Advances in Immunology Volume 36

ADVANCES I N Immunology VOLUME 36 CONTRIBUTORS TO THIS VOLUME GUIDOBIOZZI Y OLANDE BOUTHILLIER STEVENJ . BURAKOFF B...

Author: Frank J. Dixon

28 downloads 774 Views 14MB 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 I N

Immunology VOLUME 36

CONTRIBUTORS TO THIS VOLUME

GUIDOBIOZZI Y OLANDE BOUTHILLIER STEVENJ . BURAKOFF BARTONF. HAYNES ALANM. KRENSKY RICHARDALANLERNER DENISE MOUTON CAROLS. REISS CLAUDESTIFFEL MYRONR. SZEWCZUK ANDREWW. WADE OFFUWEINBERCER

ADVANCES IN

Immunology EDITED BY FRANK J. DIXON Scripps Clinic and Research foundation La Jollo, California

VOLUME 36

1984

ACADEMIC PRESS, INC. (Harcourt Brace Jovanovich, Publishen)

Orlando San Diego

New York

London

Toronto Montreal Sydney Tokyo

COPYRIGHT @ 1984, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR A N Y INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITINO FROM THE PUBLISHER.

ACADEMIC PRESS,INC.

Orlando, Florida 3 2 8 8 7

United Kingdom Edition pirblislied by ACADEMIC PRESS, INC. ( L O N D O N ) LTD. 24/28 Oval Road, London NWl 7 D X

LIBRARY OF CONGRESS

CATALOG CARDNUMBER:

I SBN 0-12-022436-4 PRINTED IN THE UNITED STATES OF AMERICA 84858687

9 8 7 6 5 4 3 2 1

6 1 - 1 705 7

CONTENTS

CONrHIBUTORS .............................................................. PREPACE ....................................................................

Vii ix

Antibodies of Predetermined Specificity in Biology and Medicine

RICHARD ALANLERNER

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

I1. The Nature of Immunogen Determinants of Intact Proteins . . . . . . . . . . .

. The Loop of Lysozyme Experiments ..............................

I11 IV . V. VI . VII . VIII IX. X XI . XI1 .

. .

XI11 . XIV XV . XVI .

.

XVII .

On the Number of Antigenic Determinants in Proteins . . . . . . . . . . . . . . . Antibodies of Predetermined Specificity ........................... Detection of the Products of Nucleotide Sequences . . . . . . . . . . . . . . . . . . Sets of Antibodies and Antibodies to Protein Domains . . . . . . . . . . . . . . . Structure-Function Studies ...................................... Antibodies to Proteins Encoded by Alternative Reading Frames . . . . . . . Exon Usage ................................................... The Chemistry of Virus Neutralization ............................. Synthetic Immunogens Representing Idiotypes, Allotypes. and Growth Factors ................................................ The Structure of an Antigenic Determinant in a Protein . . . . . . . . . . . . . . Technical Aspects: The Only Rule Is That There Is No Rule . . . . . . . . . . Theoretical Aspects ............................................. The Repertoire Should Be Tapped Further: Concept of Immunological Catalysis ......................................... Antibody Template Directed Organic Synthesis ..................... References ....................................................

1 2 5 6 7 9 11 14 15 16 17 21 26 31 32 35 38 39

A Molecular Analysis of the Cyiolytic Lymphocyte Response

STEVEN J . BURAKOFF. OFRAWEINBERGER. ALAN M . KRENSKY. CAROLS . REISS

AND

Introduction ................................................... Specificity of the CTL Response .................................. The Use of Liposomes to Study the Generation of the CTL Response . . The Use of Liposomes to Study the Helper T Cell Pathway . . . . . . . . . . . The Use of DNA-Mediated Gene Transfer of Cloned MHC Genes to Study CTL Specificity........................................... VI . The Use of Monoclonal Antibodies to Define Functional CTLAntigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Conclusions ................................................... References ....................................................

I. I1. 111. IV. V.

V

45 46 48 52 63 67 75 78

vi

CONTENTS

The Human Thymic Microenvironment

BARTON F. HAYNES

I . Introduction ................................................... 11. Anatomy and Histology of the Thymic Microenvironment ............. 111 Evidence for the Participation of the Thymic Microenvironment in Promoting T Cell Maturation ..................................... IV. Heteroantisera and Other Markers That Define Components of the Human Thymic Microenvironment ................................ V Monoclonal Antibodies That Define Components of the Human and Rodent Thymic Microenvironment ................................ VI Ontogeny of the Human Thymus Microenvironment . . . . . . . . . . . . . . . . . VII . The Human Thymic Microenvironment in Diseases of Abnormal T Cell Maturation ................................................ VIII . Summary ...................................................... References ....................................................

.

. .

87 89 94 97 99 119 126 134 136

Aging. Idiotype Repertoire Shifts. and Compartmentalization of the Mucosal-Associated lymphoid System

ANDREWW. WADEAND MYRONR . SZEWCZUK

I . Introduction ................................................... 11 Aging in the Immune System .................................... 111. Changes in the Expression of Available and Functional Repertoires .................................................... IV . Compartmentalization of the Mucosal Immune System with Age . . . . . . . References ....................................................

.

143 144 156 170 181

A Major Role of the Macrophoge in Quantitative Genetic Regulation of lmmunoresponsiveness and Antiinfectious Immunity

GUIDOBrozzr. DENISEMOUTON.CLAUDE STIFFEL.AND YOLANDE BOurHILLIER

I . Introduction ................................................... I1. Genetic Regulation of Immunoresponsiveness ...................... 111. Modifications of Macrophage Functions ............................ IV. Cell-Mediated Immunity in High and Low Antibody Responder Lines . .

189 192 197 215

V Modifications of Innate and Immune Resistance to Infections in High and Low Antibody Responder Lines ............................... VI . Considerations on the Genetic Control of Antiinfectious Immunity . . . . . VII . Conclusion .................................................... References ....................................................

218 226 231 232

.

INDEX ..................................................................... COwrEwrs OF PREVIOUS VOLUMES............................................

235 239

CONTRIBUTORS

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

GUIDOBIOZ‘LI(18Y), U 125 ZNSERM, E R 060070 CNRS, Znstitut Curie-Section de Biologie, 75231 Paris C6dex 05, Frunce

(18Y),u 125 INSERM, I!?R060070 C N R S , Institut Curie-Section de Biologie, 75231 Paris Ctdex 05, France

YOLANDE BOUTHILLIER

STEVENJ. BURAKOFF (45),Department of Pediatrics, Harvard Medical School, Dana-Farber Cancer Institute, and the Children’s Hospital Medical Center, Boston, Massachusetts 021 15

BARTONF. HAYNES (87),Department of Medicine, Division of Rheumatic and Genetic Diseases, and the Department of Microbiology and Immunology, Duke University School of Medicine, Durhnm, North Carolina 27710 ALANM. KRENSKY(45), Department of Pediatrics, Harvard Medical

School, Dana-Farber Cancer Institute, and the Children’s Hospital Medical Center, Boston, Massachusetts 02115

RICHARDALAN LERNER (l), Department of Molecular Biology, Research Znstitute of Scripps Clinic, La Jolla, California 92037 DENISE MOUTON(189), U 125 ZNSERM, E R 060070 CNRS, Znstitut Curie-Section de Biologie, 75231 Paris C6dex 05, France CAROLS. REISS (45), Department of Pediatrics, Harvard Medical School, Dana-Farber Cancer Institute, Boston, Massachusetts 02115 CLAUDE STIFFEL(189), U 125 ZNSERM, E R 060070 CNRS, Znstitut Curie-Section de Biologie, 7*5231Paris Ctdex 05, France

MYRONR. SZEWCZUK (143), Department of Microbiology and Zmmunology, Queen’s University, Kingston, Ontario K7L 3N6, Canada ANDREWW. WADE(143), Department of Microbiology and Zmmunology, Queen’s University, Kingston, Ontario K7L 3N6, Canada OFRAWEINBERCER( 4 3 , Department of Pediatrics and Pathology, Harvard Medical School, Dana-Farber Cancer Institute, and the Children’s Hospital Medical Center, Boston, Massachusetts 021 15 vii

This Page Intentionally Left Blank

PREFACE

The reviews contained in this volume reflect the many different branches of investigation in immunology today. They range from submolecular analysis of immunogenic determinants to age-related changes in the entire immunologic system. The approaches vary: structural chemistry is used to study immunogens, molecular biological tools to define the basis of the cytotoxic T lymphocyte response, immunohistochemistry based on monoclonal antibodies to analyze the thymic microenvironment, cellular immunologic technology to assess the effects of aging on the immune system, and immunogenetics to ascertain the role of macrophages in regulating immunoresponsiveness. One of the strengths of immunology is its diversity derived from the many related scientific fields that have benefited from and adapted immunologic technology and thought. The resultant broad scope of such research provides a wide range of information that continues to shed light on central immunologic problems. The subjects presented in this volume are good examples of this diversity and progress. Perhaps the most significant advance in the humoral antibody field since the advent of hybridomas and monoclonal antibodies has been the development of peptide technology, which has made possible the production of antibodies with predetermined specificity. In the first chapter Dr. Richard Lerner, a pioneer in this field, discusses the use of small peptides derived from the amino acid sequences of large proteins to elicit antibodies reactive not only with the immunizing peptides but also with the specific sites on the intact original proteins. The implications of this work for our understanding of antigen-antibody interactions as well as protein structure are significant. In addition, the precision possible with peptide antigens has provided a new level of understanding of viral antigenicity and viral neutralization by antibody. Also, the substitution of small peptides derived via nucleic acid technology for hard to isolate proteins is permitting study and manipulation of many biological systems including the idiotype network, endocrine-lymphokine controls, and tumor antigens. Dr. Lerner’s review of this considerable development is accompanied by an intriguing and challenging projection of things yet to come. The general immunologic and immunogenetic parameters of the cytolytic T lymphocyte (CTL) response have been established over the past several decades, but only recently has examination of CTLs at a molecular level been possible. The latter studies employing liposome constructs, cloning and transfer of MHC genes and monoclonal ix

X

PREFACE

antibodies to the molecules involved in the CTL are reviewed in the second chapter by Drs. Steven Burakoff, Ofra Weinberger, Alan Krensky, and Carol Reiss, all important contributors to this field. Via use of liposomes, it has been possible to demonstrate that allogeneic MHC Class I and I1 antigens alone can stimulate CTLs and that in triggering virus specific CTLs both viral and MHC antigens if not actually physically interacting must be at least very closely associated. Identification of the particular domains of the Class I molecules essential for recognition by CTLs has been achieved by exon shuffling in the construction of recombinant MHC genes. Finally, monoclonal antibodies capable of inhibiting CTL reactions are being used to identify the molecules involved in CTL interactions. These studies indicate that in addition to the primary antigen-receptor union a series of less specific accessory interactions may serve to stabilize the CTLtarget union. The complex developmental scheme of the thymus, particularly its unique microenvironment, is precisely and succinctly presented by Dr. Barton Haynes in the third chapter. The interactions of local mesodermal, endodermal, and ectodermal cells plus blood-borne thymocyte precursors in forming the thymus are traced by using monoclonal antibodies specific for the different developmental stages of the several cell lineages. The functions of the thymus in T cell repertoire generation and MHC restriction as well as in endocrine production are correlated with this structural analysis and put in perspective. Finally, the abnormalities in thymic microenvironment and T cell maturation associated with immunodeficiency and autoimmune disease are discussed. Although a decline in immunologic function with age is generally recognized, there is as yet no satisfactory explanation of the phenomenon. This subject is put in a new light by Drs. Andrew Wade and Myron Szewczuk. In the fourth chapter, they contrast changes during aging in the systemic lymphoid system with those in the mucosal lymphoid system. With age, most effector functions of the systemic immune system are compromised including reduced performance of antigen presenting cells and T helper cells as well as decreased IL-2 production and B cell responsiveness. Apparently only T suppressor activity and antiidiotype responses increase. However, two compartments of the immune system do not appear to be affected by age: the stem cells of the bone marrow and the mucosal lymphoid tissues, which retain full potential and function throughout life. This suggests that the age-related changes in the systemic lymphoid system may result from factors extrinsic to the cells themselves rather than intrin-

PREFACE

xi

sic, programmed declines in performance. The existence of two distinct immunologic compartments, one with impaired and the other with normal circuitry, may allow further insights into immune regulation. Of the various factors controlling the immune response, none is so directly related and readily demonstrated as macrophage function. The manner of handling antigens by macrophages does much to determine the magnitude and effectiveness of both humoral and cellular immune responses. In the final chapter, Drs. Guido Biozzi, Denise Mouton, Claude Stiffel, and Yolande Bouthillier characterize macrophage function relative to immunity by drawing on their extensive sxperience with mice inbred for high or low antibody responses. Mice with hyperactive macrophages, i.e., rapid antigen catabolism, tend to be poor antibody responders, while those that destroy antigen less rapidly make high antibody responses. Interestingly, cell mediated immunity is quite independent of the antibody response and tends to parallel the vigor of macrophage function, particularly with intracellular microbial agents. The authors discuss this view of macrophage Function and antibody formation in the general context of immune resistance to infections and propose a general theory dealing with host-pathogen interaction. FRANK J. DIXON


ADVANCES IN

Immunology VOLUME 36


ADVANCES IN IMMUNOLOGY, VOL. 36

Antibodies of Predetermined Specificity in Biology and Medicine RICHARD ALAN LERNER Depahnent of Molecular Biology, Research Institute of Scripps Clinic, La Jolla, California

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

111. IV.

V. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. XV. XVI.

nogen Determinants of Intact Proteins . . . . . . . . . . . . The Loop of Lysozyme Experiments . . . . . . ................ On the Number of Antigenic Determinants in Proteins. . . . . . . . . . . . . . . . Antibodies of Predetermined Specificity . . . ................ Detection of the Products of Nucleotide Sequences. . . . . . . . . . . . . . . . . . . Sets of Antibodies and Antibodies to Protein Domain .......... Structure-Function Studies. . . . . . . . . . . . . . . . . . . . . . .......... Antibodies to Proteins Encoded by Alternative Readi es.. ...... Exon U s a g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Chemistry of Virus Neutralization. . . . . . . . . . . ................ Synthetic Innnunogens Representing Idiotypes, All Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Structure of an Antigenic Determinant in a Protein . . . . . . . . . . . . . . . Technical Aspects: The Only Rule Is That There Is No Rule. . . . . . . . . . . Theoretical Aspects.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Repertoire Should Be Tapped Further: Concept of Immunological

1

2 5

6 7 9

11 14 15 16

17 21 26 31 32

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

XVII. Antibody Template Directed Organic Synthesis. ..................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38 39

1. Introduction

The diversity and exquisite specificity of antibodies has captured the imagination of scientists since the time of Erlich. And since it is possible to make use of the vast antibody repertoire to make antibodies specific for virtually any protein, it might have been reasonable to ask what more does one want? As is usually true in science, w e are learning that there is indeed much, much more. The first new major advance was the development by Kohler and Milstein (1975) of the hybridoma technology. This development allowed the purification of individual specificities from the set of responses that an animal makes upon being presented with a collection of epitopes in one or even a mixture of molecules. Thus in one elegant stroke we were capable of having unlimited quantities of antibodies reactive with a single epi1 Copyright 0 1984 by Academic Press,lnc. A11 rights of reproduction in any form reserved. ISBN 0-12-022436-4

2

RICHARD ALAN LERNER

tope. But, what epitope? It turns out that even with monoclonal antibodies this is a difficult question to answer and one which can only be approached by an often complex protocol of biochemical or genetic experiments. The basic problem, of course, is that the inducing immunogen is complex and the sorting out of specificities is a retrospective exercise. Recently, it has become clear that antibodies to most regions of a protein can be induced by immunizing with short synthetic peptides. This advance portends a change in the way we generate and think about antibodies from a retrospective to a prospective science. Indeed, the synthetic immunogenic technology has already had impact on diverse areas of biology and medicine. First, since the region in the protein with which antipeptide antibodies react is known in advance to the experimenter, they can be said to be of predetermined specificity. This has become useful in the search for protein products predicted from the nucleic acid sequence of newly described nucleic acid clones. Also, the predetermined specificity of these antibodies allows protein chemists and cell biologists to carry out precise structure-function experiments, and to orient proteins in cellular compartments and subcellular organelles. Second, protein reactive antipeptide antibodies are being used to study the chemistry and structure of antigen-antibody union. A subset of the general problem of the chemistry of antigen-antibody union is the chemistry of virus neutralization. The synthetic immunogen technology has already led us a long way toward answers in this area. Third, and perhaps in the long run the most important, the frequency with which protein-reactive antipeptide antibodies can be generated has led to a conceptual and experimental merger between the antibody and protein dynamics problems. The details surrounding these three broad issues form the subject of this review. The search of the literature for this article ended in December of 1983. As for reviews in any area of rapidly progressing science, I must apologize in advance for not citing papers which have escaped my attention. II. The Nature of lmmunogen Determinants of Intact Proteins

As we will see, the nature of protein reactive antibodies induced by small synthetic peptides is probably very different from the antibodies induced by intact proteins. Thus, this section will be brief and will serve only to orient the reader to certain key conceptional issues. The issue of protein immunogenicity has been previously authoritatively

ANTIBODIES OF PREDETERMINED SPECIFICITY

3

reviewed by Crumpton (1974) and more recently by a consortium of experts in the field (Benjamin et al., 1984). Before beginning the discussion we need to deal with one definition which concerns the difference between antigenicity and immunogenicity. Immunogenicity refers to the ability of a substance to induce an antibody, whereas antigenicity simply refers to the ability to be recognized by an antibody. Sometimes, when considering proteins, these terms blur. However, the differences can be crystalized when considering the polysaccharide coat of the pneumococcus. If a rabbit is immunized with intact pneumococcal bacteria an antibody to the polysaccharide coat is made. Thus under these circumstances, the polysaccharide is immunogenic (will induce antibody) and antigenic (will bind to antibody). If the experiment is repeated with the isolated, purified, polysaccharide no antibody is induced. Thus the purified polysaccharide is not immunogenic but is still antigenic in that it will react with antibody induced by the intact organism. This distinction will become important when antipeptide antibodies are considered since peptides can induce protein reactive antibodies which cannot be induced by immunizing with the intact protein. Thus, there are regions of a protein which are antigenic but not immunogenic. Indeed, as we will see, this is one of the most powerful aspects of the synthetic immunogen technology. For as long as the antibody problem has existed, there have been two obvious routes to a solution. One could study either the antigen or the antibody. To understand the way that studies proceeded, one needs to reflect on the scientific climate of immunology in the late 1940s and early 1950s. There were a number of things we knew (reviewed by Edelman, 1973). Working with haptenes, Landsteiner had provided evidence for molecular complimentarity in antigen-antibody union (Landsteiner, 1936), and there was evidence that most antibodies were multivalent (Marrack, 1938).Tiselius (1937) had shown that antibodies were heterogeneous in charge and others had demonstrated heterogeneity in their binding constants ( b r u s h , 1962). But little was known about the structure of antibodies and it was not until the late 1950s that the structure of immunoglobulins began to be revealed (Edelman, 1973; Porter, 1973). Considering this lack of knowledge about the chemistry of antibodies it is no surprise that prior to 1950 studies in immunology centered around the antigens. In general, the approach which most workers took was to make antibodies to an intact protein and then determine the effect of proteolysis or denaturation on its antigenicity. As we will

4

RICHARD ALAN LERNER

detail in this section, these studies were to lead to two general conclusions about the antigenicity and immunogenicity of proteins. These were that only a limited portion of an intact protein is immunogenic and that antigenic determinants of proteins were most often constructed from amino acids distant from each other in the sequence but brought into proximity by the tertiary folding of the protein chain. These determinants were often called “conformational” or “discontinuous’’ and more recently have been referred to as assembled topographical determinants (Benjamin et al., 1984). As he had done for simple haptenes, Landsteiner was to lead the way into the study of the antigenicity of proteins. Writing in 1936, Landsteiner was to thus comment on the serological specificity of proteins, “In view of the imperfect state of protein chemistry, it is not surprising that neither the observations on natural antigens nor those on modified proteins suffice definitely to interpret the specificity of protein reactions in terms of chemical structure. Progress in this direction may come from pursuing the investigation of protein split products.” He continues, “Clearly the highly selective action of the immune sera precludes specificity being determined by simple structures as single amino acids, and even reacting groups composed of di- or tripeptides could not furnish a sufficient number of combinations. Irrespective, therefore, of any particular hypothesis concerning their constitution, the specificity of proteins must be referable to complicated structures-possibly multiple, like groups in one moleculeor to several groupings whose affinities have to be satisfied before a visible reaction can occur, in which event the spatial arrangement of the reacting groups may be significant.” Based on these thoughts, a genre of experiments designed to understand the nature of antigenicity was born. If the essential nature of an antigenic determinant is complex, what is the basis of this complexity? The most telling experiments were those carried out on globular proteins where it was soon learned that upon denaturation their reactivity with homologous antibodies was almost completely eliminated (reviewed, Crumpton, 1974). Brown et al. (1959) showed that oxidation of all four of the disulfide bonds of ribonuclease eliminated its reactivity with an antibody against the native molecule. This initial finding was confirmed by many others using a variety of proteins (Freedman and Sela, 1966; Neumann et al., 1967; Shapira and Arnon, 1969; Arnon and Neurath, 1970; Young and Leung, 1970; Goetzl and Peters, 1972). Further evidence for the role of protein conformation in antigenicity came from the reciprocal experiments in which the ability of anti-


5

bodies prepared against denatured proteins to react with the native protein was studied. Again, the importance of conformation for antigenicity was demonstrated. Antibodies to denatured ribonuclease (Brown, 1962), lysozyme (Young and Leung, 1970; Arnon and Maron, 1971), or bovine serum albumin (Habeeb and Atassi, 1971) reacted only poorly if at all with the native molecule. The totality of experiments on denatured molecules was to be summarized in 1975 by Reichlin (Reichlin, 1975): “The essential meaning of these observations is that the native and fully denatured forms of proteins are for the most part entirely different structures from the immunological point of view. The other constellation of experiments designed to understand the complexity of protein antigens were the fragmentation studies. As for denatured proteins, the general conclusion of such studies was that enzymatic or chemical digestion of proteins is accompanied by greatly diminished antigenicity (reviewed, Crumpton, 1974). For example Nisonoff et ul. (1970) were not successful in obtaining immunologically reactive fragments of ribonuclease. ”

111. The Loop of Lysozyme Experiments

The experiments carried out in the late 1960s and early 1970s by Arnon, Sela, and Anfinsen and their collaborators on residues 64-82 (the “loop”) of lysozyme were of singular influence in crystallizing the concept that the antigenic determinants of proteins were conformation dependent (Arnon et al., 1971). Thus, these experiments are discussed separately here. Based on the X-ray crystallographic data of Canfield and Liu (1965), Arnon and Sela (1969) noted that residues 64-80 of lysozyme formed a loop held together at the neck as it were, by a disulfide bond between Cys 64 and Cys 80. Thus, this structure offered a crisp experimental system to test the postulate that the majority of antigenic determinants of globular proteins were dependent on complex conformations. The question concerned the comparative immunogenicity and antigenicity of the open (disulfide bonds reduced and carboxymethylated) versus the closed (intrachain disulfide bonds intact) loop. The results showed that antibodies to intact lysozyme as well as antibodies to the closed loop bound to the closed loop, whereas oxidation of the disulfide bridge of the isolated loop resulted in a molecular species with a drastic decrease in ability to bind either of these antibodies (Arnon and Sela, 1969; Maron et al., 1971; Arnon and Maron, 1971). I n the broad sense these experiments impacted on two fronts. First, by direct experimentation, including chemical syn-

6

RICHARD ALAN LERNER

thesis of antigenic determinants of intact proteins, they were to again show the conformational nature of antigenic determinants of proteins. But, of more significance to this review, the experiments suggested that in order to synthesize immunogens which mimicked those of intact proteins, one needed to build complex conformations, a possibility difficult at best and only feasible where the X-ray structure of the protein had been determined. IV. On the Number of Antigenic Determinants in Proteins

Given that antigenic determinants of globular proteins were of the conformational type, the next question considered was how many of these there were for a given protein. The general conclusions were that the number of antigenic sites was limited and on the average there was about one site for each 5000 daltons of protein (reviewed, Crumpton 1975; Benjamin et al., 1984). In fact some authors argued that the complete antigenic structures of myoglobin (Atassi, 1975) and lysozyme (Atassi, 1978) were known. In myoglobulin, for example, Atassi suggested that the complete “immunological anatomy” consisted of five antigenic sites comprising residues 15-22, 56-62, 9499, 113-119, and 145-151. In lysozyme Atassi and Lee (1978) proposed only three sites formed respectively by the discontinuous residues 5,7,13,14; 33,34,113,114; and 62,87,89,93,96,97.Recently the notion that a few sites determined the entire antigenicity of a protein has come under considerable attack (reviewed, Benjamin et al., 1984). In a particularly telling series of experiments, Ibrahimi et al. (1979) and White et al. (1978) studied lysozymes with evolutionary substitutions at positions outside the antigenic sites proposed by Atassi and colleagues. Such lysozymes were shown to be antigenically distinct, thus showing that there were more antigenic sites than previously suggested. It is, of course, still possible that a change in the molecule at one site could alter one of the original determinant sites. The suggestion that there are sites additional to those previously proposed (Atassi, 1978) was given further support by the detailed studies of Smith-Gill and colleagues (1982). These investigators made a monoclonal antibody to chicken lysozyme C. This antibody reacted completely with lysozymes of seven different species of galliform birds, partially with two other galliform species and not all with duck lysozyme. The site of antibody binding was determined by comparing the amino acid sequences of these different lysozymes (again assuming that antigenic changes due to substitution of amino acids during evolution do not result from long-range effects) and by showing that Biebrick Scarlet, a dye which binds to the catalytic site of lysozyme,


7

inhibits the binding of antibody to the enzyme. The antibody binding site was postulated to involve the Arg 68-Arg 45 complex and extend into the cleft between Arg 45 and Arg 114 giving an overall dimension for the site of at least 13 x 6 x 15A. The important point relative to the present discussion is that this site is outside those defined by Atassi and Lee (1978). Perhaps the lesson in all this is that it is unrealistic to assume that one can define the complete immunogenicity and antigenicity of a protein on the basis of a single antiserum and that what is immunogenic may depend as much on the mode of presentation of the antigen and on the species immunized as the protein structure itself. From the point of view of this review, however, the important point is that together, the two general assumptions that antigenic determinants were discontinuous and relatively few in number did not bode well for a general technology in which antibodies to any region of a protein could be generated. V. Antibodies of Predetermined Specificity

We now come to the main substance of the review which is the concepts and applications which surround antipeptide antibodies of predetermined specificity. At the outset it is necessary to understand a generality which will become refined as the arguments proceed. This is that these antibodies are in principle very different from those made against intact proteins. Antibodies induced by intact proteins are generated by largely ordered arrays of atoms and then, depending on the experiment, tested against an ordered or experimentally disordered protein target. Conversely, antibodies of predetermined specificity are made against a disordered array of atoms and tested against an always more ordered protein target. In 1980, signals began to appear which suggested the possibility for a new technology. Upon completion of the nucleotide sequence of the replication competent murine Maloney leukemia virus we were faced with an open reading frame which predicted a protein which we did not understand. The reading frame was part of the envelope gene and extended to the 3’-most end of the coding region of the gene and thus predicted an unexplained C-terminal segment of a protein (Sutcliffe et al., 1980). We chemically synthesized a peptide predicted from the last nucleotide of the reading frame and made antibodies to it. The antibody precipitated two proteins from infected and transferred cells which corresponded to the envelope precursor protein, gp85, and a precursor to the membrane anchor protein, p15E, which is now called pre-15E (Green et al., 1981). As it turned out, the reason that the predicted protein sequence did correspond with what we previously

8

RICHARD ALAN LERNER

knew about biochemistry of p15E is that during viral budding this protein undergoes a maturation step which involves cleavage and removal of a C-terminal fragment. Accordingly, the antibody which is site specific for the C-terminus detects the precursor but not the product molecule (Green et al., 1981).Thus the antibody solved two problems at once each of which was dependent on its predetermined nature. The product of an open reading frame was established and a small site on a protein was specifically followed during a biological process. Walter et al. (1980) made antibodies to synthetic peptides corresponding to the carboxy- and amino-terminal regions of the simian virus 40 large tumor antigen. These antisera reacted appropriately with the intact large T antigen. These two studies signaled the possibility of a new technology, but there was a large conceptual hurdle to pass. This had to do with the fact that both studies involved the termini of proteins which often may be disordered and perhaps more easily mimicked by short peptides. Indeed, there had been some success in the past using the termini of proteins as immunogens (Anderer and Schlumberger, 1965; Arnon et al., 1976) but little attempt was made to move beyond this. Thus, because of the argument about the ends of molecules and the general tenure of past arguments (see above) indicating that generally one needed to construct conformations to generate protein reactive antibodies, there was little certainty that the technology could be used broadly or in particular extend beyond the termini of proteins. If the argument about the termini of proteins held, synthetic immunogens would b e of limited use and certainly not a general way of inducing site-specific immunological reagents for the study of proteins. However, Green and her colleagues carried out an experiment which suggested one could use synthetic immunogens to generate antibodies of predetermined specificity which were reactive with virtually any region of a protein (Green et al., 1982). They used the influenza virus hemagglutinin as a test object because the complete nucleotide sequence of its gene was available (Min Jou et al., 1980) and its crystallographic structure was known at high resolution (Wilson et al., 1981). A series of peptides covering 75% of the HA1 chain were chemically synthesized and antibodies were made to each. Antibodies to almost all (18 of 20) peptides reacted with the intact molecule (Green et al., 1982). Since in its folded state the HA1 molecule displays a number of secondary structures including a-helix, extended chains, and BFIG.1. Sites on the surface of the hemagglutinin molecule to which antibodies bind (A) during infection with the whole virus, (B) using antipeptide antibodies. The acarbon tracing of HA1 is represented by green, blue is the a-carbon tracing of HA2, and the solvent accessible surface of the antigen binding sites is expressed by purple dots.

FIG. 1A.

Frc:. 1B.



9

sheets, it was clear from this study that reactivity of antipeptide antibodies is independent of secondary structure or location in the molecule. In Fig. 1A the portion of the molecule reactive with antipeptide antibodies is contrasted to the antigenic sites thought to be important (at least as far as virus neutralization is concerned) when the intact protein is presented as an immunogen (Fig. 1B). Unless we are in for a considerable surprise, one expects only antigenic determinants located on the surface of proteins to be reactive with antibodies. Cherenson and Wilson (unpublished observations) used an algorithm of Connolly (1983) to calculate the solventaccessible surface of all the synthetic peptides used in the study of Green et al. (1982) as free peptides as well as part of the trimeric form of the hemagglutinin. Even in the trimer, a percentage of the peptide was accessible to solvent, and in every case some of each peptide was exposed. Thus, the studies on the HA of the influenza virus made two main points. First, the use of chemically synthesized immunogens to generate antibodies of predetermined specificity was general and could be applied to most regions of proteins. Second, the immunogenicity of an intact protein is less than the sum of the immunogenicity of its pieces (Lerner, 1982). VI. Detection of the Products of Nucleotide Sequences

Because of the rapid advances in gene cloning and nucleotide sequencing, the usual course of genetic experimentation is often reversed and one may have a complete nucleotide sequence of a gene before there is any information about the protein which it encodes. This is a problem which we have called genotype in search of phenotype. Antipeptide antibodies have proved extremely useful in detecting the protein products of cloned genes. The largest amount of work has been carried out in viral systems, especially with replication defective viruses which encode oncogenes (Wong and Goldberg, 1981; Sefton and Walter, 1982; Tamura and Bauer, 1982; Boyle et al., 1983; Sen et al., 1982; Gentry et al., 1983; Robbins et al., 1983; Devare et al., 1983; Niman, 1983; Kloetzer et al., 1983; Stanker et al., 1983). In all of these experiments, the approach has been similar. After cloning and sequencing an oncogene, antibody to a peptide predicted from the nucleotide sequence was made and used to find the gene product. In some instances the antibodies were used to localize the protein in cells and inhibit its enzymatic function, points to which we will return later when other uses of antipeptide antibodies are considered. As an exemplary experiment, we can look further into the study by Baluda and his colleagues (Boyle et al., 1983).The replication-defective avian

10

RICHARD ALAN LERNER

myeloblastosis virus contains as part of its genetic information an inserted cellular sequence which is responsible for acute myeloblastic leukemia in chickens, but the protein encoded by the oncogene had not been found. They made antisera to three small synthetic peptides out of the putative 265 amino acids predicted from the long open reading frame of the virus. The use of multiple peptides from one protein is of significance for two reasons. First, since these antisera are site-specific, antisera to only one peptide might miss intermediates in a processing cascade when the part of the protein to which the antibody is directed is removed. Second, if antisera to more than one region of a protein detect the same products in cells one can be certain of the immunological specificity of the reaction, thus avoiding some of the usual problems with immunological reagents. Furthermore, if the immunological reactivity of the individual antisera is inhibited by the homologous but not heterologous peptides any possibility for spurious binding is eliminated. The three antipeptide antisera of Baluda and his colleagues behaved in concert and detected a 48,000-dalton protein from leukemic myeloblasts. Importantly the same three antisera precipitated a 110,000-dalton protein from normal hematopoietic tissue but not leukemic myeloblasts. These studies allowed the authors to conclude that ~ 4 8 ~ is " "the oncogenic relative of a differentiation specific normal cellular homolog ( p l 10Pr"t"-An'v). These studies also raise the specter of the nature and function of the "nononcogenic" part of the 110,000-dalton molecule which might be imagined to play a role in myeloblastoid differentiation. In another series of experiments, distinguished because they were the first to relate an oncogene to a normal protein of known function, Aaronson and his colleagues (Robbins et al., 1983; Devare et al., 1983) and Niman (1983) used antipeptide antibodies to demonstrate the suspected relationship between the oncogene of the simian sarcoma virus to the platelet derived growth factor. Again, one was able to gain confidence in the results since several different peptides from the amino- and carboxyl-termini (Robbins et al., 1983; Devare et al., 1983) and an internal fragment (Niman, 1983) were used to generate the antisera. Antipeptide antibodies have been used extensively to investigate the proteins of the DNA containing transforming viruses. A number of different antisera to the SV40 and polyoma T antigens have been made (Walter et al., 1980; Green et al., 1983a; Feldman and Nevins, 1983; Luka et al., 1983; Lucher et al., 1983; Yee et al., 1983).Additionally in the SV40 system, antipeptide antibodies were used to locate a previously uncharacterized "agno-protein" protein which accumu-

ANTIBODIES O F PREDETERMINED SPECIFICITY

11

lates late in a lytic infection and is encoded in the late leader region of the gene (Cosman et al., 1982a).As for the SV40 and polyoma systems, structural and transformation associated proteins have also been studied for other DNA containing transforming viruses (Luka et al., 1983; Green et al., 1983b; Lucher et al., 1983). Sutcliffe, Milner, Bloom, and their colleagues have embarked on an interesting adventure which exemplifies the use of antipeptide antibodies to find the products of newly discovered genes and probably portends studies in other systems (Sutcliffe et al., 1983). These investigators are interested in finding new proteins specific to certain areas of the brain. The strategy they have adopted is to clone DNA copies of brain messenger RNA and then by a process of hybridization to RNA from different tissues and different regions of the brain select clones specific for certain areas of the brain. When an interesting gene is found, it is sequenced and antibodies are prepared to several regions of the predicted protein sequence. In their first study, these investigators found two new proteins of interest. One was found in the cytoplasma and dendrites of large neurons distributed throughout the brain while the other marked a novel series of neuronal pathways and may be a precursor for a new neurotransmitter used by the pathways. The findings are probably just the beginning and one can expect major advances from this approach. VII. Sets of Antibodies and Antibodies to Protein Domains

The special feature of all the antibodies under consideration is the fact that their specificity is predetermined and in this sense they are region or site-specific. In some studies (i,e., where one simply wants to detect a protein) the site at which the antibodies bind is not too important, but in others knowing where they bind is the essence of the study. Again, much power is gained when the antibodies are used as sets. As mentioned above, the influenza virus hemagglutinin is a trimeric structure composed of three molecules each of HA1 and HA2. These proteins are synthesized from a single precursor which is then cleaved to form the two smaller molecules. This cleavage is a prerequisite for formation of infectious virus and part of the process by which the fusion function of the virus is activated. The process of cleavage, of course, generates new C- and N-termini of proteins, which are thought to be involved in fusion. To study these neo-termini we prepared a set of antibodies according to the strategy shown in Fig. 2. These antibodies are now being used to study the process of viral fusion with host cell mem-

12

RICHARD A L m LERNER

FIG. 2. Illustration of the power of sets of peptide antisera used to track distinct regions of a protein simultaneously. (A) The influenza hemagglutinin precursor (HAO) and the eventual products (HA1 and HA2). The synthetic peptides a to e, located within the HA0 sequence as shown (NH2- and COOH-terminal, flanking and spanning the HA1-HA2 junction), were coupled to a carrier and used to immunize rabbits as described (5-7). (B) The resulting antisera (aa, ab, ac, a d , a e ) and a positive control serum (antibody to X:47 influenza, aX:47) and a negative control serum (normal rabbit serum, N R S ) were used to precipitate extracts of [35S]methione-labeled influenza X:47 virusinfected cells. The five antisera to the peptides and aX:47 precipitated the HA0 molecule (this precursor is not appreciably cleaved during infection of cell lines), whereas the other precipitated proteins were nonspecific (sticky) since they also appeared in the normal control lane.


13

branes, but for our present purposes they serve to illustrate the generation of a set of site-specific antibodies to approach a biological problem. Hui and his colleagues also wished to make antibody to the new termini of proteins in the fibrinogen fibrin system in order to generate an antibody which would detect blood clots by reacting with fibrin but not fibrinogen (Hui et al., 1983).They synthesized two peptides representative of the newly exposed amino-termini of the fibrin a-and pchains resulting from cleavage by thrombin. They were able, by this approach, to generate fibrin-specific antibodies, a reagent which heretofore had not been available. Similarly, Sue and Sytkowski (1983) were able to generate antibodies to NH2-terminal regions of erythropoietin. Schneider and his colleagues (1983) have carried out elegant studies designed to demonstrate the external orientation of the low-density lipoprotein (LDL) receptor in fibroblasts. The receptor for the LDL is the major plasma protein involved in the transport of cholesterol into cells. After binding to its receptor, LDL enters into cells in coated pits via the process of receptor-mediated endocytosis (Goldstein et al., 1979; Pastan and Willingham, 1981; Pearse and Bretscher, 1981; Brown et ul., 1983; Drickamer, 1981; Drickamer et al., 1984; Beisiegel et ul., 1981; Anderson et ul., 1982). Because of its signal role in cholesterol metabolism it is critical to learn more about LDL receptors, particularly their orientation on cells. Schneider et al. (1983) made antibodies to a chemically synthesized peptide representative of the first 16 amino acids of the receptor. These antibodies reacted with the LDL receptor in intact, unfixed, skin fibroblasts thereby demonstrating that the NH2-terminus of the molecule is exposed on the external surface of the plasma membrane. One of the medically important as well as technically advantageous aspects of the LDL system is that fibroblasts from subjects with homozygous familial hypercholesterolemia lack the receptor. The antipeptide antisera did not stain these receptor-negative cells, thus conclusively demonstrating the specificity of the reaction. Maloy and colleagues (1984) were interested in a Class I H-2 molecule which is secreted rather than membrane bound (Cosman et ul., 198213; Kress et ul., 1983). The structure was predicted from a cDNA sequence which suggested the presence of a molecule which was homologous to other Class I proteins except for the C-terminal region which had multiple base substitutions and a 13-base pair deletion. These genetic changes altered the translational reading frame resulting in a prematurely terminated protein which lacked the cytoplasmic segment and had sufficient substitutions of hydrophilic amino acids

14

RICHARD ALAN LERNER

substitutes in the transmembrane domain to abrogate its membrane spanning function. These investigators made antibodies to a region near the C-terminus of the predicted protein and have detected a new and intriguing Class I molecule (Maloy et al., 1984).The sera detected a molecule of 40,000 daltons associated with ,&-microglobulin in the sera of six different strains of mice representing five distinct H-2 haplotypes. In addition the protein was concentrated in the liver which is presumably the source of its synthesis. This study represents a combination of approaches in which a new gene product is searched out by making antibodies directed to a unique site predicted by the nucleic acid sequence. In this case the frame-shift in the gene assures a unique antigenic determinant and allows one to get around the usual argument that Class I molecules in the serum are simply sloughed off cell surfaces or from degraded cells. Since gene cloning of the MHC has revealed 30-36 Class I genes (Hood et al., 1983), the success of Maloy and colleagues can be expected to signal many similar studies to prepare antibodies to novel MHC gene products. Semler et al. (1982), Baron and Baltimore (1982), and Morrow and Dasgupta (1983)prepared antibodies against synthetic peptides corresponding to the genome-linked protein of poliovirus (VPg). The antibodies reacted with virus-specific proteins and suggested that a membrane bound form of the viral protein, P3-9, donated VPg to viral RNA (Semler et al., 1982). Johnson and Elder (1983) prepared an antibody to a peptide predicted from the nucleic acid sequence of a replication competent, recombinant murine leukemia virus. Since the peptide was specific to a nucleic acid region acquired during the process of recombination, the antibody probe was specific for one particular recombinant envelope protein (gP70). Interestingly, this protein was found in murine thymocytes but not splenic T cells or bone marrow cells, thus giving further credence to the idea that the thymus plays a specific role in generating recombinant leukemogenic retroviruses. VIII. Structure-Function Studies

Since one can make site-specific antibodies to virtually any region of a protein, the possibility has been opened for the fine structure mapping of regions of proteins involved in biological or enzymatic functions. This approach is exemplified by the thorough studies on the transforming proteins of Rous sarcoma virus (pp60"'").Gentry et al. (1983) used antibodies to positions 498-512 and 521-526 of pp60 to study the kinase site of the molecule. Immunoprecipitates prepared


15

using anti-498-512 failed to phosphorylate exogenously added substrates, whereas precipitates prepared with antibody to position 521526, a region only six amino acids away, retained the enzymatic activity. I n a somewhat reciprocal experiment, Tamura et al. (1983) showed that in the immune kinase reaction, antibodies to peptides 103-108 and 155-160 are mostly unphosphorylated, whereas antibodies to 315-321,409-415, and 500-506 are phosphorylated to an extent similar to conventional polyclonal sera. Thus, both of these studies suggest that binding around amino acid 500 perturbs the function of the molecule, a finding consistent with the hypothesis that this region of the molecule is part of the active site. Similarly, other antipeptide antibodies have been used to inhibit the functions of proteins. Antisera to the feline sarcoma virus ( f e s )gene sequence (Sen et al., 1982) and the middle T sequence of polyoma virus inhibit protein kinase activity (Schaffhausen et al., 1982). Antisera to several regions in the NHz-terminal half of the Moloney leukemia virus polyprotein inhibit reverse transcriptase enzymatic activity whereas antibodies to peptides from the COOH-terminal portions inhibit the virus-associated endonuclease activity (Sutcliffe et al., 1983). Baron and Baltimore (1982) used antipeptide antibodies to p63, a core protein of poliovirus predicted from the replicase gene, to inhibit the replicase and polyuridylic acid polymerase activities indicating that both activities reside in p63. IX. Antibodies to Proteins Encoded by Alternative Reading Frames

A variation on the idea of using antipeptide antibodies to find the products of new genes is their use in the assignment of reading frames and the detection of the products of overlapping genes. The most comprehensive study of this nature has been carried out by Mariothini and colleagues (1983)for the human mitochondria genome. When the sequence of the mammalian mitochondrial DNAs was completed, there was the discovery of eight reading frames, each over 200 nucleotides long, which did not encode any of the known translation products (Anderson et al., 1981, 1982; Bibb e t al., 1981; Grosskopf and Feldman, 1981).Mariothini et al. (1983)were interested in the protein product of an unidentified reading frame (URF A6L) which was 207 nucleotides long and overlapped in an alternative frame the ATPase genes by 46 nucleotides. Antibodies against two peptides from the putative amino- and carboxy-terminus of URF A6L precipitated a protein of 4500 daltons from HeLa cell mitochondrial translation products.

16

RICHARD ALAN LERNER

The other use of antibodies to alternative reading frames is to study frameshift mutations. Such antibodies can be thought of as anti“ wrong” reading frame antibodies. S. Sen (personal communication) used the nucleotide sequence of several transforming viruses to predict the protein sequence which would correspond to +1 and +2 frameshifts. Sequences were selected far enough to the 3’-end of the genome so that termination codons did not occur. These antibodies were indeed capable of detecting the frameshifts and the studies so far carried out indicate that a surprisingly high rate of frameshifting occurs. These approaches should now be extended into a variety of systems including human mutations. Aside from their uses for theoretical and medical studies, antibodies to alternative reading frames should prove useful in following the fate of transfected genes altered by site-specific mutagenesis. By using sets of antibodies one could compare the fate of wild-type and mutant proteins in the same cells. X. Exon Usage

The genes of higher cells are a conglomerate of coding regions (exons) with interspersed noncoding regions (introns). By a process of RNA splicing, messenger RNA is assembled from larger transcripts. This lack of a direct one-to-one relationship between the DNA sequence and the final gene product sometimes makes it difficult to determine which regions of DNA encode the protein in question. Also, as for the immunoglobulin genes, which encode the secreted and membrane-bound forms of the molecule, different exons may be used at different times. Antibodies of predetermined specificity are ideally suited to follow exon usage during gene expression. Shinnick and Blattner (personal communication) followed exon usage in the IgD system by using antibodies to peptides representative of exons specific for the secreted or membrane-bound form of the molecule. The exon usage in the adenovirus-2 E1A transcription unit has also been studied (Feldman and Nevins, 1983; Green et d.,1983b). The E 1A region encodes functions which both regulate the expression of other genes and play a role in cellular transformation (Berk et al., 1979; Jones and Shenk, 1979a,b; Graham et al., 1978).The large E1A transcript is processed into two overlapping messenger RNA molecules (12 S and 13 s) which differ by 138 nucleotides. The two messenger RNAs are in the same reading frame and translation is initiated at the common first AUG. Thus, it has been assumed that the two proteins encoded by these messages have a common N- and C-terminus and differ only by the 46 amino acids encoded by the 138 nucleo-


17

tides unique to the larger message. Feldman and Nevins (1983) made an antibody to a peptide corresponding to the most hydrophilic residues of the putative 46 amino acids unique to the large protein. Indeed, this antibody reacted only with the larger of the two proteins, thus providing a reagent which should be useful in sorting out the role of E1A encoded proteins in transformation and control of transcription. XI. The Chemistry of Virus Neutralization

Although immunological prevention of viral disease is one of the most time honored processes in medicine, the chemistry of virus neutralization is hardly understood. The basic problem has been that not all antibodies that bind to viruses neutralize them and thus it has been difficult to use populations of antibodies to sort out the problem. For example, something as seemingly simple as the serology of viral type specificity has not been easy to comprehend at the molecular level. The advent of monoclonal antibodies offered some approaches to the problem. The usual way in which monoclonal antibodies are used to understand virus neutralization is to use an approach which can be referred to as antibody escape experiments. Basically, these experiments involve growing virus in the presence of a neutralizing monoclonal antibody to select for variants which escape neutralization. The relevant proteins of the variants and wild-type strains are then sequenced (usually by nucleic acid sequencing) and the observed changes are then said to be part of the antigen binding site for the antibody. This approach has been used successfully for viruses like polio and influenza and continues to yield interesting information. The alternative approach, of course, is to use antipeptide antibodies. Since the site on the virus or viral proteins to which these antibodies bind is known, the chemistry of viral subtype specificity and neutralization can be learned directly. One of the first systems in which a thorough study was carried out was for the hepatitis B virus. Since this virus does not replicate in vitro alternative approaches were not available. The important protein target of neutralizing antibody in this system is the product of the S gene. The S gene product [hepatitis B surface antigen (HBsAg)] is a single polypeptide, of 226 amino acids, the sequence of which has been determined from the nucleotide sequence of the S gene (Tiollais et al., 1981; Valenzuela et ul., 1979; Galibert et ul., 1979; Pasek et al., 1979). The HBsAg protein consists of a group-specific (a) and two sites of subtype-specific determinants (d/y, w/r) so that four types of viruses

18

RICHARD ALAN LERNER

exist (adw, adr, ayw, and ayr) (LeBouvier, 1971; Bancroft et al., 1972; Gold et al., 1976; Shih et al., 1978). The protein sequence of HBsAg between position 110 and 137 is the most variable and thus a candidate for a type-specific determinant (Tiollais et d., 1981; Valenzuela et al., 1979; Galibert et al., 1979; Pasek et al., 1979; Lerner et al., 1981). Gerin and colleagues showed that chemically synthesized peptides corresponding to region 110-137 could duplicate the d/y specificities and induce subtype-specific antibodies even in chimpanzees which are the relevant human model of the disease (Gerin et al., 1983). Initially, they synthesized peptides of 28 amino acids which differed by seven residues. By using shorter peptides the site of the y determinant was localized to one or more amino acids at positions 127, 131, and 134 of the S gene product. In other words, by using antibodies of predetermined specificity it was learned that the y/d serology is played out over no more than 3 of the 226 amino acids of the proteins. There have been a number of other experiments using synthetic immunogens to study the HBsAg. Dreesman et al. (1982) studied cyclic peptides between 117-137 and 122-137 and concluded that the cyclic form was beneficial in eliciting antibodies reactive with native HBsAg. Although cyclization may generate novel specificities it is obviously not necessary to generate protein reactive antibodies since linear peptides workas well (Lerneret al., 1981; Gerin et nl., 1983).The issue of cyclization of peptides comes up often, probably because of the loop of lysozyme experiments. It is necessary to remember that a loop structure worked in lysozyme because the region in question of the native protein was, in fact, a loop. However, lacking structure evidence for a loop structure, there is no evidence that cyclization of peptides is generally helpful. The studies of Prine and colleagues stand in conflict with all the other studies of the subtype specificities of synthetic HBsAg immunogens (Prine et al., 1982). They concluded that a peptide encompassing residues 138-149 contained the d determinant, whereas Bhatnagar et al. (1983) concluded that the same region was an essential part of the a determinant. There is nothing to say that a region of a protein cannot represent two serological determinants, but in view of the agreement between the Gerin and Dreesman studies it is probably correct to assign the d/y determinants to the Cterminal portion of peptide 110-137. The synthetic immunogen technology has led to a thorough understanding of the chemistry of subtype specificity and virus neutralization for the foot-and-mouth disease (FMD) virus. In fact, the chemistry of virus neutralization is better understood for this than any other virus. FMD is a highly contagious affliction of cloven-hooved animals


19

which is caused by a single-stranded positive sense RNA genome of about 8000 nucleotides. The virus has seven distinct serotypes which can be further divided so that at least 60 subtypes of the 7 serotypes are known. Thus, this is a viral system of enormous antigenic variation which can be contrasted, for example, to another piconovirus, polio, in which there are only 3 important strains. In FMD, infection with one serotype does not confer protection against any others. Thus this viral system is, at once, an interesting model and a challenge for the synthetic immunogen technology. Basically, it combines antibodies of exquisite specificity with a viral target capable of great variability. The major target of neutralizing antibody in the FMD system is the VPI protein which occurs in 60 copies on the virus capsid. Bittle et al. (1982) synthesized a number of peptides predicted from the nucleotide sequence of Kurz et al. (1981) and tested the antibodies to each for their ability to bind to and neutralize the virus. Whereas antibodies against all peptides bound to the virus, only those to region 141-160 neutralized the virus well. Pfaff et al. (1982)reasoning on the strength of a predicted surface helix of VPI found that a hexapeptide between residues 144 and 149 elicited high titers of neutralizing antibodies. Furthermore, these authors showed that the site represented by peptide 144-149 is also a major immunogen of intact virus since it could be used to absorb out a significant amount of antibody made against the intact FMD virus. Together, these studies make a significant point about the chemistry of virus neutralization. Simply stated, there is a difference between antibody binding and neutralization and only some of the antibodies which bind to viruses are capable of neutralization. This, of course, makes great sense because if all or even a majority of antibodies to the surface of a virus were capable of neutralizing then a virus could not escape the immune system b y changing one or even a few amino acids. Furthermore, as we will next see, subtype specificity is a hand-in-hand companion of a viral evasion of the immune system and this game is played out over few amino acids. Recently an experimental system emerged which allowed an even closer look at the chemistry of subtype specificity and viral neutralization for FMD. Although, in general, antipeptide antibodies neutralized the appropriate serotypes of FMD virus, the sera against type A from the USA neutralized the Pirbright strains of A type virus poorly. The pedigree of these two A strain viruses is interesting. The Pirbright A12 virus, the parent of that used by American workers (Kleid et al., 1981), had been passaged only once in BHK cells whereas the American (USA) virus had been passaged and plaque purified multiple times. When the USA virus was returned to the

20

RICHARD ALAN LERNER

Pirbright laboratories in Surrey a proper match was achieved, in that the antipeptide antibodies based on its sequence neutralized the USA virus (Rowlands et al., 1983). In an attempt to understand this situation, nucleotide sequencing was carried out on the USA and Pirbright viruses (Kleid et al., 1981). The nucleotide sequence of the USA virus agreed exactly (except for a Leu codon at position 212) with that previously reported (Bittle et al., 1982) whereas the sequence of the Pirbright virus was ambiguous at nucleotides representing amino acids at positions 148 and 153 of the virus. These findings were of considerable interest from four points of view. First, they suggested that more than one strain of virus was present in the Pirbright stock. Second, since the virus was only one passage away from the natural host, a detailed study of the virus might offer insight into the epidemiology of the FMD strain variation. Third, the variation was largely confined to positions 148 and 153 which, it should be recalled, fall into the site (141-160) previously found to induce strain-specific neutralizing antibodies (Bittle et al., 1982; Pfaff et al., 1982). Fourth, a combined immunological and biochemical study of the presumed variants might offer insight into processes by which these viruses evade the immune system as well as the chemistry of virus neutralization. Accordingly, the Pirbright virus stock was plaque purified, 10 separate clones were isolated, and their nucleotide sequence was determined. Of the 10 clones three had serine at 148 and leucine at 153, five had leucine at 148 and proline at 153, and two had serine at both 148 and 153, whereas the USA virus had Phe at 148 and Pro at 153 (Brown et al., 1983). Peptides corresponding to positions 141-160 for each of the viruses were chemically synthesized and antisera prepared against these. In each case, the antisera neutralized the homologous better than the heterologous viruses. Furthermore if antisera to “nonsense” peptides were prepared (i.e., leucine at 148 and 153 and phenylalanine at 148 and leucine at 153) they did not neutralize any of the viruses (Pfaff et al., 1982).These results emphasize the significance of positions 148 and 153 as components of strain variation and show their critical importance in reshaping their cognate regions of VPI during virus escape from immune attack. Obviously, some changes in these positions are more effective than others in allowing the virus to escape the immune system. For example, a leucine + proline change at 153 might effect the main chain direction and alter neutralization. Indeed, the synthetic peptides could be shown to be a good mimic of the natural serology since antibodies against the intact viruses, like the antipeptide antibodies, neutralized the homologous viruses much bet-


21

ter than the heterologous viruses. Finally, these studies on FMD viruses have significance for our understanding of virus variation in the field. It was surprising to find that the FMD variants were in approximately equal number, especially since the variations involved only two amino acids. This should be contrasted to the situations where mutants are selected in the laboratory under host range or immunological pressures. In these situations the frequency of selection of variants is on the order of lo4. It is not yet understood how these variants are maintained in FMD but their presence in equal numbers suggests that the changes are not disadvantageous to viral replication and are probably advantageous to evasion of the host immune system. The particulars of the FMD system aside, it is heartening to understand the chemistry of virus neutralization in such detail. The detailed X-ray structures of this and other picornoviruses such as polio (Hogle, 1984) are eagerly awaited so we can understand even more about the structure and chemistry of virus neutralization. Using antibodies of predetermined specificity we can already say what amino acids change during virus escape. When the structures are in we will be able to turn the rachet one more notch and speak of shape changes in proteins during immunological escape. At that point vaccine production will move from an empirical to a design science. There has been success in using synthetic peptides which generate antitoxin antibodies (Audibert et al., 1981; Boquet et al., 1982) in the diphtheria system. Very interesting results have recently been achieved in the malaria system. Curiously, the sporozoites of the Plusmodium knowlesi strain and Plasmodium falciparum strain of malaria have surface proteins which consist of 12 repeating 12-mers and 23 repeating ll-mers, respectively (Coppel et al., 1983; Godson et al., 1983).In both systems antipeptide antibodies reacted with the surface of sporozoites, offering hope for a synthetic vaccine against sporozoites. XII. Synthetic lrnrnunogens Representing Idiotypes, Allotypes, and Growth Factors

Although the studies described above have detailed the ability to make antibodies to variable portions of proteins, we deal separately here with the variable regions of proteins which are of special interest to the immunologist. Also included in this section are studies using antipeptide antibodies to study lymphokines. Perhaps the most telling study concerning the ability of antipeptide antibodies to achieve fine specificity is that of Alexander et al. on the

22

RICHARD ALAN LERNER

Thy-1 glycoproteins (Alexander et aZ., 1984). The Thy-1 glycoprotein exists in two allotypic forms which differ from each other by a single residue (Reif and Allen, 1964). The protein with an arginine residue at position 89 is thought to correspond to the Thy-1.1 allotype and a glutamine residue to Thy-1.2 (Williams and Gagnon, 1982). Since the two proteins differ by only a single amino acid they offered a good test system to study the ability of antipeptide antibodies to discriminate between two closely related proteins. Based on the protein sequence (103) six peptides were synthesized. Four peptides corresponded to the constant region of the molecule, whereas two spanned the region between amino acids 79 and 98 which correspond to the only known differences between the 1.1and 1.2 alloantigenic forms of the Thy-1 molecule. Antibodies to the peptides from the conserved regions of the protein were reactive with both Thy-1.1 and 1.2, whereas antisera to peptides spanning the variable regions which were predicted to be from Thy-1.2 and 1.1 molecules showed the corresponding preference for thymus extracts of the C57BL/6J (Thy-1.2) and AKR/J (Thy-1.1) mice. The peptides utilized in this study are of sufficient size that each could contain multiple antigenic determinants. Indeed, these peptides have 18 amino acids in common and differ with respect to only one. Nevertheless, the presence of a single distinguishing amino acid is sufficient that a significant portion of the immune response toward each peptide is specific, resulting in preferential reactivity of the antisera with their inducing peptides. Hence, the free peptides closely mimic the differential immunogenicity of the native Thy-1 alloantigens, and confirm the suggestion that amino acid 89 is involved in Thy-1 alloantigenicity. Of the two chemically synthesized sequences, the one with Arg at position 89 elicits antibodies with a higher degree of specificity than the one with Gln. Thus, amino acid changes can either render or abolish differential immunogenicity. In the present study replacement by GZn by Arg at position 89 diminishes differential immunogenicity and the resultant peptide appears now to be composed of a collection of more equally potent antigenic determinants. It is interesting to point out that the same phenomenon has been observed in other systems as well. In the hepatitis B surface antigen (HBsAg)system we constructed a peptide in the region of amino acids 110-137 which induced antibodies with a strict specificity for the adw subtype of HBsAg. The substitution of two amino acids within this peptide (asparagine and phenylalanine at positions 131 and 134 were


