center, D. M., and Beer, D. J. (1983). Submitted. Birch, R. E., and Polmar, S. H. (1981). Cell. Zmmunol. 57,455. Birch, R. E., and Polmar, S. H. (1982). Clin. E x p . Zmmunol. 48,218. Birch, R. E., Rosenthal, A. K., and Polmar, S. H. (1982). Clin. E x p . Zmmunol. 48, 231. Black, J. W., Duncan, A. M., Durant, C. J., Ganellin, C. R., and Parsons, E. M. (1972). Nature (London)236,385. Bourne, H. R., Melmon, K. L., and Lichtenstein, L. M. (1971). Science 173, 743. Bourne, H. R., Lichtenstein, L. M., Hennery, C. S., Melmon, K. L., Weinstein, Y., and Shearer, G. M. (1974).Science 184, 19. Brostoff, J., Pack, S., and Lydyard, M. L. (1980). Chin. E x p . Zmmunol. 39, 739. Bryant, D. H., and Kay, A. B. (1977). Clin. Allergy 211. Busse, W. W., and Lantis, S. D. H. (1979).J. Invest. Dermatol. 73, 184. Busse, W. W., and Sosman, J. (1976). Science 194, 737. Busse, W. W., and Sosman, J. (1977).J. Clin. Znoest. 59, 1080. Busse, W. W., Cooper, W., Warshauer, D. M., Dick, E. C., Wallow, I. H. L., and Albrecht, R. (1979). Am. Rev. Respir. Dis. 119, 561. Busse, W. W., Cooper, W., and Anderson, C. (1980).Agents Actions 10, 15. Cantor, H., and Gershon, R. K. (1979). Fed. Proc. Fed. Am. SOC. E x p . Biol. 38,2058. Center, D. M., and Cruikshank, W. W. (1982).J. Zmmunol. 128,2563. Center, D. M., Cruikshank, W. W., Berman, J. S., and Beer, D. J. (1983).]. Zmmunol. In press.

Cerottini, J.-C., and Brunner, K. T. (1974). Ado. Zmmunol. 18,67. Chisari, F. V., and Edington, T. S. (1974).J. E x p . Med. 140, 1122. Clark, R. A. F., Gallin, J. I., and Kaplan, A. P. (1975).J. E x p . Med. 142, 1462. Clark, R. A. F., Sandler, J. A., Gallin, J. I., and Kaplan, A. P. (1977).J. Zmmunol. 118, 137. Clark, R. A., Szot, S., Venkatasubramanian, K., and Schiffman, E. (1980).J. Zmmunol. 124,2020. Cohen, S., Ward, P. A., Yoshida, T., and Burek, C. L. (1973). Cell. Zmmunol. 9,363. Cruikshank, W. W., and Center, D. M. (1982).J.Zmmunol. 128,2569. Daly, J. W. (1975). Life Sci. 18, 349. Damle, N. K., and Gupta, S. (1981).J. Clin. Zmmunol. 1,241. DeCock, W., Decree, J., and Verhegen, H. (1977). Znt. Arch. Allergy Appl. Zmmunol. 54, 176.

INFLUENCE OF HISTAMINE

265

DeCock, W., Decree, J., and Verhegen, H. (1978). Clin. Immunol. Immunophathol. 11, 1. Diaz, P., Jones, D. G., and Kay, A. B. (1979,). Nature (London)278,454. Diaz, P., Jones, D. G., and Kay, A. B. (197913).Clin. E x p . Immunol. 35, 462. Engleman, E. G., Goldsby, R. A., Osborne, B. A., and Benike, C. (1979). Clin. Res. 27, 324 (Abstr.). Festen, H. P. M., DePauw, B. E., Smeulders, J., and Wagener, D. J. Th. (1981).Clin. Immunol. Immunopathol. 21,33. Flaxman, B. A., and Harper, R. A. (1975).J . Infiest. Dermatol. 65, 52. Foon, K. A., Wahl, S. M., Oppenheim, J. J., and Rosentreich, D. L. (1976).J.Immunol. 117, 548. Friedman, D. L. (1976). Physiol. Reo. 56, 652. Friedman, D. L., Irogoyen, 0. H., Gay, D., and Chess, L. (198O).J.Immunol. 124,2930. Garovoy, M. R., Reddish, M. A., and Rocklin, R. E. (1983).J. Immunol. 130, 357. Gershon, R. K., Eardley, D. D., Dornm, S., Grun, D. R., Shen, F. W., Yamauchi, K., Cantor, H., and Murphy, D. B. (1981).J.E x p . Med. 153, 1533. Gifford, R. R. M., Sr., and Schmidtke, J. R. (1979). Surg. Forum 30, 113. Gifford, R. R. M.,Sr., Hatfield, S. M., and Schmidtke, J. R. (1980). Transplantation 29, 143. Goodwin, J. S., Messner, R. P., and Williams, R. C., Jr. (1979). Cell. Immunol. 45,303. Hadden, J. W., Hadder, E. M., Haddox, M. K., and Goldberg, N. D. (1972). Proc. Natl. Acad. Sci. U S A . 69, 3024. Hadden, J. W., Coffey, R. G., Hadden, E. M., Lopez-Corrales, E., and Sunshine, G. H. (1975).Cell. Immunol. 20, 98. Hall, T. J., Lydyard, P. M., and Brostoff, J. (1982). Int. Arch. Allergy Appl. Immunol. 69, 380. HCbert, J., Beaudoin, R., Aubin, M., and Fontaine, M. (1980). Cell. Immunol. 54,49. HCbert, J., Beandoin, R., Fontaine, M., and Fradet, G. (1981). Cell. Immunol. 58, 336. Henney, C. S., and Lichtenstein, L. M. (1971).J. Immunol. 107, 610. Henney, C. S., Bourne, H. S., and Lichtenstein, L. M. (1972).J . Immunol. 108, 156. Herman, J. J., and Colten, H. R. (1980).J. Allergy Clin. Immunol. 66, 274. Herman, J. J., Brenner, J. K., and Colten, H. R. (1979).Science 206, 77. Ignarro, L. J., and Columbo, C. (1973). Science 180, 1181. Jorizzo, J. L., Sans, W. M., Jr., Jegasothy, B. V., and Olansky, A. J. (1980).Ann. Intern. Med. 92, 192. Kahlson, G., and Rosengren, E. (1968).Physiol. Reo. 48, 155. Kaliner, M . (1978). Am. Rev. Respir. Dis. 118, 1015. Kaplan, A. P., Horakova, Z., and Katz, S. I. (1978).J. Allergy Clin. Immunol. 61, 350. Kay, A. B. (1979).J . Allergy Clin. Immunol. 64, 90. Kedar, E., and Bonavida, B. (1974).J. Immunol. 113, 1544. Kiessling, R., and Wigzell, H. (1979). Immunol. Reo. 44, 165. Knop, J., Stremmer, R., Neumann, C., DeMaeyer, E., and Macher, E. (1982). Nature (London) 296, 757. Kowalezyk, D., Zenibala, M., and Litwora, E. (1981). Cell. Immunol. 57, 361. Kownatzki, E., Till, G., Gagelman, M., Terwart, A,, and Gemsa, p. (1977). Nature (London) 270,67. Lappin, D., and Whaley, K. (1980). Clin. E x p . Immunol. 41, 497. Lappin, D., Moseley, H. L., and Whaley, K. (1980).Clin. E x p . Immunol. 42, 515. Lett-Brown, M.A., and Leonard, E. J. (1977).J.Zmmunol. 118, 815. Lichtenstein, L. M . (1975).J . Immunol. 114, 1692.

