ADVANCES IN
Immunology VOLUME 41
This Page Intentionally Left Blank
ADVANCES IN
Immunology EDITED BY
FRANK J. DI...
14 downloads
926 Views
27MB 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
ADVANCES IN
Immunology VOLUME 41
This Page Intentionally Left Blank
ADVANCES IN
Immunology EDITED BY
FRANK J. DIXON Scripps Clinic and Research Foundation La Jolla, California
ASSOCIATE EDITORS
K. FRANK AUSTEN LEROYE. HOOD JONATHAN W. UHR
VOLUME 41
ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers Orlando San Diego New York Austin Boston London Sydney Tokyo Toronto
COPYRIGHT 0 1987 BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART O F THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER
ACADEMIC PRESS, INC. Orlando, Florida 32887
United Kingdom Edition published by
ACADEMIC PRESS INC. (LONDON) LTD. 24-28 Oval Road, London NWI 7 D X
LIBRARY O F CONGRESS CATALOG
ISBN 0-12-022441-0
CARD NUMBER:
(alk. paper)
PRINTED IN THE UNITED STATES OF AMERICA
87 88 89 90
9 8 7 6 5 4 3 2 I
61 -1 7057
CONTENTS
Cell Surface Molecules and Early Events Involved in Human T Lymphocyte Activation ARTHURWEISSA N D J O H N B. IMBODEN
.................................. .....................
I. Introduction
11. Cell !surface Molecules Involved in T Cell Activation
111. Synergy between Ca2+ Ionophores and Phorbol Esters in T Cell Activation . IV. Receptor-Mediated Signal Transduction during T Cell Activation . . ... V. Role of Intracellular Signals Other Than Ca2+ and pkC . . . . . . . . . . ... VI. Effects of Early T Cell Activation Events upon Gene Regulation . . . . . . . . . . . . VII. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............................................. References
1 1 15 19 26 26 30 31
Function and Specificity of T Cell Subsets in the Mouse JONATHAN S P R E N T A N D S U S A N
R.
WEBB
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Cell !surface Molecules Controlling T Cell Specificity and Function . . . . . . . . . 111. H-2-Restricted Recognition of Antigen by Mature T Cells IV. Recognition of H-2 Alloantigens by Mature T Cells . . . . . . . . . . . . . . . . . . . . . . . V. Consequences of T Cell Contact with H-2 Molecules in the Thymus . . . . . . . . VI. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . ......................................
39 40 51 78 95 110
113
Determinants on Major Histocompatibility Complex Class I Molecules Recognized by Cytotoxic T Lymphocytes JAMES
FORMAN
I. Introduction . . . . . . . . . . . . 11. Exon Shuffling to Produce ................. 111. Recognition of HLA Class IV. Role of Carbohydrate Moieties in Determining CTL Recognition of Class I Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Role of Pz-Microglobulin in T Cell Recognition VI. Use of Monoclonal Antibodies to Block CTL Recognition of Class I Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Class I Heavy Chains Bearing Defined Amino A Polymorphic Determinants Recognized by CTL .................. tic Cell Class I VIII. CTL Recognition of Monoclonal Antibody-Select ................................................ Variants V
135 138 149
156 158 165
vi
CONTENTS
IX . Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
....... .......
167 171
Experimental Models for Understanding B Lymphocyte Formation
PAULW . KINCADE
I . An Introductory Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . Organization of Lymphohemopoietic Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 111. Resolution of B Cell Precursors . . . . . . . . . . . ................. IV . Rearrangement and Utilization of Immunoglo
...................... V . Population Dynamics . . . . . . . . . . . VI . Long-Term Bone Marrow Cultures . . . . . . . . . . . . . . . . . . . . . . . . . . ............................... VII . An Inducible Cell Line ............ VIII . ........................................... IX . Soluble Mediators ................... X . Synthesis and Conclusions . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
181 185 188 203 205 208 220 223 232 235 239
Cellular and Humoral Mechanisms of Cytotoxicity: Structural and Functional Analogies
JOHNDING-EYOUNG
AND
ZANVIL A . COHN
I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Nature of Cytotoxicity Mediated by CTL and NK Cells . . . . . . . . . . . . . . . . . . . 111. Cytolytic Mechanisms Proposed in the Past and the Concept of Secretion and Colloid Osmotic Lysis .......................................... IV . Granule Proteins in Cell-Mediated Killing ................... V . Membrane Attack Complex of Complemen ................... VI . Other Cytolytic Pore-Forming Proteins ....................... VII . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . ................