23

replaced by threonine and tyrosine, respectively) resulted in a peptide with a broader immunological specificity. The findings in the Thy-1 system are of theoretical interest from several points of view. First, single amino acid changes can “control” the immunogenicity of relatively long peptides. This is, perhaps, not too surprising for the induction of antibodies which react with the free peptide, but when we realize that the antibodies also discriminate between the intact folded proteins some interesting concepts emerge. Obviously, since allotypic forms of Thy-1 exist, one already knew that the immune system could differentiate the single amino acid change in the context of the folded protein, but this could have been due solely to complex changes in conformation caused by the Gln and Arg change. The data show that the differential immunogenicity of Thy1.1 and Thy-1.2 can be due rather to a local perturbation in structure effected by this change. An important point which emerges from this study is that even a single amino acid substitution can result in the formation of allotypic antigens where the allotype can be detected by antisera in different ways. For example, the antipeptide antibodies efficiently detect both the synthetic peptides and the isolated allotypic Thy-1 glycoproteins, whereas monoclonal anti Thy-1 allotypic antibodies are efficient in the detection of these intact alloantigens, but d o not detect these synthetic peptides (H. Alexander et al., personal communication). Another difference between these antipeptide antisera and monoclonal anti Thy-1 antibodies is that the latter are much more efficient in the detection of cell-surface Thy-1 alloantigens. This is not surprising since, classically, monoclonal and monospecific anti Thy-1 alloantibodies were selected for their ability to bind to the cell surface of Thy-1. There are advantages to the antipeptide antibodies for certain practical uses. It is theoretically possible to use this technique to produce antibodies to other single aniino aciddefined polymorphic and mutant polypeptide sequences and to use such antibodies to detect gene sequences encoding them. For example, there is preliminary evidence that the antipeptide antisera readily detects Thy- 1 antigens in a bacteriophage expression system into which a T cell-derived cDNA has been inserted, while in contrast, the monoclonal antibodies fail to do so (Rosen et al., unpublished data). Technical details aside, the important question is why should a single amino acid change similarly affect the immunogeiiicity of a peptide in solution and the cognate structure in the internal region of a folded protein, where its conformational flexibility is likely to be constrained by bonds with neighboring structures. One possibility is

24

RICHARD ALAN LERNER

that the peptide in solution has a favored conformation which closely resembles that of the folded protein. This is, however, inconsistent with the body of evidence which suggests that in aqueous environments, even relatively long peptides have very little ordered structure (reviewed in Goetzl and Peters, 1972). This and other theoretical aspects of the problem will be discussed below. Altman and his colleagues prepared antipeptide antibodies reactive with human interleukin 2 (Altman et al., 1984). They chemically synthesized 8 peptides, each consisting of 13-15 amino acids. Antibodies to four of the eight peptides reacted with native IL-2 as judged by biochemical and immunological criteria. Furthermore, the antisera were able to react in situ with IL-2 producing cells. Cells stimulated with PHA but not control cells were stained with affinity purified rabbit antibodies against the synthetic peptides corresponding to IL-2. It seems reasonable to suspect that the IL-2 story is only the tip of the iceberg as far as factors which regulate immune functions are concerned and thus one might expect antibodies of predetermined specificity to play an increasing role in unraveling the functions of these factors. This is especially true for two reasons. First, it is likely that the factors will come to light via nucleotide sequencing, and thus predicted protein sequences will be known in advance of any formal protein chemistry. Second, as for other hormone systems there are likely to be polyproteins involved and these antibodies will be useful in sorting out precursor-product relationships. McMillan et al. (1983), Chen et al. (1984a), and Seiden et al. (1984) have chemically synthesized idiotypic determinants of mouse and human immunoglobulins. Working with the third hypervariable region of murine antidextran antibodies, McMillan and colleagues synthesized peptides corresponding to sequences in the M104 and 5558 myeloma proteins. Rabbit antisera to these peptides reacted only with heavy chains and appropriately discriminated between the M 104 and 5558 proteins. Chen et al. (1984a) studied the monoclonal IgM rheumatoid factor (Sie) using the sequence of Andrews and Capra (1981).A peptide corresponding to amino acid residues 99-111 reacted with the IgM Sie protein but not with four other IgM rheumatoid paraproteins or with pooled human IgG. These results indicate that in this system the antipeptide antibodies are reacting in this system with a private idiotype on the heavy chain of the Sie rheumatoid factor. In a separate study Chen et al. (198413) chemically synthesized peptides corresponding to the “Wa” marker on the complementarity determining region (Kunkel et al., 1973,1974) of human monoclonal IgM rheumatoid factor (RF) autoantibodies. The Wa marker is of particular


25

interest because it is related to light chains (Kunkel et ul., 1974; Andrews and Capra, 1981; Carson and Fong, 1983)and is found in 60% of monoclonal IgM-RF proteins. The antipeptide sera were specific for the light chains, the IgM-RF paraproteins, Sie, Glo, and Teh but not Lay, confirming the reactivity with the “Wa” idiotypic determinant. The experiments of Seiden and colleagues have yielded perhaps the most surprising results concerning synthetic idiotypes. They synthesized a 16 amino acid peptide corresponding to a sequence in the JH1 segment of the murine anti-a (1-3)-dextran immunoglobulin family. Of the 20 anti-a (1-3)-dextran antibodies so far sequenced 13 contain JH1, 5 contain JHz, 1 JH3, and 1JH4 (Clevinger e t al., 1980a, 1981; Shilling et al., 1980). Seiden thus expected to have an antibody which uniquely bound to JH1 containing antidextran antibodies. The surprising result was that all 16 antidextran antibodies tested bound to the antibody whereas 14/15 antibodies with different antigenic specificities did not (the exception was HOPC-1). These authors labeled this idiotype JH-Dex. The reason for the JH-Dex specificity with the antipeptide antibodies is not yet clear, but Seiden et ul. have suggested that it may be related to the very short D segments in the antidextran antibodies. Some support for these notions can be gleaned from computer graphics analysis of the crystal structures of solved immunoglobulin molecules (A. Olson, unpublished). Usually, the D segment amino acids would “cover” the J region, and thus the shortening of the D segment in antidextran antibodies may make more of J accessible to antiidiotypic antibodies (Seiden et al., 1984). Alternatively, any change in D could alter in other ways the conformation and/or the availability of the J region so as to make union with antibody more efficient. In toto, the four experiments discussed above indicate that the sitespecific nature of antibodies made against synthetic idiotypes will allow one to work out the chemical basis of idiotypic markers. Sitespecific antiidiotypes could also be used to perturb the antigen binding functions of antibodies and thus lead to a better understanding of the structural correlates of antigen-antibody union. Finally, synthetic idiotypes could be of considerable medical importance. For example, in those autoimmune diseases where the offending antibody is of restricted origin, it may be possible to use synthetic immunogens to modulate or even eliminate the clones producing the antibody. Alternatively, some antiidiotypes may block or obstruct the binding pocket so as to preclude union between the injurious antibody and its antigen. Several points illustrated in the above studies are significant as far as the potential medical applications of synthetic idiotypes are

26

RICHARD ALAN LERNER

concerned. First, antiidiotypes induced by synthetic immunogens react with native immunoglobulins. Second, synthetic idiotypes offer a degree of specificity which could not be obtained by immunizing with the intact immunoglobulins. Third, and perhaps most important, because synthetic peptides can induce antibodies to regions of a protein not ordinarily recognized (see above), the possibility exists for terminating tolerance to restricted regions of self-proteins. The finding and synthesis of a common idiotype for a human autoantibody may open up an interesting route to therapy of diseases of monoclonal origin that are not under “antigen-drive.” The difficulty in dealing with these diseases under “antigen-drive’’ is that if you could use synthetic peptides to immunize against a clone with one idiotype, upon its disappearance the immune system may reward you with an even more harmful clone (i.e., one with a slightly different antibody of higher affinity. XIII. The Structure of an Antigenic Determinant in a Protein

One of the most powerful aspects of the synthetic immunogen technology is that it offers a means to understand the chemistry and structure of antigenic determinants in proteins. The key phrase is antigenic determinants in proteins as opposed to immunogenic determinants of proteins. Almost all previous studies of protein immunogenicities depend on defining antigenic determinants after immunization with the whole protein. Consequently, it is difficult to localize precisely the antibodycombining site and to define the extent and number of amino acids directly involved in antibody-antigen union. Changes of amino acids in antigenic determinants which alter antibody binding and specificity may not always be easy to interpret in the context of complex structures. For example, changes in one region of the protein may in some circumstances alter the conformation of quite distant regions. The generation of antipeptide antibodies which have predetermined sequence specificity would seem to offer an alternative approach to the problem of understanding the general nature and structure of antigenic determinants and antibody-antigen interaction. The study of antigenic determinants in a protein defined by antipeptide antibodies differs from the study of immunogenic determinants of proteins which depend on immunization with the intact protein, in that the latter determinants are often more complex and less amenable to structural study. Wilson and colleagues used monoclonal antipeptide antibodies to study the chemical nature and structure of antigenic


27

determinants in the influenza virus hemagglutinin (Wilson et al., 1984). This study is detailed here as a prototype for the structure of an antigenic determinant in a protein. Parts of the text which follows and Figs. 3-8 are reproduced with permission from Cell. Monoclonal antibodies with predetermined sequence specificity were generated by immunizing mice with a synthetic peptide coupled to KLH. The synthetic peptide represents residues 75-110 in the HA1 chain of A/Victoria/3/75 (X:47, H3 subtype) hemagglutinin (Min Jou et al., 1980) when aligned with the A/Aichi/2/68 (X:31, H3 subtype) sequence (Verhoeyen et al., 1980). Twenty-one different monoclonal antibodies were raised as described in Niman et al. (1983). The majority of these antibodies (16 out of 21) cross-reacted strongly with X:47 influenza virus. Thus these antibodies recognized the peptide against which they were raised as well as the whole virus. The synthetic peptides which specify the binding sites (1-4) are shown in Fig. 3. Wilson et al. (1984) synthesized several smaller, overlapping peptides to localize the sequences to which antibodies bind with binding titers approximately equivalent to that of the parent protein. Eight of the 18 monoclonals reacted with peptides localized to one sequence of the parent molecule. These hybridomas bind to residues in the 36 amino acid synthetic peptide that contains HA1 residues 98106. Three hybridomas also bound in this region but could not be localized further than to sequence 88-1 10. Three antibodies could be localized to bind to sequences on the amino-terminal end of the longer peptide. These binding sites overlap and are identified as site 1, 75-86, site 2, 79-86, and site 3, 83-92. The immunodominant region, site 4, which corresponds to sequence 98-106 is shown in Fig. 4. This antigenic determinant corresponds to no more than 25% of the total amino acids in the sequence 75-1 10 and in this conformation has an accessible surface area of 763 A' (24% of the total). Site 2, residues 79-86, has an accessible surface area of 800 when calculated in the conformation it has in the native protein. The dimensions of sites 2 and 4 are approximately 13 x 22 x 12 and 12 x 22 x 14 A, respectively. These structural renderings represent an idealized situation for envisaging recognition by antibody of both the free peptide and the intact protein. As discussed by Wilson and colleagues, analysis of the structure of the antigenic determinants of the synthetic peptide is hampered at present by uncertainty about the conformation of the free peptide in solution. However, many of these difficulties are abrogated by analyzing the structure of the peptide and the antigenic determinants in the

Wz

28 PEPTIDE

SEQUENCE #

RICHARD ALAN LERNER

75

80

85

90

95

100

105

110

HCDGFQNEKWDLFVERSKAFSNCYPYDVPDYASLRS

# HA 76-110 52- 88

I 2

Y

3

75-

4

76-

5 8 7

75757983-

8 9

82 84 88

89

87 92

83-1 10 87- 94 88-110

LO 11

I2 13

91- 99 93-1 10

14

95-102

16

97- 108 98- 110 98-104

I8 17

98- 105 98-108 99-108 100-108 100- LO8 103-110 104-139

I8 19

20 21 22 23

24

SITE SITE SITE SITE

1 2 3 4

75- 88 7983-

88 92

We-108

FIG. 3. The amino acid sequences of the HA1 75-110 peptide and the 23 nested fragment peptides used in the experiments are shown relative to each other. Peptides 2 and 24 actually extend below residue 75 or beyond residue 110,respectively. Three peptides (Nos. 2,3,6) have a tyrosine prepended or appended to the fragment, designated with a Y. Sites 1-4 correspond to the locations where the vast majority of monoclonal antibodies bind.

native protein. Since the antibodies have approximately equivalent binding titers with the free peptide and intact protein, it is reasonable to assume that the antigenic determinants are shared by both. The location in the hemagglutinin molecule corresponding to the 36 amino acid chemically synthesized peptide is central in the globular head region of the HA1 polypeptide chain (Fig. 3, Wilson et al., 1981). This peptide consists of six pieces of extended polypeptide chain connected by a series of bends. The amino- and carboxyl-ends


29

of the chain contain three bends and a helix. The peptide contains residues which are either on the surface (75-83,91-96), buried in the monomer (83-91,97-99,108-110) or in the trimer interface (100-107) (Fig. 5). The sequence 83-91 contains the largest stretch of residues of the peptide inaccessible to ligands in its native conformation. When considered as part of the native monomeric protein, the peptide has an accessible surface area of 1293 A2, which corresponds to 41% of that of the free peptide. The accessible surface area of the peptide in the trimeric hemagglutinin molecule (Fig. 6) is even more limited. On the outside of the molecule, the peptide thus presents a continuous surface (Fig. 8, blue surface), where residue 75 is close to residues 96 and 97, such that residues 96-97 and 75-83 form a continuous belt of accessible surface. The residues which are in the trimer interface (100-107, Fig. 8, white surface) are accessible only to water molecules, small ligands or molecules with small protruding surfaces. Chemical identification of the residues involved in recognition of the synthetic peptide by the monoclonal antibodies together with analysis of the location and conformation of the peptide in the native structure were combined to describe the structure of the antigenic determinants in the protein. The peptide 98-106 is detailed here as this is the sequence which appears to be immunodominant. This sequence encompassing residues 98-106 lies in the trimer interface and has accessible surface only to small ligands for residues Tyr 100, Asp 101, Pro 103, Asp 104, Tyr 105, and Ala 106 (Fig. 7). The peptide binding studies indicate that the binding can probably be localized to residues 100-106. These residues form a fairly flat, slightly convex surface, embedded in the rest of the protein, with approximate dimensions 16 x 15 x 7 A. This peptide contains two prolines, two aspartates, and two tyrosines and may be more conformationally restricted due to the more limited torsional angles of proline residues. Unless the antibody hypervariable loops can protrude into the cavity in the trimer interface, the structure and binding data indicate that the sequence would be more accessible to an antibody binding to the monomer. Wilson et al. (1984) suggested that antibody binding occurs to a structure in which the hemagglutinin monomeric heads are exposed and which represents a conformation different from the native hemagglutinin trimer. Thus, an important conclusion from these and previous studies is the suggestion that antipeptide antibodies can have access to structures other than the native conformation, and the protein displays

30

RICHARD ALAN LERNER

conformational mobility at least in local regions. Indeed, conformational changes in the hemagglutinin molecule have been identified previously. For example, at low p H (pH 5.0-pH 5.5, Skehel et al., 1982) the molecule becomes accessible to cleavage by trypsin to obtain monomer head structures with altered antigenic activity (Daniels et al., 1983). In addition, unless amino acid substitutions affect distant regions in the molecule, one of the sites in the hemagglutinin to which antibodies against the intact molecule bind (site D, Wiley et al., 1981) would also require a conformational change in the hemagglutinin. Site D is buried in the trimer interface and, in terms of its inaccessibility, is similar to site 4 in the trimer and site 3 in the monomer. The complete determination of the structure of these and other antigenic determinants awaits the determination not only of the peptide conformation but also of the conformation of the peptide and intact protein when complexed with an antibody. Such studies are in progress in Wilson’s as well as other laboratories.

FIG.4. The solvent-accessible surface of the peptide 75-110 is shown on a raster display unit, The grey transparent surface reveals the ball-and-stick representation of the atoms of residues 98 to 106 of the peptide. The conformation of the peptide is that it would have in the native hemagglutinin. The carbon, nitrogen, and oxygen atoms are colored green, blue, and red, respectively. The nontransparent yellow surface covers the rest of the 75-110 region. FIG.5. The location of the peptide sequence 75-110 is shown on the schematic diagram of the X-31 influenza hemagglutinin monomer. The HA1 chain is shown in blue, the HA2 chain is shown in cyan. The peptide 75-110 is shown in yellow with the subsequences 79-86 and 98-106 highlighted in orange. The numbers of the a-carbons are also shown. The figure, as with figures 3-8, was generated using the coordinates provided by Wilson, Skehel, and Wiley (unpublished). FIG.6. The hemagglutinin trimer is shown looking down the 3-fold axis of symmetry from the end furthest from the membrane with the solvent accessible surface of the peptide 75-110 represented by dots. HA1 is colored blue, HA2 cyan. The surface of subsequence 98-106 is shown in red while the rest of the peptide’s surface is in yellow. FIG.7. The solvent-accessible surface around the subsequence 98-106 in the HA1 chain of hemagglutinin is shown in a closeup view on a raster display terminal. The grey transparent surface reveals the ball-and-stick representation of the atoms of residues 98 to 106 of the peptide. The carbon, nitrogen, and oxygen atoms are colored green, blue, and red, respectively. The nontransparent blue surface covers the rest of the 75-110 region, the yellow surface covers the HA1 chain around the 75-110 sequence and the cyan surface covers the nearby HA2 chain. FIG.8. The solvent-accessible surface of the hemagglutinin dimer highlighting the position of peptide 75-110 and its subsequence 98-106 is shown on a raster display unit. The left monomer shows HA1 in red, HA2 in grey; the right monomer shows HA1 in yellow and HA2 in cyan. On both monomers, residues 98-106 are colored white and the rest of the residues of peptide 75-110 are colored blue.

FIG.4.

FIG.6.

FIG.7.

FIG 8.


31

It is interesting to consider the nature of antigenic determinants in terms of what is known about antigen-combining sites of immunoglobulin molecules. The structural determination of multiple myeloma proteins (reviewed by Amzel and Poljak, 1979) of intact IgG molecules or fragments such as Fab have shown the hypervariable regions to be located around a cleft at the distal ends of the molecule. Binding studies with multiple myeloma proteins have shown that small ligands such as vitamin K (Amzel et al., 1974) or phosphorylcholine (Segal et al., 1974) bind to antibodies in this region but occupy only a small portion of the potential antigen-binding site. Characterization of the binding sites of hybridoma antibodies specific for (1-6) linked dextran showed that an antigen of about 5-7 hexoses was complementary to the binding site of IgA and IgG antibodies (Sharon et al., 1982).Wilson et al. calculated the accessible surface of the pocket formed b y the hypervariable residues of F e b Neu (Amzel et ul., 1974) and estimated the potential binding site to be 44 x 35 x 34 A in dimension with an area of 2400 A.2 The antigenic determinants of the free peptide reported by them have dimensions in site 1 of 18 X 23 X 14 A, site 2 of 13 x 22 x 12 A, site 3 of 29 x 13 x 11 A, and site 4 of 12 x 23 x 14 A and with accessible surface areas in the protein of 850-1160 A?-. Thus, these determinants would occupy a sizable portion of the antigen-binding site of the antibody molecule. XIV. Technical Aspects: The Only Rule Is That There Is No Rule

The question often arises as to which peptides should be made to generate protein reactive antibodies. It is going to be difficult to come up with generalities since what is learned about one peptide will not necessarily relate to a different peptide or even a similar peptide in the context of a different protein. My own view is that with certain exceptions (see below) the main issue which should guide selection of the peptide is the precision with which a given antibody can answer the experimental question at hand. As the technology has moved from simply a way to access a protein predicted by a nucleic acid sequence to a site-specific methodology, the experimenter often has a favored site in the protein against which hetshe wishes antibody to be generated. There are now enough examples where this has worked to suggest that one should simply work exactly with the regions of interest. So even if an area adjacent to the primary candidate seems more attractive (i.e., more hydrophilic) one should (within reason) serve the needs of the experiment rather than attempt to guess about the limited to nonexistent “rules” governing immunogenicity. Having

32

RICHARD ALAN LERNER

said that, we can consider a few hints. The incorporation of hydrophilic amino acids in the peptides is useful from two points of view. First, peptides containing such amino acids are likely to be soluble and thus easier to work with. Second, as mentioned above, even if one could generate antibodies to very hydrophobic peptides, the cognate sequence in the protein is likely to be buried and thus inaccessible to antibody. Another good strategy is to design the synthesis so that one obtains two immunogens for each study. Unless there are other constraints, we generally make peptides of 15 amino acids in length. During synthesis we take one-half the yield at 8 amino acids and then continue until we reach 15-thus two immunogens are generated in one synthesis. Since synthesis proceeds in the C-terminus + N-terminus direction, we put a cysteine on the C-terminus so both the 8-mer (9-mer if the artificial cysteine is counted) and the 18mer can be coupled using the same C-terminal cysteine. Sometimes it is necessary to use very short peptides. For example, in the work on synthetic idiotypes, the peptide immunogens had to be sufficiently short so as not to extend into the constant regions of the molecule. In cases like this it is often good to extend the peptide by using spacer amino acids. In the idiotype work (McMillan et al., 1983), spacers containing proline and glutamic acid residues were found to be suitable. Many different coupling methods and carrier proteins have been used in the generation of antipeptide antibodies. We usually couple the synthetic peptide to KLH through a cysteine residue but there is no evidence that this is superior to a variety of other procedures. If one is going to use the antibodies to react with the products of cultured cells it is probably best to avoid using BSA as a carrier, as the antialbumin antibodies can give spurious results since so many ligands and proteins bind to the albumin contained in the bovine serum in the culture of cells.

XV. Theoretical Aspects

As more and more data concerning antipeptide antibodies have been gathered, a theoretical construct has begun to emerge. At first, there were two rather simplistic notions to explain the results. The first idea could be called the side chain constellation theory. Basically, the notion was that antipeptide antibodies “saw” a collection of side chains as if they were a set of haptenes on a protein backbone. It was as if antibodies read the sequence of a protein. Given that any type of protein-protein binding, whether it involves antibodies or enzymes, is an interaction between specific shapes, this was never a strong idea.


33

However, to test this notion, we synthesized the C-terminal20 amino acid of the influenza virus HA1 chain in four different ways: from Lamino acids in an N + C and C + N direction and from D-amino acids in the N + C and C + N direction (Houghten and Lerner, unpublished). The D isomer polymer in the C + N direction (retro-D) is particularly interesting in that, except for the ends, the relative positions of the side chains approximate the L isomer in the N + C direction but the positions of the carbonyl group and the amide bonds are reversed, In other words, w e have a polymer with the same sequence but the wrong shape. Antibodies were made to each of the four peptides and tested for reactivity with the peptides as well as the folded HA1 protein, All four peptides raised antipeptide antibodies, but each reacted only against the peptide against which they were generated; only the antibody against the L, N + C peptide reacted with the HA1. The absolute lack of cross-reactivity between any of these four antipeptide antibodies speaks strongly against notions that depend on antipeptide antibodies reacting with a constellation of side chains. The second theory evoked a stochastic model. Here, the idea was that after peptide immunization, antibodies were made against the multiple peptide conformations but only the small fraction of antibodies against the conformation shared between the peptide and protein was reactive with the folded protein. Thus, the success of the technology was postulated to be more a testimony to the sensitivity of immunological assays which could detect a small percentage of proper antibodies than to something more fundamentally interesting. However, peptides in solution have thousands to hundreds of thousands of conformations and as more and more success for different peptides was achieved, the stochastic idea became less tenable. In other words, the scientific community was doing the statistics and the collective answer was against the stochastic model. To test the stochastic model in a formal way, Niman et al. (1983) used monoclonal antipeptide antibodies as a way of estimating the frequency with which small peptides induce antibodies that react with folded proteins. They made monoclonal antibodies to six chemically synthesized peptides from three proteins. The frequency with which the peptides induced protein-reactive antibodies was at least four orders of magnitude greater than expected from previous experimental work and vastly different from what would be predicted by calculating the possible number of peptide conformers in solution. These results suggested that sufficient structural information is contained in peptides as small as 13 amino acid residues to induce protein reactive antibodies at a high frequency. The key point was that the frequencies observed were inconsistent with any stochastic model.

34

RICHARD ALAN LERNER

If the stochastic model is also not correct, then, with some caveats which we will discuss below, we must begin to think in terms of antibodies reacting with conformations in proteins which are different from that of the native. These notions open up the exciting possibility of a merger between the protein dynamics and the antibody problems. In the first instance one can imagine that a given site in a protein is mobile and that as it passes through a conformation which the antibody recognizes, union takes place. A second possibility is that the antibody-antigen union, itself, induces a shape change. The end result of both models is the same in that antibody is bound to a protein with the site of union in a different conformation from native. But the two models are vastly different insofar as their implications for protein structure are concerned. The first model views proteins as highly dynamic structures, whereas the latter makes no such assumptions and assumes that antibody-antigen union has sufficient energy to distort the target structure. The difficulty with the latter model is that it is somewhat circular in that antibodies can only distort that which they can recognize, and distant conformations would not be efficiently recognized. Thus, all facts considered, it seems likely that for some percentage of the time, local disorder occurs on short segments of the protein, allowing reaction with antipeptide antibodies. This local disorder model suggests that peptides that elicit antibodies recognizing intact proteins are located in areas of relatively great conformational mobility in the intact protein. Whichever model one favors, it is important to note that peptide immunogens do not have the same possibility for induced fit that occurs in systems with a limited number of receptors (i.e., peptide hormones) because the immune system is a system of diversity, and unless other factors pertain (see below), “fixing” the correct conformation would seem to have no better probability than fixing the incorrect one. As mentioned above, there are a number of caveats which should be considered. It can be argued that the immune system may recognize preferentially the conformation of a peptide shared by native proteins. Although this seems to be highly unlikely at first consideration, most proteins coevolved with the immune system, and “preexposure” is, in fact, possible, particularly in the case of the viral proteins. In other words, existing proteins may have played a role in shaping the immunological repertoire. The argument against this view is, of course, that qne can make antibody against virtually any nonprotein chemical. But this could simply be a “side show” due to the cross reactivity between proteins and chemical antigens so that any antichemical antibody is in


35

reality an antiprotein antibody where the binding constant for the protein would be much higher if the correct determinant could be found. Factors involved in the presentation of the peptide antigens may also restrict the range of conformations available to the peptides, making the native conformer far more common than predicted for the free peptide in aqueous solution. For example, the membranes of presenting cells may provide an environment that limits peptide conformation. The peptides are usually coupled to carrier proteins, and interaction of the peptide with the surface of the carrier could be expected to greatly restrict the conformations accessible to the peptide. Peptides could have a more stable native-like structure than has been suggested from previous theoretical and experimental studies with model peptides. For example, peptides may contain particularly stable local structures like those shown by others (Brown and Klee, 1976; Bierzynski and Baldwin, 1982; Kim and Baldwin, 1982; Bierzynski et al., 1982). Stable local structures could serve as initiation sites in the folding of nascent or denatured proteins. If true, such a model would have important implications for the general problem of protein folding (Tanford, 1968, 1970; Anfinsen and Scheraga, 1975; Nemethy and Scheraga, 1977; Creighton, 1978; Privalou, 1979; Jaenicke, 1980; Ptitsyn and Finkelstein, 1980; Thomas and Schechtor, 1980; Wetlaufer, 1981; Richardson, 1981; Rossman and Argus, 1981; Kim and Baldwin, 1982; Levinthal, 1968). XVI. The Repertoire Should Be Tapped Further: Concept of Immunological Catalysis

The advent of antibodies of predetermined specificity will condition us to think in terms of binding to specific sites in proteins rather than to just proteins. One wonders if the next step can be taken and antibodies can be produced which bind to the same structures in proteins as do enzymes. Since the effectiveness of enzymes depends upon the stabilization of minor equilibrium states we might expect antibodies recognizing these same states to carry out catalytic functions. This is a wonderful possibility since one could fish in the immunological repertoire for any kind of enzyme so long as the substrate were sufficiently large to be immunogenic. The basic question is whether the diversity of the immune system is any match for the millions of years of molecular design which go into the evolution of an enzyme. I, for one, would bet on the diversity of the immune system. There is a real impetus to test those notions because, since w e can

36

RICHARD ALAN LERNER

now make antibodies to almost any position on a protein, success would be tantamount to having site specific enzymes for proteins. Such enzymes would be analogous to restriction endonucleases except that the specificity would be induced by the experimenter. But, how to accomplish this? Two notions come to mind. The first assumes that, as previously discussed, antipeptide antibodies work via an induced fit mechanism and thus distort the site at which they bind. If one could supply additional energy the protein might hydrolyze at the site of binding. Although possible, this approach seems unlikely to succeed easily. An alternative approach is to make antibodies to intermediates in catalysis to tip the equilibrium in favor of hydrolysis of the peptide bond. The principle, then, is one of immunological catalysis. Antibodies and enzymes are both proteins whose function depends on their ability to bind specific target molecules. Enzymatic reactions differ from immunological reactions in that the binding of substrate to enzyme leads to chemical catalysis. Enzymes catalyze the hydrolysis of proteins, by combining with the protein in a transition state of the reaction. It is generally assumed that an enzymatic reaction is accelerated with respect to the nonenzymatic reaction because of the enzyme’s ability to reduce the free energy of the transition state, and thus, the free energy of activation, of the reaction (Jencks, 1975; Pauling, 1948). The enzyme might accomplish this by binding a transition state geometry more strongly than the corresponding substrate(s) or product(s). This means that an enzyme’s intrinsic binding energy is much greater than can be measured from the binding of substrates or products. Essentially, the enzyme’s binding energy is utilized to perform the chemical reaction (Jencks, 1983). The basic idea behind immunological catalysis contemplates the design of antibodies of predetermined specificity that will stabilize transition states of peptide bond hydrolysis upon binding to the specified antigen. This should result in a reduction in the activation energy for the hydrolysis reaction, thus meeting a criterion for catalysis. Antibodies which display this property might be obtained by immunization with synthetic peptide analogs that are chemically modified to resemble the bonding characteristics of a substrate undergoing peptide bond hydrolysis-that is, transition state analogs of this reaction. The mechanism by which an antibody might catalyze the hydrolysis of a bound substance may be thought of in terms of an “induced fit” model. As loosely bound substrate distorts to conform to the binding geometry of the antibody, stress can be relieved by chemical reorganization of a single amide bond such that this reorganization leads toward hydrolysis of the bond.


HS-Yo

37

HS-ro A

FIG.9.

Toward the goal of designing immunological catalysts, we have begun to synthesize chemical analogs of polypeptides and depsipeptides which incorporate phosphonamidate or phosphonate moieties at specific sites to mimic the transition state for hydrolysis of the amide or ester bond (A. Tramontano, Lin, Bor-Sheng, and Lerner, unpublished results). These are reasonable candidates for this experiment because it is known that such phosphonamidates are, in fact, transition state analogs in the inhibition of proteolytic enzymes (Bartlett and Marlowe, 1983). Initially, we plan to investigate a system designed for the hydrolysis of a simple p-nitrophenyl ester. To illustrate some of the ideas, we can consider some sample compounds which we have prepared. Compound A is being prepared to act as the transition state analog of p-nitrophenyl ester (Compound B) in an immunological experiment (Fig. 9). Antibodies generated to Compound A bound to a carrier can be isolated and screened in an assay which tests for catalytic hydrolysis of ester (Compound B). The liberation of the colored p-nitrophenolate in this reaction will facilitate the detection of catalytically active antibodies. The following synthesis scheme is being used: diethyl p-aminobenzyl phosphonate is condensed with 3,3’dithioproprionic acid to provide the dimeric amide. This is transesterified at phosphorous to give the di-p-nitrophenyl ester, mild base hydrolysis and reduction of the disulfide provides substance A. The ester (Compound B) is prepared by a similar sequence, starting with ethyl p-aminophenyl acetate. In the immunological experiment the phosphonate/phosphonamidatemay be coupled to carriers through a disulfide bond, though other functional groups may be employed. Hydrolysis of the amide bond of polypeptides or proteins will require analogs which bear the phosphonamidate moiety (Fig. 10). Methods for the synthesis of these compounds are being explored. Phosphonamidates described for the inhibition of certain proteases (Bartlett and Marlowe, 1983; Jacobsen and Bartlett, 1981) can also be modified for induction of immunological catalysts. Since short poly-

38

RICHARD ALAN LERNER

0

R

0

Phosphonamidate

O

R

Polypeptide

FIG.10.

peptide chains can induce antibodies which recognize the protein at specific sites, we may expect that if an antibody specified for a transition state analog catalyses the hydrolysis of a short polypeptide chain, it will also catalyze the cleavage of a protein with that particular sequence of amino acids somewhere along its length. The implication of these expectations is that we could confer the activity of certain proteases to immunoglobulins. Furthermore, the antibody’s activity may be directed to any site at will by designating the amide bond to be cleaved with the phosphonamidate center in the analog used for immunization. Thus, a method would be available for the selective proteolysis of any protein whose local sequence conforms with that of the polypeptide targeted. The applications of such a method in protein chemistry, biochemistry, and medicine are without limit. For example, instead of engineering the immune system to simply bind to viruses or tumor antigens, we could aim at evoking antibodies which catalyze specific protein cleavages which inactivate viruses or kill cells. In essence, one evokes antibodies which act directly and do not depend on help from accessory factors such as complement, or complex processes like opsinization. Regardless of whether these concepts are correct in detail, one can expect antibodies of predetermined specificity to soon take on roles which may transcend their simple binding functions. (See Note Added in Proof.) XVII. Antibody Template Directed Organic Synthesis

It is a curious paradox that although almost all biological processes are mediated by proteins, very few proteins are useful as therapeutics. There are, of course, good reasons for this. Proteins are labile, often destroyed by the acid pH of the stomach, unable to cross cell membranes or the blood-brain barrier, etc. Even a lifesaving protein such as insulin could hardly be classified as an ideal drug, in that it requires daily injection and as currently given does not prevent the many complications of diabetes. Also, important peptides such as hypothalamic releasing factors cannot easily be turned into drugs because of their limited ability to be absorbed after ingestion or cross the blood-brain

ANTIBODIES O F PREDETERMINED SPECIFICITY

39

barrier. So, what one really wants is to construct nonpeptidyl organic mimics of proteins and peptides. One knows that this is possible because, for example, the endorphins and the opiates share the same receptor-it is just that when we look at the two structures we do not understand why the receptor “sees” them as mimics of each other. In other words, we do not know the rules by which nonpeptidyl organic compounds mimic proteins. If, however, these rules could be deciphered one would see a new approach to rational design of drugs. In an interesting way, antibodies of predetermined specificity have opened a wedge which could lead to a better understanding of how to make organic mimics of proteins. The way in, of course, is to learn by making organic mimics of antigens. As long as one was dealing at the whole protein level, there was little hope for such design. But, now that the problem can be reduced to only a few amino acids, it is probably fair to say that the problem has moved from impossible to formidable. The process one can use to design organic mimics of proteins is that of antibody template directed organic synthesis. In this process one begins with a monoclonal antibody and a peptide mimic of a protein antigen (i.e., as described above for the influenza system). This is essentially the same as having, in organic chemical terms, a pure host and a pure guest. In practice, each time a chemical change is made in the peptide ligand, its ability to react with the antibody is checked. If the ligand still reacts, the chemical step was permitted and the next step can be taken. If the antibody does not bind the ligand a different modification must be made. By continuing this process, one would hope to evolve a compound which no longer resembles the peptide but still shares immune reactivity. Indeed, nonpeptidyl organic antigens would be useful, but the real goal is to learn the rules for chemical mimicking of proteins. If this could be accomplished, chemicals which mimic the functions of proteins, such as insulins and even intracellular regulators, might be a reality.

REFERENCES Alexander, H., Johnson, D. A., Rosen, J., Jerabek, L., Green, N., Weissman, I. L., and Lerner, R. A. (1983).Nature (London) 306,697-690. Alhiian, A., Cardenas, J . M., Houghten, R. A., Dixon, F. J., and Theofilopoulos, A. N. (1984).Proc. N a t l . Acad. Sci. Z1.S.A. 81, 2176-2180. Amzel, L. M., and Poljak, R. J . (1979).Antiti. Rev. Biochem. 48, 961-997. Amzel, L. M., Poljak, R. J., Saul, F., Varga, J. M., and Richards, F. (1974). Proc. Natl. Acad. Sci. U.S.A.71, 1427-1430. Anderer, F. A., and Schlumberger, H. D. (1965).Biochim.Biophys. Acta 97,503-509. Anderson, S., Bankier, A. T., Barrel], B. G., de Bruijn, M . H. L., Coulson, A. R., Drouin, J., Eperon, I . C., Nierlich, D. P., Roe, B. A., Sanger, F., Schreier, P. H., Smith, A. J. H., Staden, R., and Yonng, I . G. (1981).Nature (London) 290,457-465.

40

RICHARD ALAN LERNER

Anderson, R. G. W., Brown, M. S., Beisiegel, U., and Goldstein, J. L. (1982).J.Cell Biol. 93,523-531. Andrews, D. W., and Capra, J. D. (1981). Proc. Natl. Acad. Sci. U.S.A.78, 37993803. Anfinsen, C. B., and Scheraga, H. A. (1975). Adu. Protein Chem. 29,205-300. Arnon, R.,and Maron, E. (1971).J . Mol. Biol. 61,225. Arnon, R., and Neurath, H. (1970). Immunochemistry 7,241. Arnon, R.,and Sela, M. (1969). Proc. Natl. Acad. Sci. U.S.A.62, 163-170. Arnon, R.,Maron, E., Sela, M., and Anfinsen, C. B. (1971). Proc. Natl. Acad. Sci. U.S.A. 68,1450-1455. Arnon, R., Bustin, M., Calef, E., Chaitchik, S., Haimovich, J., Novik, N., and Sela, M. (1976). Proc. Natl. Acad. Sci. U.S.A. 73,2123-2127. Atassi, M. Z. (1975). lmmunochemistry 12,423-438. Atassi, M. Z. (1978). lmmunochemistry 15, 909-933. Atassi, M. Z., and Lee, C.-L. (1978). Biochem. J . 171,429-434. Audibert, F., Jolivet, M., Chedid, L., Alouf, J. E., Boquet, P., Rivaille, P., and Siffert, 0. (1981). Nature (London) 289,593-594. Bancroft, W. H., Mundon, F. K., and Russell, P. K. (1972).J. lmmunol. 109,420-425. Baron, M. H., and Baltimore, D. J. (1982).J. Virol. 43,969. Bartlett, P. A., and Marlowe, C. K. (1983). Biochemistry 22, 4618-4624. Beisiegel, U., Schneider, W. J., Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1981).J . Biol. Chem. 256, 11923-11931. Benjamin, D. C., Berzofsky, J. A,, East, I. J., Curd, F. R. N., Hannum, C., Leach, S. J., Margoliash, E., Michael, J. G., Miller, A., Prager, E., Reichlin, M., Sercarz, E. E., Smith-Gill, S. J., Todd, P. E., and Wilson, A. C. (1984). Annu. Reu. Zmmunol. 2, in press. ' Berk, A. J., Lee, F., Harrison, T., Williams, J., and Sharp, P. A. (1979).Cell 17, 935. Bhatnagar, P. K., Papas, E., Blum, H. E., Milich, D. R., Nitecki, D., Karels, M. J., and Vyas, G. (1982).Proc. Natl. Acad. Sci. U.S.A. 79,4400-4404. Bibb, M. J., Van Etten, R. A., Wright, C. T., Walberg, M. W., and Clayton, D. A. (1981). Cell 26, 167-180. Bierzynski, A., and Baldwin, R. L. (1982).J . Mol. Biol. 162, 173-186. Bierzynski, A,, Kim, P. S., and Baldwin, R. L. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2470-2474. Bittle, J. L., Houghten, R. A., Alexander, H., Shinnick, T., Sutcliffe, J. G., Lerner, R. A., Rowlands, D. J,, and Brown, F. (1982). Nature (London) 298,30-33. Boquet, P., Alouf, J. E., Duflot, E., Siffert, O., and Rivaille, P. (1982). Molecular lmmunology 19, 1541-1549. Boyle, W. J., Lipsick, J. S., Reddy, E. P., and Baluda, M. A. (1983).Proc. Natl. Acad. Sci. U.S.A. 80,2834-2838. Brown, J. E., and Klee, W. A. (1976). Biochemistry 10, 470-476. Brown, M. S., Anderson, R. G. W., and Goldstein, J. L. (1983). Cell 32, 663-667. Brown, R. K. (1962).1.Biol. Chem. 237, 1162. Brown, R. K., Delaney, R.,Levine, L., and Van Vunakis, H. (1959).J . Biol. Chem. 234, 2043. Canfield, R. E., and Liu, A. K. (1965).J . Biol. Chem. 240, 1997. Carson, D. A., and Fong, S. (1983).Mol. Zmmunol. 20, 1081-1087. Chen, P. P., Houghten, R. A., Fong, S., Lerner, R. A., Vaughan, J. H., and Carson, D. A. (1984a).J.E x p . Med. 159, 1502-1511. Chen, P. P., Houghten, R. A., Fong, S., Rhodes, G. H., Gilgerbston, T. A., Vaughan, J. H., Lerner, R. A., and Cardon, D. A. (1984b). Proc. Natl. Acad. Sci. U.S.A.81,17841788.


41

Clarke, B. E., Carroll, A. R.,Rowlands, D. J., Nicholson, B. H., Houghten, R. A., Lerner, R. A., and Brown, F. (1983). F E E S Lett. 157,261-264. Clevinger, B., Shilling, J., Hood, L., and Davie, J. M. (1980a).J. E x p . Med. 151, 10591070. Clevinger, B., Thomas, J., Davie, J. M., Shilling, J., Bond, M., Hood, L., and Kearney, J. (1980b). pp. 159-168. Academic Press, New York. Clevinger, B., Thomas, J., Davie, J. M., Shilling, J., Bond, M., Hood, L., and Kearney, J. (1981). In “Immunoglobulin Idiotypes,” pp. 159-168. Academic Press, New York. Connolly, M. (1983).J . Appl. Crystallogr. 16, 548-558. Coppel, R.L., Cowman, A. F., Lingelbach, K. R.,Brown, G. V., Saint, R. B., Kemp, D. J., and Anders, R. F. (1983). Nature (London) 306,751-756. Cosman, D., Khoury, G., and Jay, G. (1982a). Nature (London) 295,73-76. Cosman, D., Kress, M., Khoury, G., and Jay, G. (198213).Proc. Natl. Acad. Sci. U.S.A.79, 4947-4951. Creighton, T. E. (1978). Prog. Biophys. Mol. Biol. 33, 231-297. Crumpton, M . J. (1974). In “The Antigens” (M. Sela, ed.), pp. 1-78. Daniels, R. S., Douglas, A. R.,Gonsalves-Scarano, F., Palu, G., Skehel, J, J., Brown, E., Knossow, M., Wilson, I. A,, and Wiley, D. C. (1983).In “The Origin of Pandemic Influenza Viruses” (Chu and Laver, eds.). Elsevier, Amsterdam, in press. Devare, S. G., Reddy, E. P., Law, J. D., Robbins, K. C., and Aaronson, S. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 731-735. Dreesman, G. R., Sanchez, Y., Ionescu-Matiu, I., Sparrow, T., Six, H. R.,Peterson, D. L., Hollinger, F. B., and Melnick, J. L. (1982). Nature (London) 295, 158-160. Drickamer, K. (1981).J. Biol. Chem. 256,5827-5839. Drickamer, K., Mamon, J. F., Ginns, G., and Leung, J. 0.(1984).J. B i d . Chem. 259, 770-778. Edelnian, G. M. (1973). Science 180, 830-840. Feldrnan, L. T., and Nevins, J. R. (1983). Mol. Cell Biol. 3, 829-838. Freedman, M. H., and Sela, M. (1966).J. Biol. Chem. 241, 2383. Galibert, F., Mandart, E., Fitoussi, F., Tiollais, P., and Charnay, P. (1979). Nature (London) 281,646-650. Gentry, L. E., Rohrschneider, L. R.,Casnellie, J. E., and Krebs, E. G. (1983). J. Biol. Chem. 258, 11219-11228. Gerin, J. L., Alexander, H., Shih, J. W.-K., Purcell, R. H., Dapolito, G., Engle, R.,Green, N., Sutcliffe, J. G., Shinnick, T. M., and Lerner, R. A. (1983). Proc. Natl. Acad. Sci. U.S.A.80, 2365-2369. Godson, G. N., Ellis, J., Svec, P., Schlesinger, D. H., and Nussenzweig, V. (1983). Nature (London) 305,29-33. Goetzl, E. J., and Peters, J. H. (1972).J. Immunol. 108, 785. Cold, J. W. M., Shih, J. W.-K., Purcell, R. H., and Gerin, J. L. (1976).J . Immunol. 117, 1404-1405. Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1979). Nature (London) 279, 679-685. Graham, F. L., Harrison, T., and Williams, J. (1978).J. Virol. 86, 10. Green, M., Bracknlan, K. H., Lucher, L. A., and Symington, J. S. (1983a). In press. Green, M., Brackman, K. H., Lncher, L. A., Symington, J. S., and Kranier, T. A. (19831~). J . Virol. 48, 604-615. Green, N., Shinnick, T. M., Witte, J., Ponticelli, A., Sutcliffe, J. G., and Lemer, R. A. (1981).Proc. Natl. Acad. Sci. U.S.A.78, 6023. Green, N., Alexander, H., Wilson, A., Alexander, S., Shinnick, T. M., Sutcliffe, J. G., and Lerner, R. A. (1982). Cell 28, 477. Grosskopf, R.,and Feldman, H. (1981). Curr. Genet. 4, 151-158.

42

RICHARD ALAN LERNER

Habeeb, A. F. S. A., and Atassi, M. Z. (1971). Biochim. Biophys. Acta 236, 131. Hogle, J. (1984). In “Modern Approaches to Vaccines” (R. M. Channock and R. A. Lerner, eds.). Cold Spring Harbor Press, New York. Hood, L., Steinmetz, M., and Malissen, B. (1983).Annu. Reu. Zmnzunol. 1, 529-568. Hui, K. Y., Haber, E., and Matsueda, G. R. (1983). Science 1129-1131. Ibrahimi, I. M., Prager, E. M., White, T. J., and Wilson, A. C. (1979). Biochemistry 13, 2736. Jacobsen, N. E., and Bartlett, P. A. (1981).J. Am. Chem. SOC. 103,654-657. Jaenicke, R., ed. (1980). Proc. Conf. Ger. Biochem. Soc., 28th. Amsterdam. Jencks, W. P. (1975). Adu. Enzymol. 43,219-410. Jencks, W. P. (1983). Znt. Soluay Conf., 17th, Nou. Johnson, D. A., and Elder, J. (1983).J . E x p . Med. 159, 1751-1756. Jones, N., and Shenk, T. (1979a). Proc. Natl. Acad. Sci. U.S.A.76, 3665. Jones, N., and Shenk, T. (1979b). Cell 17,683. b r u s h , F. (1962). Ado. Immunol. 2, 1. Kim, P. S., and Baldwin, R. L. (1982). Annu. Reu. Biochem. 51,459-489. Kim, P. S., Bierzynski, A., and Baldwin, R. L. (1982).J. Mol. Biol. 162, 187-199. Kleid, D. G., Yansura, D., Small, B., Dowbenko, D., Moore, D. O., Robertson, B. H., and Bachrach, H. L. (1981). Science 214, 1125-1 129. Kloetzer, W. S., and Arlinghaus, R. B. (1984). Virology, submitted. Kohler, G., and Milstein, C. (1975).Nature (London) 256, 495. Kress, M., Cosman, D., Khoury, G., and Jay, G. (1983). Cell 34, 189-196. Knnkel, H. G., Agnello, V., Joslin, F. G., Winchester, R. J., and Capra, J. D. (1973). J . E x p . Med. 137,331-342. Kunkel, H. G., Winchester, R. J., Joslin, F. G., and Capra, J. D. (1974).J.E x p . Med. 139, 128-136. Kurz, C., Forss, S., Kupper, H., Strohmaier, K., and Schaller, H. (1981). Nucleic Acids Res. 9, 1919-1931. Landsteiner, K. (1936). “The Specificity of Serological Reactivities.” Thomas, Springfield, Illinois. LeBouvier, G. L. (1971). J. Infect. Dis. 123, 671-675. Lerner, R. A. (1982). Nature (London) 299,592-596. Lerner, R. A,, Green, N., Alexander, H., Liu, F.-T., Sutcliffe, J. G., and Shinnick, T. M. (1981). Proc. Natl. Acad. Sci. U.S.A.78, 3403-3407. Levinthal, C. (1968).J. Chem. Phys. 65, 44-45. Lucher, L. A., Brackman, K. H., Symington, J. S., and Green, M. (1984). Virology 132, 217-22 1. Luka, J,, Sternas, L., Jornvall, H., Klein, G., and Lerner, R. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 1199-1203. Maloy, W. L., Culigan, J. E., Barra, Y., and Jay, G. (1984).Proc. Natl. Acad. Sci. U.S.A., in press. Mariothini, P., Chomyn, A., Attardi, G., Trovato, D., Strong, D. D., and Doolittle, R. G. (1983). Cell 32, 1269-1277. Maron, E., Shiozawa, C., Arnon, R., and Sela, M . (1975). HiochemistrrJ10, 763-771. Marrack, J. R. (1938). “The Chemistry of Antigens and Antibodies.” Medical Research Council Special Report Series, No. 230. McMillan, S., Seiden, M., Houghten, R., Clevinger, B., Davie, J. M., and Lerner, R. A. (1983). Cell 35, 859-863. Min Jou, W. M., Verhoeyen, M., Devos, R., Saman, E., Fang, R., Huylebroeck, D., Fiers, W., Threlfall, G., Barber, C., Carey, N., and Emtage, S. (1980). Cell 19, 683-696. Minor, P. D. et al. (1983). Nature (London)301,674-679. Morrow, C. D., and Dasgupta, A. (1983).J. Virol. 48, 429-439.


43

Nemethy, G., and Scheraga, H. A. (1977). Q. Reu. Biophys. 10,239-352. Neumann, H., Steinberg, I. Z., Brown, J. R., Goldberger, R. F., and Sela, M. (1967). Eur. J. Biochem. 3, 171. Niman, H . L. (1983).Nature (London) 307, 180-183. Niman, H. L., lloughten, R. A., Walker, L. E., Reisfeld, R. A., Wilson, I. A., Hogle, J. M., and Lerner, R. A. (1983). Proc. Natl. Acud. Sci. U.S.A. 80, 4949-4953. Nisonoff, A., Reichlin, M., and Margoliash, E. (1970).J . Biol. Chem. 245, 940. Pasek, M., Goto, T., Gilbert, W., Zink, B., Schaller, H., Mackay, P., Leadbetter, G., and Murray, K. (1979). Nature (London) 282, 575-579. Pastan, I. H., and Willingham, M. C. (1981).Annu. Rev. Physiol. 43, 239-250. Pauling, L. (1948). Am. Sci. 36, 58. Pearse, B. M. F., and Bretscher, M. S. (1981). Annu. Rev. Biochem. 50,85-101. Pfaff, E., Mussgay, M., Bohm, H. O., Schulz, G. E., and Schaller, H. (1982).EMBOJ. 1, 869-874. Porter, R. R. (1973).Science 180, 713. Prine, A. M., Okrain, H., and Hopp, T. P. (1982). Proc. Natl. Acad. Sci. U.S.A.79, 579-582. Privalou, P. L. (1979). Adu. Protein Chem. 33, 167-241. Ptitsyn, 0. B., and Finkelstein, A. V. (1980). Q. Reu. Biophys. 13, 339-386. Reichlin, M. (1975). Adu. Immrcnol. 20, 71-123. Reif, A. F., and Allen, J. M. (1964).J. E x p . Med. 120,413-433. Richardson, J. S. (1981). Adu. Protein Chem. 34, 167-339. Robbins, K. C., Antoniades, H. N., Sushikumar, G. D., Hunkapiller, M. W., and Aaronson, S. A. (1983).Nature (London) 305, 605-608. Rossmann, M. G., and Argus, P. (1981).Annu. Reu. Biochem. 50,497-532. Rowlands, D. J., Clarke, B. E., Carroll, A. R., Brown, F., Nicholson, B. H., Bittle, J. L., Houghten, R. A., and Lerner, R. A. (1984).Nature (London) 306, 694-697. Sanger, D. V. (1979).1.Gen. Virol. 45, 1-13. Schaffhausen, B., Benjamin, T. L., Pike, L., Casnellie, J., and Krebs, E. (1982).1.Biol. Chem. 257, 12467-12470. Schild, G . C., Oxford, J. S., DeJong, J. C., and Webster, R. G. (1983). Nature (London) 303,706-709. Schneider, W., Slaughter, C. J., Goldstein, J. L., Anderson, R. G. W., Capra, J. D., and Brown, M. S. (1983).J.Cell Biol. 97, 1635-1640. Sefton, B. M., and Walter, G. (1982).J. Virol. 44, No. 2, 467-474. Segal, D. M., Padlan, E. A., Cohen, G. H., Rudikoff, S., Potter, M., and Davies, D. R. (1974). Proc. Natl. Acad. Sci. U.S.A.71, 4298-4302. Seiden, M. V., Clevinger, B., Srouji, T., Davie, J. M., McMillan, S., and Lerner, R. A. (1984).Ann. lmmunol. 135, 77-82. Sender, B. L., Anderson, C. W., Hanecak, R., Dorner, L. F., and Wimmer, E. (1982). Cell 28,405-412. Sen, S., Houghten, R. A., Sherr, C. J., and Sen, A. (1983).Proc. Natl. Acad. Sci. U.S.A.80, 1246-1250. Shapira, E., and Arnon, R. (1969).J. Biol. Chem. 244, 1026. Sharon, J., Kabat, E. A., and Morrison, S. L. (1982). Mol. lmmunol. 19, 375-388. Shih, J. W.-K., Tan, P. L., and Gerin, J. L. (1978).J . lmmunol. 12, 520-525. Shilling, J., Clevinger, B., Davie, J . M., and Hood, L. (1980). Nature (London) 283, 35-40. Skehel, J. J., Bayley, P. M., Brown, E. B., Martin, S. R., Waterfield, M. D., White, J. M., Wilson, I. A,, and Wiley, D. C. (1982). Proc. Natl. Acad. Sci. U.S.A. 79,968-972. Smith-Gill, S. J., Wilson, A. C., Patter, M., Prager, E. M., Feldman, R. J., and Mainhart, C. R. (1982).]. lmrnunol. 128,314-321.

44

RICHARD ALAN LERNER

Stanker, L. H., Gallick, G. E., Kloetzer, W. S., Murphy, E. C., Jr., and Arlinghaus, R. B. (1983).J . Virol. 45, 1183-1189. Sue, J. M., and Sytkowsky, A. J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 3651-3655. Sutcliffe, J. G., Shinnick, T. M., Green, N., Liu, F.-T., Niman, H. L., and Lerner, R. A. (1980). Nature (London) 287,801-805. Sutcliffe, S. G., Shinnick, T. M., Green, N., and Lerner, R. A. (1983). Science 219, 660-666. Tamura, T., and Bauer, H. (1982). EMBOJ. 1, 1479-1485. Tamura, T., Bauer, H., Bin, C., and Pipkorn, R. (1983). Cell 34,587-596. Tanford, C. (1968). Ado. Protein Chem. 23, 121-282. Tanford, C. (1970). Ado. Protein Chem. 24, 1-95. Thomas, K. A., and Schechter, A. N. (1980). In “Biological Regulation and Development” (R. F. Goldberger, ed.), Vol. 2, pp. 43-100. Plenum, New York. Tiollais, P., Charnay, P., and Vyas, G. N. (1981). Science 213, 406-411. Tiselius, A. (1937). Biochem. J . 31, 313. Valenzuela, P., Gray, P., Quiroga, M., Zaldivar, J., Goodman, H. M., and Rutter, W. J. (1979). Nature (London) 280,815-819. Verhoyen, M., Fang, R., Min Jou, W., Devos, R., Huylebroeck, D., Saman, E., and Fiers, W. (1980). Nature (London) 286,771-776. Walter, G., Scheidtmann, K. H., Carbone, A. P., Laudano, A. P., and Doolittle, R. F. (1980). Proc. Natl. Acad. Sci. U.S.A.77,5197-5200. Wetlaufer, D. B. (1981). Ado. Protein Chem. 4, 61-92. White, T. J., Ibrahimi, I. M., and Wilson, A. C. (1978). Nature (London) 274, 92-94. Wiley, D. C., Wilson, I. A., and Skehel, J. J. (1981). Nature (London) 289, 373-378. Williams, A. F., and Gagnon, J. (1982). Science 216, 696-703. Wilson, I. A,, Skekel, J. J., and Wiley, D. C. (1981). Nature (London) 289, 373. Wilson, I. A., Niman, H. L., Houghten, R. A., Cherenson, A., Connolly, M. L., and Lerner, R. A. (1984). Cell, in press. Wong, T. W., and Goldberg, A. R. (1981). Proc. Natl. Acad. Sci. U.S.A. 78,7412-7416. Yee, S.-P., Rowe, D. T., Tremblay, M. L., McDermott, M., and Branton, P. E. (1983). J . Virol. 46, 1003-1013. Young, J. D., and Leung, C. Y. (1970). Biochemistry 9,2755. NOTEADDED IN PROOF.A somewhat different approach to antibody-mediated ester hydrolysis has been taken by Kohen and colleagues (F. Kohen et al., FEBS Lett. 111, 427, 1980; Biochim. Biophys. Acta 629,328, 1980). Based on studies which showed that antibodies against haptens may enhance hydrolysis of labile esters, they explored the effect of antibody against 2,4-dinitrophenyl on the hydrolysis of DNP-E-aminocaproylumbelliferone. The anti-DNP antibody seemed to induce hydrolysis of an ester which was six atoms removed from the hapten but the reaction was stoichiometric. Such hydrolysis is not truly catalytic however since the product of the reaction may bind as well or better than the ester to the antibody, thus inhibiting further reaction (Kohen et al., 1980). The binding properties ofthe antibody must, therefore, take into account the characteristic features of the reaction, as enzymes do, for the protein to express a catalytic function. An additional problem is that antibodies are, in fact, capable of hydrolyzing labile esters in a nonspecific manner (L. I. Slobin, Biochemistry 5,2836, 1966). This behavior is not dependent on binding of the ester but may merely reflect the nucleophilic quality of the hydrophilic amino acids on the antibody’s surface. Although these studies are of interest, it would seem that a more general approach would be to divert antibodies to the strncture containing the bond to be hydrolyzed rather than some distant moiety.

ADVANCES IN IMMUNOLOGY,VOL. 36

A Molecular Analysis of the Cytolytic T Lymphocyte Response STEVEN J. BURAKOFF, OFRA WEINBERGER, ALAN M. KRENSKY, AND CAROL S. REISS Department of Pediatrics, Haward Medical School, fhe Dana-Farber Cancer Institute and the Children’sHospital Medical Center, Boston, Massachusetts

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Specificity of the CTL Response.. .................................. 111. The Use of Liposomes to Study the Generation of the CTL Response. . . . . A. Allogeneic CTL Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Virus-Specific CTL Response. . . . . . . . . . . . . . . . . . ............ C. Xenogeneic CTL Response.. .................................... IV. The Use of Liposomes to Study the Helper T Cell Pathway. . . . . . . . . . . . . A. Evidence for Helper T Cell Regulation of CTL Responses . . . . . . . . . . . B. Cellular Requirements for the Induction of T Helper Cells. . . . . . . . . . . C. Antigenic Requirements for the Induction of Helper T Cells . . . V. The Use of DNA-Mediated Gene Transfer of Cloned MHC Genes t Study CTL Specificity.. ........................................... VI. The Use of Monoclonal Antibodies to Define Functional CTL Antigens. .. A. Characterization of Human CTLs.. ............................... B. Specificity of Human CTLs.. .................................... C. Cell Molecules Involved in Cytolysis . . . . . . . . . VII. Conclus .................................................. References ......................................................