266

DENNIS J. BEER ET AL.

Lichtenstein, L. M., and Gillespie, E. (1973). Nature (London)244, 287. Lichtenstein, L. M., and Gillespie, E. (1975).]. Pharmacol. E x p . Ther. 192,441. Lima, M., and Rocklin, R. E. (1981). Cell. Zmmunol. 64,324. Limatibul, S., Shore, A., Dosch, H. M., and Gelfand, E. W. (1978).Clin. E x p . Zmmunol. 33, 503. MacDonald, H. R., Sordat, B., Cerottini, J.-C., and Brunner, J. (1975).J.E x p . Med. 142, 622. Marone, G., Thomas, L. L., and Lichtenstein, L. M. (1980).J . Zmmunol. 125,2277. Martin, D. E., and Glick, B. (1982). Fed. Proc. Fed. Am. SOC. E x p . Biol. 41,573 (Abstr). Martinez, J. D., Santos, J,, Stechschulte, D. J., and Abdou, N. I. (1979).J.Allergy Clin. Zmmunol. 64, 485. McCluskey, R. T., Benacerraf, B., and McCluskey, J. W. (1963).J. Immunol. 90, 466. McFadden, R. C., Cruikshank, W. W., and Center, D. M. (1983). Cell. Zmmunol. (in press). Melmon, K. L., Bourne, H. R., Weinstein, Y., and Sela, M. (1972). Science 177, 707. Melmon, K. L., Bourne, H. R., Weinstein, Y., Shearer, G. M., Kram, J., and Bauminger, S. (1974a).J. Clin. Invest. 53, 13. Melmon, K. L., Weinstein, Y., Shearer, G. M., and Bourne, H. R. (1974b).J.Clin. Invest. 53, 22. Melmon, K. L., Rocklin, R. E., and Rosenkranz, R. P. (1981). Am. J. Med. 71, 100. Mozes, E., Weinstein, Y., Bourne, H. R., Melmon, K. L., and Shearer, G. M. (1974).Cell. Immunol. 11,57. Ooi, Y. M. (1982).J. Zmmunol. 129, 200. Osband, M., and McCaffrey, R. (1979).J. B i d . Chem. 254,9970. Osband, M. E., Cohen, E. B., McCaffrey, R. P., and Shapiro, H. M. (1980).Blood 56,293. Osband, M. E., Cohen, E. B., Miller, B. R., Shen, Y. J., Cohen, L., Flescher, L., Brown, A. E., and McCaffrey, R. P. (1981). Blood 58, 87. Pick, E. (1974). Immunology 26, 649. Pierce, N. F., Greenough, W. B., 111, and Carpenter, C. C. (1971).Bacteriol. Rev. 35, 1. Plaut, M. (1979).J. Zmmunol. 123, 692. Plaut, M., and Berman, J. (1978).J. Allergy Clin. Immunol. 61, 132 (Abstr.). Plaut, M., and Lichtenstein, L. M. (1982). In “Pharmacology of Histamine Receptors” (C. R. Granellin and M. G. Parsons, eds.), pp. 392. Wright, London. Plaut, M., and Roszkowski, W. (1979). In “Histamine Receptors” (T. 0. Yellin, ed.), pp. 361. Spectrum Publications, Holliswood, New York. Plaut, M., Lichtenstein, L. M., and Henney, C. S. (1973). Nature (London) 244, 284. Plaut, M., Lichtenstein, L. M., and Henney, C. S. (1975).J . Clin. Invest. 55, 856. Plaut, M., Nordin, L., and Thomas, L. L. (1980).Fed. Proc. Fed. Am. Soc. E x p . B i d . 39, 445 (Abstr.). Powell, J. R., and Brody, M. J. (1976).Am. J . Physiol. 231, 1002. Reichman, B. L., Handzel, Z. T., Segal, S., Weinstein, Y.,and Levin, S. (1979). Clin. E x p . Immunol. 37, 562. Remold-O’Connel, E., and Remold, H. (1974).J. B i d . Chem. 249, 3615. Richelson, E. (1978a). Nature (London) 274,176. Richelson, E. (197813). Science 201, 69. Rigal, D., Monier, J. C., and Souweine, G. (1979). Cell. Immunol. 46,360. Rocklin, R. E. (1976).J. Clin. Invest. 57, 1051. Rocklin, R. E. (1977).J. Immunol. 118, 1734. Rocklin, R. E., and Haberek-Davidson, A. (1981).J. Clin. Invest. 1,73. Rocklin, R. E., and Haberek-Davidson, A. (1983). Int. J. Immunopharmacol. In press.