269 270 273 286 299 311 319 320
Biology and Genetics of Hybrid Resistance
MICHAELB E N N E ~
....................... 1. Introduction . . . . . . . . . . . . I1. Hybrid Resistance to Normal Hemopoietic Cells . . . . . . 111. Hybrid Resistance to LeukemiaLymphoma Cells ........................ IV . Effector Mechanisms of Hybrid Resistance . . . . . . V . Genetics of Antigen Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Proposed Mechanisms of A References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
333 335 358 369 397 401 411
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTENTS OF RECENTVOLUMES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
447 473
ADVANCES IN IMMUNOLOGY, VOL. 41
Cell Surface Molecules and Early Events Involved in Human T Lymphocyte Activation ARTHUR WEISS* AND JOHN 6. IMBODENt *Deportment of Medicine, Howard Hughes Medical Institute, University of California, San Francisco, California 94 143 and fDepartment of Medicine, Veterans Administration Medical Center, San Francisco, University of California, Son Francisco, California 94 143
1. Introduction
The activation of human thymus-derived (T) cells is the result of ligandreceptor interactions. Under physiologic conditions, such ligand-receptor interactions occur at the interface of the plasma membranes of an antigenspecific T cell and an antigen presenting cell (APC) or target cell. These antigen-specific and non-antigen-specific ligand-receptor binding events result in the transduction of these events into intracellular biochemical signals in the form of “second messengers.” Ultimately, such intracellular biochemical signals influence specific targeted genes receptive to these signals which can become transcriptionally active or inactive. The summation of these events is the expression of the phenotype of an activated T cell. The diverse manifestations of T cell activation include the production of lymphokines, the appearance of new cell surface proteins (which include growth factor receptors), the acquisition of cytolytic effector function, and, as a consequence of the production of growth factors and their receptors, proliferation. In this review, we will focus primarily on the structures and function of the cell surface molecules of the human T cell which appear to initiate activation. Where appropriate, data referring to the murine system will be drawn upon. The events subsequent to initial activation events, i.e., the interaction of interleukin 2 (IL-2) with its receptor and the resultant proliferative response, will not be addressed in this review. 11. Cell Surface Molecules Involved in T Cell Activation
The study of the T cell surface molecules involved in T cell activation has been facilitated through the use of homogeneous cell populations, such as T cell clones, hybridomas and leukemic lines, and the availability of monoclonal antibodies (mAb) which define antigenic epitopes expressed on an 1 Copyright Q 1887 by Academic Press, Inc. All rights of rrproductron in any form reserved.
2
ARTHUR WEISS A N D JOHN B. I M B O D E N
array of distinct molecules. By virtue of the agonist or antagonist properties of mAb reactive with these molecules, many of these cell surface molecules are felt to play a role in T cell activation, either in antigen-induced activation or in alternate pathways of activation. Most of these molecules are thought to function as cell surface receptors. The T cell antigen receptor must play a central role in antigen-driven T cell activation and has been most intensively studied, The ligands of these other putative receptors remain to be identified or confirmed. Some of these receptors can play a primary role in activation (the T cell antigen receptor, T11, Thy-1), initiating distinct biochemical events, which alone, following ligand interaction, can lead to T cell activation under appropriate conditions. Triggering of these receptors results in an increase in the concentration of cytoplasmic free calcium ([Ca2+Ii),one of the intracellular events which is generally felt to be required for T cell activation. Other receptors appear to function as accessory molecules [Tp44, T1, interleukin-1 (IL-1) receptor], which when stimulated, are able to synergize with stimuli provided by the T3/Ti complex. These receptors have little effect upon T cell activation when stimulated alone. Still other molecules have been implicated as receptors involved in increasing the overall avidity between the T cell and the APC (or target cell) (LFA-1, T4, T8). The following will attempt to summarize the structure and evidence supporting the role of some of these cell surface molecules in T cell activation. However, we will not attempt to exhaustively review all molecules involved in activation.