45 46 48 48 49 51 52 52 56 63 67 68 70 75 78

I. Introduction

The cytolytic T lymphocyte (CTL) response has been an area of intense investigation for the past 15 years. Though initially studied for the response to allogeneic major histocompatibility complex (MHC) antigens, the observations of Zinkernagel and Doherty (1974) and of Shearer (1974) have resulted in enormous interest in the CTL response to modified syngeneic MHC antigens. The availability of a reliable and easy assay for CTL activity, i.e., the chromium release assay developed by Brunner et al. (1968), resulted in rapid progress in our understanding of the CTL response. Initially, the specificity of CTL recognition could be explored because of the availability of inbred strains of mice bearing recombinations within the major histocompatibility complex. However, there has always been a major limitation to a better molecular understand45 Copyright 8 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-022436-4

46

STEVEN J. BURAKOFF ET AL.

ing of CTL regulation and specificity, this limitation results from the fact that CTLs respond to integral membrane proteins, i.e., MHC and viral antigens. As integral membrane proteins, MHC and viral antigens could not be manipulated as readily as the soluble antigens that have been used to study humoral responses. However, in recent years there have been several technological advances that have markedly enhanced our ability to study the CTL response on a molecular level. These advances include (1) the ability to isolate and purify MHC and viral proteins in a form in which they retain the ability to stimulate CTLs; (2) the cloning of MHC and viral genes and their introduction into cells by DNA-mediated gene transfer; and (3)hybridoma technology (Burakoff, 1984). We will review three areas of the CTL response where these new technologies have advanced our understanding; namely, the specificity of the CTL response, the regulation of the CTL response, and the effector antigens utilized by CTLs in the CTLtarget cell interaction.

II. Specificity of the CTL Response (Burakoff, 1981)

Until recently, the most thoroughly investigated CTL response has been the response to allogeneic MHC antigens. In murine systems, the CTLs are generated by immunizing mice of one strain (A) with the cells from an allogeneic MHC strain (B). The majority of CTLs generated in this manner (A anti-B CTLs) have been shown to recognize the H-2K and H-2D gene products on the surface of cells from strain B, however, there is evidence that some clones of CTLs are also directed to Ia antigens. Human CTLs have a similar predominant specificity for HLA-A and HLA-B antigens, though CTLs to HLA-DRYSB, and DC antigens have also been demonstrated (vida infra). Several approaches have been employed to demonstrate that the specificity of these CTLs is for MHC antigens. One approach is to use a panel of target cells from strains of mice or from individuals that carry recombinations within the MHC. Antisera directed at the MHC antigens of the target cells can be used to inhibit lysis of these cells by CTLs. It is believed that antisera inhibit cytolysis by binding to the surface antigens of the target cells, thus preventing the direct interaction between the CTL and the target cells (Nabholz et al., 1974). Competitive (cold target) inhibition experiments, in which nonradiolabeled (cold) target cells compete with 51Cr-radiolabeled (hot) target cells for receptor sites on the CTL, are also useful for assessing specificity. By mixing hot and cold target cells from strains of mice or from individuals that differ at various regions of the MHC, one can determine the specificity of the

CYTOLYTIC T LYMPHOCYTE RESPONSE

47

CTLs for the regions that differ (De Landazuri and Herberman, 1972). The most direct means of studying CTL specificity is to purify cell surface antigens and use them to stimulate CTL responses (see Section 111) and to utilize target cells that have been transfected with cloned MHC genes (see Section V). The specificity of the CTL response to xenogeneic cells has been examined by,several groups. The stimulation of rat lymph node cells by murine spleen cells results in the induction of CTLs that recognize the H-2K and H-2D antigens on murine target cells (Burakoff et al., 1977). The specificity of these xenogeneic CTLs was defined both by genetic analysis (i.e., the use of target cells from strains of mice with recombinations within the MHC) and by antisera blocking. The CTL response of human peripheral blood lymphocytes (PBLs) to murine spleen cells demonstrated a similar specificity for H-2K and H-2D antigens (Lindahl and Bach, 1975; Carnaud et nl., 1977). In all cases, there was no evidence that CTLs recognized gene products encoded by the H-21 region. In the murine response to HLA antigens, it appears that CTLs recognize not only antigens of the HLA-A and B regions but also HLA-DR antigens (Burakoff et al., 1980; Engelhard et al., 1980). The specificity of these CTLs have been studied by utilizing purified HLA antigens inserted into liposomes and these experiments will be discussed in greater detail in Section 111. During the last few years the CTL response to virally infected or chemically modified syngeneic cells has stimulated the greatest interest among immunologists. In part, this is due to the apparent biological significance of these CTL responses. The CTL response to virally infected syngeneic cells was first described by Zinkernagel and Doherty (1974). While studying the immune response to viral infections they made the following observation: CTLs could be found in the spleens of mice infected with lymphocytic choriomeningitis (LCM) virus that could only lyse those target cells that (1) were identical to the CTLs at the H-2K and/or H-2D regions and (2) were infected with LCM virus. No lysis occurred if the MHC of the target cell was different from that of the infected mice or if the target cells were uninfected or were infected with a different virus. Thus, they observed a dual restriction for virus-specific CTLs. At about the same time that Zinkernagel and Doherty made their observations, Shearer independently made a similar observation, using a different model system (Shearer, 1974; Shearer and SchmittVerhulst, 1977). Spleen cells from mice were treated with trinitrobenzene sulfonate (TNBS), a procedure that resulted in the covalent coupling of trinitrophenyl (TNP) groups to free lysine residues in proteins on the cell surface. If these TNP-modified spleen cells were

48


irradiated and then cocultured with syngeneic normal spleen cells, CTLs were induced that would lyse only target cells that were both H-2 identical and TNP modified. Experiments by Schmitt-Verhulst et al. (1976) and Burakoff et al. (1976) demonstrated that anti-TNP sera and the appropriate anti-H-2 sera inhibited lysis of target cells, thus providing direct evidence that the H-2 gene products are part of the target antigens recognized. These observations supported the proposal that CTLs were induced in viral infections because of the viral antigens expressed on the cell surface. 111. The Use of Liposomes to Study the Generation of the CTL Response

A. ALLOCENEICCTL RESPONSE With the availability of monoclonal antibodies specific for MHC antigens, it has become possible to rapidly purify large amounts of MHC antigens. Using affinity columns of the monoclonal antibody coupled to Sepharose, the MHC antigens can be readily purified from detergent-solubilized whole cells or cell membranes. H-2 (Herrmann and Mescher, 1979; Stallcup et al., 1981) and HLA-A, B, and C antigens (Parham, 1979) from a number of specificities have been purified on a relatively large scale in this way. The MHC antigens are integral membrane proteins and are, therefore, insoluble in aqueous buffers in the absence of detergent. Since detergent concentrations necessary to maintain solubility would cause cell lysis, these immunoaffinity purified MHC antigens have been introduced into liposomes by dialysis in order to assess their ability to generate CTLs in vitro (Burakoff and Mescher, 1982). Early studies demonstrated that crude membrane preparations from allogeneic cells bearing the appropriate H-2 antigens will stimulate a secondary in vitro response (Corley et al., 1975; Engers et al., 1975; Hayry and Anderson, 1976; Wagner et al., 1976). When allogeneic tumor cells were lysed and the subcellular components fractionated, the stimulating activity was found to copurify with the plasma membrane fraction (Lemonnier et al., 1978). In this same study it was found that the membranes could also stimulate a weak primary response but were ineffective in blocking the interaction between effector CTLs and target cells. Purified membranes were not as effective as intact allogeneic cells in stimulating a response. Membranes were about 20-fold less efficient than intact cells in stimulating a secondary response and 200-fold less efficient than intact cells in stimulating a primary response. Due to the greater efficiency of subcellular prepara-


49

tions in stimulating secondary CTLs, this response has been studied most extensively. Further study of the antigen requirements for CTL stimulation became possible with the demonstration that membranes from allogeneic cells could be solubilized with detergent and CTL-stimulating activity retained (Fast and Fan, 1978; Mescher et al., 1978). Removal of the H-2 antigens from the solubilized membrane proteins by precipitation with alloantisera and protein A-Sepharose removed the stimulating activity (Mescher et al., 1978). The solubilized membrane proteins stimulated a response when added directly to culture but were more effective if reconstituted membranes were formed by removing the detergent by dialysis (Fast and Fan, 1978; Mescher et al., 1978). Partially purified H-2 antigens were also found to stimulate a specific secondary CTL response when incorporated into lipid vesicles by detergent dialysis (Sherman et al., 1980) or covalently coupled to agarose beads (Fast and Fan, 1979). As mentioned, with the development of affinity purification procedures using monoclonal antibodies it has become possible to obtain relatively large amounts of purified, detergent-solubilized H-2 antigens (Herrmann and Mescher, 1979; Stallcup et al., 1981). The mild conditions used for purification result in 70-90% recovery of serologically active antigen. Purified MHC antigens were incorporated into liposomes by mixing them with lipids (obtained by chloroform : methanol extraction) from tumor cells in the presence of detergent and then removing the detergent by dialysis. Recently, it was found that the detergent-insoluble fraction of membranes, normally removed prior to reconstitution, if incorporated into liposomes resulted in the liposomes having a far greater stimulatory capacity (Mescher et al., 1981). The major components of this matrix are actin and proteins having M ,values of 70,000, 69,000, 38,000, and 36,000.

B. VIRUS-SPECIFIC CTL RESPONSE

The CTL response to virally infected or chemically modified syngeneic cells appears to be under a dual restriction, one restriction imposed by the MHC antigens and the other determined by the virus or chemical modifier. This dual restriction provides a particularly intriguing model system that can be investigated with liposome technology. Several years ago we began to explore the antigen requirements for the triggering of Sendai virus-specific CTLs. It had previously been shown that influenza virus-specific CTLs could be generated by culturing primed spleen cells with large amounts of the influenza virus hemagglutinin (Zweerink et al., 1977). Although not directly

50


demonstrated, it is presumed that these added viral antigens are recognized in association with the MHC antigens of the cells present in culture. By introducing limiting amounts of viral antigens and partially purified MHC antigens into liposomes we were able to evaluate the role played by both the viral and MHC antigens in the induction of Sendai virus-specific CTLs (Finberg et al., 1978). Specifically, plasma membranes from the murine tumor P815, which bear H-2d antigens, were solubilized in deoxycholate (DOC) and the MHC antigens were partially purified on a lentil lectin column. Inactivated Sendai virus was similarly solubilized in DOC. These plasma membrane antigens and viral antigens were then introduced into separate liposomes or into the same liposomes and these liposomes were added to culture with spleen cells from (C57BL/6 x DBAI2)Fl (H-2" x H-2") mice that had been primed to Sendai virus. The CTLs induced were then assayed on Sendai coated P815 (H-2d)or EL4 (H-2b)targets. Two significant observations were made. First, limiting amounts of viral antigens only induced CTLs when Sendai and H-2 antigens were present in the same liposomes. When Sendai and H-2 antigens were introduced into separate liposomes and both types of liposome were added to the same culture CTLs were not induced. Second, the CTLs resulting from liposome stimulation only lysed P815 targets coated with Sendai virus. This observation strongly suggests that the H-2d MHC antigens introduced into the liposome determined the CTL specificity. This suggests that the CTLs interact with these liposomes directly and it would appear unlikely that these liposomes had to be processed by macrophage for CTLs to recognize them. If there was macrophage processing one might expect the viral antigens to then associate with the H-2d and H-2b antigens of the F1 responder macrophage and trigger CTLs that lysed both P815 (H-2d)and EL4 (H-2b)targets. It was also demonstrated that partially purified H-2d antigens plus the purified hemagglutinin/neuraminadase (HN) glycoprotein from Sendai virus could induce Sendai virus-specific CTLs, thus identifying HN as one of the viral proteins that is recognized by CTLs (Mescher et al., 1979). Hale et al. (1980a) have further demonstrated that it is the H-2 antigens that are recognized by virus specific CTLs. They have shown that purified H-2Kkantigen plus Sendai HN antigen inserted into liposomes are able to induce virus specific CTLs. Loh et al. (1979) have utilized liposomes to investigate the CTL response to vesicular stomatitis virus (VSV). They have shown that VSV primed spleen cells, when cultured with liposomes containing partially purified H-2 antigens plus the purified G glycoprotein from VSV, generate virus-specific CTLs. They also found that the H-2 and


51

viral antigens had to be in the same liposomes and that the H-2 antigens inserted into the liposomes determined the specificity of the CTLs. Similar results were observed by Ciavarra et al. (1980) working with VSV and plasma membranes from P815 tumor cells. Hale et al. (1980b) have elicited VSV specific CTLs with liposomes containing purified G protein and the purified H-2Kk antigen. Lawman et al. (1981)in studying the induction of CTLs to Herpes simplex virus Type 1 (HSV-1) have made similar observations. They also observed that HSV-1 and plasma membrane antigens had to be in the same liposomes to elicit a response. In fact, unlike Sendai virus where large amounts of DOC-solubilized Sendai virus proteins alone can elicit CTLs, neither heat-inactivated nor detergent-solubilized HSV-1 alone could stimulate CTLs. This disparity may be due to the fusion (F) protein of Sendai virus which may allow HN and F-containing liposomes to fuse to macrophages if the proteins are present in sufficiently high concentration. This interpretation is supported by the findings of Hale et al. (1980~)who found that Sendai virus proteins from virus grown in Madin-Darby bovine kidney (MDBK) cells, where the F protein is inactive, will not stimulate CTLs even in large amounts unless incorporated into liposomes with H-2 antigens. Recently, liposomes have been used to begin to define the antigens involved in tumor specific CTL responses. Alaba and Law (1978, 1980) have demonstrated that the plasma membrane antigens from the RBL-5 lymphoma, which express Rauscher murine leukemia virus antigens, can stimulate syngeneic CTLs specific for RBL-5. Furthermore, the soluble antigens have been chromatographed on an ACA 34 column and the most immunogenic pool applied to a lentil-lectin column. The eluate contained all the stirnulatory activity, suggesting that the antigens are glycoproteins. Duprez et al. (1983) have generated Moloney virus-specific CTLs utilizing reconstituted membranes from the MBL-2 lymphoma, which expresses Moloney murine leukemia viral antigens. Of particular interest is the fact that to stimulate these CTLs it was necessary to reconstitute these vesicles with the detergent-insoluble fraction described by Mescher et al. (1981). C. XENOGENEIC CTL RESPONSE Liposomes have also been utilized to explore the specificity of the murine antihuman (xenogeneic) CTL response. In the initial studies, spleen cells from C57BL/6 mice primed with the EBV-transformed lymphoblastoid cell line JY (HLA-A2,2; B7,7; DR4,6) were cultured with liposomes containing HLA-A and -B antigens isolated from the JY cell line. These liposomes induced murine CTLs with specificity

52


for JY target cells (Engelhard et al., 1978). The HLA antigens were >96% pure suggesting that HLA-A and HLA-B antigens alone are sufficient for the induction of CTLs. In these studies it was observed that the optimal response was dependent upon the density of the HLA molecules in the liposomes. It has also been shown that murine CTLs could be generated to HLA-DR antigens (Burakoff et al., 1980; Engelhard e t al., 1980). In order to obtain partially purified HLA-DR antigens, deoxycholatesolubilized JY membranes were chromatographed on either an anti&-microglobulin column or on a column containing of the antibody W6/32. This antibody recognizes all known HLA-A,B allospecificities; they lysed both JY targets and Daudi targets. Daudi cells express HLA-DR antigens (and share DR6 with JY cells) but do not express HLA-A,B products because of a genetic defect (Goodfellow et al., 1975). Second, the lysis was inhibited by a xenoantiserum anti-p29,34 which recognizes all known HLA-DR antigens. Third, targets sharing DR6 antigens were lysed best by these CTLs. CTLs induced with HLA-A,B containing liposomes lysed JY targets but not Daudi; their lytic activity was inhibited by the W6/32 antibody but not by antip29,34 and they preferentially lysed targets sharing HLA-A or -B antigens with JY. Recently these results have been confirmed by using purified HLA-DR antigens from a monoclonal anti-HLA-DR column, LB 3.1. These antigens have been incorporated into liposomes and stimulate DR-specific CTLs (Gorga et al., 1984). As mentioned, it was particularly surprising that HLA-DR antigens are able to induce a strong CTL response and, in fact, the response is at least as strong as that induced by HLA-A,B antigens. In human and rat xenogeneic responses, where murine stimulators and targets were used, there was no evidence that CTLs recognized H-21 region products (Lindahl and Bach, 1975; Burakoff et al., 1977; Carnaud et al., 1977). In allogeneic CTL responses where recognition of H-21 products has been demonstrated, the CTL response to H-21 antigens is far weaker than that elicited by H-2K and H-2D antigens. The CTL response to H-21 region antigens has been estimated to be less than 10% of the H-BK/D antigen response (Billings et al., 1977). IV. The Use of Liposomes to Study the Helper T Cell Pathway

A. EVIDENCE FOR HELPERT CELLREGULATION OF CTL RESPONSES

T-T interactions have now been identified in the development of virtually all manifestations of cell-mediated immunity. Evidence to


53

suggest the cooperative interaction among subsets of T cells for cellular immunity first came from the studies of Cantor and Asofsky (1970a). In studies of the graft-vs-host (GVH) reaction, elicited by transferring parental strain thymus-derived cells from adult animals into semiallogeneic neonatal F1 recipients, synergistic interactions between parental peripheral lymph node cells and parental thymocytes was demonstrated. This synergy occurred only when both populations were obtained from donors allogeneic to the host, and was interpreted to reflect the participation of at least two cell types in GVH reactivity. In subsequent studies (Cantor and Asofsky, 1970b) these two populations were distinguished by their relative sensitivities to neonatal thymectomy and heterologous antilymphocyte serum (ALS) as selection criteria. Using appropriate mixtures of cells from either parental strain into F1 hosts, it was found that one population, which localized to the spleen and thymus and was resistant to ALS, determined the specificity of the response and contained the precursors of the effector cells (designated TI) whereas the second cell type, found in highest concentration in peripheral blood and lymph nodes and sensitive to ALS in uiuo, appeared to serve as helper cells to amplify the response (designated T2). Both T1 and T2 cell populations were sensitive to neonatal thymectomy, indicating the thymic dependence of each population. It was then demonstrated that in a primary in vitro response to alloantigens, the cytotoxic activity of limiting numbers of peripheral T cells was increased 10- to 20-fold by the addition of quantities of thymocytes which alone were insufficient to yield significant cytotoxicity (Wagner, 1973). The dissection of the subpopulations involved in CTL responses has been more difficult than studies of humoral responses in part because the system used for generation of the CTL response appears to determine the existence of observable helper effects. Cantor and Boyse (1975) provided the first evidence that helper cells are active in in vitro generation of CTLs. It was demonstrated that under suboptimal stimulating conditions, the addition of Lyt-l+ cells augmented the response of Ly-2,3+ prekiller cells to alloantigens. In fact, it had been proposed by Bach et al. (1976) that maximal CTL activity to alloantigens is achieved by the coordinate stimulation of the effector CTLs by antigens in the H-2K or D regions, and of an MLC reactive helper T cell by I region determinants. Baum and Pilarski (1978) demonstrated an antigen-specific, radioresistant helper cell in the CTL response to alloantigens. If normal spleen cells were cocultured for 2 days with allogeneic stimulator cells, then X-irradiated and transferred into a second culture with

54


normal thymocytes as responders, these X-irradiated spleen cells markedly enhanced the response. The studies of Zinkernagel et aZ. (1978) and von Boehmer and Haas (1979) using radiation chimeras have suggested a role for helper T cells in the CTL response to viral and H-Y antigens. Zinkernagel et al. (1978) observed that when bone marrow cells from an H-2 incompatible strain or H-2K and H-21 incompatible strain (A) were allowed to mature in a lethally irradiated host (B), after viral infection the mature A T cells were incapable of mounting a CTL response even when transferred into an acutely irradiated (A X B)FI recipient. From these results the authors argued that only A helper cells with a self MHC receptor for IBand a receptor for virus developed in the B recipient. These helper cells were unable to confer help because they could only encounter A prekiller cells to interact with. These results are in conflict with the data of Doherty and Bennink (1979)using acute filtration to study the CTL response to viral infection. Their data suggest that I region products may not be essential in the CTL response to vaccinia virus, If BALB/c (H-2d) spleen or lymph node cells were depleted of T cells reactive to C3H ( H-2k)by thoracic duct cannulation and were injected intravenously into A/J (KkI-AkDd) irradiated recipients, CTLs could be generated to H-2Kkplus virus and to H-2Dd plus virus. Thus, CTLs were generated in the absence of I region homology between the responder lymphocytes (BALB/c) and the stimulator environment (A/J). The results can be reconciled if it is considered that the Ik region antigens plus virus look like Id plus virus; or alternatively, but unlikely, in contrast to the H-SK/D specificity, which is determined by the recipient environment, for I region products BALB/c T cells can present virus to other BALB/c T cells. Hamaoka et aZ. (1979)presented evidence for helper and suppressor T cells in a modified syngeneic tumor system. They found that if C3H mice were immunized with TNP-modified isologous mouse y-globulin (MGG), then immunized 8 weeks later with a TNP-modified syngeneic tumor X5563, the CTL response to the unmodified tumor was augmented. The CTL response was amplified if the mice were first treated with TNP-D,GL which eliminated TNP specific suppressor cells induced by the TNP-MGG immunization. The helper cell induced by the TNP-MGG immunization was radioresistant. Other groups (Finberg et aZ., 1979; Cooley and Schmitt-Verhulst, 1979; Fujiwara and Shearer, 1980)also found evidence for a TNP-specific, radioresistant helper T cell that augments the CTL response to TNP modified syngeneic cells. The helper T cells were generated in vivo after which they were X-irradiated and then added to in vitro culture with


55

naive syngeneic spleen cells and X-irradiated TNP-modified syngeneic stimulators. Ashman and Mullbacher (1979) and Reiss and Burakoff (1981) have shown that primed helper cells added to in vitro cultures of unprimed spleen cells can allow a primary antiviral CTL response to take place. Keene and Forman (1982) demonstrated that the generation of an in vitro CTL response to the Qa-lb alloantigen required prior in vivo priming with two kinds of antigenic determinants; one served as the target antigen for the CTL and the other as a “helper determinant” whose recognition is under H-2 linked Ir gene control. Glasebrook and Fitch (1980) extended the observations of helper T cells for CTL responses with cloned T cell lines. They demonstrated that amplifier cell lines which are stimulated by mls determinants, secrete soluble factor(s) which, in turn, induce the proliferation of cytolytic T cell lines in response to antigen. It has been consistently more difficult to demonstrate a helper cell requirement for secondary CTL responses to alloantigen. Okada and Henney (1980) found that the presence of Ly-l+ cells was required for the generation of secondary cytotoxic responses to UV or heat-inactivated stimulator cells. We have found that whereas all primary allogeneic responses appear helper dependent, with a population of primed responder cells helper dependence can only be demonstrated with a suboptimal antigen dose (Weinberger et al., 1980, 1982). Just as soluble factors were found to participate in interactions between T cells, B cells, and macrophages, they have been also found to mediate T-T interactions in cell-mediated responses. Altman and Cohen (1975)were the first to describe that a nonspecific soluble factor, generated in a short-term MLC, could augment the activity of T lymphocytes against allogeneic fibroblasts in vitro as indicated by a marked increase in the subsequent CTL responses. Such findings were interpreted as indicating that, as a result of stimulation of one subset of T cells by lymphocyte-defined (LD) antigens, a soluble mediator is released which facilitates the differentiation of another subset of T cells into functional CTLs. The data, however, did not give positive proof of this interpretation for it was possible that the amplification effect by the MLC supernatants was due merely to a “conditioning effect” or improvement of the culture conditions. Subsequent to the initial studies there have been many reports of allo-antigen or mitogen-stimulated T cell-derived factors which are capable of augmenting suboptimal CTL responses (Plate, 1976, Finke et al., 1977; Wagner et al., 1980; Simon et al., 1977; Okada et al., 1979; Rulon and Talmadge, 1979; Lalande et al., 1980). It was also found that a T cell factor produced during mitogen or alloantigen stimulation of T cells is

56


capable of supporting the long-term growth of antigen-activated T cells, both for murine T cells (Glasebrook and Fitch, 1980)and human T cells (Morgan et al., 1976). The secondary CTL responses induced by supernatants of human secondary MLCs (Uotila et al., 1978) and of murine MLCs (Ryser et al., 1978) were found to be antigen nonspecific. Furthermore, it was demonstrated that this helper activity was the product of the interaction of Ly-l+ T cells and macrophages. All the murine non-antigen-specific helper activities appear to be related. In fact, Watson et al. (1979) reported that a single class of molecules of 30,000-35,000 MW and similar charge was responsible for four distinct lymphokine activities, i.e., enhancement of PFC responses, induction of primary CTL responses, augmentation of secondary CTL responses, and maintenance of T cell lines in long-term culture. This Ly-l+ T cell derived helper factor has been designated Interleukin 2 (IL-2). The biological activities of rat and human helper factors were found to be identical to murine IL-2, but their molecular characteristics were considerably different (Watson and Mochizuki, 1980). It is possible that IL-2 from each of these species may interact with common cellular receptors on or in murine cells, implying that murine, rat, and human IL-2 share structural homologies. It is also possible that IL-2 purified from these sources is an unrelated group of molecules that interact with common target cells at different receptor sites. After antigen recognition, the provision of a second signal by the helper T cell in the form of IL-2 induces clonal proliferation and growth. These proliferating CTL precursors will exhibit cytolytic effector function only if a further lymphokine, CTL-differentiation factor, CTDF, is provided (Wagner et al., 1982; Raulet and Bevan, 1982; Garman and Fan, 1983; Finke et al., 1983). €3. CELLULAR REQUIREMENTS FOR T HELPERCELLS

THE

INDUCTION OF

The utilization of populations of T cells depleted of adherent accessory cells in studies of antigen-induced proliferation in the guinea pig provided the first evidence for the involvement of accessory cells in the presentation of antigen to T cells (Seeger and Oppenheim, 1970). Subsequently, histocompatibility requirements for macrophage-T cell interactions, which mapped to the I region, were observed by Rosenthal and Shevach (1973). Later studies in the mouse confirmed these observations. Analogous results were reported by Pierce et al. (1976) in studies on cell collaboration in secondary antibody responses, and by Kappler and Marrack (1976) for the in uitro priming of helper T cells. Other investigators, however, failed to find genetic


57

restrictions of macrophage-T cell interactions in in vitro primary immune responses (Kapp et al., 1973). They demonstrated that T cells could be sensitized to antigen which was bound to allogeneic macrophages, under the appropriate conditions of minimizing alloreactivity. Evidence has accumulated to suggest that Ia+ adherent accessory cells are required for the induction of CTLs (Wagner et al., 1972; Pettinelli et al., 1979; Weinberger et al., 1980; Rouse and Lawman, 1980). This conclusion has been drawn from the observation that the removal of splenic adherent cells (SACs) abrogates the ability of prekillers to differentiate to CTLs. If the SACs are treated with anti-Ia antiserum and complement to remove Ia+ cells, the residual Ia-negative SACs are unable to stimulate a response. Furthermore, anti-Ia antiserum, directed at the Ia antigens of the SACs added to cultures, also inhibited the induction of a CTL response, suggesting that alloantigen is recognized in association with the Ia determinants on the accessory cells (Weinberger et al., 1980). We have demonstrated that the specific recognition of allogeneic Ia determinants is indeed an important element in the generation of alloreactive CTL, as originally proposed by Bach et al. (1976). Such stimulation, however, is by no means the only pathway leading to the induction of a CTL response, for even in the absence of allogeneic Ia determinants, alloantigens may be recognized in a manner similar to conventional antigens are, i.e., they are recognized in the context of syngeneic Ia determinants on antigen presenting cells (Weinberger et al., 1982). This observation has been extended to explain the strong MLC, GVH, and CTL responses by T cells of H-2b haplotype mice to the B6.C.H-2bm1mutant, in the apparent absence of any alloantigenic differences in the I region (Weinberger et al., 1983; Rock et al., 1983). A number of studies have suggested that the accessory cell involved in the presentation of antigen to T lymphocytes is of the macrophage lineage (reviewed by Unanue, 1981). Recently, however, it has been demonstrated that dendritic cells can be present in cell populations that are generally considered to be purified macrophages, especially splenic adherent cell populations (Steinman et al., 1980). These cells bear large amounts of Ia and can act as potent accessory cells for the generation of MLC and cytotoxic responses (Klinkert et al., 1980; Steinman and Nussenzweig, 1980; Sunshine et aZ., 1982). Several B cell lymphoma lines (Glimcher et al., 1982; Walker et al., 1982) as well as endothelial cells (reviewed by Lipsky and Kettman, 1982) have also been demonstrated to function as potent antigen-presenting cells.

58


C. ANTIGENICREQUIREMENTS FOR THE INDUCTION OF HELPERT CELLS Although it appears evident that SACs play a role in the generation of CTLs, it has been difficult to evaluate and precisely define this role. As mentioned, part of this difficulty has resulted from the fact that the MHC antigens responsible for the CTL response are integral cell membrane proteins. Studies of the CTL response have required that antigen be added in the form of allogeneic or modified syngeneic cells. In contrast, the analysis of the role played by SACs in the humoral response has progressed far more rapidly because the relevant antigens are soluble, and thus more easily manipulated (reviewed by Pierce and Kapp, 1976). We have utilized purified H-2Kk antigen incorporated into liposomes to better define the cells and events required for the induction of prekillers and for the activation of the helper T cell. Initially, it was observed that the generation of CTLs by liposomes containing H-2Kkantigens required the presence of Ia-positive, Thy-l-negative accessory cells which were radioresistant and adherent to plastic, glass, and nylon wool (Weinberger et al., 1980). The use of 1251-labeledH-2Kkin liposomes has allowed a further evaluation of the function of these SACs, especially their role in antigen presentation. SACs were isolated from normal CD2Fl (H-2d)mice and incubated alone or pulsed overnight with liposomes containing 1251labeled H-2Kk. After incubation, the SACs were washed and the amount of antigen present on the pulsed SACs was determined by measurement of radioactivity. Spleen cells from CD2Fl mice that had been immunized to YAC (H-2KkDd)cells were placed in culture with various numbers of CD2Fl SACs with bound H-2Kk antigen. Parallel cultures had an equivalent amount of antigen in liposomes and an equivalent number of SACs that had not been pulsed with antigen. Twenty-fold less H-2Kkantigen was necessary to generate an equivalent CTL response if the antigen was bound to SACs. Antigen presentation was not affected by the elimination of Thy-1 positive cells or by X-irradiation of the SACs. If, however, SACs were treated with anti-Ia antiserum and complement to remove Ia-positive cells, the residual Ia-negative SACs pulsed with liposomes containing H-2Kk were unable to stimulate a CTL response. Anti-Ia antiserum, directed at the Ia antigens of the SACs added to cultures, also inhibited the induction of a CTL response suggesting that H-2Kk is recognized in association with Ia (Weinberger et al., 1980). Although the requirement for antigen-presenting cells has been firmly established, the nature of the cells actual function has not been


59

as clearly defined. Whether the accessory cells passively provide a suitable surface and the necessary soluble factors for stimulation of the T cell, or alternatively actively “process” the antigen, has been a topic of much debate. Processing may be an energy-dependent step required to intercalate the antigen into the SACs membrane or it may be a more complex process in which antigen is endocytosed, broken down, and reexpressed on the cell surface of the SACs in an immunogenic form. It has been demonstrated that SACs play an active role in the presentation of H-2Kk (Weinberger e t al., 1981a). If SACs were inactivated by UV irradiation prior to incubation with H-2Kk antigen, the stimulation by these SACs was abrogated. SACs could, however, stimulate a CTL response if allowed to interact with antigen for longer than 4 hours prior to UV inactivation if phorbol myristic acetate (PMA) (known to substitute for macrophage-derived IL-1) was added to culture. We have recently obtained evidence that after SACs interact with the Kk antigen, they can be lysed by freezing and thawing, and their partially purified plasma membranes can trigger a CTL response if PMA is added to the culture. Such SAC membrane preparations were stimulatory when they were isolated from SACs after they had interacted with the Kk liposomes at 37°C for 18 hours. These results suggest that (1)SACs must actively interact with antigen, i.e., “process” the antigen and (2) elaborate a soluble factor, Interleukin 1. Thus, H-2Kkin liposomes has been utilized to show that SACs provide at least two distinct functions in the activation of the CTL response. One of the first studies thought to indicate antigen processing by the macrophage was reported in the early 1960s by Fishman and Adler (1963). They found that a fragment of the antigen could be isolated in a form associated with the RNA extract. Many objections were later leveled at this work, including the possibility that the observed association was merely an artifact of the extraction method used. In later studies on the presentation of antigen by accessory cells it was observed that their was an obligatory lag period between antigen uptake and effective presentation, seeming to indicate some “processing” of the antigen (Waldron et al., 1974; Ziegler and Unanue, 1981; Weinberger et ul., 1981a; Chestnut et al., 1982), Ellner and Rosenthal (1975) reported that trypsinization of macrophages after allowing them to interact with the antigen (DNP-GPA), thus removing cellsurface molecules, did not affect their ability to stimulate a response. Furthermore, exposure of the macrophages to DNP-specific antibody had no effect on the response. It was concluded that the antigen molecules which were relevant for T cell recognition were at an intracellu-

60


lar site, and thus resistant to trypsin and antibody. Other attempts to block the responses of T cell-macrophage interactions with antibodies also failed. Ben-Sasson et al. (1977) demonstrated that antibodies to soluble protein antigens did not inhibit the specific binding of immune lymphocytes to antigen-pulsed macrophages even when the concentration of antibody in the incubation medium was high enough to precipitate greater than 100 times the amount of antigen that remained associated with the macrophage after washing. Furthermore, Werdelin and associates (1979) reported that they were unable to block the specific binding of T cells to macrophages with a great excess of soluble antigen. These studies have been extensively reviewed by Unanue (1981). The implications from these studies were that T cells were either not recognizing native antigenic determinants on the macrophage cell surface, or else were recognizing antigen after some sort of macrophage “processing.” Accordingly, possible explanations for the inability to inhibit T cell proliferation with antiantigen antibody would be (1) the immunologically relevant antigen is in a site inaccessible to the antibody (e.g., intracellular or buried in the membrane); (2) the T cell receptor recognizes an antigen-induced alteration of macrophage Ia and not native antigen; (3) the T cell receptor recognizes antigen fragments in association with macrophage Ia in contrast to conformational determinants or the native antigen which are recognized by the B cells (Rosenthal et al., 1976; Benacerraf, 1978). It could be argued that when the T cell interacts with cell-bound antigen, it does so via a very large number of binding sites and receptors, thereby favoring the cellular interactions over interactions with soluble antigen or antibody. Another approach was used by Korngold and Sprent (1980) to demonstrate antigen processing. For responses to minor histocompatibility antigens, they found that glutaraldehyde treatment of the immunizing spleen cells prevented stimulation by H-2 incompatible cells. It was suggested that the treated cells could not be effectively processed by host macrophages and thus were not recognized in the context of syngeneic Ia, and therefore were not immunogenic. In order to probe the intracellular processing involved in antigen presentation, agents which interfere with the degradative lysosomal pathways have been used. Ziegler and Unanue (1982) have demonstrated that chloroquine and ammonia, two lysosomotropic compounds, inhibited both the degradation of Listeria organisms by peritoneal macrophages and the binding of Listeria-primed T cells to such macrophages. Grey et al. (1982) reported similar inhibition of presentation of KLH and rabbit anti-mouse Ig by chloroquine treatment of


61

macrophages. Lee et al. (1982) demonstrated that chloroquine exerts a selective inhibitory effect on the presentation of large antigens (PPD, CRBC, and polymeric flagella) without affecting the presentation of small antigens (monomeric flagellin). These studies suggest that lysosoma1 catabolism is involved in the conversion of large particulate antigens into an immunogenic form. In an attempt to define whether antigen degradation is sufficient to invoke a response, or whether antigen processing involves other modifications to the antigen as well, Shimonkevitz et al. (1983) demonstrated that for the complex antigen chicken ovalbumin, chemical or enzymatic fragmentation of the antigen was in fact both necessary and sufficient for antigen presentation. In contrast, there are studies which seem to suggest that a small amount of antigen is retained on the macrophage surface, essentially in its native configuration and these molecules are the biologically relevant ones. Unanue and Cerottini (1970) demonstrated that (1) whereas most protein antigens are interiorized and rapidly catabolized, a small amount of antigen escapes catabolism and either remains cell-associated or is released, and (2) the immunogenicity of macrophage-bound hemocyanin is sensitive to treatment with trypsin. Furthermore, Shevach et al. (1979) found that whereas addition of antiTNP antibodies immediately after conjugation of TNP blocked responses, TNP presentation by “aged” macrophages (i.e., 24 hours after pulsing with TNP) was not affected. The presentation of antigen by aged TNP-macrophages, however, could be inhibited by anti-TNP antibodies if the cells were derivatized again after 24 hours. It was concluded that a few molecules of TNP were available on the membrane but, being so few, could not be cross-linked by antibody and were not modulated, until the amount was increased by the addition of fresh antigen. Thomas (1978) has also argued against antigen processing based on the observation that glutaraldehyde treatment of TNP-modified macrophages did not abrogate stimulation of T cell proliferation. More recently, Malek and Shevach (1982)demonstrated immunochemically that essentially all macrophage cell-surface associated GLT was nondegraded and in its native conformation, suggesting that antigen presentation did not necessitate “processing” for GLT. It appears very likely that prior to immune recognition by T lymphocytes, different antigens will be functionally “processed” in different ways. One can envision many possible steps in the conversion of a native antigen into an immunogenic form: (1) antigen binding, (2) uptake and sequestration, (3) degradation or “processing,” and (4) reexpression on the cell surface or release as an antigen-Ia complex.

62


There appears to be a difference in the amount of processing required depending on the size and complexity of the antigen studied. Our interest in the analysis of induction requirements for CTL responses led us to study the role of the antigen-presenting cell in the presentation of the MHC antigens. As mentioned earlier, such studies have been difficult to implement due to their being integral membrane proteins. The utilization of liposome technology however, has now allowed us to address this issue using purified H-2Kkas the antigen. In order to determine whether antigen processing involves extensive degradation or modification of MHC antigens, or whether the integration of the antigen into the lipid bilayer of the SAC is sufficient, the following liposome construction was made: partially purified membranes of an Iad-positive antigen presenting B cell lymphoma were solubilized in detergent, and were subsequently reconstituted together with purified Kk antigen in the presence of aprotinin, a protease inhibitor. These liposomes, which contain both Iad and Kk antigens, were then cultured with primed CD2Fl responder cells for 5 days and then cytolytic activity was assessed. A cytolytic response was induced by these liposomes, if however, Iad and Kk antigens were incorporated into separate liposomes and both liposomes were added to the same culture no CTL activity was stimulated. Furthermore, the induction of this response was blocked by the addition of anti-Iad antibody to the cultures. This inhibition could be bypassed by the addition of supernatants from Con A-stimulated T cells indicating the involvement of helper T cells in the recognition of this antigen. This observation suggests that for a response to the transmembrane antigen H-2Kk,it is sufficient for the helper T cell to see syngeneic Iad plus the Kk alloantigen in the same lipid bilayer and that “processing” for this antigen is simply its incorporation into the lipid bilayer (Weinberger et al., 1981). Our earlier studies demonstrated that SACSpulsed with H-2Kkliposomes present antigen to Ly-l+ helper T cells and these helper T cells then elaborate a helper factor which provides a signal required for the prekiller cell to differentiate to a CTL. If SACS or Ly-l+ helper T cells are eliminated from culture, the addition of exogenous helper factor(s) (derived from a rat Con A supernatant) together with antigen can bypass the helper pathway and allow the prekiller cell to respond to antigen (Weinberger et al., 1981b). Studies of Lalande et al. (1980) have suggested an approach to studying the generation of each of the necessary signals separately. They found that a 12-hour exposure of precursor CTLs to alloantigen was sufficient to result in a CTL response if the second signal, normally produced by helper T cells, is


63

provided by the addition of Con A supernatant. These experiments were done by mixing effector cells and allogeneic stimulator cells, incubating for 12 hours, and then separating the two cell populations using a flow cytometer-cell sorter. Liposomes can be used very effectively for studying the two signals in this way and provide considerable advantage over the use of whole-cell stimulators and cell sorting for separation. Effector cells can be briefly exposed to alloantigen in liposomes and the antigen can then be separated from the cells by differential centrifugation. The cells can be readily pelleted under conditions where the liposomes remain in suspension. Furthermore, the amount of antigen carried over with the cells following washing can be easily measured by using radioactively labeled H-2. Using this approach we have confirmed the findings of Lalande et al. that a 12-hour exposure to antigen is sufficient to result in generation of a maximal CTL response if a second signal is provided by adding helper factor(s) (Herrman et al., 1982). A variety of control experiments show that the response cannot be due to free antigen carried over with the cells following separation from the liposomes. Depletion of SACs has no detectable effect on the response, further suggesting that uptake and presentation of antigen are not necessary for recognition by the pre-CTL. The second signal, necessary after the 12-hour antigen pulse, can be provided by helper factor(s) or by adding SACs and additional alloantigen to the culture. This presumably is due to helper T cell production of a helper factor(s) following recognition of alloantigen and Ia on the SACs. Experiments examining the requirements for timing of addition of helper factor to produce an optimum response have indicated that pre-CTL must recognize alloantigen first for exposure to helper factor(s) to be effective. Using the protocol described above, it is possible to independently study the molecular requirements for generation of each of the signals. If effector cells are pulsed with an optimum amount of alloantigen for 12 hours and then washed, the level of response obtained is dependent on generation of the second signal. This provides an excellent system for studying macrophage processing and Ia-dependent presentation of antigen, particularly since the structure of the H-2 antigen is well understood and modification of it is easily assessed. V. The Use of DNA-Mediated Gene Transfer of Cloned MHC Genes to Study CTL Specificity

Though liposomes, with incorporated H-2 antigens, have been extremely useful in studying the antigenic requirements for helper T

64


cell and CTL induction, we have been unsuccessful in utilizing liposomes to study the CTL-target cell interaction. Specifically, we have been unable to use liposomes or any form of subcellular material, including membranes, to competitively inhibit CTL-target cell interactions. Recently, the isolation of genes that encode MHC antigens and the ability to transfect them, by DNA-mediated gene transfer, into eukaryotic cells has provided a new approach to the exploration of the specificity of the CTL-target cell interaction. Briefly, to utilize this approach the gene(s) of interest must first be cloned and characterized. A specific probe is essential to identify the gene(s) from a genomic library (reviewed in Maniatus et al., 1982). The cloned DNA is inserted into a plasmid (Thomas et al., 1974), grown in quantity in bacteria, and then used in DNA-mediated gene transfer (Wigler et al., 1979). In most laboratories, the Ca(P04)z method has been used to introduce DNA into eukaryotic cells (Graham and ver der Eb, 1973). To select for cells which have incorporated the gene(s) of interest, most laboratories have utilized cells which are deficient in the thymidine kinase gene (tk-) and cotransform with the thymidine kinase gene derived from Herpes simplex virus and the gene of interest, then expose the transformed cells to medium supplemented with hypoxanthine, aminopterin, and thymidine (HAT) (Littlefield, 1964). Only cells which have received the tk gene are able to survive. In other studies, selection has been done with other markers such as neomycin resistance using bacterial phosphotransferase APH 3’1 (Berg et al., 1978). Only a proportion of the colonies of cells which survive the selection will have been cotransformed with the gene(s) of interest as well as the selective marker. The next step is to identify which colonies express the gene product on the cell surface. Generally this is done serologically by fluorescence analysis, radioimmune assay, or by enzyme-linked immunoadsorbant assay. Once cells are identified that express high levels of the products of the transferred gene(s) one can analyze them for biologic activity. We will review those studies where transfected cells have been used as targets for CTLs. The nucleotide sequence of cloned class I MHC genes demonstrates that these genes are composed of seven or eight exons separated by introns (Hood et al., 1982; Malissen et al., 1982a). The three extracellular immunoglobulin-like domains of the class I molecule are encoded by three exons termed N, C1, and C2 in the mouse ( a l , a2, a3 in man), a transmembrane (TM) domain, and an intracellular domain in mouse encoded by three exons I1,12, and I3 and by 2 exons in man.


65

These glycoproteins are expressed on the cell in association with µglobulin which appears to bind to the C2 domain (Yokayama and Nathenson, 1983). The first exon encodes the leader (L) segment which is removed posttranslationally from the protein. Recently, a number of laboratories have cloned genes for some of the alleles of the K, D, and L loci and using the technique of DNAmediated gene transfer they have been expressed in mouse L cells (H-2k).Their gene products have been shown to be effective targets for allospecific CTLs (Evans et al., 1982; Mellor et al., 1982; Orn et d.,1982; Margulies et al., 1983; Ozato et al., 1983; Reiss et al., 1983a; McLaughlin-Taylor et al., 1983), hapten-specific CTLs (Levy et ul., 1983) and for virus specific CTLs, including CTLs to lymphocytic choriomeningitis (LCM) virus (Om et al., 1982; Frelinger et al., 1983; McLaughlin-Taylor et al., 1983), vesicular stomatitis virus (VSV) (Reiss et ul., 1983a; Forman et al., 1983),influenza virus (Mellor et al., 1982; Reiss et al., 1983a), and Sendai virus (Reiss et al., unpublished observations). Other class I gene products including Qa 2, 3 antigens may also be the target of allospecific effectors following DNA mediated gene transfer ( J . A. Frelinger et al., personal communication). Both bulk CTL populations and CTL clones (Mellor et al., 1982; Reiss et al., unpublished observations) have been effective at lysing L cells expressing transferred class I gene products. Thus, these functionally active molecules expressed on L cells can be powerful tools in analyzing the specificity of the immune response. To achieve more precise localization of the site(s) recognized by CTLs recombinant H-2 genes were constructed by the technique termed “exon shuffling” in which the exons encoding the L, N, and C1 domains were exchanged between cloned Ld and Dd genes (Evans et al., 1982). L cells were transfected with these recombinant genes and expressed hybrid H-2 molecules. These transfected L cells were used as targets for CTLs. Allospecific and virus specific CTLs restricted to either Ld or Dd were generated and assayed on these L cells expressing hybrid H-2 molecules. Both VSV and influenza virus specific CTLs and allospecific CTLs were found to recognize the N and/or C1 domains of the class I molecule (Reiss et aZ., 1983a; Ozato et

al., 1983). The importance of the N and C1 domains had been suggested by the amino acid analysis of the Kb mutants (Pease et al., 1983) and the HLA-A2 variants (Krangel et al., 1983). The Kb mutants which occur in N or C1 have been shown to effect alloreactive CTL (Sherman, 1982) and antiviral CTL (Hurwitz et al., 1983; Melief et al., 1983) recognition as have the HLA-A2 variants.

66


Human class I MHC genes have also been cloned, transfected, and expressed in L cells (Barbosa et al., 1982; Lemmonier et al., 1982). When expressed on L cells, HLA antigens are associated with murine &-microglobulin unless experimentally cotransfected with a human p2 gene. Analysis of HLA transfectants with a panel of monoclonal antibodies to &-microglobulin has shown a modification of the light chain conformation (Lemmonier et al., 1983). Several groups have attempted to demonstrate CTL recognition of HLA antigens on L cells. In most cases human allospecific bulk cultures and cloned CTLs have proven ineffective at lysing the L cell transfectants, although they lysed appropriate lymphoblastoid targets (Bernabeu et al., 1983). Other studies were carried out using bulk xenospecific mouse anti-HLA effectors; these were also ineffective (Herman et al., 1983; Bernabeu et al., 1983). Even in the presence of lectins, human allospecific T cell clones were unable to lyse L cell transfectants (Bernabeu et a,!., 1983). L cell transfectants, expressing either the HLA-A2 or B7 heavy chain and human P2-microglobulin, were not lysed by a small panel of human CTL clones or murine antihuman CTL bulk cultures (C. Bernabeu, personal communication). In contrast, using a panel of murine CTL clones specific for HLA-A2 or B7, a few were found to lyse L cells expressing appropriate A2 or B7 genes almost as well as they lysed appropriate human lymphoblast targets. However, clones did not kill these L cell transformants and others were found to have intermediate activity (Herman et al., 1983). The difficulty in lysing murine L cells expressing human MHC antigens may have several explanations. Many determinants recognized by CTLs may be lost when an inappropriate P2-microglobulin associates with the class I heavy chain. In addition, posttranslational modifications such as glycosylation may be different in L cells than in EBV-transformed human B cell lines. Alternatively, mouse L cells may lack some critical accessory cell surface interaction molecules which are essential for lysis by human CTLs (see Section VI). Truncated murine class I MHC genes have also been used to transfect L cells. In one case, when the gene was deleted of all material 3' from the beginning of the C2 exon this transfected gene was found to spontaneously recombine in L cells with endogenous DNA and a full length Ld molecule was expressed (Goodenow et aZ., 1983). In other experiments the TM exon of the Ld gene was retained and various constructs involving the 3' end encoding the cytoplasmic portion of the class I molecule were made (Zuniga et al., 1983). Several constructs were studied, one in which 24 of 31 C terminal amino acids of the molecule were deleted and another where the 25 C terminal


67

residues were replaced with a unique 19 amino acid sequence. These constructs were transfected into L cells which were used as targets for allospecific and virus-specific CTLs. Allospecific and virus-specific CTLs lysed the L cells expressing the class I molecules with these altered cytoplasmic domains of Ld equally as well as target cells expressing unaltered Ld. In contrast, Murre et al. (1983) (and Reiss et al., 1983b) examined another truncated construction of Ld in which the three exons encoding the intracytoplasmic portion of the molecule were deleted and replaced by the 3’ end of the I-APd gene starting from the terminal exon. When this truncated gene was transfected into L cells and used as a target for allospecific or influenza virus-specific CTLs, the truncated molecule was indistinguishable from the wild type Ld molecule. However, a 5-fold decrease in susceptibility to lysis by VSV specific effector T cells was noted with targets expressing the truncated Ld gene product. This data suggests that the cytoplasmic portion of Ld plays a role in CTL recognition of VSV; e.g., it may contribute to a physical interaction with the VSV G protein in the formation of an immunogenic target for CTLs. Genes coding for viral antigens have been cloned and sequenced for a number of viruses, among these are the genes for influenza virus, VSV, Semliki Forest virus, SV40, and murine leukemia viruses. The influenza virus hemagglutinin (HA) gene (Gething and Sambrook, 1981; Hartman et al., 1982) and neuraminidase gene (Davis et al., 1983) have been transfected and expressed in monkey cells. The gene encoding the VSV G protein (Rose and Gallione, 1981) has been recently expressed in mouse C127 cells (Florkiewicz et al., 1983). The Semliki Forest virus capsid protein genes have been introduced into baby hamster kidney cells (Garaff et al., 1983). Several of these viral genes have been transfected into cells and the transfected cells have been used as targets for virus-specific CTLs. BALB/c 3T3 cells have been transfected with the cloned murine Moloney leukemia virus envelope gene. These cells expressed the e m gene products p15E and gp70 and these cells were lysed by Moloney virus-specific CTLs (Flyer et al., 1983). Also the gene encoding the SV40 T antigen introduced into L cells (Tevethia et al., 1983) can be recognized by murine CTLs specific for that virus (Gooding and O’Connell, 1983; S. S. Tevethia, personal communication). VI. The Use of Monoclonal Antibodies to Define Functional CTL Antigens

DNA-mediated gene transfer of cloned MHC genes is providing a powerful approach for study of the target cell antigens. The genera-

68


tion of monoclonal antibodies to cloned or pauci-clonal functional T cells and the subsequent isolation of those antibodies that block functional activity is proving to be an extremely useful approach in defining the functionally important antigens on effector T cells. In this section we will focus on the information gained for the functional antigens on CTLs. We will concentrate on the antigens on human CTL because they are well characterized and because our contributions have been in this area. The fact that the effector cell antigens are better characterized for human CTLs than for murine CTLs may be due to the ability to immunize across a greater species barrier and more readily raise monoclonal antibodies to these effector cell antigens. The use of monoclonal antibodies to detect cell surface molecules important in the murine CTL response has repeatedly demonstrated a role for Lyt-2,3 (Nakayama et al., 1979; Shinohara and Sachs, 1979; Hollander et al., 1980), and LFA-1 (Springer et ul., 1982). Recently, Dialynas et al. (1983) have defined a cell surface structure L3T4 involved in the recognition of Ia antigens. We will briefly update our current knowledge of the human cytolytic T lymphocyte response. The discoveries that normal human T lymphocytes can be maintained in continuous culture with T cell growth factor (TCGF) (Ruscetti et al., 1977) and that antigen specific CTLs could be selected, maintained, and cloned (Gillis and Smith, 1977; Bach et al., 1979; Goulmy et al., 1980; Kornbluth and Dupont, 1980; Lotze et al., 1980) have allowed major advances in the understanding of the human CTL response. Long-term continuous cell lines and clones have been used to (1)characterize human CTL with regard to phenotype, morphology, and function, (2) define T cell antigen specificity, and (3) detect function-associated T cell surface markers. A. CHARACTERIZATION OF HUMAN CTLs Reinherz and Schlossman (1980) have defined a series of antigens on human T cells that have proved useful in delineating human T cell differentiation. It appears that early thymocytes are OKTlO+ and as they mature they acquire OKT6, OKT4, and OKT8 and express the OKTlO+, OKT6+, OKT4+, OKT8+ phenotype. OKT6 is the human equivalent of TL, while OKT8 is similar to Lyt-2 and OKT4 is similar to L3T4. The most mature thymocytes lose either OKT4 or OKT8 and are OKTlO+, OKTl+, OKT3+, OKT4+, or OKT10+, OKTl+, OKT3+, OKT8+.OKTl may be similar to Lyt-1. Peripheral T cells lose OKTlO and bear either the OKTl+, OKT3+, OKT4+, or OKTl', OKT3+,


69

OKT8+ phenotype. OKTl+, OKT3+, OKT4+ cells tend to be helper or inducer T cells and OKTl+, OKT3+, OKT8+ cells tend to be cytotoxic or suppressor cells. Initially, it was thought that the OKT4 and OKT8 antigens were markers for functionally distinct subsets of T cells. Recent evidence suggests, however, that OKT4 is a receptor that recognizes HLA-DR antigens and OKT8 is a receptor for HLA-A,B antigens (vide infra). Although there are both HLA-DR specific helper T cells and cytotoxic T cells, the majority of the cells that respond to HLA-DR are helper cells giving the impression that OKT4 is a functional marker rather than an antigen recognition structure (Krensky et al., 1982a,b; Meuer et al., 1982a; Biddison et al., 1982a). OKT6 antigen is expressed by essentially all thymocytes but is not normally expressed by peripheral blood lymphocytes. Although CTLs generated from PBL are OKT6-, we have recently generated HLADR-specific CTL clones from thymocytes and found some to be OKT3+, OKT4+, OKT6+, OKT8- (Krensky et al., unpublished observation). Although resting T cells express little, if any, detectable HLA-DR antigen, the majority of activated T cells rapidly synthesize and express HLA-DR antigens (Evans et al., 1978; Charron et al., 1980). A number of investigators have shown that human CTLs express HLADR antigens (Metzgar et al., 1979; Moretta et al., 1983). Recently, Pawelec et al. (1982a) showed that CTL lines and clones express both HLA-DR and SB antigens, but that these antigens are unable to stimulate secondary proliferative responses of primed lymphocytes. With the generation of CTL lines and clones, it has become clear that CTLs are capable of functions other than cytolysis. Luger et al. (1982) and Meuer et ul. (1982b) have recently shown that both OKT4+ and OKT8+ subsets of PBL are capable of producing IL-2. Wee et al. (1982) have isolated and expanded an antigen driven, helper cellindependent cytotoxic T cell clone that mediates cytotoxicity and proliferates to alloantigen in the absence of exogenous lymphokine, suggesting that this clone produces IL-2. Matsuyama et al. (1982) induced CTL to produce interferon (IFN) upon stimulation with Con A. Kornbluth et al. (1981) performed a series of morphological and cytochemical characterizations of CTL colonies derived from 5 to 6 days agarose MLC cultures. The cells were found to be homogeneous in size and morphology; they were lymphoblastoid with basophilic cytoplasm and a large nucleous with prominent nucleoli. They were negative for peroxidase and alkaline phosphatase activity, indicating that they were not of the myeloid lineage.

70


B. SPECIFICITY OF HUMAN CTLs CTL lines and clones have been used to define the target antigens recognized by human T cells. Studies have provided new insights into the determinants and variety of MHC products recognized by CTL as well as expanding our understanding of MHC restriction. Although HLA-A,B antigens are the major target antigens for human allogeneic CTLs (Reiss et al., 1980; Spits et al., 1981; Lotze et al., 1980; Malissen et al., 1981), Class I1 antigens, DR, DC, and SB (Ia in the mouse) are surprisingly important allogeneic target antigens. Albrechtsen et al. (1979), Feighery and Stastny (1979), and Johnsen (1980) generated MLCs between HLA-A,B identical donors to show that CTLs could recognize cell surface antigens other than Class I MHC molecules, presumably class I1 antigens. We stimulated peripheral blood lymphocytes with the Daudi cell line (an EBV-transformed B cell line devoid of Class I antigens because of a defect in the expression of Pz-microglobulin) and generated CTL lines and clones specific for HLA-DR6 antigens (Krensky et al., 1982a,b). DR specific CTL lines and clones have also been generated by Meuer et al. (1982a), Spits et al. (1982), Ball and Stastny (1982), and Flomenberg et al. (1983). Other Class I1 molecules can also be recognized by CTLs. Biddison et al. (1982a) derived SB-specific CTL lines and Spits et al. (1983) and we (Krensky et al., 1983a) have recently derived DC specific CTLs. Virus specific human CTLs to numerous viruses have also been described, including CTLs to influenza virus (McMichael et al., 1977, 1980; Biddison et al., 1979, 1980a), cytomegalovirus (Sethi et al., 1980; Quinnan et al., 1981), Herpes simplex virus (Sethi et al., 1980), mumps (Kress and Kreth, 1982), measles (Kreth et al., 1979), and Epstein-Barr virus (Misko et al., 1980; Rickinson et al., 1980; Moss et al., 1981; Tsoukas et al., 1981). Human CTLs to trinitrophenyl-modified syngeneic cells have also been described but Shaw and Shearer (1978)found only 40% of the CTL activity to be HLA restricted. Charmot and Mawas (1979) were unable to document clear cut HLA restriction in unrelated individuals. Neefe et al. (1982) recently derived a number of CTL clones which were restricted by HLA-A,B antigens but the restricting element often did not correspond directly to serologically defined antigens. Biddison et aZ. (1980b, 1982b) showed that HLA-A2 variants exist where CTLs differentially recognize molecules which were serologically typed as HLA-A2. Thus, they generated CTLs specific for influenza antigens in the context of HLA-A2. They found an individual expressing serologically indistinguishable HLA-A2 who failed to


71

serve as a target for the HLA-A2 restricted influenza specific killers. Conversely, variant HLA-A2 restricted CTLs failed to lyse influenzainfected HLA-A2+targets. Biochemical analysis of these HLA-A2 variants has revealed amino acid sequence differences in the a2 domain (&angel et al., 1983). Similar studies for H-Y specific (Goulmy e t al., 1982; Pfeffer and Thorsby, 1982) and allospecific (Horai et al., 1982; van der Poel et al., 1983; Ware et al., 1983a) CTLs provide further evidence for HLA-A2 heterogeneity.