INFLUENCE OF HISTAMINE

267

Rocklin, R. E., Greineder, D., Littman, B., and Melmon, K. L. (1978).Cell. Zmmunol. 37, 162. Rocklin, R. E., Greineder, D. K., and Melmon, K. L. (1979). Cell. Immunol. 44,404. Rocklin, R. E., Breard, J., Gupta, S., Good, R. A., and Melmon, K. L. (1980). Cell. Zmmunol. 51, 226. Rocklin, R. E., Blidy, A., and Kamal, M. (1983a). Cell. Zmmunol. 76, 243. Rocklin, R. E., Kiselis, I., Beer, D. J., Rossi, P., Maggi, F., and Bellanti, J. (1983b). Cell. Immunol. 77, 92. Roder, J. C., and Duwe, A. K. (1979). Nature (London) 278,451. Roder, J . C., and Klein, M. (1979).J. Zmmunol. 123, 2785. Roszkowski, W., Plaut, M., and Lichtenstein, L. M. (1977). Science 195, 883. Ruiz-Arguelles, A., Seroogy, K. G., and Ritts, R. E., Jr. (1982). Cell. Zmmunol. 69, 1. Saxon, A., Morledge, V. D., and Bonavida, B. (1977). Clin. E x p . Zmmunol. 28,394. Scheinberg, M. A., Bandiera, 0. P. O., Cotrin, H. P., and Postes, J. F. (1978). Clin. Res. 26,325 (Abstr.). Schwartz, A., Askenase, P. W., and Gershon, R. K. (1977).J. Zmmunol. 118, 159. Schwartz, A., Askenase, P. W., and Gershon, R. K. (1980).lmmunopharmacology 2,179. Schwartz, A., Sutton, S. L., Askenase, P. W., and Gershon, R. K. (1981). Cell. Zmmunol. 60, 426. Segal, S., Melmon, K., and Weinstein, Y. (1981). Cell. Zmmunol. 59, 171. Seifert, W. E., and Rudland, P. S. (1974). Nature (London) 248, 138. Seligman, B. E., Fletcher, M. P., dntl (hlIin, 1. I . (1983).J . I n i n ~ t ~ n o130, / . 1902. Shearer, G. M., Melmon, K. L., Weinstein, Y., and Sela, M. (1972).J. E x p . Med. 136, 1302. Shearer, G. M., Weinstein, Y., and Melmon, K. L. (1974).J . Zmmunol. 113, 597. Shearer, G. M., Simpson, E., Weinstein, Y., and Melmon, K. L. (1977).J.Zmmunol. 118, 756. Siegel, J. N., Schwartz, A., Askenase, P. W., and Gershon, R. K. (1982).Proc. Natl. Acad. Sci. U.S.A.79, 5052. Singh, V., and Owen, J. J. T. (1976). Eur. J. Pharmacol. 6, 59. Smith, J. A., Mansfield, L. E., deShazo, R. D., and Nelson, H. (1980).J . Allergy Clin. Immunol. 65, 118. Spackman, D. H., and Riley, V. (1979). Science 200, 87. Staszak, C., and Goodwin, J. S. (1980).Cell. Zmmunol. 54, 351. Staszak, C., Goodwin, J. S., Troup, G. M., Pahtak, D. R., and Williams, R. C. (1980).J . Zmmunol. 125, 181. Stewart, J., and Kay, A. B. (1980). Clin. E x p . Zmmunol. 40,423. Strannegard, I.-L., and Strannegard, 0. (1977). Scand. J , Immunol. 6, 1225. Strannegard, I.-L., and Strannegard, 0. (1979). Znt. Arch. Allergy Appl. Zmmunol. 58, 167. Study, R. E., and Greengard, P. (1978).J. Pharmacol. E x p . Ther. 207,767. Subba Rao, D. S. V., and Glick, B. (1977).Ado. E x p . Med. B i d . 88, 87. Suzuki, S., and Huchet, R. (1981). Cell. Zmmunol. 62, 396. Suzuki, S., and Huchet, R. (1982). Cell. Immunol. 68, 349. Szewczuk, M. R., Campbell, R. J., and Smith, J . W. (1981). Cell. Zmmunol. 65, 152. Thomas, Y., Huchet, R., and Grandjon, D. (1981). Cell. Zmmunol. 59,268. Thueson, D. O., Speck, L. S., Lett-Brown, M. A., and Grant, A. (1979).]. Zmmunol. 123, 626. Ting, S., Dunsky, E. H., and Zweiman, B. (1980).J . Allergy Clin. Zmmunol. 65, 196 (Abstr.).

268

DENNIS J. BEER ET AL.

Turnbull, L. W., and Kay, A. B. (1976). lmmunology 31, 797. Verhaegen, H., DeCock, W., and Decree, J. (1977).J . Allergy Clin. lmmunol. 59, 266. Wang, S. R., and Zweiman, B. (1978). Cell. lmmunol. 36,28. Ward, P. A., Offen, C. D., and Montgomery, J. R. (1971). Fed. Proc. Fed.Am. SOC.E x p . Biol. 30, 172. Ward, P. A., Unanue, E. R., Goralnick, S., and Schreiner, G. F. (1977).J.lmmunol. 119, 416. Weinstein, Y., and Melmon, K. L. (1976). lmmunol. Commun. 5, 401. Weinstein, Y., Melmon, K. L., Bourne, H., and Sela, M. (1973).J.Clin. Invest. 52, 1349. Weissmann, N. G., Zurier, R. B., Spieler, P. J., and Goldstein, I. M. (1971).J.E x p . Med. 134, 149s. Wilkenson, P. C., Parrott, D. M. V., Russel, R. J., and Slcss, F. (1977). J . E x p . M e d . 145, 1. Wong, D. M., and Herscewitz, H. B. (1979).lmmunology 37, 765. Yamauchi, K., Green, D. R., Eardley, D. D., and Gershon, R. (1981).J. E x p . Med. 153, 1547. Yaranato, S., Francis, D., and Greaves, M. W. (1976).J . Inoest. Dermatol. 67, 696. Zembala, M., Kowalczyk, D., Asherson, G. L., and Noworolski, R. (1981). Cell. lmmunol. 57, 370. Zurier, R. B., Weissmann, N. G., Hoffstein, S., Kammerman, S., and Tai, H. H. (1974).J. Clin. Invest. 53, 297.