A. T CELLANTIGEN RECEPTOR Activation of the T cell induced by an antigen on the surface of an APC must involve an interaction with the T cell antigen receptor. This receptor subserves two functions in antigen-induced activation: (1)a recognitive function in which a specific antigen is recognized in the context of the appropriate major histocompatibility complex (MHC) molecules, and (2) an effector function in which the recognitive event is transmitted across the plasma membrane to the interior of the cell, with the resultant appearance of intracellular second messengers. A fundamental understanding of the structure of the receptor is useful in order to begin to understand the basis for these two functions. 1. Structure of the T Cell Antigen Receptor
The T cell antigen receptor was identified independently in several laboratories by the generation of mAb, which reacted with unique clonally distributed antigenic epitopes on T cell lines, hybridomas, or clones (clonotypic determinants) (Allison et al., 1982; Meuer et al., 1983a; Haskins et al., 1983). These antibodies react with disulfide-linked heterodimer glycoproteins (Ti)
HUMAN T LYMPHOCYTE ACTIVATION
3
of 80-90 kDa. These heterodimers are composed of an acidic Ti-a chain of 43-54 kDa and a more basic Ti$ chain of 38-44 kDa (Reinherz et al., 1983; Kappler et' al., 1983a,b). Peptide mapping studies suggested each chain has both constant and variable domains (Reinherz et al., 1983; Acuto et al., 1983; Kappler et al., 198313). Both chains are integral membrane proteins, have two to six N-linked glycosylation sites, and have an intracytoplasmic tail of five amino acids at the carboxy-terminus (McIntyre and Allison, 1984; Yanagi et al., 1984; Sim et al., 1984). The relatively short cytoplasmic tail of these chains suggests that they are by themselves not responsible for transmembrane signaling events. A detailed understanding of the structure of the human Ti has come from study of the complementary DNA (cDNA) clones and genomic clones of the Ti-a and p chains (Yanagi et al., 1984; Sim et al., 1984). The human Ti-a and -p chains have limited homology to immunoglobulin genes, suggesting a common evolutionary origin (Yanagi et al., 1984; Sim et al., 1984; Hood et nl., 1985). Both Ti-a and -p chains are assembled from gene segments which undergo rearrangements and expression during T cell ontogeny (Royer et al., 1984, 1985; Yoshikai et al., 1984; Raulet et al., 1985; Collins et al., 1985). Analogous to immunoglobulin heavy chain genes, Ti-6 chains are assembled from recornbinational events involving variable (V), diversity (D), joining (J), and constant (C) gene segments (Siu et al., 1984). The Ti-a chain genes are similarly assembled from V, J, and C segments, but, to date, no D segments have been identified (Yoshikai et al., 1985). Thus, the diverse antigen-reactive repertoire of T cells can be accounted for, in part, from the joining of different V, J, and D gene segments as well as combinatorial associations between the Ti-a and -6 chains. Transfection studies and cell fusion studies have suggested that the Ti-a and -p chains are sufficient to confer antigen and MHC specificity upon the T cell (Dembic et al., 1986), although primary sequence studies suggest that neither Ti-a nor -p chains are solely responsible for antigen or MHC specificity (Fink et al., 1986). Thus, the evidence strongly implicates Ti heterodimer in the antigen/MHC-specific recognitive events. More preliminary evidence, however, suggests that Ti-a and -p chains may not be the only chains involved in antigen recognition. In the course of attempts to isolate the Ti-a chain, another cDNA, the Ti-y chain, was isolated (Saito et al., 1984). This gene, once thought to be prefentially transcribed in cytolytic cells (Kranz et al., 1985), has now clearly been found to be expressed in helper T cells as well (Zauderer et al., 1986). The Ti-y chain gene, like the a and p chain genes, undergoes rearrangement utilizing V and J region segments linked to constant region segments (Hayday et al., 1985; LeFranc et al., 1986). Interestingly, the Ti-y chain is the first of the Ti chains to rearrange and to be expressed during T cell ontogeny (Raulet et al., 1985;
4
ARTHUR WEISS A N D JOHN B. IMBODEN