C. CELLSURFACE MOLECULES INVOLVED IN CyroLYsis As mentioned, OKT4 (Leu3) and OKT8 (Leu2) appear to be receptors for Class I1 and Class I antigens, respectively. Other cell surface antigens also appear to be functionally involved in the CTL-target interaction. Initially, most of the T cell-specific monoclonal antibodies were generated by immunizing mice with human peripheral blood lymphocytes and screening the antibodies for binding to T cell (sub)populations (Kung et al., 1979; Ledbetter et al., 1981).Recently, several groups have isolated monoclonal antibodies by their ability to block functional T cell activity. Malissen et al. (198213)immunized BALB/c mice with anti-HLA-A2specific CTL clone, derived monoclonal antibodies and selected those monoclonal antibodies that were able to block the cytotoxic activity of the immunizing CTL clone. Seven monoclonal antibodies inhibited cytolysis by binding to a CTL surface molecule. The monoclonal antibodies isolated, termed B9.1 through B9.11, appear to recognize a variety of epitopes on the previously defined OKT8/Leu2a molecule. We immunized BALB/c mice with an OKT4+, HLA-DR-6-specific CTL line and screened monoclonal antibodies for inhibition of cytolysis by this CTL line. Binding of monoclonal antibodies to four types of molecules, LFA-1, LFA-2, LFA-3, and HLA-DR, inhibited cytolysis, suggesting that these molecules participate in the CTL-target interaction (Sanchez-Madrid et al., 1982; Krensky et al., 198313). Hildreth e t al. (1983) immunized mice with EBV-transformed human lymphoblastoid B cells and generated two monoclonal antibodies, designated MHM23 and MHM24, which apparently recognize human LFA-1, as they immunoprecipitate 180- and 95-kilodalton polypeptide chains from both B and T cells. Both antibodies inhibited lysis of influenza virus infected and EBV-transformed target cells by CTLs. Meuer et al. (1983a) immunized BALB/c mice with CTL clones and

72


selected monoclonal antibodies (Ti) clonally specific for binding and the inhibition of cytolysis. These monoclonal antibodies may recognize unique determinants of the T cell antigen receptor. The T cell surface structures recognized by these monoclonal antibodies are involved in cellular recognition and/or the cytolytic event and may be classified according to paradigms based on cellular distribution and functional studies. The first class of antigens, typified by murine LFA-1 (Springer et al., 1982), are cell surface determinants that are widely distributed on both effector and target cells. Antibodies that recognize these structures appear to block lymphocyte function via interruption of cell interactions. Anti-LFA-1 monoclonal antibodies have been shown to block cell adhesion at the effector cell level. These determinants are broadly distributed and, therefore, are unlikely to be part of an antigen specific receptor. Human LFA-1 is another example of this type of antigen. It is widely distributed on lymphoid tissues and is composed of two polypeptides of 177 and 95 kilodaltons on all cell types studied (SanchezMadrid et al., 1982; Krensky et aZ., 198313). Anti LFA-1 monoclonal antibodies block antigen specific CTL-mediated cytotoxicity, NK-mediated cytotoxicity, and T cell proliferation. CTL-mediated cytotoxicity is inhibited by anti-LFA-1 monoclonal antibody binding to the effector cell rather than the target cell. By analogy with murine LFA-1, human LFA-1 appears involved in a general process underlying cell interaction, such as adhesion. Thus, LFA-1 appears to fall into a class of molecules that are required for cell-cell interactions, and not just those specialized interactions restricted to T lymphocytes. Three unique and three complex epitopes of the human LFA-1 molecule have been defined using monoclonal antibodies in a competitive cross inhibition binding assay (Ware et al., 1983b). The identification of several unique epitopes on human LFA-1 provides an explanation for the variability in inhibition of CTL-mediated cytolysis by the various anti-LFA-1 monoclonal antibodies (Sanchez-Madrid et al., 1982; Ware et al., 1983b). Lyt-2 typifies a second class of antigens, those associated with T cell functions (Cantor and Boyse, 1975). Human antigens that can be included in this class are OKT3, OKT4, and OKT8. These antigens are expressed on T cells and may be involved in specific antigen recognition. Studies of murine lymphocyte-mediated cytolysis have shown that Lyt-2 is involved in (Nakayama et aZ., 1979; Shinohara and Sachs, 1979),but not absolutely necessary for, the effector-target interaction (Kaufman et al., 1982; MacDonald et al., 1982). Kaufman et al. (1982)


73

have isolated Lyt-2- Class I restricted CTL hybridomas that are still capable of specific cell-mediated cytolysis. Peripheral blood T lymphocytes are subdivided into two populations defined by expression of the OKT4 (62 kilodalton) and OKT8 (76 kilodalton) cell surface antigens. Although early experiments suggested that OKT4+ cells were the helpedinducer subset while OKT8+ cells were the suppressor/cytotoxic cells (Reinherz and Schlossman, 1980), recent studies have shown that OKT4+ CTLs (Moretta et al., 1981) and suppressor cells (Thomas et al., 1982) exist. Anti-OKT4 and OKT8 monoclonal antibodies may recognize monomorphic sites on MHC molecules (Krensky et al., 198213; Meuer et al., 1982a; Ball and Stastny, 1982; Spits et al., 1982; Biddison et al., 1982a; Engleman et al., 1983). OKT8 (Leu2a) is the human homologue of Lyt-2 and appears to be involved in the recognition of Class I MHC determinants, whereas OKT4 (Leu3a) recognizes Class I1 MHC determinants. Engleman et al. (1981, 1983) have presented similar data regarding helper and suppressor T cells. Although the hypothesis that OKT8 may be a receptor for Class I MHC antigens while OKT4 is a receptor for Class I1 MHC antigens is attractive, exceptions to this rule have been described. Ball and Stastny (1982), Flomenberg et al. (1983), and Spits et al. (1983) have described apparent OKT8+ HLA-DR-specific CTL. It should be noted, however, that in none of these cases has OKT4 or OKT8 monoclonal antibody been shown to block cytolysis. We (Krensky et al., 1983a) and Spits et al. (1983) have recently described DC-specific CTLs. Of particular note, we found OKT8+, DC-specific CTL while Spits et al. derived OKT4+, DC-specific CTL. In neither case did either OKT4 or OKT8 inhibit cytolysis, suggesting that another T cell surface structure may be involved in the recognition of DC antigens or that OKT4/OKT8 are accessory structures that enhance the affinity of cell interactions but are not absolutely required. The sheep red blood cell receptor molecule (49 kilodalton) recognized by anti-LFA-2 (Sanchez-Madrid et al., 1982; Krensky et al., 1983b),OKTll (Palacios and Martinez-Maza, 1982), Leu 5 (Howard et al., 1981),and mAb 9.6 (Kamoun et al., 1981) appears to be another example of a T cell function specific antigen. Palacios and MartinezMaza (1982) showed that anti-OKT11 inhibited proliferation. AntiLFA-2 mAb (Krensky et al., 1983b) and mAb 9.6 (Martin et al., 1983) inhibit all T cell functions studied. We have shown that LFA-2 monoclonal antibody precipitates molecules of different molecular weights from T cells at different stages of activation and/or differentiation,

74


showing that some heterogeneity exists. Because LFA-2 is a glycoprotein, molecular weight variation may be due to differences in either glycosylation or amino acid sequence. Experiments to assess LFA-2 clonal variation with cloned human CTL lines are in progress. We have found a hierarchy of CTL inhibition by the anti-SRBC antibodies: anti-LFA-2 = Leu 5 > 9.6 > OKT11. Whether this reflects differences in antibody affinity or epitope recognition remains unknown. Martin et al. (1983) recently showed that one anti-SRBC receptor monoclonal antibody, 9.6, inhibited cytolysis well while monoclonal antibody, 35.1 which recognizes a different epitope on the same molecule did not. Despite the fact that LFA-2 (SRBC receptor) has a wider cell distribution than OKT3 (Krensky et al., 1983b), LFA-2 monoclonal antibodies block only T cell functions. This paradox may be important in understanding the relationship between CTL and NK cells. Ortaldo et al. (1981) have suggested that NK cells may represent an early stage of T cell differentiation. Both OKT3+,Lyt-3+ (Pawelec et al., 198213) and OKT3-, Lyt-3+ (Zarling et al., 1981) NK cells have been described. Large granular lymphocytes are low or negative for OKTll and Lyt-3 (9.6), but cultured NK cells are definitely positive (Ortaldo et al., 1981; Pawelec et al., 1982b; Zarling et al., 1981). It is possible that the OKT3-, Lyt-3+ NK cells that we have described (Krensky et al., 1982c) have not yet differentiated to the stage at which LFA-2 is a functional component of the lytic process. Fast et al. (1981) and Martin et a2. (1983) have reported blocking of NK activity using the mAb 9.6. These differences may result from differences in monoclonal antibody affinity, recognition of different epitopes by these antibodies, or due to differences in NK cell subpopulations studied and/or their stage of distribution. In addition to these public determinants there must exist unique (clonotypic) surface structures involved in specific antigen recognition. Meuer et al. (1983a) recently derived monoclonal antibodies against a human CTL clone and selected anticlonotypic antibodies. Such antibodies recognized a 49-, 43-kilodalton heterodimer and block specific cytotoxic function and antigen-induced proliferation by the immunizing clone. They have also shown that this clonotypic molecule is closely associated (comodulates) with OKT3 (Meuer, 1983a), another pan T cell antigen, and that anticlonotypic monoclonal antibodies selectively induce clonal proliferation and lymphokine production (Meuer, 1983b). We have generated a monoclonal antibody to a determinant termed LFA-3. This monoclonal antibody inhibits cytolysis by both OKT4+


75

and OKT8+ CTL, but not by NK cells, and inhibits proliferation to PHA and the mixed lymphocyte response (Sanchez-Madrid et al., 1982; Krensky et al., 1983b). This monoclonal antibody specifically blocks function by binding to target cells rather than to CTLs, implying that LFA-3 may represent a target ligand for an effector-specific receptor. LFA-3 is broadly distributed on both lymphoid and nonlymphoid tissues (including monocytes, granulocytes, platelets, vascular endothelium, smooth muscle, and fibroblasts), but LFA-3 is biochemically distinct from HLA-A,B, DR and other known MHC antigens. Anti-LFA-3 monoclonal antibodies block cytolysis of lymphoblastoid, endothelial, and fibroblast targets (Collins et al., 1984) indicating that LFA-3 is a new target molecule of general importance in the CTLtarget interaction. VII. Conclusions

New technological advances are providing us with the ability to begin to define the CTL response on a molecular level. There are several conclusions that can be drawn from the application of each of these new approaches and it is worth summarizing them here. First, the use of liposomes to study CTL specificity suggests that allogeneic MHC antigens alone may be able to stimulate or be recognized by CTLs. It has been proposed by Matzinger and Bevan (1977) that all CTLs recognize MHC antigens in association with a modifier X, including allospecific CTLs. From our data, it seems unlikely that allospecific CTLs require such a modifier X for purified Class I or Class I1 antigens in liposomes can trigger CTLs. Second, the use of liposomes to study virus-specific CTLs suggests that viral and MHC antigens, if not physically interacting, must be closely associated to trigger CTLs. These liposomes were between 0.03 and 0.1 p m in diameter and still were only able to trigger virus-specific CTLs when both the viral and MHC antigens were in the same lipid bilayer. This requirement for a close association between viral and MHC antigens for triggering should impose similar requirements upon the T cell receptor(s). Third, the use of liposomes has pointed out that HLA-DR antigens can be highly immunogenic for CTLs. HLA-A,B antigens may b e the dominant immunogen, however, by isolating the HLA-DR antigens from these Class I antigens, one could more clearly assess their immunogenicity. Liposomes have also provided us with insights concerning the helper T cell pathway that regulates the CTL response. First, it became quite evident that Class I antigens can be handled by antigen

76


presenting cells in a manner similar to the way most foreign non-MHC antigens are handled. Specifically, these Class I antigens are presented to T helper cells in the context of Class I1 antigens. Liposomes have also been useful in defining the requisite antigen processing step. It appears that when a Class I antigen is the foreign antigen, “ antigen processing” may only be the requirement of intercalation of the antigen into the lipid bilayer. This is based upon our evidence that purified Kk and partially purified Iad antigens in the same liposome trigger primed CD2F1 (H-2d)helper T cells. This simple requirement for insertion in the lipid bilayer, without further antigen degradation, may not be generalizable to all antigens; especially, since Class I molecules have a hydrophobic region and most other foreign antigens do not. DNA-mediated gene transfer of cloned MHC genes and recombinant genes is providing new insights into our understanding of CTL recognition. The use of “exon shuffling” to construct recombinant MHC genes suggests that the N and/or C1 domains of the Class I molecule are essential for target cell recognition by allospecific and virus-specific CTLs. Currently recombinant genes are being constructed that shuffle exons between N and C1 to determine the role played by each of these domains. Constructions that delete most of the intracytoplasmic tail of the Class I molecule suggest that this region may play a role in CTL recognition of some but not all viral antigens. Finally, hybridoma technology has advanced our understanding of the molecules involved in the CTL-target interaction. Screening monoclonal antibodies for their ability to inhibit cytolysis has allowed us to define a role for the LFA-1, LFA-2, and LFA-3 antigens. These studies point out that a series of accessory (i-e., not antigen specific) interactions may help to stabilize the CTL-target interaction. LFA-1 and LFA-2 appear to be effector cell antigens that interact with target cell determinants that are less polymorphic than that recognized by the antigen-specific receptor. LFA-3 is a new target cell antigen that is not encoded by the MHC. The molecule has a wide tissue distribution and appears to play a role in the CTL-target interaction even when endothelial cells and fibroblasts are the targets of CTL lysis. Though these new technological approaches have allowed us to begin to develop a molecular understanding of the CTL response, there is much that remains to be learned. There are some obvious questions that need to be addressed in the near future. We have been quite successful in using liposomes to stimulate primed spleen cells to differentiate to functional CTLs, but we have been unsuccessful in using liposomes to stimulate a primary CTL response. This failure


77

suggests that there may be other important parameters that, to date, we have not addressed. The eventual elucidation of these parameters may provide important insights into T cell triggering. Even though we can now insert purified viral and MHC antigens into liposomes we have yet to answer the question, must these antigens physically interact to trigger T cells? With some of the new molecular constructions and the use of fluorescence energy transfer and fluorescence photobleaching, hopefully we can begin to address this question. Though monoclonal antibodies have allowed us to define many new molecules involved in the CTL-target interaction, the role played by these molecules in the cytolytic process remains to be defined. Finally, it has become apparent to us that there is a specific strategy that is required for the further exploration of the CTL response and of cell-cell interactions in general. Raising monoclonal antibodies that effect function allows one to define the molecules involved in these cell interactions. These monoclonal antibodies can most readily be isolated if one has well-defined model systems, preferably, model systems that utilize cloned T cells. Once important cell interaction molecules have been defined, then monoclonal antibodies can be used to begin molecular studies of these molecules. Specifically, the molecules may be isolated and purified for functional studies and their amino acid composition can provide important information to construct oligonucleotide probes for the eventual isolation of the genes that encode these structures. The availability of the cloned genes may allow one to make molecular constructions (i.e., recombinant genes) which result in protein modifications that might be impossible through direct modification of the protein. Once recombinant genes have been made, DNA-mediated gene transfer of the cloned genes can be employed to determine the functional effects of the gene products of these recombinant genes. If the gene products of the recombinant genes appear to be of interest, one can then isolate and purify these new proteins in order to study them both at a functional and biochemical level. Thus, by employing well-defined in vitro model systems, monoclonal antibodies, protein chemistry, and molecular biology, we hope to gain a better understanding of the CTL response. These new insights may greatly influence our understanding of cell-cell interactions and cellular differentiation,

ACKNOWLEDGMENTS This work was supported by U.S.P.H.S. Grants A1 17258, CA 34129 (S.J.B.),and A1 18083 (C.S.R.).S.J.B. is a recipient of an American Cancer Society Faculty Research

78


Award. A.M.K. is the recipient of an American Heart Association Clinican Scientist Award and O.W. is supported by a Leukemia Society of America Postdoctoral Fellowship. We wish to thank Pat Thomason and Nancy Knittle for their assistance in the preparation of this manuscript.

REFERENCES

’

Alaba, O., and Law, L. W. (1978).J . Exp. Med. 148, 1435. Alaba, O., and Law, L. W. (1980).J . Zmmunol. 125,414. Albrechtsen, D., Arnesen, E., and Thorsby, E. (1979). Transplantation 27,338. Altman, A., and Cohen, I. R. (1975).Eur. J . Zmmunol. 5,437. Armerding, D., and Katz, D. H. (1974).J . E x p . Med. 149, 19. Ashman, R. B., and Mullbacher, A. (1979).J . E x p . Med. 150, 1277. Bach, F. H., Bach, M. L., and Sondel, P. M. (1976). Nature (London) 259, 273. Bach, F. H., Inouge, H., Hank, J. A,, and Alter, B. J. (1979). Nature (London)281,307. Bach, F. H., Alter, B. J., Widmer, M. B., Segall, M., and Dunlap, B. (1981).Zmmunol. Rev. 54, 5. Ball, E. J., and Stastny, P. (1982).Zmmunogenetics 16, 159. Barbosa, J. A., Kamock, M. E., Biro, P. A., Weissman, S. M., and Ruddle, F. H. (1982). Proc. Natl. Acad. Sci. U.S.A.79, 6327. Barnstable, C. J., Bodmer, W. F., Brown, G., Galfre, G., Milstein, C., Williams, A. F., and Zeigler, A. (1978). Cell 14, 9. Baum, L., and Pilarski, L. (1978).J . Exp. Med. 148, 1579. Benacerraf, B. (1978).J. Zmmunol. 120, 1809. Ben-Sasson, S. Z., Lipscomb, M. F., Tucker, T. F., and Uhr, J. W. (1977).J . Zmmunol. 119, 1493. Berg, D., Jorgensen, R., and Davies, J. (1978). [ n “Microbiology” (D. Schlesinger, ed.), ASM Press, p. 13. Washington, D.C. Bernabeu, C., Finlay, D., von de Rijn, M., Maziarz, R. T., Biro, P. A., Spits, H., devries, J., and Terhorst, C. P. (1983).J. Zmmunol. 131,2032. Biddison, W. E., Shaw, S., and Nelson, D. L. (1979).J . Zmmunol. 122, 660. Biddison, W. E., Payne, S. M., Shearer, G. M., and Shaw, S. (1980a).J.Exp. Med. 152, 204s. Biddison, W. E., Krangel, M. S., Strominger, J. L., Ward, F. E., Shearer, G. M., and Shaw, S. (1980b).Hum. Zmmunol. 3,225. Biddison, W. E., Rao, P. E., Talle, M. A., Goldstein, G., and Shaw, J. (1982a).J . E x p . Med. 156, 1065. Biddison, W. E., Kostyn, D. D., Strominger, J. L., and Krangel, M. S. (1982b).J . Zmmunol. 129, 730. Billings, P., Burakoff, S. J., Dorf, M. E., and Benacerraf, B. (1977).J . E x p . Med. 145, 1387. Brunner, K. T., Mauel, J., Cerottini, J. C., and Chapias, B. (1968). Zmmunology 14, 181. Burakoff, S. J. (1981). “The Major Histocompatibility Complex in Immunobiology” (M. E. Dorf, ed.), p. 343. Garland Press, New York. Burakoff, S. J. (1984). Fed. Proc. Fed. Am. Soc. E x p . Biol. 43, 266. Burakoff, S. J., Germain, R. N., Dorf, M. E., and Benacerraf, B. (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 625. Burakoff, S. J., Dorf, M. E., and Benacerraf, B. (1977).J . Zmmunol. 118, 1795. Burakoff, S. J., Engelhard, V. H., Kaufman, J., and Strominger, J, L. (1980). Nature (London)283,495.


79

Burakoff, S. J., and Mescher, M. F. (1982). Cell Su$. Rev. 8, 173. Cantor, H., and Asofsky, R. (1970a).J . Exp. Med. 131, 235. Cantor, H., and Asofsky, R. (1970b).J. Exp. Med. 135, 764. Cantor, H., and Boyse, E . A. (1975).J . Exp. Med. 141, 1390. Carnaud, C., Fadai-Ghotbi, M., Lesavre, P., and Bach, J. F. (1977). Eur. J. Zmmunol. 7, 81. Charmot, D., and Mawas, C. (1979). Eur. J. Immunol. 9, 723. Charron, D., Engleman, E.,Benike, C., and McDevitt, H. 0. (1980).J . E x p . Med. 152, 1275. Chestnut, R. W., Colon, S. M., and Grey, H. M. (1982).J . Zmmunol. 129, 2382. Ciavarra, R. P., Kang, C. Y., and Forman, J. (1980).J . Zmmunol. 125, 336. Claman, H. N., Chaperon, E. A., and Triplett, R. F. (1966).J. Immunol. 97, 828. Collins, T., Krensky, A., Clayberger, C., Gimbrone, M., Fiers, W., Burakoff, S. J., and Pober, J. (1984).J. Zmmunol., in press. Cooley, M. A., and Schmitt-Verhulst, A. M. (1979).J.Immunol. 123,2328. Corley, R. B., Dawson, J. R., and Amos, D. B. (1975). Cell. Zmmunol. 16, 92. Davis, A. K., Bos, T. J., and Nayak, D. P. (1983).Proc. Natl. Acad. Sci. U.S.A. 80, 3976. de Landazuri, M. O., and Herbermann, R. B. (1972).Nature (London)New Biol. 238,18. Dialynas, D. P., Quen, Z. S., Wall, K. A., Pierres, A., Quintans, J., Loken, M. R., Pierres, M., and Fitch, F. W. (1983). Immunol. Rev. 74,29. Doherty, P., and Bennink, J. R. (1979).J. E x p . Med. 149, 150. Duprez, V., Mescher, M. F., and Burakoff, S. J. (1983).J. Zmmunol. 130, 493. Ellner, J. J., and Rosenthal, A. S. (1975).J. Zmmunol. 114, 1563. Engelhard, V. H., Strominger, J. L., Mescher, M., and Burakoff, S. J. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 5688. Engelhard, V. H., Kaufman, J. F., Strominger, J. L., and Burakoff, S. J. (1980).J . Exp. Med. 152, 54s. Engers, H. D., Thomas, K., Cerottini, J . C., and Brinier, K. T. (1975).J . Zmmunol. 115, 356. Engleman, E. G., Benike, C. J., Gruniet, F. C., and Evans, R. L. (lUSl).J.Zmmunol. 127, 2124. Engleman, E. G . , Benike, C. J., Metzler, C., Gatenby, P. A., and Evans, R. L. (1983). J. Zmmunol. 130,2623. Evans, G. A., Margulies, D. H., Shykind, B., Seidman, J. G., and Ozato, K. (1982). Nature (London) 300, 755. Evans, R. L., Faldetta, T. J., Humphreys, R. E., Pratt, D. M., Yunis, E. J., and Schlossman, S. F. (1978).J. Exp. Med. 148, 1440. Fast, L. D., and Fan, D. P. (1978).J. Zmmunol. 120, 1092. Fast, L. D., and Fan, D. P. (1979).J.Zmmunol. 123, 1491. Fast, L. D., Hansen, J. A,, and Newman, W. (1981).J. Zmmunol. 127,448. Feighery, C., and Stastny, P. (1979).J. E x p . Med. 149, 485. Finberg, R., Mescher, M., and Burakoff, S. J. (1978).J . E x p . Med. 148, 1620. Finberg, R., Greene, M., Benacerraf, B., and Burakoff, S. J. (1979).J. Immunol. 123, 1205. Finke, J. M., Orosy, C. G., and Battisto, J. R. (1977). Nature (London) 267, 353. Finke, J. M., Scott, J., Gillis, S., and Hilfiker, M. L. (1983).j.Immunol. 30, 763. Fishman, M., and Adler, F. L. (1963).J.E x p . Med. 117, 595. Flomenberg, N., Duffy, E., Naito, K., and Dupont, B. (1983). lmmunogenetics 17, 317. Florkiewicz, R. Z., Smith, A., Bergmann, J. E., and Rose, J. K. (1984).J. Cell Biol., in press.

80


Flyer, D. C., Burakoff, S. J., and Faller, D. V. (1983). Nature (London) 305,726. Forman, J,, Goodenow, R. S., Hood, L., and Ciavarra, R. (1983). J . E x p . Med. 157, 1261. Frelinger, J. A., Om, A., Braytm, P. B., and Hood, L. (1984).Trunsplant. Proc., in press. Fujiwara, H., and Shearer, G. M. (1980).J . Immunol. 124, 1271. Gaston, J. S. H., Rickinson, A. B., and Epstein, M. A. (1983).J . E x p . Med. 158, 280. Garaff et al. (1984).J . Cell Biol., in press. Garman, R. D., and Fan, D. P. (1983).J . Zmmunol. 30, 756. Gething, M. J., and Sambrook, J. (1981). Nature (London) 293, 620. Gillis, S., and Smith, K. A. (1977). Nature (London) 268, 154. Glasebrook, A. L., and Fitch, F. W. (1980).J. Exp. Med. 151, 876. Glimcher, L. H., Kim, K. J., Greene, I., and Paul, W. (1982).J . E x p . Med. 155,445. Goodenow, R. S., Stroynowski, I., McMillan, M., Michalson, M., Eakle, K., Sher, B. T., Davidson, N., and Hood, L. (1983). Nature (London) 301,388. Goodfellow, P. N., Jones, E. A., van Heyningen, V., Solomon, E., Bobrow, M., Miggiano, V., and Bodmer, W. F. (1975). Nature (London) 254, 267. Gooding, L. R., and O’Connell, K. A. (1983).J . Immunol. 131, 2580. Gorga, J., Foran, J., Burakoff, S. J., and Strominger, J. L. (1984). In “Methods in Enzymology.’’ Academic Press, New York, in press. Goulmy, E., Blokland, E., van Rood J., Channot, D., Malissen, B., and Mawas, C. (1980). J . E x p . Med. 152, 1825. Goulmy, E., van Leeuwen, A., Blokland, E., van Rood, J. J., and Biddison, W. E. (1982). J . E x p . Med. 155, 1567. Graham, F. L., and van der Eb, A. J. (1973). Virology 54, 536. Grey, H. M., Colon, S. M., and Chestnut, R. W. (1982).J . Immunol. 129,2389. Grossi, C. E., Zicca, A,, Cadoni, A., Mingari, M. C., Moretta, A., and Moretta, L. (1983). Eur. J . Immunol. 13,670. Hale, A. H., Ruebush, M. J., and Harris, D. T. (1980a).J. Immunol. 125, 428. Hale, A. H., Ruebush, M. J., and Harris, D. T. (1980b). In “Liposomes and Immunobiology” (B. H. Tom, and H. R. Six, eds.), p. 211. Elsevier, Amsterdam. Hale, A. H., Lyles, D. S., and Fan, D. P. (1980~). J . Zmmunol. 124, 724. Hamaoka, T., Fujiwara, H., Teshima, K., Aoki, H., Yamamato, H., and Kizagawa, M. (1979).J . E x p . Med. 149, 185. Hartman, J. R., Nayak, D. P., and Fareed, G. C. (1982). Proc. Natl. Acad. Sci. U.S.A. 79,233. Hayry, P., and Anderson, L. C. (1976). Scand. J . Immunol. 5,391. Herman, A,, Parham, P., Weissman, S. M., and Engelhard, V. H. (1983). Proc. Natl. Acad. Sci. U.S.A.80,5056. Herrmann, S. H., and Mescher, M. F. (1979).]. Biol. Chem. 254,8713. Herrmann, S. H., Weinberger, O., Burakoff, S. J., and Mescher, M. F. (1982). J . Immunol. 128, 1968. Hildreth, J. E. K., Gotch, F. M., Hildreth, P. D. K., and MeMichael, A. J. (1983). Eur. J . Immunol. 13,202. Hollander, N., Pillemer, E., and Weissman, I. L. (1980).J . Exp. Med. 152, 674. Hood, L., Steinmetz, M., Goodenow, R. (1982). Cell 28, 685. Horai, S., van der Poel, J. J., and Goulmy, E. (1982). Immunogenetics 16, 503. Howard, F. D., Ledbetter, J. A., Wong, J., Bieber, C. P., Stinson, E. D., and Herzenberg, L. A. (1981).J.Immunol. 126, 2117. Hurwitz, J., Pan, S., Wettstein, P. J., and Doherty, P. C. (1983). lmmunogenetics 17,79. Johnson, H. E. (1980). Tissue Antigens 15, 159.


81

Kamoun, M., Martin, P. J., Hansen, J. A., Brown, M. A., Siadak, A. W., and Nowinski, R. C. (1981).J. E x p . Med. 153,207. Kapp, J. A., Pierce, C. W., and Benacerraf, B. (1973).J. E x p . Med. 138, 1121. Kappler, J. W., and Marrack, P. (1976). Nature (London) 262, 797. Kaufman, Y., Golstein, P., Pierres, M., Springer, T. A., and Eshhar, Z. (1982). Nature (London) 300,357. Keene, J., and Forman, J. (1982).J. E x p . Med. 155, 768. Klinkert, W. E. F., Labadie, J. H., O’Brien, J. P., Beyer, C. F., and Bowers, W. E. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 5414. Kornbluth, J., and Dupont, B. (1980).J. E x p . Med. 152, 1645. Kornbluth, J., Silver, D. M., and Dupont, B. (1981). Immunol. Rev. 54, 112. Korngold, R., and Sprent, J. (1980).J. Exp. Med. 151,314. Krangel, M. S., Taketani, F., Biddison, W. E., and Strominger, J. L. (1983).Biochemistry 21, 6313. Krensky, A. M., Reiss, C. S., Mier, J. W., Strominger, J. L., and Burakoff, S. J. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79,2365. Krensky, A. M., Clayberger, C., Reiss, C. S., Saominger, J. L., and Burakoff, S. J. (1982b).J. Immunol. 129, 2001. Krensky, A. M., Auk, K. A., Reiss, C. S., Strominger, J. L., and Burakoff, S. J. (1982~). J . Immunol. 129, 1748. Krensky, A. M., Clayberger, C., Greenstein, J., Crimmins, M., and Burakoff, S. J. (1983a).J . lmmunol. 131,2777. Krensky, A. M., Sanchez-Madrid, F., Robbins, F., Nagy, J. A., Springer, T. A., and Burakoff, S. J. (198313).J. Immunol. 131,611. Kress, H. G., and Kreth, H. W. (1982).J. Immunol. 129,844. Kreth, H. W., ter Meulen, V., and Eckert, G. (1979).Med. Microbiol. Immunol. 165,203. Kung, P. C., Goldstein, G., Reinherz, E. L., and Schlossman, S. F. (1979). Science 206, 347. Lalande, M. E., McCutcheon, M. J., and Miller, R. G. (1980).J. E x p . Med. 151, 12. Lawman, M. J. P., Naylor, P. T., Huang, L., Courtney, R. J., and Rouse, B. T. (1981). J. Immunol. 126,304. 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. Lee, K. C., Wong, M., and Spitzer, D. (1982). Transplantation 34, 150. Lemmonier, F., Mescher, M., Sherman, L., and Burakoff, S. (1978).J. Immunol. 120, 1114. Lemmonier, F. A., Malissen, M., and Goldstein, P. (1982). lmmunogenetics 16, 355. Lemmonier, F. A., LeBorteiller, P. P., Malissen, B., Yalstein, P., Malissen, M., Mishal, Z., Caillol, D. H., Jordan, B. R., and Kourilsky, F. M. (1983).J.Immunol. 130, 1432. Levy, R. B., Richardson, J. C., Margulies, D. H., Evans, G. A., Seidman, J. G., and Ozato, K. (1983).J. Immunol. 130,2514. Liew, F. Y., and Parish, C. R. (1974).J. E x p . Med. 139, 779. Lindahl, K. F., and Bach, F. H. (1975). Nature (London) 254,607. Lipsky, P. E., and Kettman, J. R. (1982). Immunol. Today 3,36. Littlefield, J. W. (1964). Science 145, 709. Loh, D., Ross, A. H., Hale, A. H., Baltimore, D., and Eisen, H. N. (1979).J. E x p . Med. 150, 1067. Lotze, M. T., Strausser, J. L., and Rosenberg, S. A. (1980).J. Immunol. 124, 2972. Luger, T. A,, Smolen, J. S., Chused, T. M., Steinberg, A. D., and Oppenheim, J. J. (1982).J. Clin. Inuest. 70, 470.

a2


MacDonald, H. R., Glasebrook, A. L., Bron, C., Kelso, A., and Cerottini, J. C. (1982). Immunol. Rev. 68, 89. Malek, T. R., and Shevach, E. M. (1982). Eur. J. lmmunol. 12, 825. Malissen, B., Kristensen, T., Goridis, C., Madsen, M., and Mawas, C. (1981). Scand.1. Immunol. 14, 213. Malissen, M., Malissen, B., and Jordan, B. R. (1982a). Proc. Nutl. Acad. Sci. U.S.A. 79, 893. Malissen, B., Rebai, N., Liabeut, A., and Mawas, C. (1982b). Eur. J . lmmunol. 12, 739. Maniatus, T., Fretsch, E. F., and Sambrook, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Margulies, D. H., Evans, G. A., Ozato, K., Camerini-Otero, R. D., Tanaka, K., Appella, and Seidman, J. G. (1983).J. lmmunol. 130,463. Martin, P. J., Longton, G., Ledbetter, J. A., Newman, W., Braun, M. P., Beatty, P. G., and Hansen, J. A. (1983).J . lmmunol. 131, 180. Matsuyama, M., Sugamura, K., Kawade, Y., and Hinuma, Y. (1982). J. Immunol. 129,450. Matzinger, P., and Bevan, M. J. (1977). Cell lmmunol. 29, 1. McLaughlin-Taylor, E., Woodward,J. G., Macchi, M. J., McMillan, M., and Frelinger, J. A. (1983). Immunogenetics 19,2056. McMichael, A. J., Ting, A,, Zweerink, H. J., and Askonas, B. A. (1977).Nature (London) 270, 524. McMichael, A. J., Parham, P., Brodsky, F. M., and Pilch, J. R. (1980). J. Exp. Med. 152, 195. Melief, C. J., deWaal, L. P., Sterkart, M. J., Kast, W. M., and Melvold, R. W. (1983). In “Ir Genes: Past, Present and Future” (C. W. Pierce et al., eds.), p. 377. Humana Press, Clifton, New Jersey. Mellor, A. L., Golden, L., Weiss, E., Bullman, H., Hurst, J., Simpson, E., James, R. F. L., Townsend, A. R. M., Taylor, P. M., Schmidt, W., Fergula, J., Leben, L., Sanat Maria, M., Atfield, G., Fistenstein, H., and Flavell, R. A. (1982).Nature (London) 298, 529. Mescher, M., Sherman, L., Lemonnier, F., and Burakoff, S. J. (1978).J . Exp. Med. 147,946. Mescher, M. F., Finberg, R., Sherman, L., and Burakoff, S. J. (1979). In “T and B Lymphocytes” (F. H. Bach, B. Bonavida, and E. Vitetta, eds.), p. 623. Academic Press, New York. Mescher, M. F., Jose, M. J. L., and Balk, S. P. (1981). Nature (London) 289, 139. Metzgar, R. S., Bertoglio, J., Anderson, J. K., Bonnard, G. D., and Ruscetti, F. W. (1979). J . Immunol. 122, 949. Meuer, S. C., Schlossman, S. F., and Reinherz, E. L. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79,4395. Meuer, S. C., Hussey, R. E., Penta, A. C., Fitzgerald, K. A., Stadler, B. M., Schlossman, S. F., and Reinherz, E. L. (1982b).J . Immunol. 129, 1076. Meuer, S. C., Fitzgerald, K. A., Hussey, R. E., Hodgdon, J. C., Schlossman, S. F., and Reinherz, E. L. (1983a).J . Exp. Med. 157, 705. Meuer, S. C., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1983b).J . E x p . Med. 158, 988. Misko, I. S., Moss, D. J., and Pope, J. H. (1980).Proc. Natl. Acad. Sci. U.S.A.77, 4247. Moretta, A., Pantaleo, G., Moretta, L., Cerottini, J. C., and Mingari, M. L. (1983).J . E x p . Med. 157, 743.


83

Moretta, L., Mingari, M. C., Sekaly, P. R., Moretta, A., Chapuis, B., and Cerottini, J. C. (1981).J.E r p . Med. 154, 569. Morgan, D. A., Ruscetti, F. W., and Gallo, R. C. (1976). Science 193, 1007. Moss, D. J., Wallace, L. E., Rickinson, A. B., and Epstein, M.A. (1981).E u r . 1 . Immunol. 11, 686. Murre, C . , Reiss, C. S., Chen, L. B., Bernabeu, C., Burakoff, S. J., and Seidman, J. (1983).Nature (London)307,432. Nabholz, M., Vives, J., Young, H. M., Meo, T., Miggiano, V., Rijnbeek, A., and Shreffler, D. C . (1974). E u r . J. Immunol. 4,378. Nakayama, E., Shiku, H., Stockert, E., Oettgen, H. F., and Old, L. J. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 1977. Neefe, J. R., Sullivan, J. E., and Hartzman, R. J. (1982).J. Immunol. 128,227. Okada, M., and Henney, C. S. (1980).J . Immunol. 125, 850. Okada, M., Klimpel, C . R., Kuppers, R. C., and Henney, C. S. (1979).J.Zmmunol. 122, 2527. Om, A,, Goodenow, R. S., Hood, L., Brayton, P. R., Woodward, J. G., Harmon, R. C., and Frelinger, J. A. (1982). Nature (London) 297,415. Ortaldo, J. R., Sharrow, S. O., Timonen, T., and Herberman, R. B. (1981).J. Immunol. 127,2401. Ozato, K., Evans, L. A,, Evans, G. A., Shykind, B., Margulies, D. H., and Seidman, J. G. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 204. Palacios, R., and Martinez-Maza, 0. (1982).J . Zmmunol. 129, 2479. Parham, P. (1979).J . Biol. Chem. 254, 8709. Pawelec, G., Blaurock, M., Schneider, E. M., Shaw, S., and Wernet, P. (1982a). Eur. J . Immunol. 12, 967. Pawelec, G. P., Hadam, M. R., Schneider, E . M., and Wernet, P. (1982b).J. Zmmunol. 128,2271. Pease, L. R., Schulze, D. H., Pfaffenbach, G. M., and Nathenson, S. G. (1983). Proc. Natl. Acad. Sci. U.S.A.80,242. Pettinelli, C. B., Schmitt-Verhulst, A. M., and Shearer, G. M. (1979). J . Immund. 122, 847. Pfeffer, P. F., and Thorsby. (1982). Transplantation 33,52. Pierce, C. W., and Kapp, J. A. (1976). In “Immunobiology of the Macrophage” (D. S. Nelson, ed.), p. 1. Academic Press, New York. Pierce, C. W., Kapp, J. A., and Benacerraf, B. (1976). In “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (D. H. Katz and B. Benacerraf, eds.), p. 391. Academic Press, New York. Plate, J. M. D. (1976).Nature (London)260, 329. Quinnan, G. V., Kirmani, N., Esber, E., Saral, R., Manischewitz, J. F., Rogers, J. L., Rook, A. H., Santos, G. W., and Burns, W. H. (1981).J.Immunol. 126,2036. Raulet, D. H., and Bevan, M. (1982).Nature (London) 296, 754. Reinherz, E. L., and Schlossman, S. F. (1980). Cell 19, 821. Reiss, C. S., and Burakoff, S. J. (1981).J. E x p . Med. 154, 541. Reiss, C. S., Hemler, M. E., Engelhard, V. H., Mier, J. W., Strominger, J. L., and Burakoff, S. J. (1980). Proc. Natl. Acad. Sci. U.S.A.77, 5432. Reiss, C. S., Evans, G. A,, Margulies, D. H., Seidman, J. G., and Burakoff, S. J. (1983a). Proc. Natl. Acad. Sci. U.S.A. 80,2709. Reiss, C. S., Evans, G. A., Murre, C., Margulies, D. H., Seidman, J. G., and Burakoff, S. J. (1983b). Fed. Proc. Fed. Am. SOC. E r p . Biol., in press. Rickinson, A. B., Wallace, L. E., and Epstein, M. A. (1980). Nature (London) 283,865.

84


Rock, K., Barnes, M. L., Gennain, R. N., and Benacerraf, B. ( 1 9 8 3 ) ~Immunol. . 130,457. Rose, J. K., and Gallione, C. J. (1981).J . Virol. 39, 519. Rosenthal, A. S., and Shevach, E. M. (1973).]. Exp. Med. 138, 1194. Rosenthal, A. S., Blake, J. T., Ellner, J. J., Greineder, D. K., and Lipsky, P. E. (1976). In “Immunobiology of the Macrophage” (D. S. Nelson, ed.), p. 131. Academic Press, New York. Rouse, B. T., and Lawman, M. J. (1980).J. Immunol. 124,2341. Rulon, K., and Talmage, D. W. (1979).Proc. Natl. Acad. Sci. U.S.A.76, 1994. Ruscetti, F. W., Morgan, D. A., and Gallo, R. C. (1977).J . Immunol. 119, 131. Ryser, J. E., Cerottini, J. C., and Brunner, K. T. (1978).J . Immunol. 120,370. Sanchez-Madrid, F., Krensky, A. M., Ware, C. F., Strominger, J. L., Burakoff, S. J., and Springer, T. A. (1982). Proc. Natl. Acad. Sci. U S A . 79, 7489. Schmitt-Verhulst, A. M., Sachs, D. H., and Shearer, G. M. (1976).J.Exp. Med. 143,211. Seeger, R. C., and Oppenheim, J. J. (1970).J . Exp. Med. 132,44. Sethi, K. K., Stroehamann, I., and Brandis, H. (1980). Nature (London) 286,718. Shaw, S., and Shearer, G. M. (1978).J . Immunol. 121,290. Shearer, G. M. (1974). Eur. J . Immunol. 4, 527. Shearer, G. M., and Schmitt-Verhulst, A. M. (1977).Ado. Immunol. 25, 55. Sherman, L. (1982). Nature (London) 297,511. Sherman, L., Burakoff, S. J., and Mescher, M. F. (1980). Cell. Immunol. 51, 141. Shevach, E. M., Chan, C., Thomas, D. W., and Clement, L. T. (1979). In “T and B Lymphocytes Recognition and Function” (F. H. Bach, B. Bonavida, E. S. Vitetta, and G. F. Fox, eds.), p. 405. Academic Press, New York. Shimonkevitz, R., Kappler, J., Marrack, P., and Grey, H. (1983).J. Exp. Med. 158, 303. Shinohara, N., and Sachs, D. H. (1979).J . E x p . Med. 150, 432. Simon, P. L., Farrar, J. J., and Kind, P. D. (1977).J. Immunol. 118, 1129. Spits, H., deVries, J. E., and Terhorst, C. (1981). Cell. Immunol. 59, 435. Spits, H., Borst, J., Terhorst, C., and deVries, J. E. (1982).J . Immunol. 129, 1563. Spits, H., Ijssel, H., Thompson, A., and deVries, J. E. (1983).J . Immunol. 131, 678. Springer, T. A., Davignon, D., Ho, M.-K., Kurzinger, K., Martz, E., and Sanchez-Madrid, F. (1982). Immunol. Rev. 68, 172. Stallcup, K., Springer, T., and Mescher, M. F. (1981).J. Immunol. 127,923. Steinman, R. M., and Nussenzweig, M. C. (1980). Immunol. Rev. 53, 127. Steinman, R. M., Chen, L. L., Witmer, M. D., Kaplan, G., Nussenzweig, M. C.,Adams, J. C., and Cohn, Z. A. (1980). In “Mononuclear Phagocytes, Functional Aspects” (R. van Further, ed.), p. 1718. Martinus Nijhoff, The Hague. Sunshine, G. H., Katz, D., and Czitrom, A. (1982). Eur. J . Immunol. 12, 9. Tevethia, S. S., Tevethia, M. J., Lewis, A. J., Reddy, V. B., and Weissman, S. M. (1983). Virology 128, 319. Thomas, D. W. (1978).J . Immunol. 121, 1760. Thomas, M., Cameron, J. R., and Davis, R. W. (1974). Proc. Natl. Acad. Sci. U.S.A.71, 4579. Thomas, Y. L., Rogozinski, L., Irigoyen, H., et al. (1982).J . Immunol. 128, 1386. Tsoukas, C. D., et al. (1981).J . Immunol. 126, 1742. Unanue, E. R. (1972). Adu. Immunol. 15, 95. Unanue, E. R. (1981). Adu. Immunol. 31, 1. Unanue, E. R., and Cerottini, J. C. (1970).J . Exp. Med. 131, 711. Uotila, M., Rode, H. N., and Gordon, J. (1978). Eur. J . Immunol. 8, 133. van der Poel, J. J., Molders, H., and Ploegh, H. (1983). Immunogenetics 17,609. van Wauwe, F. P., DeMay, J. R., and Goossener, J. G. (1981).J. Immunol. 124,2708.


85

Von Boehmer, H., and Haas, W. (1979).J.Exp. Med. 150, 1134. Wagner, H. (1973).J. E x p . Med. 138, 1379. Wagner, H., and Boyle, W. (1972). Nature (London) 240,92. Wagner, H., Feldman, M., Boyle, W., and Schrader, J. W. (1972).J. Erp. Med. 136,331. Wagner, H., Hess, M., Feldmann, M., and Rollinghoff, M. (1976). Transplantation 21,282. Wagner, H., Rollinghoff, M., Pfizenmaier, K., Hardt, C., and Johnscher, G. (1980). J . Zmmunol. 124, 1058. Wagner, H., Hardt, C., Rouse, B. T., Rollinghoff, M., Schenrich, P., and Pfizenmaier, K. (1982).J . E r p . Med. 155, 1876. Waldron, J., Horn, R., and Rosenthal, A. S. (1974).J. Zmmunol. 112, 746. Walker, E., Warner, N. L., Chestnut, R., Kappler, J., and Marrack, P. (1982).J . Zmmunol. 128,2164. Ware, C . F., Krangel, M. S., Pious, D., Burakoff, S. J., and Strominger, J. L. (1983a). J . Zmrnunol. 131, 1312. Ware, C. F., Sanchez-Madrid, F., Krensky, A. M., Burakoff, S. J., Strominger, J. L., and Springer, T. A. (1983b).J.Zmmunol. 131, 1182. Watson, J., and Mochizuki, D. (1980).Zmmunol. Reo. 51,257. Watson, J., Cillis, S., Marbrook, J., Mochinzuki, D., and Smith, K. A. (1979).J . E x p . Med. 150, 849. Wee, S.-L, Wu, S., Alter, B. J., and Bach, F. H. (1982). Hum. Zmmunol. 3,45. Weinberger, O., Herrman, S. H., Mescher, M. F., Benacerraf, B., and Burakoff, S. J. (1980). Proc. Natl. Acad. Sci. U.S.A. 77,6091. Weinberger, O., Herrmann, S. H., Mescher, M. F., Benacerraf, B., and Burakoff, S. J. (1981a). Proc. Natl. Acad. Sci. U.S.A.78, 1796. Weinberger, O., Herrmann, S. H., Mescher, M. F., Benacerraf, B., and Burakoff, S. J. (1981b). Eur.J. Zmmunol. 11, 405. Weinberger, O., Germain, R. N., Springer, T. A., and Burakoff, S. J. (1982).J.Zmmunol. 129, 694. Weinberger, O., Germain, R. N., and Burakoff, S. J. (1983). Nature (London) 302,429. Weinberger, O., Herrmann, S. H., Greenstein, J., Mescher, M. F., and Burakoff, S. J. (1984). In preparation. Werdelin, O., Braendstrup, O., and Shevach, E. M. (1979).J.Zmmunol. 123, 1755. Wigler, M., Pellicer, A., Silverstein, S., Axel, R., Urlaub, G., and Chasin, L. (1979).Proc. Natl. Acad. Sci. U.S.A. 76, 1373. Yokayama, K., and Mathenson, S. G. (1983).J.Zmmunol. 130, 1419. Zarling, J. M., Clouse, K. A., Biddison, W., and Kung, P. C. (1981).J. Zmmunol. 127, 2575. Ziegler, K., and Unaniie, E. R. (1981).1. Ztnmunol. 127, 1869. Ziegler, K., and Unanue, E. R. (1982).Proc. Natl. Acad. Sci. U.S.A.79, 175. Zinkernagel, R. M., and Doherty, P. C. (1974). Nature (London) 248,701. Zinkernagel, R. M., Callahan, G., Althage, A., Cooper, S., Steilein, J., and Klein, J. (1978).J. E x p . Med. 147, 897. Zuniga, M. C., Malissen, B., McMillan, M., Brayton, P. R., Clark, S. S., Forman, J., and Hood, L. (1883).Cell 34, 535. Zweerink, H. J., Askonas, B. A., Millican, D., Courtneidge, S. A., and Skehel, J. J. (1977). Eur. J . Zmmunol. 7,630.


ADVANCES IN

IMMUNOLOGY,VOL. 36

The Human Thymic Microenvironment BARTON F. HAYNES Department of Medicine, Division of Rheumatic and Genetic Diseases, and the Depahent of Microbiology and lmmunolog~ Duke Univeniiy School of Medicine, Durham, North Carolina

I. Introduction

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

11. Anatomy and Histology of the Thymic Microenvironment . . . . . . . . . . . . . ,

111. Evidence for the Participation of the Thymic Microenvironment in Promoting T Cell Maturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Normal T Cell Maturation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. T Cell Leukemogenesis and Autoimmune Disease . . . . . . . . . . . . . . IV. Heteroantisera and Other Markers That Define Components of the Human Thymic Microenvironment . , . . . . . . . . . . . . . . . . . . . A. Heteroantisera against Thymic Hormones. . . . . . . . . . . . . . B. Polyclonal Antikeratin Antibody . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Anti-Group A Streptococci Antibody. . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Human or Murine Autoimmune Disease Sera . .. . . .. .. .. . V. Monoclonal Antibodies That Define Components of the Human and Rodent Thymic Microenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Endocrine Thymic Epithelium.. . . . . . . . . . . . . . . . . . . . . . . . . . . B. Mesodermal-Derived Thymic Stroma (TE-7) . . . . . . . . . .. . . . . C. Cortical Thymic Epithelium (TE-3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Subcapsular Cortical Thymic Epithelium. . . . . . . . . . . . . . . . . . . . . . . . . E. Antibodies against Class I (Anti-HLA, A, B, and C) and Class I1 (Anti-Ia-like or Ia) Antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Other Monoclonal Reagents Reactive with Human Thymic . Epithelium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Ontogeny of the Human Thymus Microenvironment . . . . . . . . . , . , . . , , VII. The Human Thymic Microenvironment in Diseases of Abnormal T Cell Maturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Severe Cellular Immunodeficiency Diseases . . . . . . . . . . . . . . . . . . . . . B. Thymoma. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Myasthenia Gravis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . .. . . VIII. Summa. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

. . . .. .

. . ..

.

.. .

87 89 94 94 96

.

.. . . ... .

.

..

.

.. . .

.

.

.

I

. . . .

. .

.

.

. . .. ...

97 98 98

99 99 111 114 116 117 117 119 126 126 131 132 134 136

1. Introduction

The thymic microenvironment is a complex specialized tissue derived from at least three sources-endoderm of the third pharyngeal pouch, ectoderm of the third brachial cleft, and mesenchymal stromal cells derived from embryonic mesoderm (Weller, 1933; Norris, 1938; Auerbach, 1960, 1961; Cordier and Haumont, 1980; Patten, 1968). Pharyngeal pouch endodenn and brachial cleft ectoderm give rise to 87 Copyright 8 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-022436-4

88

BARTON F. HAYNES

epithelial components, while mesodermal-derived mesenchymal cells form the thymic capsule, vessels, and interlobular septae (Weller, 1933; Norris, 1938; Cordier and Haumont, 1980). Early in thymic ontogeny, mesodermal-derived connective tissue induces epithelial cell maturation and fetal thymic lobulation (Auerbach, 1960, 1961; LeDouarin and Jotereau, 1975). Ultrastructurally, the nonlymphoid postnatal thymic microenvironment is composed of macrophages (Kendall, 1981; Beller and Unanue, 1977, 1978), fibroblasts, and dark and pale epithelial cells (also called epithelial reticular cells) (Kendall, 1981; Rouse and Weissman, 1981). In postnatal human thymic medulla, Hassall’s bodies, or keratinized epithelial swirls, are frequently present (Norris, 1938). Microscopists have long appreciated the dendritic nature of thymic epithelial cells (Kissel and Karden, 1979, reviewed in Kendall, 1981). Cortical and medullary epithelial cells have long, thin processes which connect with other epithelial cells via desmosomes, and form a complex interdigitating meshwork filled with thymocytes (Kissel and Karden, 1979; Kendall, 1981). Some thymic epithelial cells have been shown to express Class I and Class I1 major histocompatibility (MHC) antigens (Janossy et al., 1980; Bhan et al., 1980; Jenkinson et ul., 1981), and functional data have demonstrated that the thymic microenvironment is responsible for conferring to developing thymocytes the ability to recognize self-class I and Class I1 MHC antigens in certain in vitro T cell assay systems (Fink and Bevan, 1978; reviewed in Zinkernagel and Doughtery, 1979; Kruisbeek et al., 1981a,b; Singer et al., 1982).Kruisbeek et al. (1983)have suggested that during early T cell development, education of Class I MHC restricted T cells is conferred by different thymic elements than those involved in the education of Class I1 MHC restricted T cells. Moreover, a subset of epithelial cells within the thymic microenvironment contains a variety of thymic hormones which are postulated to induce certain stages of T cell maturation (Schmidt et al., 1980; reviewed in Goldstein et al., 1981; Haynes et al., 1983b; Beardsley et aZ., 1983). Thus, the differentiation of immature T lymphocytes is thought to be effected by thymic epithelial cells and factors secreted by these cells. In addition, thymic macrophages and their products have also been implicated in inducing thymocyte maturation (Beller and Unanue, 1977, 1978). However, in spite of the important role the thymic microenvironment plays in the maturation of T cells, little is known regarding antigenically definable subsets of cells within the nonlymphoid component of the human thymic microenvironment.

HUMAN THYMIC MICROENVIRONMENT

89

Recently we have described complex ganglioside antigens on human and rodent endocrine thymic epithelium, and demonstrated thymopoietin- and thymosin al-containing endocrine epithelial cells are present in two distinct regions of the thymus-the subcapsular cortex and medulla (Haynes et al., 1983b,c). Moreover, we have produced a panel of murine monoclonal antibodies raised against human thymus (TE-3, TE-4, TE-7, TE-8, TE-15, and TE-19) which define specific regions or components of the human thymic microenvironnient (Haynes et al., 1983a; Haynes, 1984; Lobach et al., 1984; McFarland et al., 1984). In this article we will review aspects of current knowledge regarding the role of the thymic microenvironment in promotion of T cell maturation. In addition, recent work using monoclonal antibodies to define subsets of epithelial and stromal cells within the human thymic microenvironment will be presented. Finally, the use of monoclonal antithymic epithelial antibodies as investigative probes in the study of normal thymus ontogeny and in the study of aberrant T cell and thymic epithelial cell differentiation will be discussed. II. Anatomy and Histology of the Thymic Microenvironment

The human nonlymphoid portion of postnatal thymus consists of two main components: (1)connective tissue and vessel compartment, and (2) epithelial cells and their specialized structures, Hassall’s bodies. Also present are variable numbers of macrophages (also called mesenchymal reticular cells), and rare myoid cells, plasma cells, and eosinophils (reviewed in Kendall, 1981; Singh, 1981). Thymic lobules are separated from each other by connective tissue partitions or interlobular septae. The connective tissue of the thymic capsule extends into the gland and constitutes a perivascular space that is technically extraparenchymal (Kendall, 1981). Interlobular septae penetrate the parenchyma of thymic lobules with accompanying blood vessels (Pereira and Clermont, 1971). The connective tissue compartment is devoid of epithelial cells and is demarcated from the epithelial compartment by a layer of flattened epithelial cells and their basal laminae (Pereira and Clermont, 1971; Kendall, 1981).The connective tissue compartment is rich in reticular fibers and consists of two main zones: (1)the inner medulla surrounding vessels, and (2) interlobular septae surrounding blood and lymphatic vessels entering or leaving the thymic lobule via the cortex (Weller, 1933; Pereira and Clermont, 1971; Cordier and Haumont, 1980). It has been suggested

90

BARTON F. HAYNES

that the connective tissue of the thymic interlobular septae could be a major route for the traffic of lymphocytes both into and out of the thymus (Kendall, 1981). Thymic epithelial reticular cells show great diversity of form. Generally epithelial cells possess long, branching cytoplasmic processes which form an interconnecting meshwork with lymphocytes occupying the interstices. Ultrastructurally, epithelial cells contain tonofilaments (Fig. 1)which frequently insert into desmosomes (Bearman et al., 1978; in Singh, 1981). Whereas processes of epithelial cells are frequently joined by desmosomes (Fig. l ) , no desmosomes are present between thymocytes and thymic epithelium.

FIG.1. Transmission electron micrograph of two adjacent human thymic epithelial cells grown in long-term cultures in the presence of epidermal growth factor (K. Singer, E. A. Harden, D. F. Lobach, and B. F. Haynes, unpublished observations). The nuclei (N) of adjacent cells can b e seen at the top and lower of right area. Thymic epithelial cells contain tonofilaments (dark arrows) and are joined to other thymic epithelial cells by desmosomes (white arrows) (bar = 5 Hm).