Index A Adoptive immunization, nieaning of assay, 107-108 cytolytic T cells and, 108-115 irradiated tumor-bearing recipients and, 118- 120 local passive transfer assay, 120-122 suppressor T cells and, 115-116 T cells generated i n oitro, 117-118 Adoptive immunity, specificity and suppression of, 97-100 Anti-phosphorylcholine antibodies clonotypes of, 5 diversity in D segment, 25 JII segment, 23-25 V,, segment, 23 evolution of T15 gene family, 30-31 heavy chain variable regions, 7 somatic diversification of, 7-15 J K gene segments and functional diversity, 26 hidden diversity of, 5-6 junctional diversity in V,,, D and J11 segments, 22-23 light chain variable regions, 18-19 somatic diversification of genes, 1920 pattern of variation by somatic hypervariation, 20-22 molecular basis of T15 idiotype, 29 N region of, 25-26 selection of variants, 27-29 somatic diversification of heavy chain extent of, 15-16 mutation and, 16-17 structural diversity in, 31-33 structure of variable regions and molecular basis of diversity, 6 summary of diversity of, 26-27 T15 VII gene family and, 29-30 Anti-phosphorylcholine response, restricted nature of, 4-5

Antitumor immunity, 92-93 by adoptive immunization against established tumors cyclophosphamide and, 100-105 m-irradiation and, 105-107 specificity and suppression of, 97100 tumor-induced T cell-mediated immunosuppression and, 93-97

B Basophils, histamine modulation of, 215-218 B cells, immunoglobulins as antigen receptors on, 42-43 B lymphocytes, histamine modulation of function non-suppressor cell mechanisms, 253256 suppressor cell mechanisms, 246-253

C Concomitant immunity, analysis of decay associated with suppressor T cells, 131-133 kinetics and decay measured by adoptive immunity, 127-131 paradox of passive transfer of innnunity with T cells from a donor with an established tumor, 122127 Cyclic nucleotides, histamine receptormediated changes in, 212-213 Cyclophosphamide, adoptive immunotherapy of tumors and, 100-105

E Eosinophils, histamine modulation of, 218-219

269

270

INDEX

F FcsR affinity for IgE and structure on lymphocytes and macrophages, 62-65 induction and function on lymphocytes and macrophages, 81-85 rosette assay for detection of, 65-67 FcsR-, eosinophils, 80-81 FcsR+ B and T cells in atopic patients, 71-74 B and T cells in nonatopic healthy humans, 69-71 cultured lymphocytes, macrophages and FcER- leukemic lymphocytes,

modulation of immune effector cells B lymphocytes, 246-256 expression of lymphocyte receptors, 256-259 monocyte/macrophage function,

259-262 T lymphocytes, 223-246 modulation of polymorphonuclear inflammatory cells basophils, 215-218 eosinophils, 218-219 neutrophils, 219-222 Histocompatibility antigens, fetal, ontogeny and development of, 170-172

68-69 lymphocytes in normal and parasitically infected rats and mice, 74-

77 monocytes in nonatopic and atopic humans, 77-79 rat and mouse macrophages, 79-80 Fetal-maternal balance, overview, 158-

159 Fetus expression of histocompatibility antigens, ontogeny and development of, 170-172 immunocompetence of, 172-177 proposed mechanisms for maintenance of fetal contributions to prevent allogeneic rejection, 168-169 maternal contribution to maintenance, 159-164 placental contribution, 164-168

H Histamine as an autocoid, 209-210 histamine receptor-mediated changes in cyclic neucleotides,

212-213 methods for detecting histamine receptors on leukocytes, 213-

215 pharmacology of histamine, 210-212

I Immune effector cells, histamine modulation of B lymphocytes, 246-256 expression of lymphocyte receptors,

256-259 monocyte/macrophage function, 259-

262 T lymphocytes, 223-246 Immunoglobulin(s) as antigen receptors on B cells, 42-43 coexpressed p and S mRNAs coded by very complex transcription units,

52-55 developmental regulation of heavy chain gene expression, 55-56 heavy chain genes, membrane gene segments of, 46-47 heavy chain M regions, model for transmembrane insertion of, 47-50 membrane and secreted heavy chain gene, mRNAs coded by complex transcription units, 50-52 membrane and secreted p heavy chains, different mRNAs and, 43-

46 structure, 40-42 Immunosuppression, tumor-induced T cell-mediated, adoptive immunity and, 93-97 y-Irradiation, adoptive immunotherapy and, 105-107

271

INDEX

L Leukocytes, methods for detecting histamine receptors on, 213-215 Lymphocytes affinity for IgE arid structure of FceR on, 62-65 histamine receptors, expression of,

256-259 immunologic basis of interactions between mother and fetus immunocompetence of fetus, 172-

177 immunocompetence of mother, 177-

178 permeability of placenta to lynlphocytes, 179-180 Lymphocyte interaction human maternal and neonatal assay of suppression by cord T lymphocyte subsets, 188-192 characterization of newborn lyniphocyte subset responsible for suppression, 184-188 helper activity of cord blood T lymphocytes, 196-197 mechanisms of cord T lymphocytemediated suppression, 192- 196 regulatory T lymphocytes in newborn arid adults, 197-199 suppressor activity of lymphocytes from human newborns, 180-184

M Macrophages, affinity for IgE and structure of F ~ E R on, 62-65 Monocyte/macrophage, function, histamine modulation of, 259-262 Mother, immunocompetence of, 177-178 Mutation somatic in anti-PC antibody light chains immunoglobulin class and, 22 of heavy chain variable region of anti-PC antibodies mechanism 16-17 VII gene and, 17-18

N Neutrophils, histamine modulation of,

219-222

P Passive transfer assay, local, meaning of,

120-122 Placenta, permeability to lymphocytes,

179-180 Polymorphonuclear inflammatory cells, histamine modulation of basophils, 215-218 eosinophils, 218-219 neutrophils, 2 19-222

R Ribonucleic acids messenger coding of membrane and secreted heavy chain by complex transcription units, 50-52 coexpressed p and 6 coded by very complex transcription units, 52-

55 encoding membrane and secreted p heavy chains, 43-46

S Suppressor activity assay of cord T lymphocyte subsets, characterization of newborn lymphocyte subsets responsible for, 184-

188 of lymphocytes from human newborns,

180-184 niechanisms of cord T lymphocytemediated suppression, 192-196 T

T cells cytolytic, adoptive immunity and,

108-125 generated in vitro, adoptive immunotherapy and, 117-118

272

INDEX

suppressor, inhibition of cytolytic response by, 115-116 T lymphocytes, histamine modulation of function cell-mediated cytotoxicity, 223-228 lymphocyte proliferation, 228-231

lymphokine-producing cells, 231-234 motility, 244-246 suppressor cell firnction, 234-244 Tumor immunotherapy, 133-135 by immunofacilitation, 138-142 by immunopotentiation, 135-138

CONTENTS OF PREVIOUS VOLUMES Volume 1

Cellular Genetics of Immune Responses G . J. V. NOSSAL

Transplantation Immunity and Tolerance hl. HAWK,A. LENGEROVA, A N D T.

Antibody Production by Transferred Cells

CHARLES G . COCHRANE ANI) FRANK J.

HRABA

DIXON

Immunological Tolerance of Nonliving Antigens

Phagocytosis

DERRICK ROWLEY

RICHARDT. SMITH Functions of the Complement System

ABRAHAM G. OSLER

Antigen-Antibody Reactions in Helminth Infections

E. J. L. SOULSBY In Kfro Studies of the Antibody Response ARRAMB. STAVITSKY

Embryological Development of Antigens

REED A. FLICKINGER Duration of Immunity in Virus Diseases

J. 11. HALE

A w r H o R INDEX-SUBJECT INDEX

Fate and Biological Action of AntigenAntibody Complexes

WILLIAM0. WEIGLE Delayed Hypersensitivity to Simple Protein Antigens

P. G. H. CELLAND B. RENACERRAF

Volume

In Kfro Studies of the Mechanism of Anaphylaxis

K. FRANK AUSTENAND JOHN H. HUMPHREY

The Antigenic Structure of Tumors

P. A. GOREH AUTHORINDEX-SUBJE(:T INDEX

3

The Role of Humoral Antibody in the Homograft Reaction

CHANDLER A. STETSON

Volume 2

Immune Adherence D. S. NELSON

Immunologic Specificity and Molecular Structure

Reaginic Antibodies

D. R. STANWORTH

FREUKARUSH Heterogeneity of y-Globulins

JOHN L. FAHEY The Immunological Significance of the Thymus J. F. A. P. MILLER,A. €1.E.

MARSHALL, AND R. G. WHITE

Nature of Retained Antigen and its Role in Immune Mechanisms

DAN H. CAMPBELL AND JUSTINES. GAR\'EY

Blood Groups in Animals Other Than Man W. H. STONEAND M. R. IRWIN

273

274

CONTENTS OF PREVIOUS VOLUMES

Heterophile Antigens and Their Significance in the Host-Parasite Relationship

PHILIPY. PATERSON

C. R.JENKIN

AUTHORINDEX-SUBJECT

Experimental Allergic Encephalomyelitis and Autoimmune Disease

INDEX

The Immunology of Insulin

c. C. POPE

Tissue-Specific Antigens

D. C. DUMONDE

Volume 4 Ontogeny and Phylogeny of Adaptive Immunity ROBERTA, GOODAND BEN W.

PAPERMASTER

AUTHOR INDEX-SUBJECT INDEX

Volume 6

Cellular Reactions in Infection

EMANUEL SUTERAND HANSRUEDY RAMSEIER

Ultrastructure of Immunologic Processes JOSEPH

Experimental Glomerulonephritis: Immunological Events and Pathogenetic Mechanisms

EMILR. UNANUEAND FRANK J. DIXON

D. FELDMAN

Cell Wall Antigens of Gram-Positive Bacteria

MACLYNMCCARTYAND STEPHENI. MORSE

Chemical Suppression of Adaptive Immunity ANN E. GABRIELSON AND ROBERTA.

GOOD Nucleic Acids as Antigens OTTO

Structure and Biological Activity of Immunoglobulins

SIDNEYCOHENAND RODNEYR. PORTER Autoantibodies and Disease

H. G. KUNKELAND E. M. TAN

1. PLESCIA AND WERNER BRAUN

In Vitro Studies of Immunological Responses of Lymphoid Cells RICHARDW. DUTTON

Developmental Aspects of Immunity JAROSLAV STERZLAND ARTHURM.