91

Several investigators have described ultrastructural differences between cortical and medullary epithelial cells (Hirokawa, 1969; Mandell, 1968a; Gaudecker and Muller-Hermelink, 1978; Singh, 1981), as well as various subtypes of medullary epithelial cells (Kendall and Frazier, 1979; Singh, 1981). Dark and pale epithelial cells show many morphologic variations, and are located both in the thymic cortex and medulla. Dark epithelial cells are commonly found associated with extracellular collagen fibers. Pale epithelial cells contain membrane bound vacuoles which on electron microscopy (EM) contain amorphous grey material or small vesicles (Clark, 1963; reviewed in Clark, 1973).These membrane bound vacuoles are regarded as characteristic features of pale epithelial cells and are thought to contain thymic hormones (Clark, 1973). Pale epithelial cells have been postulated to be hormone-secreting cells in the thymus, and in the thymic medulla are thought to form Hassall’s corpuscles (Singh, 1981). In the thymic cortex, two morphologic types of epithelial cells have been identified: flattened and stellate (Pereira and Clermont, 1971). Flattened epithelial cells border on connective tissue and form a continuous layer of cells along the inner surface of thymic capsule, interlobular septae, and along perivascular connective tissue spaces surrounding blood and lymphatic vessels (Pereira and Clermont, 1971). Mandel (1970) has suggested that in mice and guinea pigs thymic epithelial cells undergo differentiation and mitosis during early fetal development. Moreover, it has been postulated that postnatal murine thymus contains undifferentiated and dividing epithelial stem cells that repopulate the differentiated thymic epithelial cell component (Mandel, 1970). Thymic macrophages can be found in the capsule, interlobular septae, and perivascular spaces as well as in the cortex and medulla. Macrophages appear in greatest concentrations at the thymic corticomedullary junction (Kendall, 1981).Although macrophages have been shown to bind with immature thymocytes spontaneously, the physiologic significance of this binding is not clear (Wu and Thomas, 1983). Beller and Unanue (1977,1978) have suggested macrophages induce certain stages of thymocyte maturation. Another cell type found in the medulla and deep cortex of the thymus is called the interdigitating reticular cell (Ewijk e t al., 1974; Janossy et at., 1980) and is similar in morphology to cells found in T cell-dependent areas in mouse spleen (Steinman and Cohen, 1973; Steinman et aE., 1980, 1981).These Ia+ cells may be of the monocytemacrophage lineage (Janossy et al., 1980; Kendall, 1981).A characteristic feature of these cells is the long antler-like processes of cyto-

92

BARTON F. HAYNES

FIG. 2. Light micrograph of normal thymus from a l-year-old subject. (A) Outer portion of thymic lobule showing interlobular septum (ILS), thymic capsule (C), subcapsular cortex (SCC) and inner cortex. (B) Medullary area of thymus with Hassall’s , and eosin). body (H) ( ~ 4 0 0hematoxylin

plasm that wrap around thymocytes and in some studies are phagocytic (Kendall, 1981). Whether thymic interdigitating reticular cells are of the same lineage as dendritic cells in lymph node and spleen is not known (Janossy et al., 1980; Steinman et al., 1981). Other rare cell types in the thymic microenvironment include myoid cells, which contain striated filaments of actin and myosin, and plasma cells, eosinophils, and basophils (reviewed in Kendall, 1981). Anatomically the thymus can be divided into four regions: (1)the subcapsular cortex, (2) the inner cortex, (3) the medulla, and (4) the fibrous capsule and interlobular septae (reviewed in Clark, 1973). The subcapsular cortex is the chief site of thymic lymphopoiesis and is the zone of greatest mitotic activity (Fig. 2A). Thymocytes in the subcapsular cortex are large lymphocytes which proliferate quickly (6-9 hours/cycle) (Metcalf and Wiadrowski, 1966). It has been postulated


93

FIG.28.

that the subcapsular cortex is where bone marrow stem cells enter the thymus gland and begin maturation. Mandel (1969, 1970) has suggested that subcapsular cortical epithelial cells stimulate thymocyte mitosis. Interestingly, human subcapsular cortical epithelium has recently been shown to be antigenically different from other thymic regions (Ritter et al., 1981; Haynes et al., 1983c)and to contain thymic hormones that other regions do not (Hirokawa et aE., 1982; Goldstein et al., 1981; Haynes et al., 1983~). The inner thymic cortex is filled with small nonproliferating lymphocytes interspersed with large mitotic cells (Clark, 1973) (Fig. 2A). In murine systems, cortical thymocytes have been demonstrated to move slowly toward the thymic medulla (Weissman, 1973; reviewed in Cantor and Weissman, 1976). Clark (1973) has viewed the inner cortex as an assembly line for sequential differentiation of thymo-

94

BARTON F. HAYNES

cytes, but little is known concerning either the sequence or location of such differentiation. Umiel and colleagues (1982) have recently demonstrated a minor human cortical thymocyte population which was immunocompetent. Most workers agree that the majority of cortical thymocytes arise intrathymically from precursors in the subcapsular cortex, are immunoincompetent, and that greater than 95%of cortical thymocytes are short-lived and die intrathymically (Weissman, 1973; Scollay et al., 1980). Only a minority of cortical thymocytes are exported to peripheral lymphoid organs, or migrate to the thymic medulla (Cantor and Weissman, 1976). The thymic medulla contains more epithelial cells and fewer thymocytes than are found in the cortex (Fig. 2B). Medullary thymocytes account for approximately 10 to 15%of all human thymic lymphocytes (Reinherz et al., 1980; Haynes et al., 1983d). In contrast to cortical thymocytes, medullary thymocytes are immunocompetent, express high density HLA antigens, and express mature T cell phenotypes (Cantor and Weissman, 1976; Reinherz et al., 1980, Haynes et al., 1983d). In vivo and in vitro labeling studies in mice have suggested that a minor population of medullary thymocytes arise from subcapsular cortical prothymocytes (reviewed in Clark, 1973). However, recent studies suggest that the medullary thymocyte pool is a resident population of mature T cells, the majority of which may not be descended directly from cortical thymocytes, and are not the main source of shortlived thymocytes destined for export to the periphery (Eliot et al., 1971; reviewed in Stutman, 1978; Naparstek et al., 1982). Rather, short-lived thymocytes exported to the periphery may be derived from foci of maturing thymocytes scattered about the cortex or located at the corticomedullary junction (Umiel et al., 1982). 111. Evidence for the Participation of the Thymic Microenvironment in Promoting T Cell Maturation

A. NORMALT CELLMATURATION From experimental data in rodent models, as well as data from studies of human immunodeficiency diseases, it appears that passage through the thymus is an obligatory step for the generation of competent T cells (reviewed in Stutman, 1978). It is beyond the scope of this review to discuss in detail the evidence for participation of the thymic microenvironment in promoting T cell maturation. Readers are referred to reviews on this point (Cantor and Weissman, 1976; Stutman, 1978).What follows, therefore, is a brief summary of studies regarding


95

potential mechanisms of thymic microenvironment induction of various stages of T cell maturation. First, since hematopoietic stem cells migrate to thymus, embryonic thymic epithelium is thought to send out chemical signals to hematopoietic T cell precursor cells (reviewed in LeDouarin and Jotereau, 1981). Once in the fetal or neonatal thymic microenvironment, hematopoietic T precursor cells undergo cell division as well as a series of functional and phenotypic changes. Further, a portion of intrathymic T cells emigrate from the thymus and appear in lymph nodes and thoracic duct lymph. Emigration of these postthymic T cells appears to be unidirectional to nodes and thoracic duct (Stutman, 1978), although recent evidence suggests that in some cases, long-lived memory T cells can migrate back to the thymic medulla (Naparstek et al., 1982). A series of experiments from a large number of investigators has demonstrated that intrathymic T cell maturation requires a viable and accessible thymic microenvironment and cannot be replaced by thymus tissue within cell impermeable diffusion chambers (reviewed in Stutman, 1978). Thus, the original notion that the main maturational events within the thymus were mediated by thymic humoral factors has been modified. It is believed that direct contact between developing thymocytes and thymic microenvironment components, as well as effects of thymic hormones (thymosin cul, thymopoietin, thymulin), Interleukin 1 (IL-1), and Interleukin 2 (IL-2), mediate a complex sequence of events leading to normal T cell maturation (reviewed in Bach and Papiernik, 1981). It has recently become clear that, at least in murine systems, postthymic T cell precursors exist which are immunologically incompetent, thymus-processed lymphocytes exported from the thymus to peripheral lymphoid organs (Stutman, 1978). In addition, data suggest that stages of T cell maturation occur in neonatal animals prethymically, in bone marrow, as well as intrathymicall y and postthymically (Stutman, 1978; reviewed in Morrissey and Singer, 1983). Thus, while passage through the thymic microenvironment at early stages of immune system development is requisite for full development of the T cell repertoire, it appears that prethymic events in bone marrow and postthymic events in a variety of peripheral sites (tonsil, lymph node, spleen, skin) are important for various stages of T cell development. Bach and Papiernik (1981) have postulated that relevant factors for intrathymic T cell maturation are derived both from thymic hormones, as well as from a maturational effect of direct contact between epithelial cells and thymocytes. Regarding postthymic maturation, extrathymic microenvironments as well as circulating thymic hormones and IL-2 have all been impli-

96

BARTON F. HAYNES

cated as important factors, although specific sites of T cell maturation and specific molecular mechanisms of postthymic T cell differentiation, particularly in humans, are not known. B. T CELLLEUKEMOGENESIS AND AUTOIMMUNEDISEASE It has been known since 1944 that the thymus is important in regulating the expression of spontaneous leukemia in AKR mice (McEndy et al., 1944). Experiments by a number of investigators (reviewed in Parkinson and Waksal, 1978) have suggested that thymic epithelium in the leukemic AKR mouse has a dual role: (1)to bring the target T cell into a state of differentiation such that it is susceptible to malignant transformation and (2) to provide additional factors, such as retroviruses needed to initiate leukemogenesis. Recently it has been shown that thymic epithelium grafts from young preleukemic AKR animals placed in nonleukemic mice induced both amplified expression of retroviral genes and leukemic changes in patterns of thymocyte differentiation (Zielinski et al., 1982). I n contrast, thymic epithelial grafts of nonpreleukemic mice did not induce these changes (Zielinski et al., 1982). Morphological abnormalities of AKR thymic epithelium have been noted to precede the development of T cell leukemia (Kyewski et al., 1981). Leukemogenic potentiation of whole AKR thymic epithelium has recently been produced by an insoluble fraction of thymic epithelial membranes (Fournier and Potworowski, 1983). Regarding thymic involvement in the pathogenesis of autoimmune diseases such as systemic lupus erythematosus and myasthenia gravis (MG), no definite pathophysiologic mechanisms have yet been delineated. However, since the thymic microenvironment is responsible for educating maturing T cells in some aspects of self-MHC reactivity, it is logical to investigate the thymic microenvironment and mechanisms of normal and aberrant T cell maturation for clues to autoimmunity. Defects in T cell number and function have been observed in rodent models and in humans with systemic lupus erythematosus, juvenile onset (Type I) diabetes mellitus, and myasthenia gravis (Fauci et al., 1978; Jackson et al., 1981,1982; reviewed in Haynes and Eisenbarth, 1983). Autoantibodies against various components of thymic epithelium have been found in rat and human autoimmune diabetes mellitus and myasthenia gravis (Eisenbarth et al., 1983; Buse et al., 1983; Dardenne et al., 1983). Moreover, the thymus has been found to b e necessary for both expression of SLE-like disease in MRL mice and to regulate B cell hyperreactivity in autoimmune BXSB mice (Theofilopoulos et al., 1981; Smith et al., 1983). Thymic factors have been implicated in the pathogenesis of MG because of frequent thy-


97

mic histologic abnormalities (B cell germinal centers, thymic medullary hyperplasia, or thymoma) and because of the clinical improvement often seen in MG following thymectomy (reviewed in Haynes and Eisenbarth, 1983).

IV. Heteroantisera and Other Markers That Define Components of the Human Thymic Microenvironment

Prior to the development of monoclonal antibodies, the thymic microenvironment was studied using a variety of heteroantisera, patient sera, and other markers. A. HETEROANTISERA AGAINST THYMIC HORMONES The intrathymic location of thymic hormones has been determined using heteroantisera raised against crude, purified, or synthetic preparations of thymus-derived factors (soluble thymic factor, thymuiin, thymosin a l , thymopoietin) (Mandi and Glant, 1973; Goldstein, 1974; Potworowski, 1977; Schmidt et al., 1980; Goldstein et al., 1981; Haynes et al., 1983b). Soluble thymic factor, thymulin, and thymopoietin have all been localized to epithelial cells in the thymic cortex and medulla (Mandi and Glant, 1973; Goldstein, 1974; Potworowski, 1977; Schmidt et al., 1980; Goldstein et al., 1981; Haynes et al., 1983b). Hirokawa et al. (1982), Goldstein et al. (1981), and Haynes et al. (1983b,c) have shown that normal thymosin al-containing human thymic epithelium exists in two regions-the subcapsular cortex and the medulla. In contrast, thymosin p3 and thymosin p4 (other thymusderived though not thymus-specific polypeptides) are present only in the subcapsular cortical epithelial region and not in the medullary epithelial region (Hirokawa et al., 1982; Haynes et al., 1983~).In addition, thymosin p4 is produced by thymic macrophages (Hannappel et al., 1982). Thus, while not all thymotrophic hormones isolated from thymus are thymus specific, within the thymus these factors have generally been localized to the epithelial component of the thymic microenvironment.

B. POLYCLONAL ANTIKERATIN ANTIBODY Studies of thymus tissue with polyclonal antibodies against human keratin have shown either that Hassall’s bodies alone (Sun et al., 1979; Viac et al., 1980) or nests of medullary epithelial cells surrounding Hassall’s bodies contained keratin (Takigawa and Imamura, 1977). With most polyclonal antikeratin antibodies, cortical epithelial areas were reported as nonreactive.

98

BARTON F. HAYNES

C. ANTI-GROUP A STREPTOCOCCI ANTIBODY Rabbit antiserum against group A streptococci has been used to identify human thymic epithelium (Lyampert et al., 1976).Anti-group A streptococcal antibodies against A-polysaccharide reacted with the basal layer of keratinocytes in skin, sclera, and with the outer cortical and medullary epithelial areas of thymus. While epithelial cells surrounding Hassall’s bodies were reactive, the concentric layer of cells within Hassall’s bodies were not reactive. Lyampert et al. (1976) made the important postulate that during a humoral immune response to streptococcal antigens, autoantibodies might arise that react with thymus, interfere with T cell development, and produce immunoregulatory abnormalities. D. HUMAN OR MURINEAUTOIMMUNE DISEASE SERA A number of human diseases or murine models of human disease have been shown to have serum antibodies reactive with components of the thymic microenvironment. In addition to providing interesting antigenic associations between skin keratinocytes and components of thymus, studies using these human and mouse sera suggest that in a number of diseases, antithymic epithelial antibodies are present and potentially could disrupt normal intrathymic T cell maturation. 1 . Graft-versus-Host Disease (GVHD) GVHD is seen in recipients of allogeneic bone marrow transplantation and in patients with primary immundeficiencies following whole blood or maternal fetal tranfusions (Hood et al., 1977). In both acute and chronic GVHD, skin involvement is a prominent feature of the disease. Thymic involution following graft-versus-host disease has been described in both humans and murine systems (Seemayer et al., 1977, 1978; Seemayer and Bolande, 1980). Thymic epithelial injury has been described in murine GVHD with loss of thymic epithelial cells containing thymotrophic hormones (Potworowski et al., 1979). Interestingly, GVHD sera frequently contain antibodies that react both with skin epidermal cells and thymic Hassall’s bodies (Didierjean and Saurat, 1980).

2. Pemphigus Pemphigus is an autoimmune blistering disease of several clinical varieties. In pemphigus, autoantibodies are made which bind to intercellular substance of skin keratinocytes (Beutner and Jordan, 1964). Hashimoto et al. (1983) have demonstrated that pemphigus IgG au-


99

toantibodies activate plasminogen activator. Plasminogen activator in turn activates plasminogen to plasmin which is responsible for disruption of epithelial cell to cell contact and ultimately clinical skin lesions. Similar to GVHD serum, pemphigus serum has been shown to react with human thymic Hassall’s bodies (Pertshuk, 1974; Beletskaya and Gnezditskaya, 1974; Takigawa and Imamura, 1977). 3. Rodent Models of Type 1 Diabetes Mellitus Type I diabetes mellitus is associated with severe T and B cell immunoregulatory abnormalities in rodents and man (reviewed in Eisenbarth et al., 1983). Antibodies against facteur thymique serique (FTS) (also called thymulin) have been described in the mutant diabetic (db/db) mouse, with associated loss with age of FTS-containing thymic epithelial cells (Dardenne et al., 1983). BB Wistar rats have spontaneous autoimmune diabetes mellitus, and helper T cell lymphopenia (Jackson et al., 1981). Buse et al. (1983)have recently found that these animals have antithymic epithelial antibodies. These observations are of interest in that we have recently described complex ganglioside antigens shared by both thymic endocrine epithelial cells and pancreatic islet cells of in both man and rodents (Haynes et al., 1983b). Thus, a number of human diseases have been identified in which antibodies are made which potentially could react in vivo with epithelial cells within the thymic microenvironment. Interestingly, most of these diseases (GVHD, Type I diabetes mellitus, pemphigus) are diseases associated with immunoregulatory T cell abnormalities. V. Monoclonal Antibodies That Define Components of the Human and Rodent Thymic Microenvironment

During the past 3 years we have made or characterized a number of monoclonal antibodies that react with various components of the human thymic microenvironment. These reagents are summarized in Table I. A. ENDOCRINE THYMIC EPITHELIUM 1 . Complex Ganglioside Antigens Defined by Monoclonal Antibody A2B5 and Receptors for Tetanus Toxin Eisenbarth et al. have recently characterized a murine monoclonal antibody (A2B5) that reacts with a complex neuronal ganglioside expressed on the cell surface of neurons, neural crest-derived cells, and

TABLE I MONOCLONALANTIBODIESAGAINST HUMANTHYMIC MICROENVIRONMENT COMPONENTS

$

Monoclonal antibody

Fusion parent

Ig isotype

Reactivity pattern in human thymus

Reactivity pattern in other human tissues

Human thymic stroma Human thymic stroma

P3X63/Ag8

IgGz

Cortical epithelium

P3X63/Ag8

IgM

SCC' + M endocrine epithelium

Human thymic stroma

653

IgGl

TE-8*

Human thymic stroma

653

IgG%

Mesodermal-derived thymic stroma and vessels Hassall's bodies

TE-15

Human thymic stroma

NS-1

IgM

Hassall's bodies

Many epithelial tissues Basal layer of squamous epithelium Mesodermalderived stroma, vessels of all tissues Granulosa layer of squamous epithelium Stratum corneum of squamous epithelium

TE-7

TE-3 TE-4

Immunogen

Expressed on rodent tissues

Antigen"

No

NKb

No

NK

No

NK

No

NK

No

NK

I

z

TE-16

Human thymic stroma

653

IgC,

Hassall's bodies

TE-19

Human thymic stroma

NS-1

IgM

Hassall's bodies

A2BS

Chick retinal cells

P3X63/Ag8

IgM

SCC + M endocrine epithelium

BB TECS"

Rat autoantibody from BB Wistar Rat

P3X63/Ag8

IgCz

SCC + M endocrine epithelium Hassall's bodies

Anti-HTLV PIP

HTLV

NS-1

IgC,

+

SCC + M endocrine epithelium

Granulosa layer of squamous epithelium Squamous epithelium, erythrocytes Other neuroendocrine tissues Basal and granulosa cell layer of squamous epithelium Basal layer of squamous epithelium

No

NK

Yes

A blood group antigen on erythrocytes Complex ganglioside

Yes

Yes

NK

No

NK

BB TECS antibody, rat monoclonal autoantibody made from fusion of P3X63/Ag8 mouse cells and BB Wistar rat spleen cells. NK, not known. SCC, Thymic subcapsular cortex; M, medulla. TE-8, 15, 16, and 19 are expressed on a variety of normal human epithelial tissues as well as skin and thymus. Anti-pl9 defines an HTLV 19,000-dalton internal core protein. The molecule to which anti-pl9 binds in human thymus has not been characterized.

102

BARTON F. HAYNES

peptide-secreting endocrine cells (Eisenbarth et al., 1979, 1981, 1982). Tetanus toxin (TT), which binds to GD and GT gangliosides, also binds to this neuroendocrine family of cell types (Eisenbarth et al., 1982). These observations, coupled with previous studies demonstrating that a portion of thymic epithelium in fowl is neural crest derived (LeDouarin and Jotereau, 1975), led us to determine if antibody A2B5 and/or TT bound to cells in human and rodent thymus. We demonstrated that human and rodent thymic epithelial cells expressed antigens recognized by antibody A2B5 and as well bound TT. A2B5+ and TT+ thymic epithelial cells were found in two discrete locations-the subcapsular cortex and medulla (Fig. 3A and B). Moreover, using monoclonal antibody A2B5 and rabbit antithymopoietin or antithymosin a1 antibodies, we showed that A2B5+, TT+ thymic epithelial cells contained thymopoietin and thymosin a1 (Haynes et al., 1983b,c). Using monoclonal antibody AE-1 (antikeratin) (Woodcock-Mitchell, 1982)we have shown all A2B5+ thymic epithelial cells contain keratin (Haynes et al., 1984).All of the cells belonging to the amine precursor uptake and decarboxylation (APUD) series of cells previously studied, including pancreatic islet cells, anterior pituitary cells, adrenal medulla, as well as medullary carcinomas of the thyroid, melanomas, and neuroblastomas reacted with antibody A2B5 and TT (Eisenbarth et al., 1981, 1982). Similar to the recent studies of neuron-specific enolase, a series of nonneuronal APUD cells have been found to express what was initially considered neuronal antigens (Tapia et al., 1981). Though the majority of cells reacting with antibody A2B5 are of neural crest derivation, not all A2B5+ cells derive from the neural crest. In particular, pancreatic islet cells may not be of neural crest origin (LeDouarin and Jotereau, 1975; Kissel et al., 1981; Pictet et al., 1976), yet these cells and their tumors bound antibody A2B5 and TT. Therefore, the expression of complex gangliosides by cells may reflect a common endocrine function rather than common embryonic derivation.

2. Antibody TE-4 Antibody TE-4 is the product of a murine hybridoma from a fusion using spleen cells from an animal immunized with human thymic stroma (Haynes et al., 1984). In indirect fluorescent assays antibody TE-4 defined the A2B5+, thymosin al-containing thymic epithelial cells located in the subcapsular cortex and medulla (Haynes et al., 1984) (Fig. 4A). Unlike A2B5, TE-4 did not react with any other endocrine tissues such as islet cells of pancreas, anterior pituitary, or adrenal medulla. However, like A2B5, antibody TE-4 reacted with basal


103

keratinocytes of squamous epithelium in skin, tonsil, upper esophagus, and conjunctiva (Haynes et al., 1984). Several lines of evidence suggest that TE-4 does not bind to known keratin-like molecules. While all TE-4+ cells contain keratin as defined by antibody AE-1, not all keratinized cortical epithelial cells

FIG.3. (A) Reactivity pattern of A2B5 antibody on normal thymus. Using indirect immunofluorescence, antibody A2B5 identified two distinct regions of normal thymus epithelium-the subcapsular cortical area (SCC) and the medullary area. Cortical stroma cells, all thymocytes, and the capsule of the thymus (C) were nonreactive with A2B5 (~400). (B) A2B5 in normal human thymus medulla stained cells in a reticular pattern. Note long interconnecting dendritic-like processes (arrows) of A2B5-reactive epithelial cells (~400) (Haynes et al., 1983b,c).

104

BARTON F. HAYNES

FIG.3B.

were TE-4+, nor were AE-1+ Hassall’s bodies reactive with TE-4. Moreover, TE-4 had surface reactivity with thymic epithelial cells, and skin and thymic epithelial cells grown in culture frequently are TE-4- yet contain keratin (AE-1+) (K. Singer, E. Harden, and B. Haynes, unpublished observations). TE-4, though having a similar reactivity pattern to anti-pl9, differed from anti-pl9 in that TE-4 was conserved in ontogeny beginning at 7 weeks, and did not react with HTLV-infected T cells (HUT 102 cells) (see Sections V,A,3 and VI below).


105

FIG.4. Reactivity patterns of antithymic epithelial antibodies in normal thymus. (A) Antibody TE-4 reacted with dendritic epithelial cells in thymic medulla (arrow) and subcapsular cortex (not shown). H labels a TE-4 nonreactive Hassall’s body. (B) Antibody BB TECS reacted with dendritic epithelial cells in thymic medulla (arrow) and subcapsular cortex (not shown). H labels a BB TECS reactive Hassall’s body. (C) Antibody TE-8 in thymus reacted only with Hassall’s bodies (H). (D) Antibody TE-15 in thymus reacted only with granular material within Hassall’s bodies (H) (arrow) (x400).

3. Antihuman T Cell LeukemialLymphoma Virus (HTLV) p l 9 A novel human retrovirus, HTLV, has recently been isolated from a variety of adult T cell leukemias and lymphomas that have so far generally exhibited a particular geographic distribution in the Southern United States, West Indies, and Southern Japan (Poiesz et al., 1980, 1981; Gallo et al., 1982). Specific antibodies to HTLV proteins have been found in virus-positive cases, often in family members, and much less frequently in the general population (Kalyanaraman et al., 1981, 1982; Posner et al., 1981; Robert-Guroff et al., 1982a,b; Blattner

106

BARTON F. HAYNES

et al., 1982). To further study the incidence of HTLV in large patient populations, a monoclonal antibody (12/1-2) to the HTLV structural protein p19 has been used to examine T cells of patients and normal donors for evidence of HTLV expression (Robert-Guroff et al., 1981). The monoclonal anti-pl9 antibody immunoprecipitates a 19,000dalton structural protein of HTLV. The antigen is present in HTLVinfected T cells and is not detected in non-HTLV-infected malignant or normal T cells of any type (Robert-Guroff et al., 1981). During a recent screen of anti-p19 reactivity with a large number of human tissues, we observed that anti-pl9 reacted strongly with a surface membrane epithelial component of normal human thymus (Haynes et al., 1983~).We characterized the specificity of anti-pl9 binding to human thymus, and sought other HTLV antigens in normal thymic tissue. We demonstrated the extent of reactivity of anti-pl9 with thymic epithelium to be a function of age of donor of the thymus, and showed that all anti-p19+ thymic epithelial cells were A2B5+ and contained thymosin a1 and thymopoietin. Moreover, no other HTLV proteins or proviral HTLV DNA were demonstrated in anti-p19+ thymic epithelium (Haynes et al., 1983~).Thus, in normal human thymus, the antigen defined by anti-pl9 is specific for the neuroendocrine component of thymic epithelium and is acquired during normal thymic epithelial ontogeny (see Section VI below). Our results indicated that thymic epithelium is not generally infected with HTLV. These observations may simply reflect a cross-reactivity between HTLV and thymic epithelial antigens, or suggest that HTLV p19 is a host-encoded protein selectively expressed in normal thymus, induced in HTLV-infected T cells, and incorporated into the virus. The only normal human tissue in addition to thymus reactive with anti-pl9 was the basal layer of squamous epithelium (Haynes et al., 1983e). Whatever the physiologic relevance of p19 antigen expression by human endocrine thymic epithelium, the p19 antigen has provided a unique marker for stages of human thymic epithelial maturation. It is important to note that during HTLV infection in adult T cell leukemia patients, high titers of serum antibodies are made against the same p19 molecule and against the same antigenic determinant with which monoclonal anti-pl9 antibody reacts (Robert-Guroff et al., 1982~). Thus, as in pemphigus, GVHD, and Type I diabetes mellitus, antibodies are made during HTLV infection which potentially could react in vivo with a surface antigen of human thymic epithelium. 4 . BB Wistar Rat Autoantibody BB TECS Buse et al. (1983) fused splenocytes from a diabetes-prone BB rat with P3X63/Ag8 mouse myeloma cells, and produced a rat-mouse


107

hybridoma secreting an IgG, myeloma protein, BB TECS. Characterization of BB TECS reactivity demonstrated that, like anti-pl9, TE-4, and A2B5, BB TECS identified the thymosin a1 containing, AE-1+ (antikeratin), epithelial cells in the thymic subcapsular cortex and medulla (Fig. 4B) (Buse et al., 1983, 1984). Unlike antibodies antip19, TE-4, and A2B5, antibody BB TECS also reacted with all cells within Hassall’s bodies (Fig. 4B). Similar to TE-4 and anti-pl9, BB TECS reacted with keratinocytes in squamous epithelium (see Section V,F below).

5. Monoclonal Antibodies That Selectively Bind to Hassall’s Bodies Monoclonal antibodies TE-8, TE-15, TE-16, and TE-19 were made by fusion of murine myeloma cells with murine spleen cells immunized with human thymic stroma (Haynes, 1984; Lobach and Haynes, 1984). While these antibodies all reacted with Hassall’s bodies, their reactivity patterns were not identical. TE-8, TE-16, and TE-19 reacted with all cells within thymic Hassall’s bodies, while TE-15 reacted with granular material within Hassall’s body swirls (Fig. 4C and D). In addition to reacting with thymic Hassall’s bodies, TE-19 reacted with a carbohydrate antigen on group A human erythrocytes (Telen Haynes, unpublished observations). While TE-8, TE-16, and TE-15 were not expressed on rodent thymus, TE-19 was expressed on rat thymus (M. Telen, E. Harden, and B. Haynes, unpublished observations). Since rat thymus contains few if any Hassall’s bodies, it is interesting to note that TE-19 stained scattered foci of rat thymic medullary cells. 6 . Patterns of Extrathymic Reactivity of Antithymic Epithelial Antibodies ( B B TECS, Anti-HTLV p19, A2B5, TE-4, TE-8, TE-16, and TE-15) The extrathymic reactivity patterns of thymic epithelial monoclonal antibodies are summarized in Table I. It is particularly interesting to compare the reactivity of these reagents in thymus and skin because of the known differentiation pathway of squamous epithelium keratinocytes from basal to the outer layer of stratified squamous epithelium (Doran et al., 1980; Watt and Green, 1982). Moreover, histologic similarities between skin and thymic epithelium have been described (Gaudecker and Schmale, 1974; Mandell, 1968a,b). In postnatal thymus, A2B5, TE-4, and (regarding postnatal thymuses greater than 36 months) anti-HTLV p19, all have the same reactivity pattern and define the thymosin a l + , keratin+, endocrine subcapsular cortical, and

108

BARTON F. HAYNES

medullary thymic epithelium. This pattern is represented in Fig. 4A by TE-4 reactivity. Importantly, the central swirl of Hassall’s bodies is TE-4-, A2B5-, and anti-pl9-. BB TECS, like TE-4, A2B5, and antip19, defines the endocrine subcapsular cortical and medullary thymic epithelial component, but also reacts with all cells within Hassall’s body (Fig. 4B). As mentioned TE-8 and TE-16 react with the entire Hassall’s body (Fig. 4C) while TE-15 reacts with linear and granular material within some Hassall’s bodies (Fig. 4D). In skin, antibodies TE-4, A2B5, and anti-pl9 react only with the basal layer of keratinocytes (represented by TE-4 reactivity in Fig. 5A). Antibody BB TECS, which reacts with both thymic epithelium and Hassall’s bodies, reacts with both the basal layer and the stratum granulosum of skin, while the stratum spinosum and the stratum corneum (SC) are BB TECS nonreactive (Fig. 5B). Antibodies TE-8 and TE-16, which react with Hassall’s bodies, react in squamous epithelium with only the stratum granulosum (represented by TE-8 reactivity in Fig. 5C). Finally, antibody TE-15, which selectively reacts with granular material inside Hassall’s bodies, reacts only with the stratum corneum (SC) in skin (Fig. 5D). Regarding shared expression of TE-4, p19, A2B5, and BB TECS antigens of thymic epithelial cells with basal squamous epithelium, skin, and tonsil epithelium have both been implicated in promoting extrathymic T cell maturation or activation (Rubenfeld et al., 1981; Osteroom et al., 1981; Sauder et al., 1982). Further, basal keratinocytes in skin have recently been shown to contain both thymopoietinlike and thymulin-like thymotrophic hormones (Kato et al., 1981;Chu et al., 1983). In the tonsil, A2B5+,TE-4+ epithelium forms a dendritic network of cells at the base of tonsillar crypts. Therefore, the presence of TE-4, A2B5, and p19 antigens on both thymic epithelium and basal keratinocytes of squamous epithelium might signify common T cell trophic functions, and identify human extrathymic as well as intrathymic T cell inductive environments. Skin basal keratinocytes, or stem cells within the basal cell layer, give rise to each succeeding superficial layer of keratinocytes (Watt and Green, 1980; reviewed in Lavker and Sun, 1983). Using our panel of antithymic epithelial cell reagents, we have defined antigenic changes which occur during squamous epithelial keratinocyte maturation (Fig. 5). As basal cells mature into the stratum spinosum cell layer, TE-4, BB TECS, A2B5, and p19 antigens are lost; as the stratum spinosum cell layer matures to the stratum granulosum layer, the antigens defined by antibodies TE-8, TE-16 are expressed, and the antigen defined by BB TECS is reexpressed. As keratinocytes in the stra-


109

FIG.5. Reactivity patterns of antithymic epithelial antibodies in normal skin. (A) Antibody TE-4 reacted with keratinocytes in the basal layer (arrow). Some suprabasal keratinocytes were weakly TE-4 reactive while cells within stratum granulosum and stratum corneum (SC) were nonreactive with TE-4. (B) Antibody BB TECS reacted with the basal layer and stratum granulosum (arrows) of skin while suprabasilar keratinocytes and stratum corneum (SC) were nonreactive. (C) Antibody TE-8 selectively reacted in skin with the stratum granulosum (arrow). (D) Antibody TE-15 selectively reacted in skin with only the stratum corneum (SC). Arrow identifies the epidermal-dermal junction ( ~ 4 0 0 ) .

tum granulosum layer become the stratum corneum, TE-8, TE-16, and BB TECS antigens are lost and the TE-15 antigen is expressed. These changes and the antigenic patterns of skin and thymic epithelium are summarized in Fig. 6. Given that human thymic epithelial cells and skin keratinocytes both have near identical growth requirements in vitro (K. Singer, E. Harden, and B. Haynes, unpublished observa-

110

BARTON F. HAYNES

FIG.6. Schematic summary of reactivity of antithymic epithelial antibodies in thymus (A) and skin (B). Areas in skin and thymus that coexpress antigens were identically coded. Thus, TE-7+ areas in thymus (see Section V,B) were vessels, capsule, and interlobular septae; in skin TE-7 reacted with dermal fibrous connective tissue. TE-4, A2B5, p19, and BB TECS reacted with subcapsular cortex and medulla thymic epithelium, and with basal layer keratinocytes in skin. BB TECS also reacted with Hassall’s bodies in thymus and the stratum granulosum in skin. Antibody TE-3 (see Section V,C) in thymus reacted with cortical thymic epithelial cells but with no skin epithelial keratinocytes. Antibody TE-8 reacted with the concentric layer of cells within thymic Hassall’s body swirls, and in skin, selectively reacts with cells with the stratum granulosum. Antibody TE-15 reacted with material within cells inside Hassall’s bodies, and in skin selectively reacted with the stratum corneum.

tions), contain keratin, and share similar surface antigens (Figs. 4, 5, and 6), it is reasonable to propose a maturation pathway for human thymic epithelial cells similar to epidermal keratinocytes. Indeed, Mandell (1968a,b, 1970)and Gaudecker and Schmale (1974) have suggested from morphological comparisons of skin and Hassall’s bodies that medullary thymic epithelial cells have a potential for differentiation similar to that of skin keratinocytes.


111

TE-4+ subcapsular cortical and medullary thymic epithelial cells most likely give rise to Hassall’s bodies, losing TE-4, A2B5, and p19 antigens but continuing to express BB TECS antigen. During thymic epithelial maturation that leads to Hassall’s body formation, antigens TE-8, TE-16, and TE-19 are acquired. Finally within the concentric epithelial Hassall’s body swirl, antigen TE-15 is expressed. Therefore, one can envision a dynamic maturation process of thymic epithelium from TE-4+, TE-8-, TE-16- epithelial cells to TE-4-, TE-8+, TE-16+ Hassall’s bodies. Itoh et al. (1982) have presented evidence that cultured rodent endocrine thymic epithelial cells (IT-26 R21) form Hassall’s bodies in uitro. We have found these differentiated rat epithelial cells to be TE-19+ (E. Harden and B. Haynes, unpublished observations). Similar maturation of human thymic epithelium in vitro has been reported (Pyke and Gelfand, 1974). Acquired loss of Hassall’s bodies in chronic disease states or such as is found in the acquired immunodeficiency syndrome (AIDS) (Elie et al., 1983) could result from interruption of thymic epithelial cell maturation (Haynes, 1984). As has been alluded to in the discussion of cross reactive antigens of group A streptococci (Lyampert et al., 1976) and HTLV (Haynes e t al., 1983e) with endocrine thymic epithelium, one potential mechanism of interruption of thymic epithelial maturation could be formation of antithymic epithelial antibodies. It is interesting that some patients with AIDS are infected with HTLV (Gallo et al., 1983; Essex et al., 1983). From data in the literature and from comparison of reactivity of monoclonal reagents with thymus and skin, it appears that endocrine thymic epithelium, at least in the thymic medulla, gives rise to Hassall’s bodies. Moreover, it is likely that in the pre- and postnatal human thymus, epithelial cells undergo a continuous maturation process. Therefore, a potentially important mechanism of disruption of T cell maturation could be either intrathymically or extrathymically at the site of epithelial-T cell contact via disruption of epithelial cell maturation. B. MESODERMAL-DERIVED THYMIC STROMA (TE-7) Antibody TE-7 is the IgGl product of a lymphocyte hybridoma produced by a fusion between 653 murine myeloma cells and BALB/c spleen cells from an animal immunized with human thymic stroma (Haynes et al., 1984). In human thymus, antibody TE-7 reacted with all structures within the interlobular septae, with the thymic capsule, and with coarse bands of stroma penetrating into fhe cortex and medulla (Fig. 7).

112

BARTON F. HAYNES

FIG.7. Reactivity of TE-7 with normal thymus. Antibody TE-7 did not react with epithelium in the subcapsular cortex (S), cortex (C) or medulla, but rather reacted with thymic capsule (arrow) and the fibrous tissue and vessels in interlobular septae (ILS) (x400).

In double fluorescence studies using directly fluoresceinated TE-7 with antibody TE-4, the TE-4+ and TE-7+ stromal elements were found to be mutually exclusive components of the thymic microenvironment (Haynes et al., 1984) (Fig. 8A and B). It was observed, however, that TE-4+ and TE-7+ cells were frequently in close contact (small open arrows, Fig. 8A and B). TE-7 reactive thymic stroma was uniformly thymosin al-, keratin-, A2B5-, TE-4-, and p19-. Tissue screen of reactivity of TE-7 antibody demonstrated that it reacted with fibrous connective tissue of virtually all tissues tested and was not

FIG.8. Direct comparison of reactivity of directly fluoresceinated TE-7 and indirectly rhodaminated TE-4. (A) Shows TE-4' dendritic cells in normal 3-month-old thymic medulla (solid small arrows), and TE-4- adjacent stroma (large arrow). (B) Shows TE-7+ adjacent stroma (large arrow) with dendritic TE-4+, TE-7- cells (solid small arrows). Although TE-4+ and TE-7+ cells were mutually exclusive, it was seen that TE-4+epithelial cells and TE-7+ stromal cells were frequently in close physical contact (small open arrows-A and B) ( X 4 0 0 ) (Haynes et al., 1984).

114

BARTON F. HAYNES

thymus specific. Of malignant tissues tested, TE-7 reacted only with malignancies of fibrous tissue, e.g., fibrosarcoma (Haynes et al., 1984). While antibody TE-7 was specific in the adult for fibroblasts and vessels in all organs tested, it reacted with cartilage, vessels, and mesenchymal interstitial cells in early fetal tissues. Notocord, neural crest ectoderm, spinal cord, and all endodermal tissues were TE-7-. Thus, TE-7 identified cells of mesodermal origin in fetal and adult tissues (Haynes et al., 1984).

THYMIC EPITHELIUM (TE-3) C. CORTICAL Whereas TE-4, A2B5, anti-pl9, and BB TECS delineated keratinized endocrine subcapsular cortical and medullary thymic epithelium, keratinized cortical epithelial cells are present that are nonreactive with antibodies TE-4, A2B5, p19, and BB TECS (Haynes et al., 1984; McFarland et al., 1984). The staining pattern of antithymosin al, and antithymopoietin antisera (i.e., nonreactivity with the majority of cortical epithelial cells) suggested that cortical epithelial cells did not contain these thymotrophic hormones (Haynes et al., 1983b,c). To investigate cortical epithelium further, we have developed a monoclonal antibody, TE-3, that selectively reacts with cortical thymic epithelium (McFarland and Haynes, 1984) (Fig. 9). TE-3 is an IgGg murine monoclonal antibody raised against human thymic stroma. In thymus, the location of reactivity of TE-3 corresponded closely with the anatomical thymic cortex (dotted line, Fig. 9). In addition to cortical thymic epithelium, thymic macrophages and large dendritic cells in the medulla reacted with TE-3 (arrows, Fig. 9). Using double staining fluorescent assays with directly fluoresceinated TE-3 and antibodies A2B5, AE-1 (antikeratin), TE-7, and rabbit antithymosin al, we found that TE-3-reactive cortical epithelium was TE-7-, contained keratin, did not contain thymosin al, and generally was A2B5-. Similarly, all thymic medullary epithelium was TE-3-, A2B5+, contained keratin and thymosin al. In the subcapsular cortex, however, a subset of epithelium was both TE-3+, A2B5+, TE-4+, and as well contained keratin and thymosin al. As expected, TE-3+ thymic macrophages contained neither keratin nor thymosin a l , and were nonreactive with antibodies TE-4, A2B5, and TE-7 (McFarland et al., 1984). Extrathymically TE-3 reacted with tissues derived from a variety of embryonic origins. Thyroid, pancreas, anterior pituitary cells, and brain neurons were all reactive with antibody TE-3, while skin epidermal keratinocytes and adrenal gland were not reactive. Whereas BB TECS, A2B5, p19, and TE-7 all reacted with cell surface mem-


115

FIG.9. Reactivity of TE-3 with normal thymus. Antibody TE-3 selectively reacted with cortical epithelium (cortex) and not epithelium in medulla. Hassall’s bodies (H) were TE-3 nonreactive. Dotted line shows cortical-medullaryjunction. Large dendritic, TE-4-, keratin-, thymosin al- cells were frequently seen in the thymic medulla (arrow) (X400).

brane antigens, antibody TE-3 reacted with a cytoplasmic antigen. Therefore, in the thymus, antibody TE-3 defined the TE-4, A2B5, BB TECS, and anti-pl9-nonreactive epithelial component of the inner cortex. Since antibody TE-3 reacts with thymic macrophages and a number of extrathymic tissues, it is not a thymic epithelial-specific differentiation antigen. Rather, TE-3 more likely defines an intracellular molecule that reflects a specialized function of cortical thymic epithelium. It is known from tritiated thymidine labeling studies of thymic explants that subcapsular cortical thymocytes constitute the

116

BARTON F. HAYNES

most rapidly dividing pool of thymocytes (Clark, 1973). In contrast, the majority of cortical thymocytes are not dividing; rather the great majority of thymic lymphocytes are thought to die intrathymically. Whether the termination of thymocyte proliferation once subcapsular cortical thymocytes enter the inner TE-3+ cortical epithelial zone is a programmed event as part of thymocyte differentiation or is a function of the cortical microenvironment is not known. Interestingly, factors which inhibit lymphocyte DNA synthesis (a-globulins) have been isolated from bovine thymus (Carpenter et al., 1971). D. SUBCAPSULAR CORTICAL THYMIC EPITHELIUM As mentioned (Section IV,A above) the subcapsular cortex differs from the medulla in that antithymosin p3 (Hirokawa et aZ., 1982) and antithymosin p4 (Haynes et al., 1983c)antisera react with subcapsular cortical but not medullary epithelium. Ritter and colleagues (1981) demonstrated that a monoclonal antibody against the human homolog of the murine Thy-1 molecule (McKenzie and Fabre, 1981) reacted in human thymus with the subcapsular cortical thymic epithelium. Wekerle and colleagues (1980) described a specialized cell type of murine thymic epithelium, the thymic nurse cell (TNC), and demonstrated that this cell expressed H-2K, D, and I region determinants. Moreover, the TNC is an unusually large keratinized cell (Vakharia, 1983), and routinely contains 20 to 40 small thymocytes, many of which are dividing (Wekerle et al., 1980). In addition to expressing both polymorphic and nonpolymorphic determinants of the HLA A, B, and C molecule, human TNC are Ia+ but do not express T cell antigens (Ritter et al., 1981). Using Thy-1 monoclonal antibody Ritter et al. (1981) localized human TNC in vivo to the outer thymic cortex. Vakhaira et al. (1983) have suggested that cortical thymocytes within TNC are viable and functionally mature. It has been postulated that the TNC plays a pivotal role in T cell maturation by mediating direct stromal-induced thymocyte maturation (Wekerle et al., 1980; Vakharia, 1983). It is interesting that Thy-1 epithelial cells in the subcapsular cortex are in the area to which thymocyte precursors first migrate and proliferate after arrival in the thymus. Thy-l+ thymic epithelium may play an important role in the early stages of thymocyte maturation via cell to cell contact and/or local thymic hormone production (Clark, 1973; Bach and Papiernik, 1981). What role the Thy-1 molecule might be playing in thymocyte maturation is not known, although sequence homology of this molecule with regions of the immunoglobulin molecule has been demonstrated (Cohen et al., 1981).Thy-1 antigen in rats


117

has been postulated to be involved in cell-cell/cell-connective tissue adhesion (Ritter and Morris, 1980; Morris and Ritter, 1980). E. ANTIBODIESAGAINST CLASS 1 (ANTI-HLA,A, B, AND c)AND CLASS 11 (ANTI-Ia-LIKE OR Ia) ANTIGENS Both Class I and Class I1 MHC antigens have been shown to be expressed on epithelial cells of human and rodent thymic microenvironment (Janossy et al., 1980; Bhan et al., 1980; Jenkinson et al., 1981). In general, studies in humans have shown that subcapsular cortical, cortical, and medullary epithelial cells express both HLA and Ia (Janossy et al., 1980; Bhan et al., 1980). To determine human MHC antigen expression by thymic microenvironment components defined by antibodies TE-4 or TE-7, either directly fluoresceinated anti-HLA A, B, and C monoclonal antibody (3F10) (Haynes et al., 1982) or anti-Ia monoclonal antibody (L-243) (Lampson and Levy, 1980) was sequentially incubated on thymus sections which had been previously labeled with either TE-4 or TE-7 antibodies and rhodaminated goat anti-mouse immunoglobulin (Haynes et al., 1984). In these double-labeling experiments, we found TE-4+ thymic epithelium strongly expressed HLA and Ia antigen (Fig. 10A-D). While more cells were HLA+ and Ia+ than were TE-4+, all TE-4+ cells were clearly HLA+ and Ia+. In contrast, TE-7+ thymic stroma was uniformly Ia-. Areas of TE-7+ stroma were also HLA unreactive, while foci of TE-7+ thymic stroma reacted weakly with antiHLA antibody (Fig. 10E-H). In similar double fluorescent assays, TE-3+ cortical epithelium was also strongly reactive with anti-HLA and anti-Ia monoclonal antibodies. Table I1 summarizes the phenotype of TE-4+, TE-3+, and TE-7+ human thymic microenvironment components (Haynes et al., 1984; McFarland and Haynes, 1984).

F. OTHERMONOCLONAL REAGENTSREACTIVE WITH HUMAN THYMIC EPITHELIUM Naito et al. (1983) have recently described monoclonal antibody 3-40, an IgM, rnurine antibody produced by a hybrid cell line from NS-1 myeloma cells and murine spleen cells immunized with MOLT-4 human leukemic T cells. Regarding normal and malignant hematopoietic cells, antibody 3-40 has the interesting property of reacting with only T cell acute lymphoblastic leukemia cells. The 3-40 antigen is a protein of 35,000 to 40,000 daltons. Naito and colleagues (1983) noted in frozen tissue sections that antibody 3-40 reacted with normal thymic epithelial cells and Hassall’s bodies, as well as with colon, ovarian, and melanoma tumors. We have further characterized the

FIG.10. Expression of MHC antigens by endocrine thymic epithelium (TE-4+)and thymic stroma (TE-7+). Either directly fluoresceinated anti-HLA (3F10) or anti-Ia (L-243) was sequentially incubated on normal 3-month-old thymus sections previously labeled with either TE-4 or TE-7 and rhodaminated goat anti-mouse Ig. TE-4+ epithelium (A and C, arrows) strongly expressed HLA antigens (B, arrows) was either weakly HLA+ or HLA- (F,arrow) while all TE-7+ stroma was Ia- (H, arrow) (X400) (Haynes et al., 1984).

119


TABLE I1 CHARACTERISTICS OF TE-4+, TE-7+, AND TE-3+ COMPONENTS OF HUMAN THYMIC MICROENVIRONMENT TE-4 Localized in subcapsular cortex (SCC) Localized in cortex Localized in medulla Localized in thymic capsule, interlobular septae Contained thymosin a1 Reactivity with Anti-HLA (3F10) Anti-Ia (L-243) A2BS Anti-pl9 Antikeratin (AE-1) TE-4 TE-3 TE-7

TE-7

TE-3

Yes No" Yes No

No No No Yes

Yes Yes No No

Yes

No

SCC only

Positive

Weakly positive to negative Negative Negative Negative Negative Negative Negative Positive

Positive

Positive Positive Positive Positive Positive SCC only Negative

Positive SCC only SCC only Positive SCC only Positive Negative

While the majority of cortical epithelial areas are TE-4-, cells can be seen in cortical areas.

reactivity of antibody 3-40 (Haynes and Dupont, 1984). In thymus, 3-40 reacted with the TE-4+,thymosin al-containing endocrine thymic epithelium. However, unlike A2B5, TE-4, and anti-pl9, 3-40 reacted with Hassall's bodies, portions of mesodermal TE-7+ thymic stroma, and the endothelium of thymic vessels. Like A2B5,3-40 reacted with pancreatic islet cells, anterior pituitary, and brain. Therefore, the 3-40 antigen, like HTLV p19 (Haynes et al., 1983e), is an example of a T cell tumor-associated antigen that shares an antigenic determinant expressed on endocrine thymic epithelium. Finally, Janossy and colleagues (1983) have described monoclonal antibody RFD-4 which has a near identical reactivity pattern in thymus as antibody A2B5. VI. Ontogeny of the Human Thymus Microenvironment

Data suggest that at least three separate tissues come together during fetal development to form the thymic microenvironment. Mesodermal cells give rise to fibroblasts and other connective tissue cells. Thymic epithelial cells have been postulated to be derived from brachial cleft ectoderm and third pharyngeal pouch endoderm (Weller, 1933; Norris, 1938). Moreover, neural crest ectoderm has also

120

BARTON F. HAYNES

been suggested to contribute to the thymic microenvironment (LeDouarin and Jotereau, 1975). In both mouse (Cordier and Haumont, 1980) and man (Weller, 1933; Norris, 1938), it has been shown that the thymus develops primarily from the third endodermal pharyngeal pouch and the third ectodermal cleft. In man (4-5 weeks) and mouse (day l l ) , proliferation of the ectoderm of the third brachial cleft covers the endodermally derived epithelium (Weller, 1933; Norris, 1938; Auerbach, 1960, 1961; Cordier and Haumont, 1980; Patten, 1968). The endodermal thymic epithelial component appears to chemically attract immigrant bloodborne stem cells (Auerbach, 1960; LeDouarin and Jotereau, 1981) which, in man, first appear in the thymus at weeks 8-9 (Papiernik, 1970). Thus, the epithelial component of mature (neonatal) normal thymic microenvironment is thought to be composed of a central endodermal region, and a peripheral ectodermal region (Norris, 1938); cortical epithelial cells are thought to be ectodermal in origin while medullary epithelial cells are endodermal in origin (Norris, 1938; Cordier and Haumont, 1980). Auerbach (1960, 1961) has demonstrated that the mesodermal component of the thymus microenvironment (fibroblasts and vessels) surrounds the thymic primordia in the human at 6-7 weeks gestation. Thus, from the characterization of the reactivity of TE-4, TE-7, A2B5, and anti-pl9 monoclonal antibodies, we expected TE-7 to identify the outer mesenchymal connective tissue of the preseptated thymic primordia (7-9 weeks gestation), and to identify intralobular septa1 tissue, vessels, and thymic capsule of lobulated fetal thymus (greater than 10-12 weeks gestation). In contrast, TE-4, BB TECS, p19, and A2B5 antigens, if conserved during ontogeny, should be present on epithelial components of thymic epithelium. An ontogeny study of reactivity of antibodies A2B5, p19, TE-4, and TE-7 was performed on 7 fetal and 5 infant thymuses of the following ages: fetal-7,9, 10, 13, 15, 16, and 24 weeks gestation; postnatal-3, 6,12,36, and 48.5 months (Haynes et al., 1984).At 7 weeks, the thymic rudiments were still paired tubular structures approximately 160 pm in length ventral to the carotid artery, which had not yet fully migrated caudally and fused at the midline (Weller, 1933; Norris, 1938; Patten, 1968). Thymic epithelial rudiments until 9 to 11 weeks were devoid of lymphoid cells entirely (Patten, 1968; Papiernik, 1970). Figure 11A demonstrates a high power ( ~ 4 0 0view ) of the 7 week thymic rudiment. In Fig. 11B the A2B5+ thymic epithelium is depicted (arrows), while Fig. 11C shows TE-4+ thymic epithelium. Figure 11D shows the outer rim of TE-7+ mesodermally derived connective tissue (arrows) surrounding TE-7- epithelial thymus. The entire epithelial

FIG.11. Phenotypic characterization of human 7 week fetal thymus. (A) High power view ( ~ 2 0 0of) 7 week thymus (hematoxylin and eosin); (B) demonstrates the entire rudiment stained with A2B5 in a granular pattern; (C)demonstrates that a central portion of the thymic rudiment was TE-4+; and (D) demonstrates that while the entire epithelial component of the thymic rudiment was TE-7-, TE-7+ thymic skoma surrounded the thymus. In each picture (A-D) arrows demarcate the thymic rudiment (x200) (Haynes et al., 1984).

122

BARTON F. HAYNES

FIG.12. Human 12 week fetal thymus. (A) Shows TE-4+ epithelium in a dendritic pattern (arrows) in the 12 week fetal thymus which is now populated with E-rosette+, 3A1+ thymocytes (T cell markers not shown). (B) Demonstrated TE-7+fibrous capsule, vessels (V), and interlobular septum (arrows) in the lobulated 12 week fetal thymus (x400) (Haynes et al., 1984).

thymus at 7 weeks was p19- and portions of the central epithelium were BB TECS+. From 9 to 13 weeks TE-4+, A2B5+ epithelium was arranged in lobulated zones surrounded by TE-7+ areas with no discernible cortical-medullary junction (Fig. 12A and B); no Hassall’s bodies were seen. After 9 weeks gestational age, numerous thymocytes were present that expressed the T cell antigens 35.1 (E-rosette receptor) and 3A1 (Haynes et al., 1980; Haynes, 1981). By 15 weeks a cortical-medullary junction was present and the TE-4+ component had compartmentalized into the subcapsular cortical and medullary areas. Furthermore, at 15 weeks Hassall’s bodies were present as seen


123

FIG.12B.

in normal infant thymuses. Anti-pl9 defined an antigen on endocrine (TE-4+)epithelium that was acquired during ontogeny, first appearing in the subcapsular cortex region at between 12 and 15 weeks fetal gestation (Table 111)(Haynes et al., 1983e). Figure 13 summarizes the ontogeny of reactivity of A2B5, BB TECS, TE-4, anti-pl9, and TE-7 antibodies with human thymic microenvironment components beginning at 10 weeks of fetal gestation. Whether the TE-4+ population of subcapsular cortical and medullary epithelial cells constitutes the endodermal or the ectodermal component of thymic epithelium at present remains speculative. Our

124

BARTON F. HAYNES

TABLE 111 REACTIVITY OF ANTI-^^^ ANTIBODY WITH NORMAL THYMUS TISSUE AS DETERMINED BY INDIRECT IMMUNOFLUORESCENCE~ Thymus number

Age

Pattern of reactivity of anti-pl9 antibody

1 2 3

7 weeks gestation 12 weeks gestation 15 weeks gestation

4

24 weeks gestation

All areas of thymic epithelium were nonreactive All areas of thymic epithelium were nonreactive Most cortical and medullary areas were nonreactive. Occasional linear arrays of p19+ dendritic cells were located in the subcapsular cortical region of rate thymic lobules Most cortical and medullary areas were nonreactive. Occasional linear arrays of p19+ dendritic cells were located in the subcapsular cortical region of rare thymic lobules All subcapsular cortical areas of each lobule rimmed by linear arrays of p19+ dendritic cells. Occasional p19+ cells were scattered about upper cortex. Most medullary areas contained no p19+ cells All subcapsular cortical areas of each lobule were rimmed by linear arrays of p19+ dendritic cells. Occasional p19+ cells were scattered about upper cortex. Only occasional to few medullary areas contained p19+ cells All subcapsular cortical areas of each lobule were rimmed by linear arrays of p19+ dendritic cells. Occasional p19+ cells were scattered about upper cortex. Most medullary areas contained many p19+ cells Streaks of p19+ epithelium were present throughout thymic fibrofatty stromal remnant

5

4 months

(postnatal)

12 months

8

30 months

9

60 years

From Haynes et al. (1983e).

ontogeny study did not include embryos prior to 7 weeks gestation necessary to resolve this question. The most likely origin for TE-4+ epithelium, based on its physical location in the thymus (medulla and subcapsular cortex) is pharyngeal pouch endodenn (Norris, 1938). However, the shared expression of TE-4 antigen by thymic epithelial cells with cells of known ectodermal origin (basal layer of squamous epithelium of conjunctiva, esophagus, skin, and tonsil) suggests possible ectodermal origin of TE-4+ thymic epithelium. In thymus, antibody TE-7 identified the mesodermal-derived fibrous septae and vessels. Thus, the results of our ontogeny study agreed with light microscopic descriptions of the relationship of thy-

125

HUMAN THYMIC MICROENVIRONMENT Thymic

Antigen

Location

-

Fetal Gestational Age in Weeks

10 I

12

15

16

24

Postnatal Age in Months

36

3

12

24

36

SUBCAPSULAR CORTEX

A285

SUBCAPSULAR CORTEX

TE-4

SUBCAPSULAR CORTEX

BB TECS

~19

TE - 7

WBCAPSULARCORTEX

w

H

NON-NEUROENMXRINE STROMA

FIG.13. Acquisition of human thymic epithelial antigens during ontogeny. Whereas

A2B5, TE-4, BB TECS, and TE-7 antigens were preserved throughout ontogeny, the p19 antigen was selectively acquired during ontogeny (adapted from Haynes, 1984).

mic stroma to thymic epithelial rudiments from 7 weeks gestation through birth. In the quail-chick hybrid system, LeDouarin and Jotereau (1975) demonstrated mesenchyme from neural crest mesectoderm gave rise to thymic connective tissue and vessels. On the other hand, it has been suggested that the neural crest contribution to the thymus forms the endocrine epithelial component, based on the known common function of neuroendocrine secretory cells (Kissel et al., 1981), and on the common expression of the complex ganglioside antigen defined by monoclonal antibody A2B5 on thymic epithelium and neural and neural crest-derived human and rodent tissues (Eisenbarth et al., 1979, 1981, 1982; Haynes et al., 1983b). We have shown all TE-4+ cells that contained thymosin a1 and keratin were A2B5+. Conversely, all TE-7+ mesenchyma-derived connective tissue cells were A2B5-. Thus, the neural crest mesectodermal component of the thymic microenvironment proposed by LeDouarin and Jotereau most likely corre-

4a

126

BARTON F. HAYNES

sponds conceptually to the TE-7+,TE-4- component of human thymic stroma. In the adult thymus, the TE-7+ component of thymic stroma was located in the interlobular septae and throughout the cortex and medulla as fibrous bands of Ia negative and HLA weakly positive connective tissue. The importance of TE-7 as an antigenic marker of the mesodermal-derived component of thymic stroma is brought to light by the observations of Auerbach (1960, 1961) and others (LeDouarin and Jotereau, 1975) that mesodermal stroma may play an important role in inducing murine thymic epithelial differentiation and fetal thymic lobulation at the stage of fetal development just prior to lymphoid cell population of the thymic rudiment. VII. The Human Thymic Microenvironment in Diseases of Abnormal T Cell Maturation

Using the reagents described in this review as investigative probes, we have begun to study the thymic microenvironment in disorders of abnormal T cell maturation. A. SEVERE CELLULAR IMMUNODEFICIENCY DISEASES Severe cellular immunodeficiency diseases comprise a spectrum of heterogeneous disorders characterized by profound T and B lymphocyte dysfunction, and are complicated by bacterial and viral infections, usually leading to death within the first 2 years of life (Buckley, 1982). Severe combined immunodeficiency disease (SCID) and cellular immunodeficiency with normal serum immunoglobulins (Nezelof syndrome) are two clinical syndromes within this spectrum of diseases (Buckley, 1982; Dosch et al., 1978; Lawlor et al., 1974). In these conditions the thymus is frequently dysplastic; histopathologic study has usually revealed a small thymic rudiment, with scanty lymphoid cells and islands of thymic epithelium scattered throughout a fibrofatty stroma (Borzy et al., 1979). We have studied the thymic rudiments from five children with congenital severe cellular immunodeficiency diseases. Two children had thymic dysplasia with normal immunoglobulins (Nezelof syndrome) and three children had severe combined immunodeficiency disease (SCID). A number of abnormalities of thymic epithelium were found (Haynes et al., 1983c) (Table IV). First, the thymic epithelium was not divided into subcapsular cortex, cortex, and medullary regions, and no Hassall’s bodies were present (Fig. 14A) (Haynes et al., 1983~). Thus, in patients 1, 2, 3, and 5, TE-4, A2B5, and BB TECS identified the

TABLE IV CHARACTERIZATION OF THYMIC EPITHELIUM IN SEVERE CELLULAR IMMUNODEFICIENCY DISEASES" Patient

Diagnosis

A2B5

TT

Anti-pl9

1

SCIDd SCID SCID Nezelofsyndrome Nezelofsyndrome

+ + +

-

+/+/-

2 3 4

5 a

-

+

ND'

+

ND

+, All thymic epithelial cells reactive;

+

-

+/-

TE4

BB TECS

+ + + +

+ + + +

TE-8b TE-lSb +/+/-

-

ND

-

+I+/-

+/-

-

TE-7'

Thymosin a1

Thymosin p4

+ + ND + +

+ + ND +

+ + ND +

ND

ND

-, all thymic epithelial cells nonreactive; +I-, subset of thymic epithelial cells reactive.