SILVERSTEIN Effect of Bacteria and Bacterial Products on Antibody Response

J. MUNOZ

Anti-antibodies

PHILIPG. H. CELLAND ANDREWS. KELUS

AUTHORINDEX-SUBJECT INDEX Conglutinin and lmmunoconglutinins

P. J. LACHMANN

Volume 5 Natural Antibodies and the Immune Response STEPHEN V. BOYDEN Immunological Studies with Synthetic Polypeptides MICHAELSELA

AUTHORINDEX-SUBJECT INDEX

Volume 7 Structure and Biological Properties of Immunoglobulins

SYDNEY COHEN AND CESAR MILSTEIN

275

CONTENTS OF PREVIOUS VOLUMES

Genetics of Immunoglobulins in the Mouse

MICHAEL POTTER AND ROSE AN LIEBERM Mimetic Relationships between Group A Streptococci and Mammalian Tissues

JOHNB. ZARRISKIE Lymphocytes and Transplantation Immunity

DARCYB. WILSONAND R. E. BILLINCHAM Human Tissue Transplantation

JOHNP. MERRILL

AUTHOR INDEX-SUBJECT INDEX

AUTHOR INDEX-SUBJECT INDEX

Volume 10 Cell Selection by Antigen in the Immune Response

GREC;OHY W. SISKIND AND BAHUJ BENACERKAF Phylogeny of Immunoglobulins

HOWARD M. GREY Slow Reacting Substance of Anaphylaxis R o a m r P. ORANGE A N D K. FRANK AUSTEN

Volume 8 Chemistry and Reaction Mechanisms of Complement

HANSJ. MULLER-EBERHARD Regulatory Effect of Antibody on the Immune Response JONATHAN

The Immunology and Pathology of NZB Mice J . B. HOWIEAND B. J. HELYER

W. UHRA N D G ~ R A N

MOLLER The Mechanism of Immunological Paralysis D. W. DRESSERAND N. A. MITCHISON In Wfro Studies of Human Reaginic Allergy ABRAHAMG. OSLER,LAWRENCE M.

Some Relationshipsamong Hemostasis, Fibrinolytic Phenomena, Immunity, and the Inflammatory Response

OSCARD. RA~NOFF Antigens of Virus-Induced Tumors KARL

I9ARE1,

Genetic and Antigenetic Aspects of Human Histocompatibility Systems

D. BERNARD AMOS AUTHORINDEX-SUBJECT INDEX

AND DAVID A. LEVY LICHTENSTEIN.

AUTHOR INDEX-SUBJECT INUEX

Volume 11 Electron Microscopy of the Immunoglobulins

N. MICHAEL GREEN

Volume 9 Secretory Immunoglobulins

THOMAS B. TOMASI, JR., A N D JOHN BI ENENSTOCK Immunologic Tissue Injury Mediated by Neutrophilic Leukocytes

CHARLES G. COCHHANE The Structure and Function of Monocytes and Macrophages ZANVIL A. COHN

Genetic Control of Specific Immunity Responses H U G H 0.MCDEVtI'T AND BARLIJ BENACERRAF

The Lesions in Cell Membranes Caused by Complement

JOHNH. HUMPHREY AND ROBERT R. DOURMASHKIN Cytotoxic Effects of Lymphoid Cells in Wfro

PETER PERLMANN AND

GdRAN

HOLM

276

CONTENTS OF PREVIOUS VOLUMES

Individual Antigenic Specificity of Immunoglobulins JOHN E. HOPPERA N D ALFRED

Transfer Factor

H. S. LAWRENCE Immunological Aspects of Molaria Infection

NISONOFF

IVOR N. BROWN

AUTHORINDEX-SUBJECT INDEX

In Vitro Approaches to the Mechanism of Cell-Mediated Immune Reactions

BARRYR. BLOOM Immunological Phenomena in Leprosy and Related Diseases

Volume 12 The Search for Antibodies with Molecular Uniformity

RICHARD M. KRAIJSE Structure and Function of yM Macroglobulins

HENRYMETZCER Transplantation Antigens R. A. REISFELT AND B. D.

~ H A N

The Role of Bone Marrow in the Immune Response NABIHI. ABDOUAND MAXWELL

RICHTER

J . L. TURK ANI) A. D. M. BRYCESON Nature and Classification of ImmediateType Allergic Reactions

ELMER L. BECKER AUTHOR

INDEX-SUBJECT INDEX

Volume 14 lmmunobiology of Mammalian Reproduction ALANE. BEER A N D R. E. BILLINGHAM Thyroid Antigens and Autoimmunity

Cell Interaction in Antibody Synthesis

D. W. TALMAGE, J. RADOVICH, AND €1. HEMMINCSEN The Role of Lysosomes in Immune Responses

GERALDWEISSMANN AND PETER DUKOR

SIDNEY SHULMAN Immunological Aspects of Burkitt's Lymphoma

GEORGEKLEIN Genetic Aspects of the Complement System CHESTER A. ALPER AND FRED S.

ROSEN Molecular Size and Conformation of Immunoglobulins

KEITH J. DORRINCTON A N D CHARLES TANPORD AUTHOR

INDEX-SUBJECT INDEX

The Immune System: A Model for Differentiation in Higher Organisms

L. HOODAND J. PRAHI. AUTHOR INDEX-SUBJECT INDEX

Volume 13

Volume 15

Structure and Function of Human Immunoglobulin E

The Regulatory Influence of Activated T Cells on B Cell Responses to Antigen

HANSBENNICH JOHANSSON

AND

s. GUNNAR 0.

DAVIDH. b T BENACERRAF

Z AND B A R U J

277

CONTENTS OF PREVIOUS VOLUMES

The Regulatory Role of Macrophoges in Antigenic Stimulation

E. H. UNANUE Immunological Enhancement: A Study of Blocking Antibodies JOSEPH

D.

FELUMAN

Genetics and Immunology of Sex-linked Antigens

L. GAbSEH AND WILLYS K. SrLvens DAVII)

Current Concepts of Amyloid

EDWAHU c. F R A N K L I N A N D DOROTHEA ZUCKEH-FRANKLIN

AUTHORINDEX-SUBJECT INDEX

Volume 16 Human Immunoglobulins: Classes, Subclasses, Genetic Variants, and ldiotypes

J. B. Nnrvic: A N D H. G. KUNKEL Immunological Unresponsiveness WIL1,lAM 0. WEIGLE Participation of Lymphocytes in Viral Infections

FHEIIEHICK WHEELOCKA N U STEPHEN T. TOY E.

Immune Complex Diseases in Experimental Animals and Man c. G . COCHRANE A N D 1). KOFFLER

The lmmunopathology of Joint Inflammation in Rheumatoid Arthritis N A T H A N J. ZVAIYLEH

In Kfro Studies of Immunologically Induced Secretion of Mediators from Cells and Related Phenomena

ELMERL. BECKEHAND PETERM. HENSON Antibody Response to Viral Antigens KtiL?.H

M. COWAN

Antibodies to Small Molecules: Biological and Clinical Applications

VINCENTP. BUTLER, JH., BEISER

AND

SAMM.

AUTHOR ~NDEX-SWMJECT INDEX

Volume 18 Genetic Determinants of Immunological Responsiveness

DAVIII L. GASSEHAND WILLYSK. S II.\wns Cell-Mediated Cytotoxicity, Allograft Reiection, and Tumor Immunity

JEAN-CHARLES CEROTTINI THEOUORE BHUNNER

AND

K.

Antigenic Competition: A Review of Nonspecific Antigen-Induced Suppression

H U G HF. PROSS

AND

DAVIDEIUINGEH

Effect of Antigen Binding on the Properties of Antibody

HENRYMETZGEH lymphocyte-Mediated Cytotoxicity and Blocking Serum Activity to Tumor Antigens

HELLSTHOM AND INGEGERD HELLSIXOM

b H L ERIK

Aur HOK INUKX-S LIUJ EC:T INUEX

Aumioi%INDEX-SUBJECT INDEX

Volume 17

Volume 19

Antilymphocyte Serum

Molecular Biology of Cellular Membranes with Applications to Immunology S . J. SINGER

EUGENEM. LANCE, P. B. MEIIAWAH, AND H o m w N. TAUU

278

CONTENTS OF PREVIOUS VOLUMES

Membrane Immunoglobulins and Antigen Receptors on B and T Lymphocytes

NOELL. WARNER

Thymus-IndependentB-Cell Induction and Paralysis ANTONIOCOUTINHO AND GORAN

MOLLER Receptors for Immune Complexes on Lymphocytes

SUBJECT INDEX

VICTOR NUSSENLWEIC Biological Activities of Immunoglobulins of Different Classes and Subclasses

HANSL. SPIECELBERC SUBJECT INDEX

Volume 22 The Role of Antibodies in the Rejection and Enhancement of Organ Allografts

CHARLES B. CARPENTEH, ANTHONYJ. F. D'APICE,AND ABUL K. ABBAS

Volume 20

Biosynthesis of Complement

Hypewariable Regions, Idiotypy, and Antibody-Combining Site

J. DONALD CAPHA AND J. MICHAEL &HOE

Structure and Function of the J Chain

MARIANELLIOTTKOSHLAND Amino Acid Substitution and the Antigenicity of Globular Proteins

HARVEY R. COLTEN Graft-versus-HostReactions: A Review

C. GREBEAND J. WAYNE STEPHEN STREILEIN Cellular Aspects of Immunoglobulin A

MICHAELE. LAMM Secretory Anti-Influenza Immunity

YA. S. SHVARTSMAN AND M. P. ZYKOV

MORRIS REICHLIN The H-2 Major Histocompatibility Complex and the I Immune Response Region: Genetic Variation, Function, and Organization

DONALD C. SHHEFFLER AND CHELLA S . DAVID Delayed Hypersensitivity in the Mouse

ALFREDJ . CROWLE SUBJECT INDEX

SUBJECT INDEX

Volume

23

Cellular Events in the IgE Antibody Response

KIMISHICEISHUAKA Chemical and Biological Properties of Some Atopic Allergens

T. P. KINC

Volume 21 X-Ray Diffraction Studies of Immunoglobulins

.

ROBERTO J POLJAK Rabbit Immunoglobulin Allotypes: Structure, Immunology, and Genetics

THOMAS J. KINDT Cyclical Production of Antibody as a Regulatory Mechanism in the Immune Response

WILLIAM 0. WIECLE

Human Mixed-Lymphocyte Culture Reaction: Genetics, Specificity, and Biological Implications Bo DUPONT, JOHN A. HANSEN, AND

EDMOND J. YUNIS lmmunochemical Properties of Glycolipids and Phospholipids

DONALD M. MARCUSAND GERALD A. SCHWARTINC SUBJECT INDEX

CONTENTS OF PREVIOUS VOLUMES

279

H-2 Mutations: Their Genetics and Effect on

Volume 24

Immune Functions The Alternative Pathway of Complement Activotion

0. G O T Z E

AND

H. J.

MuLLER-EBEHnARD

Membrane and Cytoplasmic Changes in B Lymphocytes Induced by Ligand-Surface Immunoglobulin Interaction

GEORGE R. SCHHEINER AND EMILR. UNANUE

Lymphocyte Receptors for Immunoglobulin

HOWARD B. DICKLEH

JAN

KLEIN

The Protein Products of the Murine 17th Chromosome: Genetics and Structure ELLEN S. VITETTA A N D J. DONALD CAPRA

Expression and Function of ldiotypes on Lymphocytes

K. EICHMANN The 6-Cell Clonotype Repertoire

NOLANH. SICALAND NORMANR. GINMAN

Ionizing Radiation and the Immune Response ROBERTE. ANDERSONAND NOEL L. WERNER SURJEW

INDEX

Volume 25

SUBJECT

Volume 27 Autoimmune Response to Acetylcholine Receptors in Myasthenia Gravis and Its Animal Model JON

Immunologically Privileged Sites

CLYDEF. BAKERAND R. E. BILLINGHAM Major Histocompatibility Complex Restricted Cell-Mediated Immunity

GENEM. SHEARER AND ANNE-MARIE SCHMIT T-VE RHULSI .