'Although epithelial cells were present in some thymus rudiments that were TE-8 and TE-15 reactive, no formed Hassall's bodies were present. TE-7 in all dysplastic thymus was reactive with connective tissue that unlike neonatal thymus, did not comprise interlobular septae and capsule but rather was diffusely intermingled with the neuroendocrine thymic epithelial component. In patients 1,2, and 4, thymic rudiments were tested for reactivity with anti-Ia (L-243) and anti-HLA (3F10)and found to be strongly positive. SCID, Severe combined immunodeficiency disease. ND, not done.

128

FIG.14. Reactivity of thymic rudiments from patients with severe cellular immunodeficiency diseases with antithymic epithelial mono) SCID thymus (hematoxylin and eosin stain). Arrows show epithelial clonal antibodies. (A) Low power light microscopic view ( ~ 1 0 0of rudiment devoid of lymphocytes. Antibody BB TECS (B) reacted with entire epithelial component of SCID thymus (arrows) (x400). Anti-pl9 (C) reacted with only a subset of SCID thymic epithelial cells. Open arrows in C show p19 reactive epithelium and closed arrows show p19 nonreactive epithelium. Granular fluorescent areas represent autofluorescent pigment. (D) Antibody TE-15 reacted with epithelial cells arranged concentrically in thymic rudiment from a patient with Nezelof syndrome (arrows).

130

BARTON F. HAYNES

entire epithelial component of the thymus (Fig. 14B). As expected from the age of these children (all less than 3 years of age) only a subset of thymic epithelium was p19 reactive (Fig. 14C). Whereas patients 1 and 2 (SCID) were A2B5+,unlike normals, their thymuses did not bind TT. Conversely, patient 4 (Nezelof's) did not react with A2B5 but bound TT (Haynes et aZ., 1983~).The A2B5-, TT+ phenotype is not specific for Nezelof syndrome, since thymus from a second patient with Nezelof syndrome (patient 5) reacted with A2B5. Evidence that some thymic epithelial maturation was occurring in dysplastic thymuses comes from the observation that four thymuses contained scattered TE-8+ epithelial cells, and two thymuses contained foci of TE-15+ cells in aggregates suggestive of early Hassall's body formation (Table IV) (Haynes et al., 1983a) (Fig. 14D). Finally, the spatial relationship of TE-7 thymic stroma to TE-4+ endocrine epithelium was abnormal. Unlike normals, in the four dysplastic thymuses tested, TE-7 stroma was completely intermixed with the TE-4+, A2B5+ epithelium (Haynes et aZ., 1983a). In contrast to SCID thymus rudiments which were essentially devoid of T cells, Nezelof thymus contained numerous (though fewer than normal) lymphocytes with mature T cell surface antigens By indirect immunofluorescence, thymosin a l , (Haynes et al., 1983~). thymopoietin, and thymosin p4 thymic hormones were all present in thymic epithelial cells of all three patients tested (Table IV). By radioimmunoassay, thymosin a1 levels per gram of tissue were markedly decreased in SCID thymuses, whereas thymosin a1 was present at approximately one-half normal level in the Nezelof thymus (patient 4) (Haynes et al., 1983~). The clinical and immunologic heterogeneity of severe cellular immunodeficiency disease has long been appreciated. The discovery that some cases were associated with lymphocyte deficiencies of adenosine deaminase (ADA) or nucleoside phosphorylase has helped segregate two groups of the patients on biochemical grounds: those with purine metabolism enzyme deficiencies and those with no known enzyme deficiencies (Hirshorn, 1977). This distinction has been important in that it appears, particularly with ADA deficiency, that the immunologic lesion is primarily in the ADA-deficient bone marrow stem cell. Further, infusion of ADA-deficient SCID patients with ADA+ erythrocytes has restored lymphocyte functional integrity in some (Parkman et al., 1975; Chen et al., 1978). Of great interest has been the observation that the thymus gland in ADA-deficient SCID patients, though severely dysplastic, and in many cases near identical in morphology of ADA+ SCID thymus, may occasionally contain Has-


131

sall’s corpuscles (Borzy et al., 1979). Moreover, on reconstitution with ADA+ erythrocytes, thymic epithelium from ADA- subjects can be induced to release thymic hormones (thymopoietin) (Lewis et al., 1977). Thus, the presence of a hypoplastic thymus in ADA-deficient SCID, plus the observation that enzyme replacement and correction of a biochemical abnormality in lymphoid stem cells can induce normal thymic epithelium endocrine function, has given rise to the notion that there is normally an inductive interaction of bone marrow stem cells with thymic epithelium for normal thymic epithelial cell maturation and function (Pyke et al., 1975; Pahwa et al., 1977, 1978). In 1975 Pyke and colleagues demonstrated in a patient with ADA+ SCID that peripheral blood and bone marrow T cell precursors could be induced to mature in vitro by normal thymus epithelium. Likewise, SCID patient dysplastic thymic epithelium could differentiate in vitro and induce the differentiation of normal BM cells to mature T cells (Pyke et al., 1975). Pahwa et al. (1978) have shown in one ADA+ SCID patient that, although SCID thymic epithelium induced maturation of normal bone marrow cells, it could not induce maturation of SCID bone marrow T cell precursors, whereas normal thymic epithelium could induce the same SCID marrow cells to acquire T cell markers. Pahwa and colleagues (1977)also observed in another ADA+ SCID patient that neither thymus nor fetal liver alone reconstituted the patient, whereas a combination of the two transplants did. Taken together, these observations suggested that in some forms of ADA+ SCID, there may be both stem cell defects and thymic epithelial cell defects, or alternatively, there may be a defect in SCID thymic epithelium-stem cell interaction. Thus, using these probes of human thymic epithelium, we have demonstrated heterogeneous defects in thymic epithelial surface marker expression in severe primary cellular immunodeficiency diseases. These defects presumably reflect abnormalities of in vivo thymic epithelial maturation. B. THYMOMA The term thymoma described tumors of the thymus gland composed mainly of lymphocytes and epithelial cells in varying proportions (Rosai and Levine, 1976; Levine and Rosai, 1976). However, they are always considered to be neoplasms of the epithelial component (Levine and Bensch, 1972): the lymphoid component has been regarded as secondary, either induced by the epithelial cells or reactive. Thus, using antibodies A2B5, TE-4, TE-7, and anti-pl9, we have studied the surface antigen characteristics and regional distribution of

132

\

BARTON F. HAYNES

endocrine thymic epithelium in six thymomas (Mokhtar et al., 1983; Haynes et al., 1983c; Harden and Haynes, 1984). Like normal thymus, all six thymoma specimens contained epithelium that was A2B5+. Unlike normal thymus, A2B5+ thymic epithelium in thymomas was homogeneously distributed throughout the entire thymoma tissue with loss of the normal subcapsular corticalmedullary distribution of endocrine epithelium (Mokhtar et al., 1983). We have recently studied five additional thymomas at Duke University (Harden and Haynes, 1984). In all five thymomas, A2B5+, TE-4+ thymic epithelium was randomly distributed throughout the tumor amid thymocytes that were primarily of immature cortical thymocyte phenotype (T6+,3A1+,p80-, T1-, T3-). Like SCID and Nezelof syndrome dysplastic thymuses (and unlike normal thymus), all thymomas contained TE-7+ stroma homogeneously intermixed with the endocrine epithelium. All TE-4+, A2B5+ thymic epithelial areas in thymomas contained thymosin al. Interestingly, as seen before (Mokhtar et al., 1983), only a subset of A2B5, TE-4-reactive thymic epithelial cells were p19+. In one of five thymomas, p19 was completely nonreactive. Since all patients in whom thymomas developed were adults, the lack of reactivity of thymoma endocrine epithelium suggested a phenotypic similarity of thymoma endocrine epithelium with fetal or neonatal TE-4+, anti-pl9- thymic epithelium. These observations give rise to the hypothesis that in some thymomas, the thymic epithelium is either immature or dedifferentiated. C. MYASTHENIAGRAVIS Myasthenia gravis (MG) is a disorder of neuromuscular transmission. Evidence indicates that it is an autoimmune disorder affecting nicotinic acetylcholine receptors (Fambrough et al., 1973).Antibodies against acetylcholine receptors (AChR) are present in the serum of most MG patients (Patrick and Lindstrom, 1973; Almon et al., 1974) and administration of serum anti-AChR antibodies can produce a myasthenic syndrome (Toyka et al., 1977). Additionally, injections of human monoclonal anti-AChR antibodies into rodents induces the myasthenic state (Lennon and Lambert, 1980). Because of the frequent histologic abnormalities in thymuses of patients with MG and the sometimes dramatic clinical improvement seen following thymectomy (Papatestas et al., 1976; Olanow et al,, 1982), thymic factors have been implicated in the pathogenesis of the disease.


133

1 . Effect of Thymectomy on Circulating Peripherul Blood T Cell Subsets We have used monoclonal antibodies which define antigens on functional subsets of normal T cells to study the effects of thymectomy on the circulatory kinetics of peripheral T cells in MG (Haynes, 1981). We found that MG patients with thymic atrophy but not MG patients with a hyperplastic thymus had a selective decrease in the 3A1+, T4+ subset of T cells and, after thymectomy, the number of T4+ cells normalized (Haynes et al., 1983f). Decreased suppressor cell activity has been reported in peripheral T cells in myasthenic patients prior to thymectomy (Koethe et al., 1981; Mischak and Dau, 1981; Berrih et al., 1981).We found no abnormality in suppressor cell number, as defined by the T8 monoclonal antibody, either before or after thymectomy. Rather, there was a significant decrease in the number of T4+ cells in MG patients with atrophic thymus which normalized following thymectomy (Haynes et al., 1983f). These data suggest that the atrophic thymus exerts a selective inhibition of circulation of T4+ cells which is removed after thymectomy. The mechanism for this inhibition is not known. Of particular interest is the work of Dalakas et at. (1981) implicating thymosin a1 in the pathogenesis of MG. Using an immunofluorescence technique, they demonstrated that thymic epithelial cells containing thymosin a1 are present in all patients with MG, including those with atrophic histology, and further suggest that this may be the reason thymectomy is beneficial in all age groups. The thymus has been shown to interact with pituitary, adrenal, and testicular function in animals (Deschaux et al., 1979; Rebar et al., 1981). In our studies, thymectomy resulted in no change in circulating androgen levels. There was also no change in pre- and postthymectomy serum cortisol levels in myasthenic patients with hyperplastic histology. MG patients with atrophic histology, however, did show a significantly lower postthymectomy plasma cortisol level, compared with their prethymectomy levels (Haynes et ul., 1983f). These data demonstrate in MG patients with an atrophic thymus that thymectomy has an effect on the number of circulating T cells, and in particular, on those T cells expressing antigens 3A1 and T4. This effect may in part be mediated by changes in plasma adrenal corticosteroid levels after thymectomy or may be due to a factor produced by atrophic thymuses in MG-(Haynes et al., 1983f).

134

BARTON F. HAYNES

2 . Phenotypic Characterization o j Normal and MG Thymocytes Thymocyte suspensions were prepared from portions of thymuses obtained from 15 MG patients at therapeutic thymectomy and from 7 control subjects at the time of cardiac surgery (Harden and Haynes, 1983). Antibody NA1/34, which defines the cortical thymocyte T6 antigen, and antibody AlG3, which defines a medullary thymocyte antigen (p80) were used to study MG thymocyte subsets (Haynes et al., 1983d; McMichael et al., 1979). The antigens defined by antibodies Na1/34 and A1G3 are reciprocally expressed on normal thymocytes (Haynes et al., 1983d).We found an increase in AlG3+ (medullary) thymocytes in both hyperplastic and atrophic MG thymuses as compared to normals (37.1 2 6 and 35.8 f 3%,respectively, in MG compared to 23.6 9% in normals p < 0.001). Additionally, while the range of percentage medullary thymocytes (AlG3+) was narrow in normal thymuses (18-27%), the range in MG thymuses was broad (19-60%). Twelve of 16 MG thymuses had elevated (27%) numbers of medullary thymocytes. Tissue sections stained with AlG3 confirmed the presence of an expanded medulla in these patients. A significant reciprocal decrease in cortical (NA1/34, T6+) thymocytes was also seen (Harden and Haynes, 1983).

*

3. Endocrine Thymic Epithelial Patterns in Normal and MG Thymuses Analysis of MG thymus tissue sections stained with the panel of monoclonal antibodies indicated that not only was there an expanded medullary thymocyte compartment in many patients with MG, but also there was an abnormal spatial relationship of thymic epithelium to medullary thymocytes (Harden and Haynes, 1983). In normal thymus, A2B5+ thymic epithelium was present throughout the thymic "medulla such that most medullary thymocytes were in close proximity to A2B5+ thymic epithelium. In contrast, in three of seven (two hyperplastic, one atrophic) MG thymuses studied, A2B5+ epithelium existed only in islands within the medulla, with foci of medullary thymocytes isolated from contact with endocrine epithelium. The role of thymic epithelial-lymphoid interaction in T cell maturation in MG is an important area for further investigation. VIII. Summary

Several major points should be emphasized that provide directions for future research. First, using monoclonal reagents we have been able to phenotypically identify four major regions of the human thy-


135

mus microenvironment: (1) the thymic capsule, interlobular septae and stroma (TE-7+),(2) the subcapsular cortex (TE-4+,Thy-l+, A2B5+, anti-p19+, BB TECS+, TE-3+),(3)the cortex (TE-3+),and (4) the medulla (TE-4+,A2B5+,anti-p19+, BB TECS+).TE-4+ and TE-3+ thymic epithelium constitute HLA+, Ia+ subsets of thymic epithelium that are candidates for cell types of the human thymic microenvironment that might participate in conferring MHC restriction to maturing T lymphocytes. TE-7+ stroma most likely represents the mesodermalderived thymic component that early in development induces thymic epithelial differentiation. Second, whereas TE-4, anti-pl9, and BB TECS antibodies may be thymic epithelial lineage markers, they all react with the basal layer of squamous epithelium of various organs. In particular, in the tonsil, A2B5+,TE-4+ epithelium splays out in the base of tonsillar crypts and morphologically appears similar to thymic medullary epithelial cells. Therefore, these markers of endocrine thymic epithelium may also identify extrathymic areas of T cell differentiation. Third, the concept that thymic epithelium is constantly differentiating in the developed thymus is suggested by the coexpression of TE-4, TE-8, TE-16, and TE-15 antigen by layers of squamous epithelial keratinocytes and by thymic epithelium. That there is a TE-4/TE8RE-15 keratinocyte maturation pathway in skin gives credence to the notion that a similar pathway exists from TE-4+, TE-8-, TE-15endocrine medullary epithelial cells to TE-4-, TE-8+, TE-15+ Hassall’s bodies. Fourth, from the literature and the work presented in this review, three phases of thymic microenvironment development can be defined. The first phase is during early fetal development (4 to 8 weeks in humans) when mesodermal-derived fibrous tissue induces endodermal and ectodermal-derived thymic epithelium to proliferate and mature. TE-7+ mesenchymal stroma invaginates TE-4+ thymic epithelium and effects thymic lobulation. The second phase occurs between 9 and 15 weeks fetal development when the thymic primordia is colonized by blood-borne thymocyte precursors. Presumably during this stage, thymic epithelium promotes bone marrow cell colonization of thymus by producing chemoattractant molecules. During this second phase, the thymic epithelium continues to develop such that by 15 weeks fetal gestation, thymocytes as well as TE-4+ thymic epithelium and TE-7+ mesodermal-derived fibrous stroma have formed subcapsular cortical, cortical, and medullary areas, and Hassall’s bodies are present. The third phase of thymic microenvironment development consists of that period of time when the thymic microen-

136

BARTON F. HAYNES

vironment contributes to various stages of intrathymic T cell functional maturation. These functions involve participation in generation ofthe T cell repertoire, generation of MHC restriction, and production of soluble thymotrophic factors such as thymosin al, thymopoietin, and thymulin. Exactly when this phase begins and ends is not known. Phase three most likely commences early on, soon after bone marrow stem cells colonize the thymic primordia. Finally, although it is clear that the thymus is critical to development of cellular immunity early in fetal development, it is not clear what role the thymus plays in later postnatal life in the promotion of normal T cell differentiation. Further, the precise roles the thymic microenvironment plays in promotion of leukemogenesis, in the induction of T cell autoreactivity in autoimmune diseases, and in promotion of malignant transformation of thymic epithelium itself are not known. Study of thymic microenvironment components in normal thymus development and in the generation of T cell diversity and MHC restriction should provide insight into cellular and molecular mechanisms of T cell leukemogenesis and the role of T cells in autoimmunity. Hopefully, the reagents presented in this review will facilitate the isolation and in vitro culture of thymic epithelial subsets from normal and abnormal thymuses and aid in the development of in vitro assays of intra- and extrathymic T cell differentiation.

ACKNOWLEDGMENTS Supported in part by National Institutes of Health Grants CA28936, K0400695, CA11265, AH-19368, and a basic research grant from the March of Dimes Birth Defects Foundation. The author acknowledges the work of Drs. Kay Singer and Elizabeth Harden in growing cultured thymic epithelial cells, Mr. David Lobach in producing electron micrographs of thymic epithelial cells, the expert technical assistance of Mr. Richard M. Scearce, Ms. Lucinda L. Hensley, and Ms. Winifred Ho, and the expert secretarial assistance of Ms. Joyce Lowery.

REFERENCES Almon, R. R., Andrew, C. B., and Appel, S. H. (1974). Science 186,55. Auerbach, R. (1960). Deu. B i d . 2,271. Auerbach, R. (1961). Dew. Biol. 3,336. Bach, J. F., and Papiernik, M. (1981). Ciba Found. Symp. 4,215. Beardsley, T. R., Piersebacher, M., Wetzel, G. D., and Hays, E. F. (1983). Proc. Natl. Acad. Sci. U.S.A.80,6005. Bearman, R. M., Levine, G. D., and Bensch, K. G. (1978).Anat. Rec. 198,755. Beletskaya, L. V., and Gnezditskaya, V. (1974). Bull. E x p . Biol. Med. 77,678. Beller, D. I., and Unanue, E. R. (1977).]. Zmmunol. 118, 1780. Beller, D. I., and Unanue, E. R. (1978).J . Zmmunol. 121, 1861.


137

Berrih, S., Gaud, C., Bach, M. A., LeGrigand, H., Binet, P., and Bach, J. F. (1981). Clin. E x p . Immunol. 45, 1. Beutner, E. H., and Jordan, R. H. (1964). Proc. SOC. E x p . B i d . Med. 117,505. Bhan, A. K., Reinherz, E. L., Poppenia, S., McCluskey, R. T., and Schlossman, S. F. (1980).J . E x p . Med. 152, 771. Blattner, W. A., Kalyanaraman, V. S., Robert-Guroff, M., Lister, T. A., Galton, D. A. G., Sarin, P., Crawford, M. H., Catovsky, D., Greaves, M., and Gallo, R. C. (1982).1nt.J. Cancer 30,257. Borzy, M. S., Schulte-Wissermann, H., Gilbert, E., Horowitz, S. D., Pellett, J., and Hong, R. (1979). Clin. Immunol. Immunopathol. 12,31. Buckley, R. H. (1982).In “Cecil Textbook of Medicine” (J. B. Wyngaarden and L. H. Smith, eds.), pp 1789-1796. Saunders, Philadelphia, Pennsylvania. Buse, J. B., Srikanta, S., Haynes, B., and Eisenbarth, G. S. (1983).Clin. Res. 31,339A. Buse, J. B., Haynes, B. F., and Eisenbarth, G. S. (1984). Submitted. Cantor, H., and Weissman, I. (1976). Prog. Allergy 20, 1. Carpenter, C. B., Phillips, S. M., and Merrill, J. P. (1971). Cell. Immunol. 2,435. Clark, S . L. (1963). A m . J .Anat. 112, 1. Clark, S. L. (1968).J . E x p . Med. 128, 927. Clark, S. L. (1973). Cont. T o p . Immunobiol. 2, 77. Chen, S. H., Ochs, H. O., Scott, C. R., Gibbett, E., and Tingle, A. J. (1978). J . Clin. Invest. 62, 1386. Chu, A. C., Patterson, J. A. K., Goldstein, G., Berger, C. L., Takezaki, S., and Edelson, R. L. (1983).J . Invest. Dermatol. 81, 194. Cohen, F. E., Novotny, J., Sternberg, M. J. E., Campbell, D. G., and Williams. (1981). Biochem. J . 195,31. Cordier, A. C., and Haumont, S. J. (1980). Am. J . Anat. 157, 227. Dalakas, M., Engel, W. K., McClure, J. E., Goldstein, A,, and Askansas, V. (1981).Ann. N.Y. Acad. Sci. 377,477. Dardenne, M., Savino, W., Gastinel, L. N., Nabarra, B., and Bach, J. F. (1983).J . Immunol. 130, 1195. Deschaux, P., Massengo, B., and Fontanges, R. (1979). Thymus 1,95. Didiejean, L., and Saurat, J. H. (1980). Clin. E x p . Dermatol. 5, 395. Doran, T. I., Vidrich, A., and Sun, T. T. (1980). Cell 22, 17. Dosch, H. M., Lee, J. W. W., Gelfand, E. W., and Falk, J. A. (1978). Clin. E x p . Immunol. 34, 260. Eisenbarth, G. S., Walsh, F. S., and Nirenberg, M. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 4913. Eisenbarth, G. S., Shimizu, K., Conn, M., Mittler, B., and Wells, S. (1981). Cold Spring Harbor Symp. p. 209. Eisenbarth, G. S., Shimizu, K., Bowring, M. A., and Wells, S. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 5066. Eisenbarth, G. S., Jackson, R., and Srikanta, S. (1983). In “Monoclonal Antibodies: Probes for the Study of Autoimmunity and Immunodeficiency” (B. F. Haynes and G. S . Eisenbarth, eds.), p. 197. Academic Press, New York. Elie, R., LaRoche, A. C., Arnoux, E., Guerin, J. M., Pierre, G., Malebranche, R., Seeymayer, T., Dupuy, T. M., Russo, P., and Lapp, W. S. (1983). New Engl. J . Med. 308, 841. Elliot, E. V., Wallis, V., and Davies, A. J. S. (1971).Nature (London)New B i d . 234,77. Essex, M., McLane, M. F., Lee, T. H., Falk, L., Howe, C. W. S., and Mullins, J. I. (1983). Science 220, 859.

138

BARTON F. HAYNES

Ewijk, W., van, Verzijden, J. H. M., Kwast, T. H., van der, and Luijex-Meijer, S. W. M. (1974). Cell Tissue Res. 149,43. Fambrough, D. M., Drachman, D. B., and Satyamurti, S. (1973). Science 182,293. Fauci, A. S., Steinberg, A. D., Haynes, B. F., and Whalen, G. (1978).J . Immunol. 121, 1473. Fink, P. J., and Bevan, M. J. (1978).J . E x p . Med. 148, 766. Fournier, M., and Potworowski, E. F. (1983). Thymus 5, 89. Gallo, R. C., Mann, D., Broder, S., Ruscetti, F. W., Maeda, M., Kalyanaraman, V. S., Robert-Guroff, M., and Reitz, M. S. (1982). Proc. Natl. Acad. Sci. U.S.A. 79,5680. Gallo, R. C., Saran, P. S., Gelmann, E. P., Robert-Guroff, M., Richardson, E., Kalyanaraman, V. S., Mann, D., Sidhur, G. D., Stahl, R. E., Zolla-Pazner, S., Leibowitz, J., and Popovic, M. (1983). Science 220, 865. Gaudecker, von, B., and Muller-Hermelink, H. K. (1978). Ado. Exp. B i d . 114, 19. Gaudecker, von, B., and Schmale, E. M. (1974). Cell Tissue Res. 151,347. Goldstein, A. S., Low, T. L. K., Thurman, G. B., Zatz, M. M., Hall, N., Chen, J., Hu, S. K., Naylor, P. B., and McClure, J. E. (1981). Recent Prog. Horm. Res. 37,369. Goldstein, G. (1974). Nature (London) 247, 11. Hannapel, E., Xu, G., Morgan, J., Hempstead, J., and Horecker, B. L. (1982).Proc. Natl. Acad. Sci. U.S.A. 79, 2172. Harden, E. A., and Haynes, B. F. (1983). I n “Monoclonal Antibodies: Probes for the Study of Autoimmunity and Immunodeficiency” (B. F. Haynes and G. S. Eisenbarth, eds.), p. 297. Academic Press, New York. Harden, E. A., and Haynes, B. F. (1984). Submitted. Hashimoto, K., Shafran, K. M., Lazarus, G. S., and Singer, K. H. (1983).J . E x p . Med. 157, 259. Haynes, B. F. (1981). Immunol. Rev. 57, 127. Haynes, B. F. (1984). In “Thymic Hormones and Lymphokines” (A. L. Goldstein, ed.). Plenum, New York. Haynes, B. F., and Dupont, B. (1984). Submitted. Haynes, B. F., and Eisenbarth, G. S. (1983). “Monoclonal Antibodies: Probes for the Study of Autoimmunity and Immunodeficiency.” Academic Press, New York. Haynes, B. F., Mann, D. L., Hemler, M. E., Schroer, J. A,, Shelhamer, J. A., Eisenbarth, G. S., Thomas, C. A., Mostowski, H. S., Strominger, J. L., and Fauci, A. S. (1980). Proc. Natl. Acad. Sci. U S A . 77, 2914. Haynes, B. F., Reisner, E. G., Hemler, M. E., Strominger, J. L., and Eisenbarth, G. S. (1982). Hum. Immunol. 4,273. Haynes, B. F., Scearce, R. M., and Hensley, L. L. (1983a). Clin. Res. 31,354A. Haynes, B. F., Shimizu, K., and Eisenbarth, G. S. (1983b).J . Clin. Inoest. 71, 9. Haynes, B. F., Warren, R. W., Buckley, R. H., McClure, J. E., Goldstein, A. L., Henderson, F. W., Hensley, L. L., and Eisenbarth, G. S. (1983~). J. Immunol. 130, 1182. Haynes, B. F., Harden, E. A,, Telen, M. J., Hemler, M. E., Strominger, J. L., Palker, T. J., Scearce, R. M., and Eisenbarth, G. S. (1983d).J . Immunol. 131, 1195. Haynes, B. F., Robert-Guroff, M., Metzgar, R. S., Franchini, G., Kalyanaraman, V. S., Palker, T. J., and Gallo, R. C. (1983e).J. Erp. Med. 157, 907. Haynes, B. F., Harden, E. A., Olanow, C. W., Eisenbarth, G. S., Wechsler, A., Hensley, L. L., and Roses, A. D. (1983f).J. Immunol. 130, 1182. Haynes, B. F., Scearce, R. M., Lobach, D. F., and Hensley, L. L. (1984).J . Erp. Med. 139, 1149. Hirokawa, K. (1969). Acta PathoLJpn. 19, 1. Hirokawa, K., McClure, J. E., and Goldstein, A. L. (1982). Thymus 4, 19. Hirshhorn, R. (1977). Prog. Clin. Immunol. 3, 67. Hood, A. F., Soter, N. A,, Rappeport, J., and Gigli, I. (1977).Arch. Dermatol. 113, 1087.


139

Itoh, T., Kasahara, S., Aizu, S., Kato, K., Takeuchi, M., and Mori, T. (1982). Cell Tissue Res. 226,469. Jackson, R., Rassi, N., Crump, T., Haynes, B. F., and Eisenbarth, G. S. (1981). Diabetes 30. 887. Jackson, R. A., Morris, M. A., Haynes, B. F., and Eisenbarth, G. S. (1982). New Eng1.J. Med. 306, 385. Janossy, G., Thomas, J. A., Bollum, F. J., Granger, S., Pizzolo, G., Bradstock, K. F., Wong, L., McMichael, A,, Geneshaguru, K., and Hofmrand, A. V. (1980). J . Immunol. 125, 202. Janossy, G., Bofill, M., Dosiewicz, T. J., and Chilosi, M. (1983). In “Current Topics in Pathology” (H. K. Muller-Himelink, ed.). Springer-Verlag, Berlin and New York. Jenkinson, E. J., van Ewijk, W., and Owen, J. J. T. (1981).J . Exp. Med. 153, 280. Kalyanaraman, V. S., Sarngadharan, M. G., Bunn, P. A., Minna, J. D., and Gallo, R. C. (1981). Nature (London)294, 271. Kalyanaraman, V. S., Sarngadharan, M. G., Nakao, Y., Ito, Y., Aoki, and Gallo, R. C. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 1653. Kato, K., Ikeyama, S., Takaoki, M., Shino, A., Takeuchi, M., and Kakinuma, A. (1981). Cell 24, 885. Kendall, M. D. (1981). “The Thymus Gland” (M. D. Kendall, ed.), p. 63. Academic Press, New York. Kendall, M. D., and Frazier, J. A. (1979).Cell Tissue Res. 199, 37. Kissel, R. G., and Kardon, R. H. (1979). “Tissues and Organs, A Text-Atlas of Scanning Electron Microscopy,” p. 72. Freeman, San Francisco, California. Kissel, P., Andrea, J. M., and Jacquier, A. (1981). In “The Neuro-Cristopathies” p. 131. Masson, New York. Koethe, S., Cook, A., McQuillen, D., and McQuillen, M. (1981). Ann. N.Y. Acad. Sci. 377,447. Kruisbeek, A. M., Hodes, R. J., and Singer, A. (1981a).J. Exp. Med. 153, 13. Kruisbeek, A. M., Sharrow, S. O., Mathison, B. J., and Singer, A. (1981b).J . Zmmunol. 127,2168. Kruisbeek, A. M., Fultz, M. J., Sharrow, S. O., Singer, A., and Mond, J. J. (1983).J. E x p . Med. 157,1932. Kyewski, B., Hunsmann, G., Friedrich, R., Ketelsen, U. P., and Wekerle, H. (1981). Hematol. Blood Transf. 26, 372. Lampson, L., and Levy, R. (1980).J . Immunol. 125,293. Lavker, R. M., and Sun, T. T. (1983).J.Invest. Dermatol. 81, 121s. Lawlor, G. J., Ammann, A. J., Wright, W. C., LaFranchi, S. H., Bilstrom, D., and Stiehm, E. R. (1974).J . Pediatr. 84, 183. LeDouarin, N. M., and Jotereau, F. V. (1975).J . Exp. Med. 142, 17. LeDouarin, N. M., and Jotereau, F. V. (1981). In “The Thymus Gland’’ (M. D. Kendall, ed.), p. 37. Academic Press, New York. Lennon, V. A., and Lambert, E. H. (1980). Nature (London) 285,238. Levine, G., and Bensch, K. (1972). Cancer 30,500. Levine, G . D., and Rosai, J. (1978). Hum. Pathol. 9,495. Lewis, V., Twomey, J., and Goldstein, G. (1977). Lancet 2, 471. Lobach, D. F., Scearce, R. M., and Haynes, B. F. (1984). Submitted. Lyampert, I. M., Beletskaya, L. V., Borodiyuk, N. A., Gnezditskaya, E. V., Rassokhina, I. I., and Danilova, T. A. (1976). Zwimuiiology 31, 47. McMichael, A. J., Pilch, J. R., Galfre, G., Mason, D. Y., Falore, J. W., and Milstein, C. (1979). Eur. J . Zmmunol. 9,205. Mandell, T. (1968a).E x p . Biol. Med. Sci. 46,755. Mandell, T. (1968b). 2. Zellforsch. 89, 180.

140

BARTON F. HAYNES

Mandell, T. (1969). Aust. J . E x p . Biol. Med. Sci. 47, 153. Mandell, T. (1970). 2.Zellforsch. 106, 498. Mandi and Glant (1973).Nature New B i d . 246,25. McEndy, D. P., Boon, M. C., and Furth, J. (1944). Cancer Res. 4,377. McFarland, E. J., Scearce, R. M., and Haynes, B. F. (1984).J . Immunol., in press. McKenzie, J . L., and Fabre, J. W. (1981).J . Immunol. 126, 843. Metcalf, D., and Wiadrowski, M. (1966). Cancer Res. 26,483. Mischak, R., and Dau, P. (1981).Ann. N.Y. Acad. Sci. 377,436. Mokhtar, N., Hsu, S. M., Lad, R. P., Haynes, B. F., and Jaffe, E. (1984). Hum. Pathol. 15,378. Morris, R. J., and Ritter, M. A. (1980). Cell Tissue Res. 206, 459. Morrissey, P. J., and Singer, A. (1983).In “Monoclonal Antibodies: Probes for the Study of Autoimmunity and Immunodeficiency” (B. F. Haynes and G. S. Eisenbarth, eds.), p. 67. Academic Press, New York. Naito, K., Knowles, R. W., Real, F. X., Morishima, Y., Kawashima, K., and Dupont, B. (1983). Blood 62, 852. Naparstek, Y., Holoshirtz, J., Eisenstein, S., Reshef, T., Rappaport, S., Chemke, J., Ben-Num, A., and Cohen, I. R. (1982).Nature (London)300,262. Norris, E. H. (1938). Contrib. Embryol. Carnegie Inst. 27, 193. Olanow, C. W., Wechsler, A. S., and Roses, A. D. (1982).Ann. Surg. 196, 113. Oosterom, R., Kater, L., and Rademakers, L. H. P. M. (1981).Clin. Immunol. Immunopathol. 19,428. Pahwa, R., Pahwa, S., Good, R. A., Incefy, G. S., and O’Reilly, R. J. (1977). Proc. Natl. Acad. Sci. U S A . 74,3002. Pahwa, R. N., Pahwa, S. G., and Good, R. A. (1978). Clin. lmmunol. Immunopathol. 11,437. Papatestas, A,, Genkins, G., Horowitz, S., and Kornfeld, P. (1976).Ann. N.Y. Acad. Sci. 274, 555. Papiernik, M. (1970). Blood 36,470. Parkinson, D. R., and Waksal, S. D. (1978).In “Antiviral Mechanisms and the Control of Neoplasia” (P. Chandra, ed.), p. 53. Plenum, New York. Parkman, R., Gelfand, E. W., Koun, F. S., Sanderson, A,, and Hirshhorn, R. (1975).New Engl. J . Med. 292, 714. Patrick, J., and Lindstrom, J. M. (1973). Science 180, 871. Patten, B. M. (1968). “Human Embryology,” p. 206. McGraw-Hill, New York. Pereira, G., and Clermont, Y. (1971).Anat. Rec. 169, 613. Pertschuk, L. P. (1974). Tissue Antigens 4,446. Pictet, R. L., Rall, L. B., Phelps, P., and Rutter, W. J. (1976). Science 191, 191. Poiesz, B. J., Ruscetti, F. W., Reitz, M. S., Kolyanaraman, V. S., and Gallo, R. C. (1980). Proc. Natl. Acad. Sci. U.S.A.77, 7415. Poiesz, B. J., Ruscetti, F. W., Gazdar, A. F., Bunn, P. A., Minna, J. D., and Gallo, R. C. (1981). Nature (London) 294, 268. Posner, L. E., Robert-Guroff, M., Kalyanaraman, V. S., Poiesz, B. J., Ruscetti, F. W., Fossieck, B., Bunn, P. A., Minna, J. D., and Gallo, R. C. (1981). J . E x p . Med. 154,333. Potworowski, E. F. (1977). Clin. E x p . Immunol. 30,305. Potworowski, E. F., Seeymayer, T. A,, Bolande, R. P., and Lapp, W. (1979).2.Immunitaetsforsch. 155,240. Pyke, K. W., and Gelfand, E. W. (1974). Nature (London) 251,421. Pyke, R. W., Dosch, H. M., Ipp, M. M., and Gelfand, E. W. (1975).New Engl. J . Med. 293,424.


141

Rebar, R., Miyake, A., Low, T., and Goldstein, A. L. (1981). Science 214,669. Reinherz, E. L., Kung, P. C., Goldstein, G., Levy, R. H., and Schlossman, S. F. (1980). Proc. Natl. Acad. Sci. U.S.A.77, 1588. Ritter, M. A,, and Morris, R. J. (1980). Immunology 39, 85. Ritter, M. A., Sauvage, C. A., and Cotmore, C. R. (1981). Immunology 44,439. Robert-Curoff, M., Ruscetti, F. W., Posner, L. E., Poiesz, B. J., and Gallo, R. C. (1981). J . Erp. Med. 154, 1957. Robert-Guroff, M., Nakao, Y., Notake, K., Ito, Y., Sliski, A., and Gallo, R. C. (1982a). Science 215, 975. Robert-Curoff, M., Kalyanaraman, V. S., Blather, W. A., Popovic, M., Sarngadharan, M. G., Maeda, M., Blayney, D., Catovsky, D., Bunn, P. A., Shibata, A., Nakao, Y., Ito, Y., Aoki, T., and Gallo, R. C. (1982b).J . E x p . Med. 157, 248. Robert-Guroff, M., Fahey, K. A., Maeda, M., Ito, Y., and Gallo, R. C. (1982~).Virology 122, 297. Rosai, J., and Levine, G. D. (1976). In “Atlas of Tumor Pathology.” Armed Forces Institute of Pathology, Washington, D.C. Rouse, R. V., and Weissman, I. L. (1981). Ciba Found. Symp. 84, 161. Rubenfeld, M. R., Silverstone, A. E., Knowles, D. M., Halper, J. P., Desota, A., Fenoglio, C. M., and Edelson, R. L. (1981).J . Znuest. Dermatol. 77,221. Sauder, D. N., Carter, C. S., Katz, S. I., and Oppenheim, J. J. (1982).J.Invest. Dermatol. 79, 34. Schmidt, D., Monier, J. C., Dardenne, M., Pleau, J. M., Deschaux, P., and Bach, J. F. (1980). Thymus 2, 177. Scollay, R., Jacobs, S., Jerabek, L., Butcher, E., and Weissman, I. L. (1980). Eur. J . Immunol. 10,210. Seemayer, T. A., and Bolande, R. P. (1980). Arch. Pathol. Lab. Med. 104, 141. Seemayer, T. A., Lapp, W. S., and Bolande, R. P. (1977). Am. J . Pathol. 88, 119. Seemayer, T. A,, Lapp, W. S., and Bolande, R. P. (1978). Am. J . Pathol. 93, 325. Singer, A., Hathcock, K. S., and Hodes, R. J. (1982).J . Erp. Med. 155, 339. Singh, J. (1981). In “The Thymus Gland” (M. D. Kendal, ed.), p. 133. Academic Press, New York. Smith, H. R., Chused, T. H., Smathers, P. A., and Steinberg, A. D. (1983).J.Immunol. 130, 1200. Steinman, R. M., and Cohen, Z. A. (1973).J . E x p . Med. 137, 1142. Steinman, R. M., Wihner, M. D., Nussenzweig, M. C., Chen, L. L., Schlesinger, S., and Cohen, Z. A. (1980).J. Invest. Dermatol. 75, 14. Steinman, R. M., Nussenzweig, M. C., Witmer, M. D., Nogueira, N., and Cohen, Z. A. (1981). In “Heterogeneity of Mononuclear Phagocytes” (0.Forster and M. Landy, eds.), p. 189. Academic Press, New York. Stutman, 0. (1978). Immunol. Reo. 42, 138. Sun, T. T., Shih, C., and Green, H. (1979). Proc. Natl. Acad. Sci. U.S.A.76,2813. Takigawa, M., and Imamura, S. (1977).J . Invest. Dermatol. 68,259. Tapia, F. J., Barbosa, A. J. A., Marangos, P. J., Polak, J. M., Bloom, S. R., Dermody, C., and Pearse, A. G. E. (1981). Lancet I, 808. Theofilopoulous, A. N., Balderas, R. S., Shawler, D. L., Lee, J., and Dixon, F. J. (1981). J . E x p . Med. 153, 1405. Toyka, K. V., Drachman, D. B., and Griffin, D. E. (1977). New Eng1.J. Med. 296, 125. Umiel, T., Daley, J. F., Bhan, A. K., Levy, R. H., Schlossman, S. F., and Reinherz, E. L. (1982).J . Immunol. 129, 1054. Vakhaira, D. D. (1983). Thymus 5, 43. Van der Geld, H. W. R., and Strauss, A. J. L. (1966). Lancet i, 57.

142

BARTON F. HAYNES

Viac, J., Schmitt, D., Staquet, M. J., and Thivolet, J. (1980). Thymus 1,319. Watt, F. M., and Green, H. (1982). Nature (London) 295,434. Weissman, I. (1973).J . E r p . Med. 137,504. Wekerle, H.,Ketelsen, U., and Ernst, M. (1980).J . E z p . Med. 151,925. Weller, G. L. (1933). Contzib. Embryol. Carnegie Inst. 22, 95. Woodcock-Mitchell, J., Eichner, R., Nelson, W. G., and Sun, T. T. (1982).J . Cell. Biol. 95, 580. Wu, S., and Thomas, D. W. (1983).J . Immunol. 131,2110. Zielinski, C . C.,Waksal, S. D., and Datta, S. K. (1982).J . Zmmunol. 129, 882. Zinkernagel, R. M., and Dougherty, P. C. (1979).Adu. Immunol. 27,51.

ADVANCES I N IMMUNOLOGY, VOL. 36

Aging, ldiotype Repertoire Shifts, and Compartmentalization of the Mucosal-Associated Lymphoid System ANDREW W. WADE AND MYRON R. SZEWCZUK’ Department of Microbiology and Immunology, Queen’s Universiv, Kingston, Ontario, Canada

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Aging in the Immune System ....................................... A. Changes within the B Cell Population.. . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Changes within the T Cell Population.. ...........................

C. Changes within the Macrophage Population.. ...................... D. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Changes in the Expression of Available and Functional Repertoires with A g e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Changes within the Hemopoietic Stem Cell Population . . . . . . . . . . . . . . B. Changes in the Expression of Splenic Cell Repertoires. . . . . . . . . . . . . . . C. Proposed Mechanisms for Changes in Available and Functional Repertoires.................................................... IV. Compartmentalization of the Mucosal Immune System. . . . . . . . . . . . . . . . . . A. Morphology.. ................................................. B. Lymphocyte Trafficking.. ....................................... C. Mucosal Immune Responses in the Gut ........................... D. Unique Cell Populations.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Aging and the Mucosal Immune System.. ......................... References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

143 144 145 148 154 156 156 157 162 167 170 170 173 175 177 178 181

1. introduction

In the last few years we have seen enormous growth in the understanding of cellular communication and the vital role of lymphokines and idiotype interactions in influencing immune function. The responsibility of these factors in the declining immune potential of aging animals is just now beginning to receive attention. Since the aging field, prior to 1980, has been extensively reviewed by Makinodan and Kay (131),this critique does not attempt to provide a comprehensive summary of the literature in studies of aging, but instead concentrates on describing, in detail, recent advances in the field with only passing reference to work performed prior to 1979. In addition, we summarize Gerontology Research Council of Ontario Scholar.

143 Copyright Q 1984 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-022436-4

144

ANDREW W . WADE AND MYRON R. SZEWCZUK

salient aspects of the mucosal immune system, evidence for its compartmentalization, and describe its change with age in the murine system. The current literature is evaluated with the intent to provide explanations for the mechanisms governing the age-associated decline in immune reactivity.

II. Aging in the Immune System Over 50 years ago the first descriptions of age-related changes in the immune system were published by Friedberger et al. (56) and Thomsen and Kettel (228), who demonstrated that serum immunoglobulins to blood group antigens and xenogeneic erythrocytes declined with age. Since then attention has focused on the immune system in studies of aging for various reasons: (a) the thymus (controlling the maturation of T lymphocytes) is the first organ to undergo changes associated with senescence, involuting at sexual maturity in all mammals (19,196);(b) genes of the major histocompatibility complex (controlling immune reactivity) when transferred from a long-lived to a short-lived mouse strain extend the lifespan of that mouse (205,242); (c) techniques amenable to the study of the molecular and cellular biological changes occurring with age are easily applied to cells of the immune system; and (d) with age there is an increased incidence of diseases associated with an impaired immune system (70,134,173, 178,212). As a result, theories have been extended to explain aging in the immune system, ranging from the Unitary Immunologic Theory of Aging where the programmed decline in the physiological competence of the individual is thought to be due to changes manifested exclusively through the immune system (reviewed in 246), to less encompassing theories ascribing the immunologic decline in terms of intrinsic or programmed cellular changes (43, reviewed in 39 and 245), or the responsibility of extrinsic and microenvironmental factors in the aging process (for example, 55,68,78,217, 218,247, reviewed in 131 and 111). At present, however, there is little agreement with regard to the relative contributions that extrinsic and intrinsic factors have in the age-associated decline in immune reactivity: Makinodan (131,179,180) and Kay (111) have placed the relative responsibility in the aging process at 10 and 90% for extrinsic and intrinsic influences, respectively. Elucidation of the mechanisms involved in senescence would help provide a possible means of correcting age-associated rises in autoimmunity, cancer (178),and life-threatening infections (70,134,173,212)as well as allow insights into the delicate balance of regulatory factors controlling the

AGING AND THE MUCOSAL-ASSOCIATED LYMPHOID SYSTEM

145

immune system. Current findings in the field of aging necessitate reevaluating the contribution of intrinsic and extrinsic factors in mediating the immunological decline associated with aging. These areas will be discussed within the context of what is known from classical aging studies.

A. CHANGES WITHIN

THE

B CELLPOPULATION

Early work of Makinodan and Peterson (130) demonstrated that murine B cell immune responses decline with age. This has been verified and confirmed in numerous, exhaustive studies of mitogen responsiveness, T lymphocyte-dependent (TD), and T lymphocyte-independent (TI) immune responses in humans, rats, and mice, using both in vivo and in vitro assay systems (1,3,24,25,27,29,38,42,43,64,

66,83, 118,125,128,132,135,151,156,157,167,178,180,188,191,192,200, 204,225,247,249, reviewed in 39,111,131,242,246). Although there is no question of the existence of this decline, the factors governing its cause are still highly controversial. 1. Subcellular Changes In assessing these age-associated decreases in B cell function, most investigators concur that relative numbers of B cells do not change appreciably with age (for example, 25,87,255). Qualitative defects have been noted in these lymphocytes (43,191,198,224,237,244,

25 1,252). Formal work in murine B cell membrane perturbations is only recently receiving attention although indirect evidence for the existence of possible defects comes from studies which show that with age there are changes in cellular Fc receptors (FcR; 198,244) with a concomitant decline in FcR-mediated regulation (224). Extensive work with rats has demonstrated that although there is no age-related decline in this type of regulation, there are decreased densities of surface antigens B, Ia, and FcR (252), as well as p, y , and (Y chains (191,251) in old animals. Capping of the former three membrane proteins was slower in old rat cells due to postulated defects in the cells’ microtubule network and decreases in membrane fluidity (252). A 20-fold decrease in cellular ribosomal RNA has been described in old rat immune spleen cell populations (237),although no differences were noted in nonimmune rat cells. This work implies that upon antigenic stimulation, lymphocytes do not respond by increasing ribosomal RNA and, hence, protein synthesis. In the murine system, Dupere and Kolodziej (43), have shown that lymphocytes from old animals synthesize 8.5%less RNA when stimulated from a resting state to an

146

ANDREW W. WADE AND MYRON R. SZEWCZUK

antigen-activated secondary response state as compared with that found for cells derived from young animals. While messenger RNA derived from cells of aged animals was as efficient as that from young animals in synthesizing protein in an in vitro translation system, total cellular levels of mRNA were reduced by 26%in antigen stimulated cells from old animals. Specific reductions of more than 40% were found in messenger RNA specific for immunoglobulin in secondarily stimulated aged B lymphocytes (43). Resting lymphocytes evidently contain the necessary cellular constituents required for normal RNA and protein synthesis but do not respond maximally to activating stimuli. While this anergy may indicate that genetic abnormalities exist within these cells, evidence has been presented which attributes these defects to endocrinological changes associated with age (57),or decreases in Interleukin 2 (IL-2) levels (31,150,225,226)in the murine system. The situation is less clear in the human and rat systems. Mitogen-induced proliferation has been almost universally described as declining in aged B cells (1,25,29,45,150,151,191,192)with at least one exception (118).This decline is noted in the rat, mouse, and human systems with Fc fragments (150,151),antiimmunoglobulin (191,192), pokeweed mitogen (PWM) (1,29), bacterial lipopolysaccharide (LPS) (25,45), and dextran sulfate (45) signals. Since the rat and human response remained, for the most part, unaffected by the addition of young T cells (29),IL-2 (191),or B cell helper factors (192) to the cultures, it appears that the observed decline in proliferation was an intrinsic property of the senescent cells. Murine B cells, however, were sensitive to IL-2 treatment (see Section II,B), indicating that the decline noted is not entirely due to intrinsic defects within this system.

2 . Functional Changes Similar reductions with age have been noted in the primary immune response to either T-dependent (TD) (3,24,27,38,43,64,83,101, 132,135,157,178,188,192,217,249)or T-independent (TI) (25,27,38,42, 66,68,128,157,204)antigens. Explanations for the decline include decreases in T helper activity (3,24,43,64,135,249),increases in suppressor activity (27,38,64,68,101,135,188,217,249),changes in the efficiency of the cellular interactions required for the response (42,135, 179,192),and/or the induction of suppressive antiidiotypic regulation (66-68,122,217,219). From studies using TI antigens, it was concluded that defects also exist in the B cells themselves (except in ref.


147

247 where no decline was found). It is interesting to note that in all of these studies, only Smith (204) has tested responses to a true TI group 1 antigen (LPS); [T-independent group 2 antigens TNP-Ficoll, type I11 pneumococcal polysaccharide (SIII) and trinitrophenylated polyacrylamide beads (TNP-PAA) were used in the other studies]. As TI group I antigens are thought to be more “T independent” than TI group 2 antigens (reviewed in 152), the examination of responses to TI-1 antigens would possibly yield clearer information with regard to changes within the B cell population. This study (204) showed that while the immune response to LPS still declined with age, its onset was delayed in mice until 20 months of age (versus anywhere from 4 to 10 months for the TI group 2 antigens). Implications of this work are that (a) regulatory cells induced with age influence TI group 2 B cell responses to a greater extent; (b) a selective functional decline occurs in lyb 3+5+7+cells before the lyb 3-5-7- B cell subpopulation; and/or (c) the B cell repertoire of cells responding to these TI-1 and TI-2 antigens undergo differential changes with age. Secondary immune responses were also found to decline with advanced age (24,38,42,64,125,128,199,200,249),with one group of investigators finding no change (133).Decreases in antibody avidity and selective losses of IgG producing cells were noted in the immune response (64), as they were in the primary response (64,217,218,247). The decline in the secondary immune response was found not to be due to changes in antigen dose or response kinetics (199), but instead was thought to be due to specific suppression (64,122,125,200)or defective T cell help (125,200). Further experiments testing the ability of LPS to suppress an ongoing immune response in vivo (250)or enhance it in vitro (118)demonstrated that the desired effects could be induced only in young mice, while old mice were completely refractory to its influences. This would indicate that not only have the B cells in the aged animals declined in their ability to respond to the mitogenic properties of LPS in an in vitro immune response, but have also become down-regulated in vivo through possible prior environmental exposure to bacterial endotoxin. As a possible mechanism, LPS in vivo has been found to stimulate the production of suppressive antiidiotypic antibody (190), which in unrelated studies has been found to be increased in aged animals (66-68,122,217,219). Studies examining changes in tolerance induction with age have shown that the B cells, in a number strains, lose their ability to be tolerized by old age (156). The apparent paradox of having this defect expressed in old mice along with increased suppressor cell activity

148


was resolved by demonstrating that a subpopulation of B cells emerge in old animals that are sensitized by tolerogens but remain relatively insensitive to suppression (3,34). This resistance to tolerance induction with age has also been correlated with an increase in a T-helper factor (type 2) which acts in an antigen nonspecific, non-MHC restricted fashion (140,141).

3. Conclusions To summarize, the above findings imply that there are a variety of regulatory mechanisms and intrinsic defects responsible for the decline in B cell function and proliferative capacity. It is apparent that suppression of antibody formation can be mediated through B cells (188) and T cells (27,38,64,101,135,188,249),idiotype-antiidiotype regulation (id-anti-id) (66-68,122,217,219), environmental influences (118,250), or defective helper cell function and cellular communication (3,24,42,43,64,135,192,249).In the rat and human systems, intrinsic defects have also been suggested to account for some of this decline, as well as reducing the proliferative capacity of these cells (1,25,27,29,38,42,43,64,66,68,128,150,151,191,192,198,204,224, 244,251,252). Studies examining the subcellular activity and mitogenic properties of lymphocytes show that resting cell function is essentially normal in aged animals but that significant differences become apparent after antigenic or mitogenic stimulation. The depressed state of activated B cells from aged mice (but not from aged rats or humans) can be partially, if not completely reversed by the addition of exogenous factors, which suggests that defects responsible for these reductions may not be intrinsically programmed in these cells. Work presented in the next section (Section 111)provides evidence that reductions in murine B cell function may well result from a decrease in the frequency of antigen-responsive and IL-2-producing cells. Further work is required to delineate the mechanisms responsible for qualitative reductions in B cells from rats and humans where the defects may well be intrinsically programmed. Regardless of the mechanism the data do show that age-associated reductions in B cell function occur in the three animal models (rat, human, and murine) used in aging studies.

B. CHANGES WITHIN THE T CELLPOPULATION The finding that the thymus was the first organ to show any change with age (19,196) and the recognition of its importance in T cell differentiation, and indirectly in immune regulation, has caused the majority of aging research to focus on the study of senescent T cell defects

AGING AND T H E MUCOSAL-ASSOCIATED LYMPHOID SYSTEM

149

(10,23,31,62,64,128,192,226,249,251)and their relationship with the involuting thymus (91,247), decrease in circulating thymic hormones (55,247), and the appearance of age-associated pathologies such as autoreactive antibody (112). Given the importance of T cells in regulating immune function, it is likely that the majority of the regulatory defects described in the immune system with age can be traced back to changes within the T cell compartment. 1. Quantitative Changes

Quantitative changes in absolute and relative numbers of T cells and their subpopulations could possibly account for the defects in immune regulation found in aged individuals. This has proven not to be the case in mice, since total numbers of Thy-l+, helpedinducer, and suppressor/cytotoxic T cells as detected by Thy-1.2, Lyt-1, and Lyt-2 alloantigen expression, respectively, do not change appreciably with age in the spleen (23,31,167). In a comprehensive study performed with aging C57BL/6J male mice in this laboratory (236), the percentage of Thy-1.2+, Lyt-l+, and Lyt-2+ cells was found to undergo variable changes in systemic areas (thymus, bone marrow, spleen, and peripheral lymph nodes), while significantly decreasing in mucosal areas (mesenteric lymph nodes, mediastinal lymph nodes, and Peyer’s patches) with age. Upon immunization of these aged mice with TNPKLH, relative numbers of the three cell populations in systemic areas did not change for the most part from those levels determined for nonimmunized animals. In mucosal areas, however, immunization of old mice resulted in an increase, from the relatively low cell numbers determined for age nonimmune animals, up to those levels found in the mucosa of young mice. From this work and the findings of Kay et al. (113), the extent of T cell subpopulation changes appears to be highly strain and tissue dependent. Gilman et aZ. (63) have studied age-related changes in T cell populations of the rat. They found that Thy-l+ cells were decreased in the spleen and cervical lymph nodes of old rats, while numbers of W3/13+ and W3/25+ cells remained essentially unchanged over the age span. The situation is even less clear in aging humans when cells bearing the corresponding membrane antigens: OKT3, OKT4, and OKT8 are measured. A number of investigators have described the OKT3-positive (87) or E-rosette-positive (98) cell number (mature T cells) as remaining constant with age, while others have reported a decrease in relative (29,155) and absolute (139) numbers of OKT3+ cells. Conflicting results have also been presented with respect to changes in the relative and absolute numbers of both helpedinducer (OKT4+)and

150


suppressoricytotoxic (OKT8+) T cell subpopulations with age (29,139); and the influence of gender on such changes (139,155). Decreases in relative and absolute numbers of OKT8+, without changes in the OKT4+ cell subpopulation have been reported (155), as have increases in the relative numbers of OKT8+ cells, and relative decreases in OKT4+-positivecells with advanced age (139). Others have described decreases for both subpopulations (29). The reasons for these divergent findings are not immediately apparent, but may reflect sample groups that are too small and the great variability within the elderly population. In addition to these quantitative changes with age, qualitative defects have also been identified within these cells (26,43,63,237).

2. Subcellular Changes Callard et al. (26) found that by mixing both young and old syngeneic murine spleen cells in culture, a mixed leukocyte reaction could be detected. Similarly, they were able to demonstrate that young animals were capable of generating a humoral immune response in vivo against syngeneic cells isolated from old donors. The results suggest that as mice age, a novel cell surface antigen appears or is expressed on the lymphocytes of these animals. Gilman et al. have described a reduction in the distribution, density and capping rate of W3i13, W3125, and Thy-1.2 antigens on senescent rat T cells (63). These defects, thought to be due to changes in the microtubule and not the microfilament network within the cells, resulted in the decline in lateral membrane mobility and the slower capping rate observed in these experiments. Evidently rat B and T cells both exhibit the same defect in their cytoskeleton with age. Much of the work previously described for B cells regarding the status of RNA levels within aged lymphocytes (43,237) applies to murine T cells as well. Therefore, levels of total, ribosomal, and messenger RNA are reduced in immune murine T lymphocytes. Differences in the various RNA levels between young and old lymphocytes are not as great as in immune cells, and normal mRNA function can be demonstrated in aged lymphocytes (43), which suggest that downregulatory forces might exist within the aged population preventing normal activation in response to antigen. T cell mitogen-induced proliferation decreases with age in the human (23,62,71,87,98,136),rat (192), and mouse (23,31,45,148)systems in response to concanavalin A (Con A) (31,45,62,136,192) and phytohemagglutinin (PHA) (23,45,62,71,98,187).In the mouse system, Callard and Basten (23) have determined that the age-associated decline


151

in proliferation was not due to changes in the response kinetics, soluble inhibitors, number of PHA binding cells, or macrophages. With regard to this latter finding, however, the authors based their conclusions on the discovery that proliferative responses of old cells were unchanged after removal of (old) adherent cells. This point may prove important since Chang et al. (31), testing the responses of three mouse strains to Con A, found that the decline in the mitogenic response was due to reduced IL-2 production by old T cells (42% decrease), possibly resulting from defective Interleukin 1 (IL-1) synthesis by macrophages (unpublished results). A similar finding has been reported in preliminary studies by Gorczynski et al. (73). Chang et al. (31)found that this age-related decline was not due to suppressor cells or factors but instead could be reversed to young response levels by the addition of exogenous unfractionated rat Con A supernatant, or increased by the addition of adherent cells from young animals. These results are in disagreement with those found using the rat model (191),where IL-2 production by young and old spleen cells was found to be equal after exogenous IL-1 addition and Con A stimulation. Furthermore, the addition of fractionated rat Con A supernatant (IL-2) to Con A-stimulated old cells produced no enhancement of either B cell responses (191) or mitogen-induced proliferation (192) as it did above. Nonelicited, young and old peritoneal exudate cells produced equal levels of IL-1 in response to Con A (192). Dissection of the proliferative response in aged humans demonstrated yet another pattern of events: IL-2 production by aged peripheral blood lymphocytes was reduced in response to either Con A or PHA, but the depressed proliferative levels could not be increased upon exogenous IL-2 addition (62). Using three different assay systems, Inkeles et al. (98) demonstrated that not only were there reduced numbers of mitogen responsive cells in aged individuals (2225% of young levels), but also that cells initially responding to the mitogen were unable to expand into a pool of proliferative cells (reduced “burst size”). This cellular defect was found not to be related to abnormal levels of cyclic nucleotides within these cells as both cyclic AMP and cyclic GMP levels were normal (136).The second (cytoplasmic) signal generated in PHA-stimulated peripheral blood leukocytes has also been found to be unchanged by old age (76). Correct interpretation of these studies may be hampered by the finding that tritiated thymidine ( [3HlThy)incorporation into aged human cells as a measure of proliferation causes cell cycle arrest in Gz or M phase and an artifactually reduced proliferative response (87). Unfortunately, all of the work cited has relied on this technique, with

152


exposure times to [3H]Thy of 4 hours (62),8 hours (136), 16 hours (31), 24 hours (98,187), and 60 hours (71). As a point of reference, 24 hours of exposure to [3H]Thy caused a 130%artificial reduction in the proliferative response of aged cells (87). The results do demonstrate that important mechanistic differences exist between the three animal models used in the study of aging. With age, human and rat T cells become refractory to IL-2 signals and exhibit qualitative defects in their proliferative capacity. Murine and human cells, however, appear to develop defects in cellular communication with age, with specific reductions in IL-2 levels. These findings might account, in part, for the different changes in functional T cell activity noted between the human and mouse systems described in the following section. 3. Functional Changes In addition to defects in mitogen responsiveness, reductions in various murine T cell functions have been found to occur with age. These include decreases in cytotoxic T lymphocyte (CTL) function (72,226), delayed type hypersensitivity (DTH) reactions (187),mixed leukocyte reactions (MLR) (226),protective mediator T cells in Listeria monocytogenes (167), and lactate dehydrogenase virus (10) infections, allogeneic effect induction (256),and in helper T cell activity (24,42,64,128, 225,249). Conversely, suppressor T cell activity has been found to be increased in these animals (24,27,38,64,122,128,200,249).These findings in the mouse are in direct contrast to those found in the human system where an increase in helper cell (29) and a decrease in suppressor cell (1,29) function have been reported. In this system, it is not entirely clear how increases in helper cell function, as measured by enhancement of immunoglobulin (Ig) production in mixed populations of young B cells and aged OKT4+ (versus young OKT4+)T cells (29), can coexist with the reported decline in IL-2 production and reactivity of these cells (62). It is suspected that the allogeneic cell mixtures used in these experiments, and in the determinations of suppressor cell activity (1,29), do not give a true representation of these suppressor and helper T cell functions. Detection of T cell changes within the murine immune system has similarly relied on mixing experiments with either whole spleen cell populations (27,38,42,64,125,256)or isolated B and T cells (24,128, 225,249) from young and old syngeneic mice, in vitro (38,42,225), in adoptive transfer experiments (24,64,249,256),or after growth in diffusion chambers (128,200). Specific (24,27,38,64,128,200,249,256)or nonspecific (27,122) suppressor cell demonstration for the most part


153

was indirectly detected by anti-Thy-1.2 or anti-Lyt-2 antibody plus complement treatment enhancing the plaque forming cell response of old T and young B cell mixtures. It is noteworthy that in one study (128), simultaneous measurements were made of helper, suppressor, and B cell function in individual mice. A unique method of detecting T helper cell activity was employed in this work and bears mention, since other studies measuring this cell function could not completely exclude the contribution of suppressor cell activity in their assays. Helper function in this study was measured by the ability of T cells, activated with carrier-anticarrier immune complexes (preferentially induces T helper cells), to restore the plaque-forming cell (PFC) response of hapten-primed, T-depleted syngeneic cells after incubation in peritoneal cavityimplanted diffusion chambers. Using such a technique, the authors (128) were able to confirm previous findings of defective T cell help in aged mice. This activity and B cell responses (to TNP-Ficoll) declined steadily after 4 months of age while suppressor cell activity increased biphasically with a rapid rise until 15 months of age, and a slow increase thereafter in the (C57BL/6xC3H)Fl mice used. The decline in immune responsiveness with age was found to precede the detection of suppressor cells in two studies (38,64), while in another, the decrease in T cell function preceded the development of autoreactive antibody (112). Therefore, there is evidence for a relationship between the various functional T cell changes and the development of autoimmunity, but the picture is less clear for the role of increased suppressor cell activity in age-related immune response reductions. The interleukins appear to play a critical role in these declining T cell functions. Several investigators have noted that murine IL-2 production is reduced with age in response to Con A (31,148,225,226),Fc fragments (150), antigen (226), PHA (225), or allogeneic (149a,226) stimulation. Active suppression was not responsible for the reduction (31,225,226),but instead the decline was suggested to be the result of a defect or decrease in Lyt-l+ T cells (31,148,149a,150,225,226)and/or the synthesis of IL-1 by macrophages (31).The addition of exogenous IL-2 to murine cell cultures was able to restore MLR (226), CTL (149a,226),Con A (31),and PFC (225) responses to young-adult levels, while antigen-induced proliferative (57) or suppressor cell (227) responses were not completely recovered by such treatment. In the human, both IL-2 (62) and leukocyte migration inhibition factor (208) have been found to be reduced by old age. It appears that reduced IL-2 production can account for much of the age-related decline in the various immune parameters of old mice.

154


Since exogenous IL-2 can recover at least some of this cell-mediated reactivity, a reduction in IL-2 receptors does not seem likely on these senescent cells. As one possible explanation, Miller (148)has recently shown that precursor frequencies of IL-2 secreting, KLH-specific helper T cells were reduced in aged animals. In limiting dilution cultures with T cell depleted peritoneal cells, IL-2 production by these cells was shown to be normal (148). At present, the immune defect in uiuo cannot be unequivocally attributed to a primary reduction in IL-1 or in IL-2 synthesis. The reduced endogenous production of these lymphokines could account, in part, for reductions in both antibody formation and the mitogenic proliferation of B and T cells. Antibody production is dependent upon the IL-2-induced proliferation of T helper cells (127,243). Mitogenic stimulation of T cells is dependent upon proliferative IL-2 signals [as mitogens are thought to simply induce IL-2 receptor expression on these cells (reviewed in 127)], and the mitogenic stimulation of B cells is now recognized to require the presence of IL-1 and a T cell-derived B cell growth factor (95). 4 . The Role of Thymic Inuolution The involution of the thymus precedes the development of functional T cell defects and therefore most likely plays a major role in these abnormalities. Support comes from a study where transplantation of thymic lobes into. syngeneic thymectomized mice (91) demonstrated that the degree of thymic involution correlated with the maturity of circulating T cells. It was further shown that with this involution, the thymus loses its capacity to influence splenic T cell mitogenic responses, helper function, and proliferation to allogeneic stimulation, all of which are found to decline in the aging animal (91). Total thymectomy accelerates the decline in specific IgG PFC and antibody affinity normally seen with age (247). Reconstitution of aged mice with thymic hormones (55,247) restores their proliferative responses (55), PFC responses to TD antigens (55,247), and increases their helper cell activity in vitro (55).In such experiments, it is suspected that the thymic hormones may effect their action by inducing the maturation of cells capable of IL-2 production. C. CHANGES WITHIN THE MACROPHAGE POPULATION Callard (22) has demonstrated that there is no change in the ability of aged murine macrophages to support and regulate immune reactions to mitogens and antigens. Similar thinking has persisted


155

throughout most of the studies presented within this review even though Price and Makinodan (179) reported in 1972 that aged antigen presenting cells prevented sensitized B and T cells from responding maximally to limiting doses of antigen. Since attention has focused primarily on B and T cell changes with age, few reports exist which describe any age-related change in the macrophage population. Recent descriptions of published and unpublished work by Chang et al. (31) has indicated that macrophages may play an important role (via decreased production of IL-1) in the reduced IL-2 production noted in aged mice. Discussion here will be limited to a brief description of two recent reports describing regulatory roles for aged macrophage s . Using a reverse plaque assay, Antonaci et al. (6) have demonstrated that a reduction occurs in spontaneous human PFC with age. However, upon removal of monocytes from the system or pretreating the cells with indomethacin, the spontaneous PFC levels were restored. This indirectly suggested a role for monocytes, and prostaglandin release from these cells in down-regulating the human immune system during aging. A more comprehensive study examining the role of macrophages in the decline in FcR-mediated regulation was performed by Scribner and Moorhead in 1982 (198). Antiimmunoglobulin (anti-Ig) stimulates the proliferation of only aged lymphocytes, while its F(ab')z fragment will stimulate cells from both young and old animals. It was found that a radiation resistant, Thy-1.2-, sIg-, plastic adherent cell isolated from aged spleen cells [as well as bone marrow (BM) from young and old mice], was able to induce a proliferative response in young spleen cells to anti-Ig. This effect could be duplicated in these young cells after 30 days residence in irradiated aged recipients. It was demonstrated, therefore, that the decline in FcR-mediated regulation with age may involve a mobile macrophage population, derived originally from young bone marrow, which eventually populates the spleen, bone marrow, and peritoneum of aged mice. The study implicates a role for a macrophage subset in the declining immune regulation of old mice and also postulates the existence of cell trafficking changes with age. Although the antigen-presenting function of macrophages may not change with age (22), the work of Chang et al. (31) and the two studies presented above (6,198) suggest that changes in the macrophage population with age may contribute significantly to altered immune reactivity. Further work is required to define its precise role in the aging process.

156


D. SUMMARY Evidence has been presented that:

1. Primary and secondary T cell-dependent antibody responses decline with age, as do T cell-independent and mitogenic B cell responses. The demonstration of defective T help, decreased cell cooperation, and increased suppressor cell activity attributes most, but not all, of this decline to impaired T cell activity. 2. Changes within T cell populations and their receptivity to IL-2 may represent a major divergence in the age-related down-regulatory mechanisms operating in the human, mouse, and rat immune systems. 3. Murine T cell-mediated immunity declines with age as does T helper cell function, while suppressor cell activity is found to be increased in old mice. IL-2 production is decreased for all cell-mediated functions in senescent mice. 4. Reconstitution of many murine T cell-mediated responses is possible with exogenous IL-2 or thymosin (a thymic hormone), strongly implicating a role for the involuting thymus and an immature or dysfunctional Lyt-l+ cell in the declining immune reactivity with age. 5. Macrophages may also play a role in mediating changes in immune reactivity and possibly IL-2 levels in old animals. 111. Changes in' the Expression of Available and Functional Repertoires with Age

The major focus of the studies presented above has been to describe the nature of effector cell populations and their changing regulation within the systemic immune system during aging. Although many changes in immune reactivity with age have been described, there is still little agreement regarding the inductive processes responsible for these alterations. Recent reports in the murine system have indicated that (a) the bone marrow (BM) population of aged mice may potentially retain the ability to express both the available and functional cell repertoires identified in the BM population of young mice (68,73,79,81,94, 179,255); (b) many immune functions of old mice can be enhanced by the addition of exogenous IL-2 or thymosin (31,55,225,226,247); and (c) the effector function of individual cells within the spleen has been shown, in several instances, to be unaffected by age (148,255). Although preliminary, these studies may indicate that any defects in immune reactivity of old mice are not largely due to intrinsic alterations within the lymphoid cells themselves. A


157

system that is relatively free of these defects would allow the elucidation of possible regulatory changes responsible for age-related defects in immune reactivity. We present evidence that (a) available and functional repertoires expressed within the BM of old mice retain the potential to be the equal to those expressed in young mice; (b) these repertoires of BM cells from old mice are induced to change upon maturation and differentiation in the aged systemic environment (and not in the young systemic environment); and (c) the repertoire changes within the spleen cell population of senescent mice are stable and irreversible upon removal from the aged microenvironment. The possible mechanisms responsible for changing these cell repertoires are discussed.

A. CHANGES WITHIN THE HEMOPOIETIC STEMCELLPOPULATION As a necessary starting point for elucidating the mechanisms governing the immunological decline with age, the status of the hemopoietic stem cells in the aged individual will be considered. Careful analysis of any results are required as irregularities found within these precursor cells which are thought to be genetically programmed events may well have resulted from environmental influences and/or defective regulation within the bone marrow. Conversely, stem cells thought to be unchanged with age may reveal inherent abnormalities upon maturation. These considerations may prove important in the interpretation of precursor T cell changes discussed below. 1 . In Situ Proliferation The idea that cells may have a genetically programmed, finite lifespan came from studies of Hayflick working with human fibroblasts (86). Subsequent work using in vitro passage (84,85,168) or in vivo (serial) adoptive transfer of bone marrow cells (2,88,147) confirmed Hayflick’s initial work and demonstrated that hemopoietic stem cells after several passages became refractory in their ability to divide or repopulate recipient animals. The implication of these findings was that the cell’s demise was due to the accumulation of copying errors in the DNA with successive divisions. Interpretations of results in these original studies which used serial adoptive transfer techniques, (2,88,147,162) depended upon three assumptions. (a) Individual members of the stem cell pool acquire approximately equal mitotic burdens; (b) all stem cells are equally transplantable; and (c) lethally irradiated recipients provide an environment in which stem cells can proliferate and self-renew normally. With the observation, however, that stem cell populations are

158


heterogeneous in their degree of “sternness” (93,253),some question has been raised with regard to the validity of these previous studies (80,81,93,162,253).Recent work has indicated that the previous findings may be an artifact of the serial transplantation procedures since stem cells committed to proliferation and not differentiation are thought to be either diluted out with successive transfers (78,82,194), destroyed, or induced to differentiate upon removal from their initial marrow environment (80,82). After in situ depletion of cycling cells (with hydroxyuridine) (184) or repopulation competition studies with BM from young and old animals (78,80) it was determined that there was no real limit in the number of times bone marrow from aged animals can repopulate lymphoid-depleted animals. When BM from young and old animals is placed in a competitive situation, BM cells from the old animal will actually repopulate the recipient to the same degree, or better than the BM cells from a young animal (80,245). Although evidence to the contrary has been presented by Kay (112), findings in the above work (78,82,194,253)indicate that BM cells from young and old mice are equally transplantable and essentially nullify the supposition that intrinsic limitations exist in the lifelong regenerative capacity of bone marrow cells.

2. Available and Functional Repertoire Expression Studies examining the functional potential of BM cells from old mice after transfer into irradiated young recipients have demonstrated that PFC responses to specific antigens are unchanged as compared to PFC levels produced by recipients reconstituted with BM from young animals (79,81). Further studies have shown that BM from young or old animals when transferred into an irradiated young environment are indistinguishable in (a) endogenous PFC levels to a number of antigens (73), (b) specific PFC responses following LPS stimulation (73), (c) frequencies of antigen-specific cells (179,255), and (d) the idiotype repertoire expressed after immunization with TNP-Ficoll (68). The findings in these studies, however, are highly dependent upon technical aspects of the adoptive transfer techniques employed; Table I illustrates some of these differences and the consequential phenotype of the functional BM cell repertoire in young and old recipients. When placed in a young environment, BM from aged animals will express a “young” repertoire provided that (a) antilymphocyte serum treated BM from aged donors was not injected with thymus cells from young animals in experiments involving relatively short (4 week) in-


159

cubation periods in irradiated young recipients (49); (b)the irradiated young recipients were not thymectomized before cell transfer (49,79,118); and (c) the donor BM cells were not allowed to mature in the presence of spleen cells from old mice in studies involving prolonged residency periods (179). Astle and Harrison have recently shown that BM from aged donors does not necessarily express this “young” repertoire when the converse experiments are performed with irradiated aged recipients (9a). In their study, irradiated aged recipients implanted with a young thymus and injected with BM cells from young mice produced PFC responses to sheep red blood cells or proliferative responses to PHA that were comparable to those of young += young chimeras (9a). On the other hand, when aged recipients were injected with BM from old animals they did not produce similar responses as the young += old chimeras (9a). Gorczynski et al. (73) examining endogenous and LPS-stimulated PFC responses in similar experiments (without a source of young T cells) demonstrated that the antigen-specific B cell repertoires expressed by stem cells of both young and old animals were adversely affected by the “aged” environment. Hirokawa et al. (92) examining primary anti-sheep red blood cell PFC, mitogenic, and allo-CTL responses in aged recipients reconstituted with young BM (with a young thymus graft 1month after reconstitution) found that the young cells after several months in an aged environment were adversely affected only with respect to their PFC responses (as compared to untreated middle aged controls). In this latter experiment, however, irradiated aged recipients reconstituted with BM from old animals were unfortunately not examined (92). These studies highlight some important differences in the functional efficacy between stem cells from young and old animals. In aged recipients, the cell repertoires of both young and old BM cells are adversely affected by the aged environment (73). Other studies have demonstrated that maturation in the presence of “young” T cells confers resistance to this aged environment for BM cells from young, but not aged animals (9a). Collectively, the above work suggests that BM from aged mice may produce the altered recognition repertoire observed in the spleen of aged animals (Section II1,B) and will do so in the splenic environment of aged (but not young) mice with or without exogenous T cells from young animals. In the splenic environment of young mice similar findings can be observed if the thymus of recipient animals is removed, or if antilymphocyte serum-resistant BM cells mature in the

TABLE I THEAGE-ASSOCIATED PHENOTYPE OF BM FUNCTIONAL CELLREPERTOIRESAFTER ADOPTIVETRANSFER EXPERIMENTS" ~~

w

8 BM donor age Old

Irradiated recipient age

Recipient thymectomy

Young

-

BM cells plus

Residency period (days)

8 30 40 49-1 19 70-304

-

~

Phenotype of functional repertoire Bb

Young Youn$ Young" Young Young

56

-

+ + +

SRBC primed T cells SRBC primed T cellsc SRBC primed T cellsc Spleen cells-old Spleen cells-old

28 8 28 30-45 7 182

Young Oldd Oldd Oldd Youngg Old

T

Reference

Old

Young

Old

Old

Thymus

40 122-243

Old’ Old

Oldf

(79)

Thymus

40 334 56 122-243

Old Old Young

Old Young-oldfSh Young” Youngf

(73) (92) (72)

Thymus

g

(94

a The literature is summarized with respect to the techniques and findings of studies using adoptive transfer experiments designed to examine the nature of the functional repertoire of cells from the bone marrow (BM) of young and old mice. A “young phenotype” denotes response or cell frequency level that BM from a young animal would give under similar experimental conditions. Similarly an “old phenotype” refers to response or cell frequency level obtained when BM from old animals is used. All experiments noted here used the technique of injecting BM into lethally irradiated recipients and examining the specified responses after the indicated residency period. As detected by anti-sheep red blood cell (SRBC) PFC responses in oioo unless denoted as such. BM was pretreated with antilymphocyte serum before transfer into irradiated recipients. Anti-SRBC PFC responses were examined in oitro. Antigen-specific endogenous PFC and lipopolysaccharide-stimulatedPFC were examined. f Frequencies of phytohemagglutinin-responsivecells were examined. g Cell frequencies were examined under limiting dilution conditions. Frequencies of alloantigen-specific CTL were examined.

*

c

(94

162


presence of T cells from young mice (49). Studies identifying a change in the available repertoire of “young” stem cells placed in irradiated old recipients without exogenous T cells (73) and in the functional repertoire of “young” stem cells placed in irradiated young recipients with splenic T cells from old animals (68) strongly suggest a major role for the aged microenvironment and systemic T cells in effecting these repertoire changes. The available information indicates that the stem cell populations from young and old mice are not identical but that the BM of aged mice fully retains the ability to express the cell repertoires of younger animals. As an addendum, Gorczynski et al. (72) have examined the status of allogeneic precursor CTL in young and old irradiated recipients of BM from young and old animals (resident period-56 days). In young + old combinations the repertoire of allo-specificities and affinities generated was similar to that found in young + young combinations, whereas old + young chimeras resulted in the generation of specificities characteristic of old + old combinations (72). The results suggest that the age-associated defect in allogeneic CTL is intrinsically programmed in the stem cell population. The results are surprising somewhat since other reports have demonstrated (a) significant improvements in PHA responses of irradiated young recipients reconstituted with BM from aged animals (81);(b) normal antigen-specific T helper cell repertoires as determined from PFC response to SRBC in irradiated young recipients of BM from old mice (49,79,81);and (c) normal (young) idiotype repertoires of T helper cells as determined by examining the generation of hapten-augmentable (HA) PFC in irradiated young recipients of BM from aged animals (68). In contrast to these studies (49,68,79,81)suggesting an unchanged potential for stem cells of aged mice to generate mitogen-reactive T cells (81),and T helper cell repertoires characteristic of young mice (49,68,79,81),the possibility exists that alterations of the potential allo-CTL repertoire of BM from aged animals are intrinsically programmed within these cells (72).

B. CHANGES IN

EXPRESSION OF SPLENIC CELLREPERTOIRES Changes in the available repertoire of splenic B and T cells with age are best studied under conditions where numbers of only the cell type in question are present in limiting concentration-with other factors held constant at saturating levels. The results of such studies (presented below) are meant only to provide information regarding the frequency of specific cell types in the spleen of aged animals generated without the influence of regulatory cell interactions. It is asTHE


163

sumed that the analysis of antiidiotypic antibody in specific immune responses is indicative of the functional idiotypic repertoire generated in the response. It is evident that the individual changes described cannot account in toto for the depressed immune responses of old animals, but collectively may account for some of the changes seen.

1 . B Cell Repertoire Shij-ts Zharhary and Klinman (255), examining the clonal B cell response to dinitrophenylated hemocyanin (DNP-Hy) in splenic fragment cultures, have found that DNP-responsive cells in aged mice yielded antibody-forming clones that were normal in both the amount and relative affinity of antibody produced, but with slightly less of a tendency to produce IgG antibody. The frequency of anti-DNP specific cells in the bone marrow of these aged mice was equal to that found in young animals, and most importantly, this frequency was reduced in the spleen of old animals. To our knowledge, only three other groups have examined the frequencies of B cell clone numbers in the spleen of aged animals using limiting dilution analysis. In one study, numbers of antigen-specific endogenous and LPS stimulated spleen cells underwent age-associated changes in old animals (73). Another study has shown that frequencies of LPS reactive spleen cells are reduced by old age (4).In a third study, Hooijkaas et al. (94) measured age-related changes in frequencies of background and LPS-induced antigen-specific PFC in the spleen and BM of a number of mouse strains. They found small frequency differences between spleen and bone marrow populations of the various aged C57BL/6J mice tested, and variable changes within the other mouse strains. In that study, however, small frequency differences could have been due to the large variability within groups and the small sample numbers (94).

2 . T Cell Repertoire Shij-ts At the clonal and stem cell level, Gorczynski et al. (72) have examined the status of allogeneic precursor CTL (pCTL) and mature CTL in aging C57BL/6J7 CBA/J, C3H/HeJ, and C3HB6 F1 mice. Using limiting dilution analysis of splenic pCTL in mixed leukocyte cultures, it was found that the frequency of clones for different antigenic determinants changed with age; specifically that the precursor frequency for allo-(H-2) antigens increased, while a decrease occurred in the frequency of anti-TNP-modified self-clones. Overall, however, the total proportion of pCTL in the spleen was unchanged as judged after

164


Con A stimulation of the cultures. Functionally pCTL isolated from aged animals have a reduced proliferative capacity (decrease in burst size), normal specificity (no third party lysis), and a reduced binding affinity for their targets. After transferring BM cells from aged mice into irradiated young recipients, it was demonstrated that the defects noted above were an intrinsic property of the stem cells. Other studies have similarly shown a decrease in the frequency of alloantigen-specific (161b) or PHA-induced pCTL (148). However, it was shown that the extent of target cell lysis (148,16113)and cell proliferation in IL-2 supplemented, CTL limiting dilution assays were normal (148). Changes with age in frequencies of splenic T cell populations have been shown with IL-%-secreting(T helper) cells responsive to Con A (148), KLH (148), or alloantigens (149b), and with proliferating cells (148). In these experiments (148), the reduction in cell frequencies was not associated with any loss per cell of cytotoxic activity, IL-2generating ability, or burst size. These studies (4,72,73,148,149b,161b,255)suggest that changes may occur in the frequencies of various clone numbers of B and T cells in the spleens of aged animals, with recent reports indicating no loss of function in the resident cells (148,255). Studies examining immune responses of young recipients reconstituted with spleen cells from old animals (under nonlimiting conditions) demonstrate that any changes within the spleen cell population of old mice are not reversed in a young environment after residency periods of greater than 50 days (see Table 11),but instead are stable upon transfer with the donor cell population (72,81,180). 3. Functional Idiotype Repertoire Shifts The existence of an age-related shift in the idiotype repertoire of spleen cells was initially demonstrated by Goidl et al. (66). When testing for the inhibition of PFC by antiidiotypic antibody in the sera of young and old mice, they found that sera from aged immune C57BL/6J mice inhibited anti-TNP PFC isolated from old mice to a greater extent than it did with PFC isolated from young mice. Conversely, sera from young immune animals could reversibly inhibit PFC from similarly aged mice, but would rarely inhibit the plaque formation of cells extracted from mice of other ages (66). Since antiidiotypic antibody was shown to be responsible for this inhibition, the results indicated that the B cell idiotype repertoire stimulated by TNP-Ficoll had undergone a change with age. Goidl et al. (68) have also studied the ability of young and old animals to produce anti-


165

TABLE I1 THE AGE-ASSOCIATEDPHENOTYPE OF SPLENIC B CELLREPERTOIRES AFTER ADOPTIVE TRANSFER EXPERIMENTS‘ Irradiated recipient age

Residency period (days)

Phenotype of the functional B cell repertoires

Reference

Old

Young

8 5 50

Oldb,c,d Youngb Youngb Oldb

(179) (79) (180) (180)

Young

Old

5-9 6

Spleen cell donor age

The literature is summarized with respect to the techniques and findings of studies using adoptive transfer experiments designed to examine the nature of the functional repertoire of splenic B cells from young and old mice. A “young phenotype” denotes response or cell frequency level that spleen cells from a young animal would give under similar experimental conditions. Similarly an “aged phenotype” refers to response or cell frequency level obtained when spleen cells from old donors are used. Except where denoted, all experiments used the technique of injecting donor spleen cells into lethally irradiated recipients. The specified response levels were determined after the indicated residency period in the irradiated animal. As detected by anti-sheep red blood cell (SRBC) plaque-forming cell (PFC) responses in vivo. As detected by anti-lipopolysaccharide PFC response levels. The frequency of antigen-specific cells was analyzed under limiting dilution conditions. Recipients were implanted with diffusion chambers containing transplanted cells.

idiotypic antibody during the immune response to TNP-Ficoll [as detected by hapten augmentation (HA) of PFC responses]. They found that BM from young or old animals, when transferred into irradiated recipients, did not produce HA PFC upon immunization. Transfer of spleen cells from young or aged donors into similar recipients demonstrated that only spleen cells (or splenic B cells alone) from old animals were capable of generating HA PFC (68). Subsequent experiments examining the spleen cell population responsible for inducing this regulatory change revealed that only splenic T cells from old mice could induce BM cells (from young or old mice) to generate HA PFC in a young recipient (68). Evidently the idiotypic repertoire expressed by cells in the BM of old mice is indistinguishable from the repertoire in the BM and spleen of young mice, but trafficking of these cells to the spleen allows for interaction with anti-

166


idiotypic T cells which in turn induce changes in the splenic B cell idiotypic repertoire. Gorczynski et al. (72) examined the idiotype repertoires of allopCTL generated from the spleen of young and old mice. These studies demonstrated that the idiotypic repertoire of cells from young mice was different from that of old mice, and furthermore, that the aged” repertoire was less heterogeneous and nonoverlapping with that of young animals. This change in repertoire corresponded with a reduction in the target binding affinity of the cells without a change in specificity (72). The results are in contrast somewhat with those obtained from the examination of syngeneic tumor-specific CTL (52) from the spleen of old mice. In this study, the CTL were found to lose their specificity with age, while also exhibiting an altered idiotypic repertoire (52). ‘I

4 . Summary The common findings in the majority of these reports are that (a) the bone marrow of aged mice fully retains the potential to produce cells with a normal available and functional repertoire once removed from the aged environment (49,68,73,79,81,179);(b) the lymphocyte population in the spleen of old mice contains stable alterations in the frequencies of antigen-specific B cells (73,255), LPSresponsive cells (4), IL-2-producing cells (l48,149b), pre-CTL (52,72,16lb), Con A-reactive cells (148), and idiotype-bearing lymphocytes (66,68); and (c) despite these changes in cell frequency within the splenic microenvironment, there is evidence that effector cells can individually display full cell function (148,255). At this point we would like to emphasize that all of the possible techniques recommended by Sigal and Klinman (202) for the analysis of clonal repertoires in the bone marrow and spleen have not been fully employed in these studies. Direct antigen binding cell, isoelectric focusing, and fine specificity analyses have not to our knowledge been performed on tissues from aging mice. Many of the studies using limiting dilution analysis (4,72,94,148,149b,161b)and splenic focus cultures (255) [two recommended techniques (202)l have suggested that repertoire shifts occur only in the spleen and not in the BM of aging animals. This study is presented, with some trepidation, as indicative of a general phenomenon occurring within the aged animal. Clearly more detailed analysis is required before definitive conclusions can be drawn. We proceed with the assumption that the available repertoire of specificities, idiotypes, and V region genes within


167

the BM is functionally indistinguishable in young and old mice and that any age-associated modification occurs during the selection of the functional repertoire. The following discussion examines a few possible mechanisms which might be responsible for this modification.

MECHANISMSFOR CHANGES IN AVAILABLEAND C. PROPOSED FUNCTIONAL REPERTOIRES The complex pathways involved in the formation of functional B lymphocytes from BM stem cell populations have recently been reviewed (116,16la,164). We approach the problem of defining how the repertoire of cells may change with age by considering the outcome of possible interactions that stem cells may have with other lymphoid cells and antibodies in the environment of the aged animal. 1 . Antigen Presentation Price and Makinodan (179), examining antigen dose-responses in the intact mouse, determined that approximately 10 times the dose of antigen was required to stimulate maximally a PFC response in old mice as compared with that required for young mice. They attributed this effect to deficiencies in antigen processing. Unfortunately, since that time, immunological aging research has primarily focused on regulatory T cell changes with age and assigned antigen-presenting cells (APC) a rather passive role in any deficiencies identified. This decision may prove to have been premature as shown by Gorczynski et al. (73). This study examined and compared various antigenpresenting properties of APC from both young and old mice. They found that (a) APC (splenic, peritoneal, or cultured from BM) from aged mice were deficient in their ability to induce a PFC response in vitro to TNP-dextran or TNP-levan. In adoptive transfer experiments, this characteristic was not shown to be reversed in a young environment. (b) Velocity sedimented peritoneal exudate, or cultured BM cells from old mice, when cultured with nonadherent TNP-KLHprimed spleen cells (young) and antigen, significantly reduced the avidity of the PFC response. Under identical conditions, APC from young mice were shown to enhance the avidity of the response. In nonaging studies, adherent cells have been shown to be (a) an essential requirement in the final stages of precursor B cell maturation (116,117) and (b) a contributing factor in the selection of T cell repertoires through antigen interaction with membrane Ia antigens (33,193,201). In the context of work performed by Callard et al. (16) which demonstrated that putative MHC-I region products were

168


changed on the surface of lymphocytes (APC not directly examined), the selection of new cell repertoires may occur at the level of antigen presentation in the context of this altered Ia antigen. It is evident therefore that changes in antigen presentation, processing, or factor production with age could potentially influence both the number and affinity of clones selected in an immune response. Clearly this important area demands further study. 2 . Effect of Antiidiotypic Antibody on Zdiotype Expression Networks of idiotype-antiidiotype interactions have been proposed to control the expression and frequency of idiotype bearing cells (18,103,233). The structure, function, and properties of idiotypes and antiidiotypes have been discussed and analyzed in detail in a number of recent reviews (17,46,181). The present discussion will be limited to an analysis of how the idiotype repertoire of B cells can be selected for after cells express immunoglobulin (Ig) receptors in the BM and subsequently enter the microenvironment of the systemic immune system. These idiotype-driven mechanisms will be presented in the context of what is known of changes occurring in idiotype-antiidiotype regulation with age. At the outset, it must be emphasized that (a) discussion of this important area could form the basis of an entire review; (b) the terms “idiotype” and “antiidiotype” are strictly operational as any idiotype can equally well be an antiidiotype and vice versa; and (c) little detailed information exists at present regarding age-related changes in defined idiotype systems or in the specific components regulating idiotype expression with age. Szewczuk and Campbell (217), studying changes in the PFC response to TNP-BGG with age in C57BL/6J mice, found that numbers of IgM, IgG, and IgA anti-TNP PFC were reduced in the spleen of old animals. This reduction was accompanied by a loss of high avidity PFC (217,218) and an increase in hapten-augmentable (HA) PFC (217). [The degree that PFC responses are enhanced by free hapten has been correlated with the reversible binding of antiidiotypic antibody to the surface of the B cells (65,67,69,216,217,219)].This phenomenon was examined further in adoptive transfer experiments, involving the transfer of isolated splenic B and T cells from young or old mice into irradiated young recipients (214). It was found that only the combination of B cells (old) with T cells (old) would result in the generation of HA PFC after immunization with TNP-BGG. This indicated that, at least for the generation of an antiidiotypic antibody response, synergy was required between the responding cells. In addi-


169

tion, these studies (214,216-219) provide further evidence that the idiotype repertoire is changed in aging mice. An increase in HA PFC with age in C57BL/6J mice has also been reported in studies employing a variety of T-independent (66,68,215), or T-dependent (215) antigens. Analysis of HA PFC generation with age in different mouse strains after immunization with TNP-BGG has demonstrated that the property of producing high HA PFC by old age is linked with genes at the Igh-C and another undetermined locus

(216). There is abundant evidence that antiidiotypic antibody is generated during an immune response to a number of antigens (35,51,60,61,67, 99,189). In old age however, the production of antiidiotypic antibody is greatly increased over that produced in young animals (66,68,214-219). Several lines of reasoning would provide explanation for this phenomenon, which may include any, all, or none of the following: (a) breakdown in the selection of altered-self reactive clones in the involuting (aged) thymus (11); (b) selection of “new” T cell idiotype repertoires at the level of altered antigen presentation (33,193,201) in terms of defective processing (73,179) or the putative altered Ia antigen found in aged mice (26);and (c) the effect of hormonal imbalances (thymic or otherwise) or other milieu changed in the aged animal. As these possibilities remain untested, we turn our attention from the inductive effects responsible, to the effects that abnormally high antiidiotypic antibody may have on the available idiotype repertoire of cells in old mice. Antiidiotypic antibody has been shown to influence significantly the functional repertoire of cells through either enhancement or suppression of cells bearing the corresponding idiotype. Kelsoe et al. (114,115) and Reth et al. (182) have demonstrated that suppression (vs enhancement) is highly dependent upon the dose of antiidiotypic antibodies injected. It is unknown at this time whether the increased levels of antiidiotypic antibodies produced during the normal immune response in aged animals reach the threshold limits (milligram range) required for the induction of suppression (14,115,182).Studies by Klinman (122) have shown that aged animals acquire the capacity to suppress the responses of syngeneic primary dinitrophenyl-specific B cells. This suppression was noted only for syngeneic BALB/c B cells because B cells of strains differing from BALB/c in the heavy chain allotype-idiotype locus were not suppressed (122). It is not difficult to perceive that during the host’s lifelong barrage of antigenic stimulation, the accumulations of idiotypic regulatory

170

ANDFWW W. WADE AND MYRON R. SZEWCZUK

“

errors” (through changes in antigenic presentation, thymic influences, etc.) could result in significant changes in the functional idiotypic repertoire. Activation of silent clones, excess production of antibody bearing minor idiotypes, and the generation of idiotypic memory” for “new” idiotypes could all be contributing pressures on the functional idiotypic repertoire to change with age. ‘I

111. Compahentalization of the Mucosal Immune System

Without exception, the work previously discussed has approached aging in terms of examining and describing changes solely within the systemic immune system (spleen, peripheral lymph nodes, and blood). While information in this regard is clearly important, the mucosal lymphoid apparatus, encompassing over a third of the body’s lymphoid tissue, and forming the first line of defense in infection and foreign antigen exposure, has largely been ignored. As will be illustrated below, this system performs a critical regulatory role in preferentially exposing the systemic immune system to selected antigens, while conferring resistance to others. It is comprised of a unique population of cells which exhibit distinctive gut homing properties, and characteristics. Evidence is also accumulating that the unique mucosal environment is capable of maintaining the growth and maturation of cells distinct from those in systemic areas, supporting the idea that this immune system, while in communication with the rest of the animal, can also remain compartmentalized and distinct. Finally its role in maintaining immune reactivity during aging in the face of an increasingly anergic systemic immune system will be discussed.

A. MORPHOLOGY The gut-associated lymphoid tissue (GALT) is comprised of lamina propria aggregates, Peyer’s Patches (PP), mesenteric lymph nodes (MLN), and lymphocytes within the mucosal epithelium (IEL), lamina propria (LPL), and thoracic duct (TDL) (reviewed in 13,169,172,238).Although each discrete area has unique cellular and functional characteristics, the system is in constant flux with continual communication within itself and, to lesser degrees, with mucosal areas of the respiratory system (BALT) (14,15,144,145),and the genital area (144,145) (the common mucosal immune system). Numerous lymphocytes are present within the mucosal epithelium itself (intraepithelial lymphocytes) (see Fig. 1). These cells lie between the enterocytes in the gut wall and are thus in intimate contact with the lumen and its contents. Many of the cells are T lymphocytes


171

NTRAFPITHELIAL LYMPH, ( T cells, granular cells, NK, & mast cells)

LAMINA PROPRIA LYMPH.

(22% plasma, 18%B, 41% T I a 10% granular cells)

cells

B cells, macrophages, 8 dendritlc cells

1 : EFFERENT

17

LYMPH.

40% B (C3R+, sIgA+M+R slgA+D+)cdls

irTDL

2.1 % plasma cells, 42% T cells, macrophages.

(60-90% T )

FIG.1. Lymphoid populations within the murine small intestine. This figure illustrates the distribution of mucosal lymphocyte populations within the epithelium, lamina propria, Peyer’s Patches, mesenteric lymph nodes (MLN), and thoracic duct lymph (TDL). Granular cells: Thy-1.2+,Lyt-1-, Lyt-2+ granulated cells within the epithelium and lamina propria. NK, Natural killer cells; M cells, microfold cells; C3R, complement receptor for C3; T (cells), thymus-derived cells.

(50),but granular cells (Thy-1.2-, Lyt-2+)(143,197),a unique population of natural killer (NK) cells (153,222,223), and mast cells (9,171) are also present. The exact function of these cells is presently unknown, although it is thought that a significant proportion are active blast cells, some of which may be engaged in trafficking to the lamina propria and beyond (138). Because of their close contact with the luminal surface, it is presumed that some priming of these cells must occur in situ. Lymphoid cells are also present in the interstitial tissue of the lamina propria, being comprised mainly of mature plasma cells together with the occasional macrophage, lymphocyte, unique granular cell (143,197), and mast cell (9,171). The function of these granular cells (Thy-l.2-, Lyt-2+), isolated from both the lamina propria and epithelium, is unknown. However, given the pleomorphic surface phenotype of cloned NK cells with at least two cloned lines found expressing a Thy-l.2-, Lyt-2+ phenotype (clone JT8) (89) or the human equivalent--OKT3-, OKT8+ (163), these cells may be NK cell precursors (or related in some way). It is clear that they’re unrelated to mast or T cells (197). The mast cells isolated from both tissues were

172


distinct from peritoneal, thymic, and connective tissue mast cells in their morphological and histochemical properties (9,171).These cells are thought to be involved in allergic reactions within the gut after interacting with IgE antibody produced by resident sIgE+ cells (44,213).Tseng (232)has described the lymphocyte population within the lamina propria as being composed of 22% cytoplasmic immunoglobulin (cIg) containing cells (96%IgA+), 18%surface Ig-positive (sIg+)B cells, 41% T cells, and 10% granulocytes. It was found that the B cells (C3R-) lost their surface Ig quickly upon culture, with many expressing sIgA or sIgM only. Due to the presence of this high number of sIgA+ IgD- B cells [phenotype of memory cells (ZO)], and the easily tolerizable sIgM+D- B cell (241),the population was markedly distinct from that found in the spleen. No features unique to the T cell population in the lamina propria were detected except possibly that a very low density of Thy-1.2 was present on their surface. The superior and inferior MLN drain the Peyer’s Patches and total gut mucosal area (28,229).These nodes do not differ morphologically from other lymph nodes, but are unique in the fact that they are constantly exposed to antigenic stimulation (172).The cell population has been determined to comprise: 40% sIg+ B cells, 2.1%cIg+ cells, 42% T cells (232) and macrophages. Its B cells differ from those of the lamina propria in that they express a less terminally differentiated phenotype: with the presence of C3 receptors and double isotypes (sIgA+M+and sIgA+D+)on their cell surface (232). MLN T cells are found to have high levels of Thy-1.2 on their surface (232). Scattered throughout the lamina propria are aggregates of lymphoid tissue ranging in size from minor cellular foci or nodules, to massive accumulations and Peyer’s Patches (reviewed in 238). The Peyer’s Patches have a follicular structure with T and B lymphocytes (20,106), macrophages (119,203),and dendritic cells (203,209,248).The luminal dome area of Peyer’s Patches is covered with a flattened epithelium composed of specialized M (microfolds) cells important in the uptake and handling of antigen (166). Short protuberances on their luminal surface form cytoplasmic bridges between the usual columnar microvillus-covered epithelial cells suggesting that some intracellular communication must occur between these cells. These M cells are also in intimate contact with the lymphocytes overlying the PP dome area. The microfolds on the cell’s surface aid in the uptake of molecules and particles from the lumen and their subsequent transfer to the subjacent lymphoid tissue. Virtually all of the lymph from the gut finally reaches the thoracic duct from which its cellular and other contents enter the systemic


173

circulation. Over half of the lymphoid cells in the thoracic duct are T cells (65-90%) (37) but there are also minor populations of T and B blast cells (137).

B. LYMPHOCYTE TRAFFICKING While it is evident that the Peyer’s Patches are the prime site for gut antigen entry through their specialized M cells, some luminal antigen is also known to gain entrances to the mucosa through pinocytosis by the enterocytes lining the gut wall (240). In certain cases, the material entering in this manner (or through the M cells) can remain undigested and be released via exocytosis into the lamina propria (240). Regardless of the entry point, antigen induces activation of resident cells which traffic through the MLN to systemic areas where some proliferation occurs before returning to the gut mucosa. A large body of work has examined the unique characteristics and specialized gut homing properties of these cells in rats, mice, and sheep (reviewed in 12,16). These studies have found that the general lymphoid system is comprised of two populations of cells; small recirculating lymphocytes and large antigen-activated blast cells (126,254). Each of these cell populations exhibits different trafficking patterns, both in themselves but also within their different subpopulations. Conflicting reports have emerged with regard to the timing and localization patterns of these cells mainly because of technical problems associated with cell isolation of LPL and collection methods for TDL (207). Following will be a summary of the salient points regarding trafficking of each population (reviewed in 170). Recirculating lymphocytes make up the bulk of cells migrating through the body at any one time. De Freitas et al. (3)reported that T cells isolated from the thoracic duct lymph showed minimal migration to the gut, but homed instead to the peripheral lymph nodes (PLN) and spleen (as in 36). The identical pattern was observed for T cells isolated from the PLN (36,37,211). B cells, however, migrated preferentially to the gut and spleen as opposed to the PLN (75,211).Analysis in short-term homing studies of T cell subpopulation shifts revealed further subdivision of the trafficking pattern of these cells based upon their lyt phenotype (124). In this work, it was reported that T lymphocytes which homed to the gut were almost entirely Lyt-2- cells, whereas equal numbers of Lyt-2- and Lyt-2+ cells homed to the peripheral lymph nodes. In a series of experiments, Smith and Ford (206,207) have demonstrated and extensively discussed the kinetics and localization of rat lymphocytes trafficking to a number of tissues in

174


syngeneic recipients (second pass in vivo). Recently, these trafficking patterns were shown to be similar in both normal and athymic, nude rats (53,54). In the above studies (37,75,124,211),the distribution patterns described for the cells were similar regardless of their tissue origin. The trafficking of these small recirculating cells, therefore, appears to be a preprogrammed event, depending solely upon the class of cell. Superimposed on the trafficking pattern of recirculating lymphocytes is the very directed, antigen-specific homing of large activated blast cells. It has been found that large dividing T and B cells isolated from either the TDL or MLN will home preferentially back to the gut mucosa (74,144,146).Similarly, antigen-activated blasts obtained from the PLN will home selectively back to their tissue of origin (12,16,74,146).This trafficking preference of mucosal cells has been attributed to regional blood supply (165), sex hormones in the case of genital mucosa homing (145), and a combination of antigen-independent (97) and -dependent mechanisms (96,176,177).Working with isolated jejunal rat intestinal loops, Husband (97) and Pierce and Cray (176,177)demonstrated that upon enteric immunization, antigen-specific plasma cells appeared throughout the gut, but were most numerous at the challenge site due to antigen-dependent memory cell proliferation. The remainder were thought to have arisen in the mucosal follicles or MLN, and migrated systemically before returning to the LP. This homing was organ specific, since blasts activated by a duodenal challenge returned preferentially to the jejunum, but not necessarily to the antigenic site. A further characteristic of these blast cells was that they did not divide in the LP upon their return, but differentiated into memory and IgA-secreting cells instead. Tseng (230) confirmed this migration using PP cells which after stimulation, traveled to the spleen, remained there for 5 days, returned, and repopulated the lamina propria where they differentiated into mature plasma cells. Removal of the spleen still resulted in IgA precursors leaving the PP before returning to the LP (230),indicating first, that spleen migration is not an obligatory step in the homing pathway, and second, that this migration of precursor cells was probably a characteristic of the cells themselves. Therefore, this type of migration appears not to be related to specific types of cells, as for recirculating lymphocytes, but is dependent instead upon the tissue where stimulation initially occurred. One of the mechanisms proposed to explain the distinctive trafficking pattern of recirculating lymphocytes is that these lymphoid cells interact preferentially with the high endothelial cells of the postcapillary venules (HEV)within the various organs (PLN, MLN, PP) (21).In


175

vitro studies of recirculating lymphocytes (124,211,234)have shown that, in accordance with in vivo findings, B cells and Lyt-2- T cells bind to the PP HEV selectively, while T cells as a whole bind better to the HEV in PLN (124,211). Factors can be isolated from lymphocytes which inhibit HEV binding (32), indicating that this directed-trafficking system can be overridden. Macrophage products released after antigen activation have also been assigned a regulatory function in lymphocyte binding of HEV (89). A possible role for lymphocyte microvilli in HEV recognition has been proposed since cells within their respective microenvironments under the scanning electron microscope appeared smooth, while recirculating B and T cells exhibit numerous microvilla on their surface. When these latter lymphocytes crossed the HEV of their target organs, they withdrew their microvilli and appeared smooth upon arrival at the lymphatic stroma (234). Spangrude et al. (210), using pertussigen and fucoidin as inhibitors, have broken down the lymphocyte-HEV interaction into two stages: (a) lymphocyte-endothelial cell recognition, and (b) activation and execution of the actual extravasational event. Gallatin et al. (58)examining the basis of HEV recognition produced a monoclonal antibody MEL-14 which selectively inhibited lymphocyte binding to PLN HEV. MEL-14 did not bind lymphoid cells which either do not recognize HEV or bind only to Peyer’s Patch HEV (58). There is direct evidence therefore that recirculating lymphocytes are comprised of specific populations which, by virture of their surface receptors, target specifically to the PP or PLN. In summary, the recirculating pool of unstimulated lymphocytes demonstrates distinctive trafficking patterns dependent only upon their cell type (independent of their organ source). B cells and Lyt-2T cells migrate preferentially to the gut (B cells traffic to the spleen as well). In contrast, T cells travel selectively to the peripheral lymph nodes. These findings correlate with the preferential binding of these lymphocyte subpopulations to the HEV of their target organs. Upon activation of these celIs, they change their migratory pattern and home specifically to the organs from which they were derived (and presumably stimulated), independently of their cell type. Cells stimulated in the gut appear to require a round of systemic migration as a necessary step in their maturation, before their return and activation within the mucosa. C. MUCOSALIMMUNE RESPONSESIN THE GUT The immune response in the mucosa is unique in the fact that it involves antibody mainly of the IgA isotype (123,231,232),under cer-

176


tain conditions can induce systemic unresponsiveness (oral tolerance) (30),stimulate reactivity solely within the mucosa ( 1 5 4 ~ 8 5or ) ~evoke systemic responses as well (40,41,195).The growing body of literature dealing with gut immunity implicates Peyer’s Patches in all three of these phenomena (8,107,108,119,129,142,154,158).Such a finding is not unexpected since PP act as the entry point for much of the antigen contacting the gut, and furthermore, contain subpopulations of B cells, (20) and Lyt-l+ T lymphocytes which effect unique gut-related functions. The latter cell population is primarily involved in generating specific IgA precursor plasma cells which mature upon systemic migration and seeding of the gut mucosa (107,119,231). Further information with regard to the unusual function of this tissue comes from a study performed by Kiyono et al. (119) who demonstrated that while Peyer’s Patches have the full complement of cells required to generate an immune response in vitro [including functional antigen presenting cells (119,203,209); although not found in ref. 1041, the tissue architecture in vivo prevents local immune response development. The Peyer’s Patches therefore appear to be essentially a factory producing activated precursor plasma cells which, through their special characteristics and homing capabilities, return and generate immune responses within the gut tissue. To effect this unique function, Peyer’s Patches contain at least two types of specialized T cells; one population of IgA FcR-positive cells which mediate a class switch in IgM and IgG bearing B cells to IgA (107,108,110), and a second population which induces further maturation of these cells specifically (48,109,120) or nonspecifically (231). It has been speculated that the novel structure of Peyer’s Patch germinal centers with their high content of Lyt-l+ T cells and B cells expressing IgA, IgG, or IgM (but not IgD) allows for this class switch and the generation of memory and precursor IgA plasma cells, but limits maturation to this point (20). It is apparent that PP are exquisitely sensitive to the dose and/or presentation route of antigen for either triggering immune responsiveness or unresponsiveness (100,129,175). Tolerance generated after oral ingestion of antigen can be mediated through suppressor T cells (40,129,142,158,175,184),B cells (8,77,100,105), or immune complexes (5). Upon activation of cells within the PP, seeding of distant systemic sites with antigen-responsive cells may (40,41,154,195)or may not (154) occur. Conversely, intraperitoneal administration of antigen can cause a mucosal immune response to be detected (175) or result in unresponsiveness instead (100,174).Mixing of the immunization protocols (parenteraVora1) to generate a secondary response in


177

the gut has been successful (174) or was shown to induce suppression (100,175). It appears therefore that complex interactions occur between the systemic and mucosal lymphoid systems in generating an immune response. The present state of knowledge allows no further conclusions to be drawn.

D. UNIQUECELLPOPULATIONS Gearhart and Cebra (59) have shown that differences exist in the B cell repertoires directed against phosphorylcholine in the Peyer’s Patches and spleen of nonimmune mice with the spleen containing larger numbers of cells bearing the TEPC-15 idiotype. Indirect support for this concept also comes from a study performed by Jackson and Mestecky (100) who demonstrated that after using a mixed immunization protocol (oral and intravenous administration of antigen) of bovine serum albumin, suppressive antiidiotypic plasma cells could be detected in mucosal areas, but not after using either immunization route alone. One can speculate that the mixed protocol causes recruitment of cells bearing different idiotypes, which circulate and prime for an antiidiotypic response in the opposing area. It is evident from studies (174) showing cooperation when using this mixed immunization protocol that overlap in the idiotype repertoires must exist in some cases. The strongest evidence that unique cell populations can exist within the mucosa comes from the elegant study of Eldridge et al. (47) who demonstrated that an entirely different B cell population is found in the mucosa of the X-linked immunodeficient (xid) CBA/N mouse, without any evidence of this population existing in systemic areas. The xid mutation results in a defective maturation of B cells, and an inability of CBA/N mice to respond to T-independent group 2 antigens. The above investigators have found that the B cell subpopulation missing in the spleen of these animals can be isolated from their Peyer’s Patches at 6-8 weeks of age. The cells were capable of responding to T-independent group 2 and T-dependent antigens, and expressed the surface phenotype of mature B cells. The mucosal T cell population from these mice has similarly been shown to have unique functional properties (121). In contrast to T cells from the spleen, isolated or cloned T cells from the PP of xid mice uniquely demonstrated the ability to provide help to B cells (xid) in generating an immune response to SRBC. In addition, it was thought that mucosal T cells were uniquely capable of enhancing receptors for T cell help on lyb 5- B cells, and possibly providing a maturational signal to these cells (121).

178


Unique gut-associated tissue has also been implicated as a primary lymphoid area in the mature and neonatal animal. Nieuwenhuis and co-workers (159,160), examining postirradiation regeneration of B follicular structures in the rabbit, concluded that germinal centers in the appendix were capable of producing virgin B cells with diverse antigen specificities. Reynolds and Morris (183) have presented compelling evidence that PP may serve as a primary lymphoid organ in the fetal and postnatal sheep. Lymphopoiesis occurred in the PP follicles in the absence of antigen during the perinatal period and continued postnatally in both the follicles and interfollicular areas. At 8 weeks of age, the PP extended along 17% of the small intestine and contained the greatest density of proliferating lymphoid cells found in the animal (183). Therefore, the unique nature of the mucosal immune system is exemplified by its ability to serve as a primary lymphoid area (in some animals) and by its ability to shield entirely separate cell populations from the systemic immune system. The available information supports the existence of not only distinct populations of isotype-specific T and B cells, but also of cell subpopulations expressing unique receptor/idiotype repertoires within the mucosa (with further evidence below). Whether this is a result of the gut environment with its incessant antigen exposure or regulatory interactions remains to be determined.

E. AGINGAND THE MUCOSALIMMUNE SYSTEM Unfortunately few studies have examined changes within the mucosal immune system during aging or have compared these changes with those in the systemic immune system (reviewed in 220). Szewczuk et al. (221) have examined the immune response in the spleen, peripheral lymph node (PLN), mediastinal lymph node (BLN), and mesenteric lymph node (MLN) of aged mice after intraperitoneal (ip) immunization with TNP-BGG in complete Freund’s adjuvant (CFA). It was found that while systemic (spleen and PLN) IgM, IgG, and IgA PFC responses declined with age, mucosal PFC responses (MLN and BLN) remained vigorous in cells bearing all three isotypes. This unimpaired mucosal response was also observed by Rivier et al. (186), who found that IgA responses to the (~(1-3) glucan determinant on dextran €31355 increased dramatically with age in the MLN. In this study, however, the splenic IgA immune response did not decline over the age span, but this may be a result of testing the BALB/c mice used up to only 17 months of age (middle age).


179

Further work from our laboratory (235) has shown that the agerelated PFC response dichotomy between splenic and mucosal-associated tissues can also be observed upon secondary ip stimulation of aged C57BL/6J male mice with either TNP-BGG or TNP-KLH. The findings indicate that while splenic PFC responses decline with age, the immune response of cells within the mucosal-associated lymph nodes of aged mice may remain unaffected by the down-regulatory mechanisms induced with age. When animals were immunized intragastrically (ig) with TNP-BGG (221,235)or TNP-KLH (235) PFC responses in both the spleen and MLN increased significantly with age. Splenic PFC responses of old animals, after ig immunization, were greater than those induced by ip immunization (235), and preliminary results (Wade and Szewczuk, unpublished observations) indicate that HA PFC were not induced in systemic or mucosal tissues of aged mice. These findings raise some interesting questions regarding the nature of the mucosal immune system. (a) Why do mucosal immune responses not decline with age? (b) Why are HA PFC not generated in either the spleen or MLN of old mice following ig immunization? (c) Do the functional cell repertoires change in this area with age as they have been shown to do in the spleen? Some answers may be provided by the following observations. We have previously discussed (Section II1,C) the results of adoptive transfer experiments showing that only splenic B and T cells from old mice were able to cooperate to produce HA PFC in an immune response. Recipients not receiving this combination of B and T cells were unable to mount such a response (214). Presumably, idiotype identity in the B and T cell populations is required for the generation of HA PFC in this system. Could our inability to generate HA PFC in the spleen of aged mice following ig immunization reflect a lack of idiotype identity between the mucosally stimulated lymphoblasts and resident spleen cells in this system? If we assume for a moment that the idiotype repertoire of cells within the mucosa (unlike systemic immune system) undergoes nonparallel changes with age, then ig immunization would stimulate mucosal cells expressing a different set of idiotypes than those expressed in the spleen. Upon trafficking of these activated blast cells to the spleen, there would be minimal recognition of these “new” idiotypic determinants and antiidiotypic antibody (HA PFC) would not be generated to any great degree. If we can speculate further, the exposure of mucosally stimulated cells to the splenic environment during a primary (ig) response might not allow sufficient time for the induction

180


and expression of any suppressive mechanisms against these cells in the aged animal. One would predict that once such (idiotype-specific) suppression was generated, it would also be effective against cells (idiotype positive) within the mucosa. This prediction is in agreement with our findings that the secondary ig PFC response declines with age in both the spleen and MLN of aged animals in response to both TNP-BGG and TNP-KLH (235).Unlike the primary (ig) response, the secondary (ig) response was found to yield elevated HA PFC numbers in the spleen of old mice (Wade and Szewczuk, unpublished observations). Although highly speculative at this time, the hypothesis that the idiotype repertoires of systemic and mucosal cells do not undergo parallel, age-associated changes is supported by the following findings. (a) The heterogeneity of a primary (ip) anti-TNP PFC response in the spleen and PLN becomes restricted during aging, with a preferential loss of high affinity clones-this heterogeneity restriction was not identified in mucosal tissues of the same mice (218); (b) intragastric boosting of old mice with TNP-BGG or TNP-KLH fails to activate memory cells generated after ip priming-in contrast to that found in young mice (235);and (c) the appearance of antiidiotype blocked, HA PFC in the spleen and PLN of aged mice following a primary (ip) immune response to TNP-BGG, was not demonstrated in the MLN or BLN of aged animals (219). In this latter study, Szewczuk and Campbell (219) have shown that antigen-specific B cells within systemic areas of old mice are recognized, bound, and reversibly inhibited by antiidiotypic antibody on their surface, whereas no such antibody could be demonstrated on the surface of mucosal lymphocytes from these same animals. The findings therefore strongly indicate that the idiotype repertoires expressed by systemic and mucosal lymphocytes of aged mice are different and nonoverlapping. Clearly, the inductive signals or repertoire changes responsible for the development of augmented systemic antiidiotypic regulation during senescence (66,68,122,217,219)are either not triggered, or not effective in the mucosal immune system. In support of this hypothesis, we have presented evidence that (a) the stem cell population of old mice retains the potential to express indistinguishable functional and idiotype repertoires as those of young mice (Section 111); (b) the mucosal immune system is comprised of specific antigen-activated, recirculating, B and T lymphocytes which participate in unique immune responses (Sections IV,AD above); and (c) highly distinct populations of T and B lymphocytes


181

exist within the GALT, and not systemic areas of xid mice (Section

IV,D). Although the available information supports our contention that the repertoires expressed on systemic and mucosal lymphocytes are nonoverlapping by old age, clarification of these events and the mechanisms responsible must await further work. Returning for a moment to the discussion of whether aging changes are a result of intrinsic or extrinsic factors, the discovery that both stem cells and mucosal immune responses remain essentially unchanged throughout the murine lifespan would argue strongly that reductions described for the immune parameters of cells within the systemic immune system may well result from factors extrinsic to the cells themselves. The idea that the immune system of the gut remains vigorous and compartmentalized with age is exciting, since this area presents the first line of defense against many antigens encountered in the environment and would therefore protect aging individuals whose systemic immune responses have declined considerably. The finding of two distinct immunological compartments in the body, one with an impaired and the other with an unimpaired immune circuitry, would allow further insights into immune regulation, its change with age, and eventually might permit a possible means of boosting systemic immune response in aged individuals with disseminated infections or cancer.