Current Status of Rat Immunogenetics

DAVIDL. GASSER

INDEX

LINDSTROM

MHC-Restricted Cytotoxic T Cells: Studies on the Biological Role of Polymorphic Major Transplantotion Antigens Determining T-cell Restriction-Specificity, Function, and Responsiveness

ROLF M. ZINKERNAGEL AND PETER C. DOHEHTY Murine Lymphocyte Surface Antigens

IANF. C. MCKENZIEAND TERRY PO.I..l'ER

Antigen-Binding Myeloma Proteins of Mice

MICHAELPOTTER Human Lymphocyte Subpopulotions L. Cnsss A N D S. F. SCHLOSSMAN

The Regulatory and Effector Roles of Eosinophils

PETERF. WELLERAND EDWARD J. GOKEL

SUBJECTINDEX

SURJEC"INDEX

Volume 26

Volume 28

Anaphylatoxins: C3o and C5a

The Role of Antigen-Specific T Cell Factors in the Immune Response TOMIO TADA AND KO OKUMURA

TONYE. HUCLIAND HANSJ. M ULLER-E BERHAHD

280

CONTENTS OF PREVIOUS VOLUMES

The Biology and Detection of Immune Complexes ARCYRIOSN. THEOFILOPOULOS AND

FRANK J. DIXON The Human la System

R. J. WINCHESTER AND H. G. KUNKEL Bacterial Endotoxins and Host Immune Responses

DAVIDC. MORRISON AND JOHNL. RYAN Responses to Infection with Metazoan and Protozoan Parasites in Mice

Volume 30 Plasma Membrane and Cell Cortex Interoctions in Lymphocyte Functions

FRANCIS LOOR Control of Experimental Contact Sensitivity

HENRYN. CLAMAN, STEPHEND. MILLER,PAULJ . CONLON,AND JOHN W. MOORHEAD Analysis of Autoimmunity through Experimental Models of Thyroiditis and Allergic Encephalomyelitis

WILLIAM 0. WEIGLE

GRAHAM F. MITCHELL SUBJECTINDEX

The Virology and lmmunobiology of Lymphocytic ChoriomeningitisVirus Infection

M. J. BUCHMEIER, R. M. WELSH,F. ] DUTKO,AND M. B. A. OLDSTONE

Volume 29 Molecular Biology and Chemistry of the Alternative Pathway of Complement

HANSJ. MULLEH-EBERHARD AND RORERTI). SCHREIBER Mediators of Immunity: Lymphokines and Monokines

Ross E. ROCKLIN, KLAUSBENDTZEN, DIRKGHEINEDER

AND

Adaptive Differentiationof Lymphocytes: Theoretical Implicationsfar Mechanisms of Cell-Cell Recognition and Regulation of Immune Responses

DAVIDH. KATZ Antibody-Mediated Destruction of VirusInfected Cells

J. G. PATRICK SISSONS AND MICHAEL B. A. OLDSTONE Aleutian Disease of Mink DAVIDD. PORTER,AUSTIN E. AND HELENG. PORTER LARSEN,

Age Influence on the Immune System

TAKASHI MAKINODAN AND MAHGUERITE M. B. K A Y SUBJECTINDEX

INDEX

Volume 31 The Regulatory Role of Macrophages in Antigenic Stimulation Part Two: Symbiotic Relationship between Lymphocytes and Macrophages

EMILR. UNANUE T-cell Growth Factor and the Culture of Cloned Functional T Cells

KENDALL A. SMITHAND FRANCIS W. RUSCETTI Formation of B Lymphocytes in Fetal and Adult Life

PAUL

w. KINCADE

Structural Aspects and Heterogeneity of Immunoglobulin Fc Receptors

JAYC. UNKELESS, HOWAHD FLErr, AND IRA S. MELLMAN The Autologous Mixed-Lymphocyte Reaction

MARCE. WEKSLEH, CHARLES E. MOODY,JR., AND R O m w W. KOZAK INDEX

28 1

CONTENTS OF PREVIOUS VOLUMES

32

Volume

Polyclonal B-Cell Activators in the Study of the Regulation of Immunoglobulin Synthesis in the Human System THOMAS A. W A L D M A N N AN]) SAM11EI.

BHODEH Typing for Human Alloantigens with the Prime Lymphocyte Typing Technique N. MORLING,€3. K. JAKORSEN, P. PLATZ, L. P. R Y m n , A. S\‘EJGAAHI>, AND

M. THOMSEN

Protein A of Staphy/ococcus aureus and Related Immunoglobulin Receptors Produced by Streptococci and Pneumonococci JOHN

J. LANGONE

Regulation of Immunity to the Azobenzenearsonate Hapten

MARKI. GREENE,MITCHELLJ. NELLES, MAN-SUNSY,ANI) ALFHEI) “SONOFF

Immunologic Regulation of Lymphoid Tumor Cells: Model Systems for Lymphocyte Function ARUL K. ARRAS

INDEX

Autoantibodies to Nuclear Antigens (ANA): Their lmmunobiology and Medicine

EN(: M. TAN The Biochemistry and Pathophysiology of the Contact System of Plosma CHARLES G. COCHRANE AND JOHN H.

G n I wI N Binding of Bacteria to Lymphocyte Subpopulations

MAHIUSTEODORESCIJ AND EUGENE P. MAYE n

INDEX

Volume 34 T Cell Alloantigens Encoded by the IgT-C Region of Chromosome 12 in the Mouse F. L. OWEN Heterogeneity of H-2D Region Associated Genes and Gene Products TF:I)H . HANSEN, KEIKOOZATO, ANI)

DAVID H.

SACHS

Human Ir Genes: Structure and Function THOMAS A. GONWA, B. MATIJA

PK~EHLIN, AND

JOHN

D. STORO

Interferons with Special Emphasis on the Immune System

R o m x r M. FRIEDMAN AND STEFANIE N. VOGEL

Volume

33

The CBNN Mouse Strain: An Experimental Model Illustrating the Influence of the XChromosome on Immunity

IRWINSCHER The Biology of Monoclonal Lymphokines Secreted by T Cell Lines and Hybridomas AMNONALTMAN AND DAVID H . KATiZ

Acute Phase Proteins with Special Reference to C-Reactive Protein and Related Proteins (Pentaxins)and Serum Amyloid A Protein

M. B.

PEI’YS AND

MARILYNL.

BArrZ

Lectin Receptors as Lymphocyte Surface Markers NAT H A N SHAHON

INDEX

This Page Intentionally Left Blank