ACKNOWLEDGMENTS This work was supported in part by grants-in-aid of research from the Medical Research Council of Canada, MA-7347, the Muscular Dystrophy Association of Canada and the Gerontology Research Council of Ontario.

REFERENCES 1. Abe, T., Morimoto, C., Toguchi, T., Kiyotaki, M., and Homma, M. (1981). Scand.J. Immunol. 13, 151. 2. Albright, J. W., and Makinodan, T. (1976). J. E r p . Med. 144, 1204. 3. Amagai, T., Nakano, K., and Cinader, B. (1982). Scand. J . Immunol. 16,217. 4. Anderson, J., Coutinho, A., and Melchers, F. (1977).J . E r p . Med. 145, 1511. 5. Andre, C., Heremans, J. F., Vaennan, J. P., and Cambiaso, C. L. (1975).J . Exp. Med. 142,1509. 6. Antonaci, S., Jirillo, E., Lucivero, G., Gallitelli, M., Garofalo, A. R., and Bonomo, L. (1983). Clin. E x p . Immunol. 52, 387. 7. Arnold, B., Wallich, R., and Hammerling, G. J. (1982).J . Exp. Med. 156, 670. 8. Asherson, G. L., Zembala, M., Perera, M. A. C. C., Mayhew, B., and Thomas, W. R. (1977). Cell. Immunol. 33, 145.

182


9a. Astle, C. M., and Harrison, D. E. (1984).J. Immunol. 132, 673. 9b. Befus, D., Pearce, F. L., Gauldie, J., Hosewood, P., and Bienenstock, J. (1982). J . Immunol. 128,2475. 10. Bentley, D. M., and Morris, R. E. (1982).J . Immunol. 128, 530. 11. Bevan, M. J. (1981). Immunol. Todag 2,216. 12. Bienenstock, J. (1982). In “Recent Advances in Mucosal Immunity” (W. Strober, L. A. Hanson, and K. W. Sell, eds.), p. 35. Raven, New York. 13. Bienenstock, J., and Befus, A. D. (1980).J . Immunol. 41, 249. 14. Bienenstock, J., Johnston, N., and Perey, D. Y. E. (1973). Lab. Inoest. 28,686. 15. Bienenstock, J., Johnston, N., and Perey, D. Y. E. (1973). Lab. Inoest. 28, 693. 16. Bienenstock, J., Befus, A. D., McDermott, M., Mirski, S., Rosenthal, K., and Tagliabue, A. (1983).Ann. N.Y. Acad. Sci. 409, 164. 17. Bona, C., and Caasenave, P. A. (1980). “Lymphocytic Regulation by Antibodies.” Wiley, New York. 18. Bona, C., and Hiernaux, J. (1981). Crit. Reo. Immunol. 2, 33. 19. Boyd, E. (1932).Am. J . Dis. Child. 43, 1162. 20. Butcher, E. C., Rouse, R. V., CofEman, R. L., Nottenburg, C. N., Hardy, R. R., and Weisman, 1. C. (1982).J . Immunol. 129, 2698. 21. Butcher, E. C., Kraal, G., Stevens, S. K., and Weissman, I. L. (1982). Ado. E x p . Biol. 149, 199. 22. Callard, R. E. (1978). Eur. J. Immunol. 8,697. 23. Callard, R. E., and Basten, A. (1977). Cell. Immunol. 31, 13. 24. Callard, R. E., and Basten, A. (1978). Eur. J . Immunol. 8,552. 25. Callard, R. E., Basten, A., and Waters, L. K. (1977). Cell. Immunol. 31, 26. 26. Callard, R. E., Basten, A., and Blanden, R. V. (1979).Nature (London) 281, 218. 27. Callard, R. E., Fazekas de St Groth, B., Basten, A., and McKenzie, I. F. C. (1980). J . Immunol. 124, 52. 28. Carter, P. B., and Collins, F. M. (1974).J. Exp. Med. 139, 1189. 29. Ceuppens, J. L., and Goodwin, J. S. (1982).J . Immunol. 128,2429. 30. Challocombe, S. J., and Tomasi, T. B., Jr. (1980).J . Exp. Med. 152, 1459. 31. Chang, M. P., Makinodan, T., Peterson, W. J., and Strehler, B. L. (1982).J . Immunol. 129,2426. 32. Chin, Y.-H., Carey, G. D., and Woodruff, J. J. (1982).J . Immunol. 129, 1911. 33. Clark, R. B., Chiba, J., Zwerg, S. E., and Sharach, E. M. (1982). Nature (London) 295,412. 34. Cinader, B. (1983). Clin. Biochem. 16, 121. 35. Cosenza, H. (1976). Eur. J. Immunol. 6, 114. 36. De Freitas, A. A., Rose, M. L., and Parrott, D. M. V. (1977). Nature (London) 270, 731. 37. De Freitas, A. A,, Rose, M., and Rocha, B. (1980). Cell. Immunol. 56,29. 38. Dekruyff, R. H., Kim, Y. T., Siskind, G. W., and Weksler, M. E. (1980).J. Immunol. 125, 142. 39. Doggett, D. L., Chang, M. P., Makinodan, T., and Strehler, B. L. (1981).Mol. Clin. Biochem. 37, 137. 40. Dolezel, J., and Bienenstock, J. (1971). Cell. Immunol. 2,458. 41. Dolezel, J., and Bienenstock, J. (1971). Cell. Immunol. 2, 326. 42. Doria, G., D’Agostaro, G., and Garavini, M. (1980). Cell. Immunol. 53, 195. 43. Dupere, S. L. R., and Kolodziej, B. J. (1983).Age 6, 11. 44. Durkin, H. G., Bazin, H., and Waksman, B. H. (1981).J. Exp. Med. 154, 640. 45. Duwe, A. K., Roder, J. C., and Singhal, S. K. (1979). Immunology 37,293.


183

46. Eichmann, K. (1978). Adv. Immunol. 26, 195. 47. Eldridge, J. H., Kiyono, H., Michalek, S.M., and McGree, J. R. (1983).J.E x p . Med. 157. 789. 48. Elson, C. O., Heck, J. A., and Strober, W. (1979).J. Exp. Med. 149,632. 49. Farrar, J. J., Loughman, B. E., and Nordin, A. A. (1974). J . Immunol. 112, 1244. 50. Ferguson, A., and Parrott, D. M. V. (1972). Clin. Exp. Immunol. 12,477. 51. Ferrandez, C., and Moller, G. (1980). Scand. J . Immunol. 11, 53. 52. Flood, P. M., Urban, J. L., Kripke, M. L.,and Schreiber, H. (1981).J . Exp. Med. 154,275. 53. Fossum, S., Smith, M. E., and Ford, W. L. (1983). Scand. J . Immunol. 17,539. 54. Fossum, S., Smith, M. E., and Ford, W. L. (1983). Scand. J . Immunol. 17,551. 55. Frasca, D., Garavini, M., and Doria, G. (1982). Cell. Immunol. 72, 384. 56. Friedberger, E., Bock, G., and Furstenheim, A. (1929). Z. Immunforsch. Exp. Ther. 64, 294. 57. Frol’kis, B. K., Bezrukov, V. V., and Muradyen, K. K. (1979).E x p . Gerontol. 14,77. 58. Gallatin, W. M., Weissman, I. L., and Butcher, E. C. (1983). Nature (London) 304,30. 59. Gearhart, P. J., and Cebra, J. J. (1979).J . E r p . Med. 149,216. 60. Geha, R. S. (1983).J. Immunol. 130(4), 1634. 61. Geha, R. S. (1982).J. Immunol. 129(1), 139. 62. Gillis, S., Kozak, R., Durante, M., and Weksler, M. E. (1981). J . Clin. Inuest. 67, 937. 63. Gilman, S. C., Woda, B. A., and Feldman, J. D. (1981).J . Immunol. 127, 149. 64. Goidl, E. A., Innes, J. B., and Weksler, M. E. (1976).J. E x p . Med. 144, 1037. 65. Goidl, E. A,, Schrater, A. F., Siskind, G. W., and Thorbecke, G. J. (1979).J . Exp. Med. 150,154. 66. Goidl, E. A,, Thorbecke, G. J., Weksler, M. E., and Siskind, G. W. (1980). Proc. Natl. Acad. Sci. U S A . 77, 6788. 67. Goidl, E. A,, Schrater, A. F., Thorbecke, G. J., and Siskind, G. W. (1980). Eur. J. Immunol. 10, 810. 68. Goidl, E. A., Choy, J. W., Gibbans, J. J., Weksler, M. E., Thorbecke, G. J., and Siskind, G. W. (1983).J . E x p . Med. 157, 1635. 69. Goidl, E. A., Hagama, T., Shephard, G. M., Siskind, G. W., and Thorbecke, G. J. (1983).J . Immunol. Methods 58, 1. 70. Good, R.A., and Yunis, E. J. (1974). Fed. Proc. Fed. Am. SOC. E x p . Biol. 33,2040. 71. Goodwin, J. S., Searles, R. P., and Tung, K. S. K. (1982). Clin. E x p . Immunol. 48, 403. 72. Gorczynski, R. M., Kennedy, M., and MacRay, S. (1983). Cell. Immunol. 75,226. 73. Gorczynski, R. M., Kennedy, M., MacRae, S., Benzing, K., and Price, G. B. (1984). Zmmunopharm., in press. 74. Griscelli, C., Vassalli, P., and McCluskey, R. T. (1969).J . Exp. Med. 130, 1427. 75. Gutman, G. A., and Weisman, I. L. (1973). Transplantation 16,621-629. 76. Gutowski, J. K., Innes, J., Weksler, M. E., and Cohen, S. (1984). J. Immunol. 132,559. 77. Hanson, D. G., and Miller, S. D. (1982).J . Immunol. 128, 2378. 78. Harrison, D. E. (1983).J . E x p . Med. 157, 1496. 79. Harrison, D. E., and Doubleday, J. W. (1975).J. Immunol. 114, 1314. 80. Harrison, D. E., and Astle, C. M. (1982).J . E x p . Med. 156, 1767. 81. Harrison, D. E., Astle, C. M., and Doubleday, J. W. (1977).J. Immunol. 118,1223. 82. Harrison, D. E., Astle, C. M., and Delaittre, J. A. (1978).J . E x p . Med. 147, 1526.

184


Harrison, D. E., Archer, J. R., and Astle, C. M. (1982).J . Immunol. 129, 2673. Hayflick, L. (1968). Sci. Am. 281, 32. Hayflick, L. (1973).Am. J . Med. Sci. 265,432. Hayflick, L. (1975). E r p . Cell Res. 371, 614. Hefton, J. M., Dutkowski, R., Darlington, G. J., and Weksler, M. E. (1983). Science 219,1335. 88. Hellman, S., Bottnick, L. E., Hannon, E. C., and Vigneulle, R. M. (1978). Proc. Natl. Acad. Sci. U.S.A. 75,490. 89. Hercend, T., Reinherz, E. L., Meuer, S., Schlossman, S. F., and Ritz, J. (1983). Nature (London) 301, 158. 90. Hendriks, H. R., and Eestermans, I. L. (1983). Eur. J . Immunol. 13, 663. 91. Hirokawa, K., and Makinodan, T. (1975).J. Immunol. 114, 1659. 92. Hirokawa, K., Sato, K., and Makinodan, T. (1982).Clin. Immunol. Immunopathol. 22, 297. 93. Holliday, R., Huschtscha, L. I., Tarrant, G. M., and Kirkwood, T. B. L. (1977). Science 198, 366. 94. Hooijkaas, H., Pressman, A. A., van Oudenaren, A., Benner, R., and Haaijman, J. J. (1983).J . Immunol. 131, 1629. 95. Howard, M., Mizel, S. B., Lachman, L., Ansel, J., Johnston, B., and Paul, W. E. (1983).J . E x p . Med. 157, 1529. 96. Husband, A. J., and Gowan, J. L. (1978).J. E r p . Med. 148, 1146. 97. Husband, A. J. (1982).J. Immunol. 128, 1355. 98. Inkeles, B., Innes, J. B., Kuntz, M. M., Kadish, A. S., and Weksler, M. E. (1977). J . E r p . Med. 145, 1176. 99. Jackson, S., and Mestecky, J. (1979).J . Exp. Med. 150, 1265. 100. Jackson, S., and Mestecky, J. (1981). Cell. Immunol. 60,498. 101. Jaroslow, B. N., Suhrbier, K. M., Fry, R. J. M., and Tyler, S. A. (1975).J . Natl. Cancer Inst. 541, 1427. 102. Jayaraman, S., and Bellone, C. J. (1982).Eur. J. Immunol. 12,272. 103. Jerne, N. K. (1974).Ann. Immunol. 125,378. 104. Kagnoff, M. F. (1975).J . E r p . Med. 142, 1425. 105. Kagnoff, M. F. (1978). Cell. Immunol. 40, 186. 106. Kagnoff, M. F., and Campbell, S. (1974).J . E x p . Med. 139, 398. 107. Kawanishi, H., Saltzman, L. E., and Strober, W. (1982).J . Immunol. 129,475. 108. Kawanishi, H., Saltzman, L. E., and Strober, W. (1983).J . Exp. Med. 157,433. 109. Kawaniski, H., Saltzman, L., and Strober, W. (1983).J. E x p . Med. 158,649. 110. Kawaniski, H., and Strober, W. (1983). Mol. Immunol. 20, 917. 111. Kay, M. M. B. (1979).J. Inuest. Dermatol. 73, 29. 112. Kay, M. M. B. (1979). Clin. Immunol. Immunopathol. 12, 301. 113. Kay, M. M., Mendoza, J., Diven, J., Denton, T., Union, N., andLajiness, M. (1979). Mech. Ageing Deu. 11, 295. 114. Kelsoe, G., Reth, M., and Rajewski, K. (1980). Immunol. Rev. 52, 75. 115. Kelsoe, G., Reth, M., and Rajewski, K. (1981).Eur. J . Immunol. 11,418. 116. Kincade, P. W. (1981).Adu. Immunol. 31, 177. 117. Kincade, P. W., Lee, B., Paige, C. J., and Schaid, M. P. (1981). J . Immunol. 127,255. 118. Kishimoto, S., Takahama, T., and Mizumachi, H. (1976).J. Immunol. 116,294. 119. Kiyono, H., McGhee, J. R., Wannemuehler, M. J., Frangakis, M. V., Spalding, D. M., Michalek, S. M., and Koopman, W. J. (1982). Proc. Nutl. Acad. Sci. U.S.A. 79,596. 83. 84. 85, 86. 87.


185

120. Kiyono, H., McGhee, J. R., Mosteller, L. M., Eldridge, J. H., Koopman, W. J., Kearney, J. F., and Michalek, S. M. (1982).J . Exp. Med. 156, 1115. 121. Kiyono, H., Mosteller, L. M., Eldridge, J. H., Michalek, S. M., and McGhee, J. R. (1983).J. Immunol. 131, 2616. 122. Klinman, N. R. (1981).J. Exp. Med. 154, 547. 123. Komisar, J. L., Fuhrman, J. A., and Cebra, J. J. (1982).J . Immunol. 128, 2376. 124. Kraal, G., Weissman, I. L., and Butcher, E. C. (1983).J . Immunol. 130, 1097. 125. Krogsrud, R. L., and Perkins, E. H. (1977).J . Immunol. 118, 1607. 126. Lance, E. M., and Taub, R. N. (1969). Nature (London) 221,841. 127. Larsson, E., Coutinho, A., and Martinez-A,, C. (1980). Zmmunol. Reu. 51,61. 128. Liu, J. J., Segre, M., and Segre, D. (1982). Cell. Immunol. 66, 372. 129. MacDonald, T. T. (1983). Eur. J . Zmmunol. 13, 138. 130. Makinodan, T., and Peterson, W. T. (1962). Proc. Natl. Acad. Sci. U.S.A. 48, 234. 131. Makinodan, T., and Kay, M. M. B. (1980). Adu. Immunol. 29, 287. 132. Makinodan, T., Chino, F., Lever, W. E., and Brewen, B. S. (1971).J . Gerontol. 26, 508. 133. Makinodan, T., Chino, F., Lever, W. E., and Brewen, B. S. (1971). Gerontology 26,515. 134. Makinodan, T., Heidrich, M. C., and Nordin, A. A. (1975). In “Immunodeficiency and Autoimmunity in Aging” (D. Bengsina, ed.), pp. 193-198. Sinauer, Stamford, Connecticut. 135. Makinodan, T., Albright, J. W., Good, P. I., Peter, C. P., and Heidrick, M. C. (1976). Immunology 31, 903. 136. Mark, D. H., and Weksler, M. E. (1982).J. Immunol. 129, 2323. 137. Marsh, M. N. (1975). Gut 16,665. 138. Marsh, M. N. (1975). Gut 16,674. 139. Mascart-Lemone, F., Delespesse, G., Servais, G., and Kunstler, M. (1982). Clin. Exp. Immunol. 48, 148. 140. Matsuzawa, T., and Cinader, B. (1983a). Gerontology 29,41. 141. Matsuzawa, T., and Cinader, B. (1983b). Gerontology 29,49. 142. Mattingly, J. A., and Waksmann, B. H. (1978).J . Immunol. 121, 1878. 143. Mayrhofer, G., and Whately, R. J. (1983). Int. Arch. Allergy Appl. Immunol. 71, 317. 144. McDermott, M. R.,and Bienenstock, J. (1979).J . Immunol. 122, 1892. 145. McDermott, M. R.,Clark, D. A., and Bienenstock, J. (1980). J . Immunol. 124, 2536. 146. McWilliams, M., Pillips-Quagliata, J. M., and Lamm, M. E. (1975).J . Zmmunol. 115,54. 147. Micklem, H. S., and Ross, E. (1978). Ann. Immunol. 129C, 367. 148. Miller, R. A. (1984).J. Immunol. 132, 63-68. 149a. Miller, R. A., and Stutman, 0. (1981). Eur. J . Immunol. 11, 751. 149b. Miller, R. A., and Stutman, 0. (1982). Lymphokine Res. 1,79. 150. Morgan, E. L., and Weigle, W. 0. (1982).J . Immunol. 129,36-39. 151. Morgan, E. L., Thoman, M. L., and Weigle, W. D. (1981). Cell. Immunol. 63, 16-27. 152. Mosier, D. E., and Subbarou, B. (1982). Immunol. Today 3,217-222. 153. Mowat, A. Mc.I., Tait, R. C., MacKenzie, S.,Davies, M. D. J., and Parrott, D. M. V. (1983). Clin. Exp. Zmmunol. 52, 191-198. 154. Muller-Schoop, J. W., and Good, R. A. (1975).J. Immunol. 114,1757-1760. 155. Nagel, J. E., Chrest, F. J., and Adler, W. H. (1981).J . Immunol. 127, 2086-2088.

186


Nakano, K., and Cinader, B. (1980).J . Immunogenet. 7, 183-190. Naor, D., Bonavida, B., and Walford, R. L. (1976).J. Immunol. 117,2204-2208. Ngan, J., and Kind, L. S. (1978).J . Immunol. 120,861-865. Niewenhuis, P., and Kauning, F. J. (1974). Immunology 26,509. Niewenhuis, P., van Nouhuys, C. E., Eggens, J. H., and Keuning, F. J. (1974). Immunologg 26,497. 161a. Niewenhuis, P. (1981). Immunol. Today 2, 104. 161b. Nordin, A. A,, and Collins, G. D. (1983).J . Immunol. 131,2215. 162. Ogden, D. A., and Micklem, H. S. (1976). Transplantation 22,287-293. 163. Ortaldo, J. R., Sharrow, S. O., Timonen, T., and Herberman, R. B. (1981).J . Immunol. 127,2401-2409. 164. Osmond, D. G. (1982).Adu. Erp. Med. Biol. 149,3-10. 165. Ottaway, C. A., and Parrott, D. M. V. (1980). Immunology 41,955-961. 166. Owen, R. L., and Jones, A. L. (1974). Gastroenterology 66, 189-203. 167. Patel, P. J. (1981).J. Erp. Med. 154, 821-831. 168. Pawelec, G., Schneider, E. M., Rehbein, A., Schaa, I., and Wernet, P. (1983). Scand. J . Immunol. 17, 147-153. 169. Parrott, D. M. B. (1976). Clin. Gastroenterol. 5,211-228. 170. Parrott, D. M. V., and Wilkinson, P. C. (1981). Prog. Allergy 28, 193-284. 171. Pearce, F. L., Befus, A. D., Gauldie, J., and Bienenstock, J. (1982).J . Immunol. 128,2481-2486. 172. Pepys, M. B. (1983). In “Virus Infections of the Gastrointestinal Tract” (D. A. J. Tyrrel and A. Z. Kapikian, eds.), pp. 89-110. Dekker, New York. 173. Phair, J. P. (1979).J . Chron. Dis.32, 535-540. 174. Pierce, N. F., and Gowans, J. L. (1975).J . Exp. Med. 142, 1550-1563. 175. Pierce, N. F., and Koster, F. T. (1980).J . Immunol. 124, 307-311. 176. Pierce, N. F., and Cray, W. C., Jr. (1981).J . Immunol. 127, 2461-2464. 177. Pierce, N. F., and Cray, W. C., Jr. (1982).J . Immunol. 128, 1311-1315. 178. Pike, M. C., Krails, M. D., Henderson, B. E., Casagrand, J. T., and Hoel, D. G. (1983). Nature (London) 303,767-770. 179. Price, G. B., and Makinodan, T. (1972).J . Immunol. 108,403-412. 180. Price, G. B., and Makinodan, T. (1972).J . Immunol. 108,413-417. 181. Rajewsky, K., and Takemori, T. (1983).Annu. Reu. Immunol. 1,569-607. 182. Reth, M., Kelsoe, G., and Rajewsky, K. (1981). Nature (London) 290,257. 183. Reynolds, J. D., and Morris, B. (1983). Eur. J . Immunol. 13,627. 184. Richman, L. K., Chiller, J. M., Brown, W. R., Harrison, D. G., and Vaz, N. M. (1978).J. Immunol. 121,2429-2434. 185. Richman, L. K., Graeff, A. S., Yarchoan, R., and Strober, W. (1981).J . Immunol. 126,2079. 186. Rivier, D. A., Trefts, P. E., and Kagnoff, M. F. (1983). Scand. J . Immunol. 17, 115-121. 187. Roberts-Thomson, I. C., Whittingham, S., Youngchaiyud, U., and MacKay, I. R. (1974). Lancet 2,368-370. 188. Roder, J. R., Duwe, A. K., Bell, D. A., and Singhal, S. K. (1978). lmmunology 35, 837-848. 189. Rodkey, L. S., and Adler, F. L. (1983).J. Exp. Med. 157, 1920. 190. Rose,.L..M., Goldman, M., and Lambert, P.-H. (1982).J . Immunol. 128, 21262133. 191. Rosenberg, J. S., Gilman, S. C., and Feldman, D. J. (1982). J . Immunol. 128, 656-660.

156. 157. 158. 159. 160.


187

192. Rosenberg, J. S., Gilman, S. C., and Feldman, J. D. (1983).J . Zmmunol. 130,17541758. 193. Rosenthal, A. S. (1978). Zmmunol. Reu. 40, 135. 194. Ross, E. A., Anderson, N., and Micklem, H. S. (1982).J. E x p . Med. 155,432-444. 195. Rothberg, R. M., Kraft, S. C., and Farr, R. S. (1967).J. Zmmunol. 98, 386-395. 196. Santisteban, G. A. (1960). Anat. Res. 136, 117-126. 197. Schrader, J. W., Scollay, R., and Battye, F. (1983).J . Zmmunol. 130, 558-564. 198. Scribner, D. J., and Moorhead, J. W. (1982).J . Zmmunol. 128, 1377-1380. 199. Segre, M., and Segre, D. (1976).J. Zmmunol. 116, 731-734. 200. Segre, D., and Segre, M. (197q.J. Zmmunol. 116,735-738. 201. Shevach, E. M., and Rosenthal, A. S. (1973).J. E x p . Med. 138, 1213. 202. Sigal, N. H., and Klinman, N. R. (1978). Adu. Zmmunol. 26, 255. 203. Sminia, T., Wilders, M. M., Janse, E. M., and Hoelsmit, E. C. M. (1983). Zmmunobiology 164, 136-143. 204. Smith, A. M. (1976).J . Zmmunol. 116,469-474. 205. Smith, G. S., and Walford, R. L. (1977). Nature (London) 270, 727-729. 206. Smith, M. E., and Ford, W. L. (1983). Zmmunology 49,83-94. 207. Smith, M. E., and Ford, W. L. (1983). Cell. Zmmunol. 78, 161-173. 208. Sohnle, P. G., Larson, S. E., Collins-Lech, C., and Guansing, A. R. (1980). J . Zmmunol. 124, 2169. 209. Spalding, D. M., Koopman, W. J., Eldridge, J. H., McGhee, J. R., and Steinman, R. M. (1983).J. Exp. Med. 157, 1646-1659. 210. Spangrude, G. J., Braaten, B. A., and Daynes, R. A. (1984). J . Immunol. 132, 354. 211. Stevens, S. K., Weissman, I. L., and Butcher, E. C. (1982). J. Zmmunol. 128, 844-851. 212. Stobo, J. D., and Tomasi, T. B. (1975).J. Clin. Dis. 28, 437-440. 213. Suemura, M., Urban, J. F., and Ishizaka, K. (1978).J. Zmmunol. 121,2413-2421. 214. Szewczuk, M. R. (1982). C a n . ] . Aging 1 and 2,3-10. 215. Szewczuk, M. R. (1983). Cell. Immunol. 82,282. 216. Szewczuk, M. R. (1984). Cell. Zmmunol. 84, 393. 217. Szewczuk, M. R., and Campbell, R. J. (1980). Nnture (London) 286,164-166. 218. Szewczuk, M. R., and Campbell, R. J. (1981).J. Zmmunol. 126,472-477. 219. Szewczuk, M. R., and Campbell, R. J. (1981). Eur.J. Zmmunol. 11,650-656. 220. Szewczuk, M. R., and Wade, A. W. (1983).Ann. N.Y. Acad. Sci. 409,333-344. 221. Szewczuk, M. R., Campbell, R. J., and Jung, L. K. (1981).J. Zmmunol. 126,22002204. 222. Tagliabue, A., Luiri, W., Soldateochin, D., and Boraschi, D. (1981). Eur. J. Zmmunol. 11,919-922. 223. Tagliabue, A., Befus, A. D., Clark, D. A., and Bienenstock, J. (1982).J. E x p . Med. 155, 1785. 224. Taylor, R. B. (1982). Zmmunol. Today 3,47-51. 225. Thoman, M. L., and Weigle, W. 0. (1981).J. Zmmunol. 127, 2102-2106. 226. Thoman, M. L., and Weigle, W. 0. (1982).J. Zmmunol. 128, 2358-2361. 227. Thoman, M. L., and Weigle, W. 0. (1983).J. Exp. Med. 157, 2184. 228. Thomsen, O., and Kettel, K. (1929). 2. Immunforsch. E x p . Ther. 63, 67. 229. Tilney, N. L. (1971).J . Anat. 109, 369-383. 230. Tseng, J. (1981).J. Immunol. 127, 2039-2043. 231. Tseng, J. (1982).J . Zmmunol. 128, 2719-2725. 232. Tseng, J. (1982). Cell. Zmmunol. 73, 324-336.

188


233. Urbain, J., Wuilmart, C., and Cazenave, P. A. (1981). Contemp. Top. Mol. lmmunol. 8, 113. 234. van Ewijk, W., Brons, N. H. C., and Rozing, J. (1975).Cell. Zmmunol. 19,245-261. 235. Wade, A. W., and Szewczuk, M. R. (1984). Submitted. 236. Wade, A. W., and Szewczuk, M. R. (1984).Submitted. 237. Wagner, A. P., Wagner, L. P., and Psarrou, E. (1982).Age 5, 113-117. 238. Waksman, B. H., and Ozer, H. (1976). Prog. Allergy 21, 1-113. 239. Walford, R. L. (1974).Fed. Proc. Fed. Am. SOC. Exp. Biol. 33,2020. 240. Walker, W. A., and Isselbacker, K. J. (1974).Gastroenterology 67,531-550. 241. Walker, S. M., and Weigle, W. 0. (1981).]. Exp. Med. 153, 653-664. 242. Wallace, D. J., Bluestone, R., and Klironberg, J. R. (1982). Bull. Rheum. Dis. 32, 13-19. 243. Watson, J. (1979).J. Exp. Med. 150, 1510-1519. 244. Weiner, H. L., Moorhead, J. W., and Claman, H. W. (1976)J. lmmunol. 116,16561661. 245. Weksler, M. E. (1980).Proc. SOC. Erp. Biol. Med. 165, 200-205. 246. Weksler, M. E. (1981).Hosp. Pruct. 16, 53-64. 247. Weksler, M. E., Innes, J. B., and Goldstein, G. (1978).J. Exp. Med. 148,996-1006. 248. Wilders, M. M., Sminia, T., and Janse, E. M. (1983).lmmunology 50, 303. 249. Wilson, D. A., and Braley-Mullen, H. (1982). Cell. Immunol. 74, 72-85. 250. Winchurch, R. A., Birmingham, W., Hiberg, C., and Munster, A. (1982). Cell. lmmunol. 67,384-389. 251. Woda, B. A., and Feldman, J. D. (1979).J. E x p . Med. 149,416-423. 252. Woda, B. A,, Yguerabide, J., and Feldman, J. D. (1979).J . lmmunol. 123,21612167. 253. Worton, R. G., McCullock, E. A., and Till, J. E. (1969).J. Exp. Med. 130, 91-103. 254. Zatz, M. M., and Lance, E. M. (1970). Cell. lmmunol. 1, 3-17. 255. Zharhary, D., and Klinman, N. R. (1983).J . Exp. Med. 157, 1300-1308. 256. Zuberi, R. I., and Katz, D. H. (1982).J . Zmmunol. 129,272-277.

A Major Role of the Macrophage in Quantitative Genetic Regulation of Immunoresponsiveness and Antiinfectious Immunity GUIDO BIOZZI, DENISE MOUTON, CLAUDE STIFFEL, AND YOLANDE BOUTHILLIER

U.125 INSERM and ER 060070 CNRS, lnstitut Curie-Section

de Biologie,

Paris, France

I. Introduction ..................................................... 11. Genetic Regulation of Immunoresponsiveness ........................ A. Specific Antigen-Restricted Regulation ............................ B. General Regulation: Selection of High and Low Antibody Responder Lines ........................................................ C. Phenotypic Expression of the Selected Genes. ..................... 111. Modifications of Macrophage Functions .............................. A. Processing of the Selection Antigen .............................. B. Processing of Unrelated Antigens: Nonspecific Effect of the Selection . C. Bactericidal Activity. ........................................... IV. Cell-Mediated Immunity in High and Low Antibody Responder Lines. . . . V. Modifications of Innate and Immune Resistance to Infections in High and Low Antibody Responder Lines .................................... A. Macrophage-Dependent Immunity ............................... B. Antibody-Mediated Immunity ................................... VI. Considerations on the Genetic Control of Antiinfectious Immunity. . . . . . . VII. Conclusion ...................................................... References ......................................................

189 192 192 193 196 197 198 202 209 215 218 219 224 226 23 1 232

1. Introduction

A century ago, Elie Metchnikoff (1884) discovered phagocytosis and recognized the importance of this phenomenon in the defense against infections. In 1908, the Nobel Prize in Physiology and Medicine was awarded jointly to Elie Metchnikoff and Paul Ehrlich who stressed the role of humoral antibodies in immunity. The award closed a longlasting controversy between the supporters of humoral or cellular factors in antiinfectious immunity, since it underlined the fundamental role of the two mechanisms and the importance of their cooperation in the defense against pathogenic microorganisms. Metchnikoff described two types of phagocytes, designated “microcytes” and “macrocytes.” In mammals, phagocytic cells are now referred to as polymorphonuclear leukocytes and macrophages. Whereas 189 Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.

ISBN 0-12-022436-4

190

GUIDO BIOZZI ET AL.

polymorphonuclear leukocytes play a chiefly peripheral role in antiinfectious defense, macrophages also operate as the central cells in the genetic regulation of immunoresponsiveness. It is clear today that the immune system is constituted by a coordinated network of perfectly integrated and interacting cells and molecules subject to strict cooperation in order to ensure the highest possible efficiency in antiinfectious immunity. A simplified scheme of the immune system in higher vertebrates is represented in Fig. 1. The enzyme equipment of macrophage phagosomes endows these cells with bactericidal or bacteriostatic activity on ingested microorganisms therefore constituting the first important mechanism in antiinfectious defense. The metabolic activity of macrophages on engulfed antigens also regulates the specific response of T and B lymphocytes through a complex process of antigen handling and antigen presentation, estab-

Immune system phagocytosis

multiplicotion and differentiation effector calls @@ cell mediated immunity

\ lymphokims

Q@

I

pbsmocytes

v+yantibodies humoral immunity

/

opson'ns

bactericidol effect

"Nothing in biology makes sense except in the light of evolution" Th. Dobzhansky

FIG. 1. Schematic representation of cellular and molecular interactions in the immune system of higher vertebrates.

MACROPHAGE I N REGULATION OF IMMUNORESPONSIVENESS

191

lishing a sort of symbiotic relationship between lymphocytes and macrophages (Unanue, 1972, 1981). The schematic representation of the immune system clearly indicates that there are two essential components in the immune response: one is specific and the other nonspecific. The specific response involves the stereospecific selective recognition, by molecular receptors on the surface of T and B small lymphocytes, of the antigen determinants presented by macrophages. This stereospecific recognition triggers the process of multiplication and differentiation producing the clone of the T-derived lymphocyte effectors of cell-mediated immunity, and the clone of the B-derived plasmocytes responsible for the production of humoral antibodies. The nonspecific aspect of the immune response includes the handling of the phagocytized antigen and the rate at which the process of multiplication and differentiation of small lymphocytes takes place. These phenomena are essentially nonspecific since all antigens, regardless of their specificity, must be phagocytized and processed by macrophages in order to trigger the process of multiplication and differentiation of specific T and B lymphocytes. We have included Th. Dobzhansky’s sentence in Fig. 1since, in our opinion, it is particularly relevant as far as the evolution of the immune system is concerned. I n fact the teleological function of the immune system, which has directed its evolution, is to grant the best protection to a genetically heterogeneous natural animal population against all types of infections. This protection is produced by three essential mechanisms which have emerged successively during evolution. The primeval mechanism, phagocytosis, is already present in such primitive organisms as amoeba. It is devoid of specificity and memory. The second mechanism, cell-mediated immunity, emerged early in invertebrates. It is endowed with memory and a definite degree of discriminating specificity. The humoral antibody response, the last mechanism produced by evolution in primitive fish, is also endowed with long-lasting memory. Different classes and subclasses of antibodies which have extremely refined specificity emerged rapidly afterward. Specific memory plays an extremely important role in immunity since it enables a quicker and stronger immune response to previously surmounted infections. The protective effect of specific vaccination is essentially based on immunological memory. The antibody molecules, according to their isotypes, play specialized defensive roles against various types of invading microorganisms, particularly in collaboration with the complement system, inducing bactericidal or opsonizing effects.

192

GUIDO BIOZZI ET AL.

These three essential immune functions became perfectly integrated and optimally coordinated in primitive vertebrates and presented a remarkable evolutive invariance for hundreds of millions of years thereafter. In fact, in its structure and functions, the immune system in primitive fish is fundamentally similar to that of the present mammals. It is obvious that the efficiency of antiinfectious protection provided by the immune system is a character of extremely high selective value. The selective value of a given character has an evolutive significance only at the level of a natural population constituted by genetically heterogeneous individuals. In view of its high selective value, the remarkable evolutive invariance of the immune system indicates that it is able to confer an optimum protection upon a natural population of animals against all types of different pathogens present in its ecological environment. The cause for this evolutive invariance could erroneously be attributed to the genetic homogeneity of animal species, at the level of the alleles regulating immunoresponsiveness, i.e., the high selective value of this character might have only retained the favorable alleles producing the strongest activity of the three operative functions of the immune system: phagocytosis, cell-mediated immunity, and antibody responsiveness. In fact, this hypothesis does not hold true since recent studies on the genetic regulation of immunoresponsiveness demonstrated the very large individual phenotypic variability of the principal immunologic parameters in the various mammalian species investigated. This large phenotypic variability may result from either a large allele polymorphism at the level of a single locus or from an additive effect of relevant alleles at several independent loci. Quantitative polygenic regulation has been demonstrated by our study on the genetic regulation of antibody responsiveness based on the selective breeding of high and low antibody responder lines of mice (Biozzi et al., 1979a, 1980; Ibanez et al., 1980). II. Genetic Regulation of lrnrnunoresponsiveness

A. SPECIFICANTIGEN-RESTRICTED REGULATION

The most important development toward the understanding of the mechanisms of evolution in the past 20 years has been the demonstration that each animal species constitutes an enormous reservoir of genetic variability. The average of structure gene heterozygosity estimated by electrophoretic mobility of protein molecules is 6.6% in vertebrates. The rate of genetic polymorphism in the mouse is close to

MACROPHAGE IN REGULATION OF IMMUNORESPONSIVENESS

193

29%of the gene pool. Each individual therefore presents a very large number of discrete differences in the primary structure of a large number of proteins. The immunogenicity of the antigenic determinants (epitopes) of a protein can be modified by small changes in the primary structure. These changes alter either continuous antigenic determinants formed by an array of contiguous amino acids, or discontinuous determinants formed by separated amino acids that are brought into proximity by the tertiary folding of the chain. As a result, the immune response to a given epitope in genetically heterogeneous individuals of the same species should present a large variability due to the phenomenon of cross-self-tolerance. Each single animal can respond only to antigenic determinants not present in the “self.” The high rate of polymorphism will therefore produce a large individual variability in the response to the same epitope. It is clear that the nonresponse due to cross-self-tolerance is specific for any given epitope. It is very likely that the genetic control of immunoresponsiveness operated by specific immune response genes (Ir genes) is due to a cross-self-tolerance mechanism. In fact, specific Ir genes control the immune response to synthetic polypeptides of restricted heterogeneity, to multideterminant natural proteins administered at threshold doses (when only the most potent antigenic determinant operates), or to alloantigens differing from those of the self only by details in the molecular structure (McDevitt and Benacerraf, 1969). The reason why the specific Ir genes are often linked to the Major Histocompatibility Complex is that this chromosomal region is characterized by an extremely large allelic polymorphism. It is obvious that the quantitative contribution of any structural locus to the regulation of individual specific immunoresponsiveness by cross-self-tolerance is directly related to its degree of allele polymorphism. The Ir genes control the specific immune responsiveness to epitopes through a complex mechanism of immunocyte interaction in which macrophages play an important role (Benacerraf, 1980; Rosenthal et al., 1980).

B. GENERAL REGULATION: SELECTION OF HIGHAND Low ANTIBODY RESPONDER LINES We have undertaken the study of the genetic regulation of immunoresponsiveiess to natural multideterminant immunogens by the method of bidirectional selective breeding of high and low antibody responder lines of mice. Five selections have been carried out so far using various antigens such as heterologous erythrocytes, bacteria (Salmonellae), or purified heterologous serum proteins (Biozzi et al.,

194

GUIDO BIOZZI ET AL.

1979a).The quantitative antibody response to all these natural immunogens is subject to polygenic regulation operated by the additive effect of alleles located at several independently segregating loci (Biozzi et al., 1980). The results reported in this review have been obtained in Selection I which has been the most extensively investigated. Selection I was initiated with a foundation population of 62 random bred albino mice. The bidirectional selective breeding was carried out for maximal or minimal agglutinin response to sheep erythrocytes for six generations. Afterward the selective breeding was continued, alternating noncross-reacting pigeon erythrocytes and sheep erythrocytes at each successive generation in order to avoid the interference of maternal antibodies passively transmitted to their progeny (Feingold et al., 1976). An optimal immunizing dose of heterologous erythrocytes was used and the phenotypic character “agglutinin titer” was measured 14 days after intravenous primary immunization when the interline difference in antibody response is maximal. The selective breeding produced a progressive divergence between the high and low responder lines during 16 consecutive generations. The maximal interline separation was then reached (selection limit) since the value of the interline difference in Fl6 could not be increased by the continuation of the selective breeding. The high and low lines at the selection limit were therefore considered as homozygous at the level of all the loci controlling quantitative antibody responsiveness. The modification in antibody response to sheep erythrocytes produced in Selection I is represented in Fig. 2 in which are shown the agglutinin responses to sheep erythrocytes in the foundation population, high and low responder lines at selection limit (F16-F36), interline F1 hybrids, Fz interline segregant crosses, and both backcrosses (F1 X high) and (F1 X low). In all these populations, the phenotype frequency distribution is close to a normal curve when the agglutinin titer is expressed as the log 2 of the highest serum dilution giving a positive agglutination. The selective breeding resulted in a very large modification in antibody responsiveness to sheep erythrocytes. The mean agglutinin titer of the high line is 230-fold higher than that of the low line. The mean response of F1 hybrids is intermediate between that of high and low lines with a small degree (0.27) of incomplete dominance of the high response character. The mean phenotypic variance due to environmental effects in the three genetically homogeneous populationshigh, low, and F1 is 1.21. The variance of the genetically heterogeneous foundation population and that of Fz segregants is larger


195

I /830

Faundotion

1/25

I15800 n

1/800

FI hybrlds 0

V-1.6

c

11400

F2 segngonts

fir

j,:

1/2700

11200

backcross

10

I

3

high

5

7

9

I1

1 3 1 5

ogglutinin titer (log 2 1

FIG.2. Frequency distribution of individual agglutinin titers in the foundation population, in high and low lines at selection limit, and in their hybrids. The theoretical curves characterizing each population were established using the mean and the standard deviation (SD) of individual agglutinin titers measured 14 days after iv immunization with 5 x loRsheep erythrocytes. V, Variance = SD2.

because it is produced by both genetic and environmental factors. It can be calculated that in a genetically heterogeneous natural mouse population such as the foundation population, 53% of the total phenotypic variability is due to genetic factors and 47% is produced by environmental effects. The mean value of the realized heritability of the character in the selective breeding is 0.20 0.08. Different methods of genetic analysis concorded in estimating that the quantitative antibody response to sheep erythrocytes is a polygenic trait regulated by the additive effect of about 10 independent loci (Biozzi et al., 197913). One of these loci is

*

196

GUIDO BIOZZI ET AL.

linked to the H-2 locus and another to the structure genes of immunoglobulins (Ig allotype) (Biozzi et at., 1979a).

C. PHENOTYPIC EXPRESSION OF THE SELECTED GENES The genes segregated in each line during the selective breeding are expressed at the level of the immune system itself. In fact the antibody response of spleen cell cultures stimulated in v i t r o with sheep erythrocytes is about 100-fold higher in the high than in the low line. The in v i t r o response of high or low line lymphocytes supplemented with macrophages of the opposite line indicates that the potentiality of both cell types is modified by the selected genes (Doria et al., 1978). In vivo cell transfer experiments summarized in Fig. 3 also demonstrate that both lymphocyte and macrophage activities are affected by the selective breeding. The results in Fig. 3A show that spleen cells from high donors produce a much more efficient restoration of antibody response in irradiated immunosuppressed mice than spleen cells from low donors. Since the immunorestoration is, due to the radiosensitive small lymphocytes, it is very likely that some of the alleles regulating anti-

FIG.3. Cell transfer experiment: restorationof responsiveness to sheep erythrocytes in immunosuppressed recipients (950 rads given 24 hours before transfer). (A) Kinetics of antibody production in irradiated outbred mice after iv injection of 4 x lo7 spleen cells from high or low donors 3 x lo8 sheep erythrocytes. (B) Kinetics of antibody production in high and low irradiated mice after iv injection of 4 x 107 spleen cells from (high x low) FI donors t 3 x lo8 sheep erythrocytes.

+


197

body responsiveness are expressed at the level of the lymphocyte. Macrophage activity is also greatly affected by the genetic selection as shown in Fig. 3B. This activity is known to be radioresistant. The same number of spleen cells from intermediate responder F1 hybrids produces a restoration of the antibody response to sheep erythrocytes which is much more efficient when transferred into the irradiated high line than into low line recipients. The most probable interpretation of these results is that high line macrophages are capable of inducing a stronger response of F1 lymphocytes than low line macrophage s. The genes segregated in each line during the selective breeding regulate the quantitative antibody response by operating on the rate of multiplication and differentiation of antibody-producing cells. This rate is much greater in the high than in the low line. The early antibody response to intravenously injected sheep erythrocytes takes place essentially in the spleen (Biozzi et al., 1972a). The cytodynamic study of the exponential phase of the immune response to sheep erythrocytes has showed that the mean doubling time of antibody producing cells in the spleen is 9 hours in the high line and 16 hours in the low line. At the end of the exponential phase, the total number of specific immunocytes in the spleen is 3,000,000 in the high line and 270,000 in the low line. The number of antibody-secreting mature plasmocytes is 720,000 and 16,500 in the high and low line spleens, respectively: a 44-fold interline difference. 111. Modifications of Macrophage Functions

Macrophages are characterized by two fundamental functions: high phagocytic activity and active catabolism of the engulfed antigens. Several findings have demonstrated that these two functions are subject to independent genetic regulations. The phagocytic activity of the reticuloendothelial macrophages is similar in the high and the low line whereas the catabolic function differs greatly. The phagocytic function of liver and spleen macrophages can be measured from the rate of blood clearance of intravenously injected particulate material and expressed by the phagocytic index K (Biozzi et al., 1953; Stiffel et al., 1970). The value of the phagocytic index K was established in the high and low lines of mice from the rate of phagocytosis of colloidal carbon particles, colloidal macromolecular aggregates of bovine serum albumin (1251-labeledCA-BSA) and sheep erythrocytes (51Cr-labeled SE) which is the antigen used for the Selection (Fig. 4).

198

1251

CA BSA I mg / mouse

Colloidal carbon 1.6 mg / mouse

-

..

0

GUIDO BIOZZI E T AL. 51Cr sheep erythrocytes 5.108 SEImouse

2

c

0

n I

High line Low line

-----

10 minutes

20

10 minutes

20

10 minutes

20

FIG.4. Phagocytic activity of liver and spleen macrophages in high and low lines. Rate of blood clearance of colloidal carbon, heat-aggregated bovine serum albumin ('9-labeled CA-BSA) and Wr-labeled sheep erythrocytes. Blood concentrations are expressed as percentage of initial concentrations.

No significant interline difference was observed in the rate of phagocytosis of these three particulate substrates by liver and spleen macrophages. The organ distribution of the phagocytized Wr-labeled sheep erythrocytes, determined after complete blood clearance, was roughly similar in the two lines: 90 and 91% in the livers and 7.4 and 6% in the spleens of high and low mice, respectively. High and low antibody responsiveness in the two lines is not therefore dependent on the amount of antigen phagocytized in the spleens.

A. PROCESSING OF THE SELECTION ANTIGEN The fate of phagocytized sheep erythrocytes is very different in high and low mice as demonstrated in Fig. 5. In this experiment, the persistence of the immunogenicity of sheep erythrocyte antigens in the spleens of high and low mice is estimated from the capacity to induce an antibody response in primed recipients challenged with irradiated homogenates of spleen removed from high and low donors at appropriate time intervals after immunization. The results show that the persistence of the antigen in immunogenic form is much shorter in low than in high line spleens. The rate of breakdown of sheep erythrocyte immunogen is an exponential function of time. The half-life of the immunogen is about five times longer in high than in low line spleens. It can be calculated that in the high line 1.5%


199

1/1280Ag half - life = 54 hours

v)

f

e .-c 0 .-

V320e

f

.-c L

.-2 c

V80Ag half life: 12 hours

.-C

.-

C

; 1/20. P B i

FIG.5. Persistence of immunogenicity of sheep erythrocytes in the spleens of high and low mice. Agglutinin titers in groups of primed outbred mice receiving irradiated homogenates of spleens removed from high and low donors at different times after immunization. (Subimmunogenic priming dose of the recipients: 105 sheep erythrocytes iv. Donor immunization: 2 x lo9sheep erythrocytes iv. Spleen irradiation: 10,000 rads. Injection of spleen homogenates: 4 spleen equivalent per recipient ip. Antibody titers in the recipients: measured on the twelfth day.) of the injected antigen is still immunogenic after 5 months whereas a comparable level is reached in about 2 weeks in the low line. The persistence of the immunogen in the spleen is a very important factor accounting for the interline difference in antibody responsiveness between high and low lines. It is very likely that the persistence of the antigen in the spleen is related to the catabolic activity of spleen macrophages, as clearly suggested by the results in Fig. 3B. In fact, as will be reported later, there is a striking difference in antigen processing between high and low line macrophages (Wiener and Bandieri, 1974; Adorini and Doria, 1981). The interline difference in the rate of antigen catabolism probably accounts for the peculiar patterns of antibody response dynamics observed in the two lines under different immunization procedures. An example is given in Fig. 6 which shows the response of the two lines to increasing doses of sheep erythrocytes injected intravenously. Low line mice require a larger threshold dose of antigen to give a detectable antibody response which moreover is of shorter duration than that of the high line. A large increase in the antigen dose is needed in the low line to obtain a substantial prolongation of antibody response. From the complete study of dose-response relationships (from

GUIDO BIOZZI ET AL.

200,

High line

el

512 0

*-a-

I****

---------------

Low line

lo9

FIG.6. Kinetics of agglutinin production in high and low mice after iv immunization with increasing doses of sheep erythrocytes.

threshold to maximal doses) it has been calculated that the dose of lo* sheep erythrocytes is required in the low line to obtain the peak agglutinin response induced by the dose of lo5 sheep erythrocytes in the high line, i.e., a 1000-fold interline difference (Biozzi et al., 1972a). The effect of the persistence of antigenic stimulation on antibody responsiveness is also clearly illustrated in Fig. 7 by comparing the kinetics of agglutinin responses to sheep erythrocytes injected either intravenously (Fig. 7A) or subcutaneously into the two hind footpaws, emulsified in Freund's complete adjuvant (Fig. 7B). The intravenously injected sheep erythrocytes are cleared from the blood in less than 1 hour, inducing a single pulse of antigenic stimulation in the spleen (see Fig. 4). On the contrary, the antigen emulsified in Freund's adjuvant persists a long time in the footpaws, producing a long-lasting stimulation of the regional lymphoid system. The results in Fig. 7 show that the interline difference in antibody levels in the advanced phase of the response is much smaller after the subcutaneous immunization because the continuous release of antigen from the local depot counterbalances to some extent the rapid breakdown of the antigen inside the macrophages. The direct evidence that the poor and short-lasting response in the low line is due to inadequate immunogenic stimulation is given by


201

A I I l0,OOO

B

1

10

20

days

30

11600

10

20

30

40

50

days

60

FIG.7. Kinetics of agglutinin production in high and low mice after (A) iv injection o f 2 x lo8sheep erythrocytes; (B) sc injection of 5 x lo7sheep erythrocytes (in 0.05 mi) emulsified in complete Freunds adjuvant in each footpaw.

the experiments summarized in Fig. 8. Our previous cytodynamic study of antibody responses to intravenously injected sheep erythrocytes in conventional mice has demonstrated that the exponential expansion of the specific immunocyte clone is abruptly stopped on the fourth day postimmunization because of lack of persistent immunogenic stimulation. In fact an antigen supply given during the exponential phase is able to prolong the exponential increase of the specific cell clone (Biozzi et al., 1968).We applied this experimental approach to the high and low lines. The results in Fig. 8 show that a second intravenous injection of antigen given 4 days after the primary immunization greatly increases the antibody response in the low line. This effect is only slightly improved by giving four antigen injections. On the contrary, the similar repetition of two or four antigen injections in the high line only produces a hardly significant increase in antibody response. These findings show that in high responders, since the antigen catabolism in

202

GUIDO BIOZZI ET AL.

15-

High line line

LOW

-----

+

+-I 4

7

4 injections

Control I injection

I4

28 days

FIG.8. Effect of repeated sheep erythrocytes injections on the kinetics of agglutinin production in high and low lines: 5 x lo8 sheep erythrocytes were injected iv on day 0 (1 injection), on days 0 and 4 (2 injections), and on days 0, 2, 4, and 6 (4 injections).

macrophages is slow, a single antigen dose induces a maximal stimulation of the immune system. They also demonstrate that the genetic defect in antibody synthesis of low responders is related to a shortage of antigenic stimulation. The above results (Figs. 6,7, and 8) converge in pointing out that the rapid catabolism of antigens in macrophages is an important factor determining the poor antibody responses of low line mice. This deficiency can be at least partially compensated by increasing the persistence of the immunogenic stimulation. This is a clear example illustrating how a genetic defect can be corrected under favorable environmental conditions.

B. PROCESSING OF UNRELATED ANTIGENS:NONSPECIFIC EFFECTOF THE SELECTION The genetic modification of antigen catabolism inside the macrophages is also very probably responsible for a remarkable consequence of the selective breeding, namely its nonspecific effect. It has been shown that the macrophage antigen processing constitutes a nonspecific component of the immune system functions (Fig. 1). Therefore the genetic regulation of antibody responsiveness in high and low lines is not restricted to the heterologous erythrocytes used as selection antigens but operates likewise upon the antibody responses to many other immunogens of distinct specificity as shown in Table I.

MACROPHAGE I N REGULATION OF IMMUNORESPONSIVENESS

203

It is clear that the superiority of high line antibody responsiveness concerns a large variety of antigens of disparate nature and specificity such as heterologous erythrocytes, bacterial and viral antigens, heterologous proteins, haptens, and also synthetic polypeptides of restricted heterogeneity currently used for the detection of specific Ir genes. It is evident that the extent of the nonspecific effect, compared with the interline separation of responsiveness to the selection antigens, presents considerable variations. The modification of antibody response to some antigens such as human y-globulin, rabbit y-globulin, and T4 bacteriophage is as large as that concerning the selection antigen. For the majority of the other antigens tested, the nonspecific effect is of intermediate degree. It should be stressed that, though very large, the nonspecific effect cannot be considered as general since there are two exceptions, levan and dextran, which induce an equivalent antibody response in high and low lines (Howard et al., 1974; Biozzi et al., 1975). The extent of interline separation from the data in Table I should not be considered as an absolute estimation since it is largely dependent on the immunization procedure, the physicochemical antigen nature, the route and number of antigen injections, and the use of adjuvants, which varied in the different experiments. Another important finding arising from the study of responses to various antigens is that the interline difference in antibody production concerns all classes and subclasses of antibody: IgA, IgM, IgG,, IgGz, and reagins (Biozzi et al., 1970; Prouvost-Danon et al., 1977; Andre et al., 1977). The importance of macrophage metabolic activity in the nonspecific effect produced by the genetic selection appears clearly from the results presented in Fig. 9 illustrating the kinetics of antibody response to bovine serum albumin (BSA) administered in two physical forms. Macromolecular aggregates of BSA obtained by controlled heating (CA-BSA) (Biozzi et al., 1957a) injected intravenously are cleared from the blood by liver and spleen macrophages in about 3 hours. Consequently, the antibody synthesis is localized in the spleen and results from a single pulse of antigenic stimulation. In low line mice the resulting antibody response is low and transient because the antigen is rapidly catabolized whereas it persists in high line macrophages, producing a strong and long-lasting antibody response (Fig. 9A). When the BSA adsorbed on an alum gel is injected intraperitoneally, the very large initial interline difference in antibody levels de-

TABLE I NONSPECIFIC EFFECTOF THE SELECTIVE BREEDING: ANTIBODY TITERSIN HIGHAND Low LINESIMMUNIZED WITH VARIOUS ANTICENS’.’’ Mean antibody response High line

Low line

Sheep erythrocytes Pigeon erythrocytes

116,000 1/11,OOo

1/25 1/32

Agglutination Agglutination

Heterologous erythrocytes

Rat Man (Group A)

1/20,OOo 1/4,000

1/2w 1/60

Agglutination Agglutination

Proteins

Human y-globulin Rabbit y-globulin Bovine serum albumin Hemocyanin Hen egg albumin Hen egg albumin Hen egg-white lysozyme Ring-necked pheasant egg-white lysozyme

1/m 1/2,000 111,000 112% 1/64 11512 4,655 9,318

1/20 118 1/20

Advances in Immunology Volume 36

Advances in Parasitology Volume 36

Advances in Agronomy, Volume 36

Advances in Catalysis, Volume 36

Advances in Immunology Volume 45

Advances in Immunology Volume 34

Advances in Immunology Volume 1

Advances in Immunology Volume 10

Advances in Immunology Volume 30

Advances in Immunology Volume 82

Advances in Immunology Volume 31

Advances in Immunology Volume 52

Advances in Immunology Volume 48

Advances in Immunology Volume 56

Advances in Immunology Volume 12

Advances in Immunology Volume 21

Advances in Immunology Volume 41

Advances in Immunology Volume 71

Advances in Immunology Volume 84

Advances in Immunology Volume 43

Advances in Immunology Volume 73

Advances in Immunology, Volume 61

Advances in Immunology Volume 14

Advances in Immunology Volume 46

Advances in Immunology, Volume 106

Advances in Immunology Volume 19

Advances in Immunology Volume 99

Advances in Immunology Volume 81

Advances in Immunology Volume 4

Advances in Immunology Volume 80

Advances in Immunology Volume 36

Advances in Parasitology Volume 36

Advances in Agronomy, Volume 36

Advances in Catalysis, Volume 36

Advances in Immunology Volume 45

Advances in Immunology Volume 34

Advances in Immunology Volume 1

Advances in Immunology Volume 10

Advances in Immunology Volume 30

Advances in Immunology Volume 82

Advances in Immunology Volume 31

Advances in Immunology Volume 52

Advances in Immunology Volume 48

Advances in Immunology Volume 56

Advances in Immunology Volume 12

Advances in Immunology Volume 21

Advances in Immunology Volume 41

Advances in Immunology Volume 71

Advances in Immunology Volume 84

Advances in Immunology Volume 43

Advances in Immunology Volume 73

Advances in Immunology, Volume 61

Advances in Immunology Volume 14

Advances in Immunology Volume 46

Advances in Immunology, Volume 106

Advances in Immunology Volume 19

Advances in Immunology Volume 99

Advances in Immunology Volume 81

Advances in Immunology Volume 4

Advances in Immunology Volume 80

Recommend Documents