Haars et al., 1986) and, thus, has been proposed to be important in thymocyte selection (Raulet et al., 1985; Garman et al, 1986). Until recently, the protein product of the Ti-y chain had not been identified. However, recent studies suggest that it is expressed as a 55-kDa glycoprotein on an small, unusual subpopulation of human peripheral T cells which fail to express T4 (CD4) or T8 (CD8) antigens (Brenner et d., 1986; Weiss et d., 1986c; Lanier and Weiss, 1986). It may exist as a non-disulfide-linked heterodimer or as a single chain in association with T3 (see below). It has also been detected on the surface of T4-/T8- thymocytes felt to represent the most immature population of the thymus (Bank et al., 1986; Lanier and Weiss, 1986). On human T cells, the Ti heterodimer or the protein product of the Ti-y chain gene is associated with three invariant peptides which comprise the T3 (CD3) antigenic complex. T3 consists of at least three distinct integral membrane proteins: The T3-6 chain, a 22-kDa glycoprotein; the T3-Echain, a 21kDa nonglycosylated protein; and the T3-y chain, a 26- to 28-kDa glycoprotein (Borst et al., 1982, 1983a; Kanellopoulos et al., 1983). The cDNAs encoding these three chains have been isolated and sequenced (van den Elsen et al., 1984; Gold et al., 1986; Krissansen et al., 1986). The expression of these T3 genes is limited to T cells. T3-6 and T3-y chains exhibit substantial homology to each other, but not to other known proteins (Krissansen et al., 1986). Homologous chains have been identified in the murine system (Allison and Lanier, 1985; Samelson et al., 1985; Oettgen et al., 1986). However, additional chains have also been identified in the mouse. These include the 5 chain, a disulfide-linked homodimer or heterodimer of 32 kDa, with monomers of 14-17 kDa (Samelson et al., 1985; Oettgen et al., 1986), and a more recently isolated p21, a disulfide-linked dimer of 42 kDa with 21kDa subunits, which is phosphorylated on tyrosine residues with activation by antigen (Samelson et al., 1986). It is likely that homologues to the T3-4 chain and p21 will be identified in the human. Thus, the T3 complex may consist of seven chains. All three of the cloned chains of T3 contain between 40 and 80 cytoplasmic residues (van den Elsen et al., 1984; Gold et al., 1986; Krissansen et al., 1986). This observation, together with the findings discussed below regarding the agonist properties of T3 mAb and the phosphorylation of T3 chains, are consistent with the notion that T3 plays a role in transmembrane signaling events. Several lines of evidence suggest that the antigen receptor exists as a molecular complex composed of Ti and T3, or, in the case of the protein product of the Ti-y chain, Ti-y and T3. The physical association of the Ti-a/P chain heterodimer was demonstrated by comodulation (Meuer et al., 1983a), coimmunoprecipitation (Reinherz et al., 1983; Borst et al., 1983b), and the chemical cross-linking (Allison and Lanier, 1985; Brenner et al., 1985). T3
H UMAN T LYMPHOCYTE ACTIVATION
5
has been linked to the protein product of the Ti-y chain by coimmunoprecipitation (Weiss et al., 1986c) and chemical cross-linking (Brenner et al., 1986; Bank et al’., 1986). Evidence suggests that the association between T3 and Ti is obligatory in that mutants of the T cell leukemic line Jurkat, which lack Ti+ chain transcripts, contain T3 proteins trapped intracellularly (Weiss and Stobo, 1984; Ohashi et al., 1985). Reconstitution of the Ti+ chain by transfection into one such mutant resulted in the reexpression of Ti and T3 (Ohashi et al., 1985). Although the close association of T3 and Ti is suggested by such studies, the exact nature of this association is not clear. Under many conditions of immunoprecipitation, T3 and Ti do not coprecipitate (Allison et al., 1982; Haskins et al., 1983; Samelson et al., 1983; Weiss and Stobo, 1984). In ;icross-linking study, the T3-y chain was chemically cross-linked to the Ti+ chain, suggesting a close association between these chains (Brenner et al., 1985). A shortcoming of this study is the observation that neither the Ti-a and -p chains nor the chains of T3 were cross-linked to each other. One striking observation has been made from the sequence analyses of the component chains of Ti and T3. All three of the Ti chains, a, p, and y, of mouse and man contain an unusually placed highly charged basic lysine residue within the putative transmembrane domain (Yanagi et al., 1984; Sim et al., 1984; Saito et al., 1984), whereas the three chains of T3 contain conserved acidic residues of aspartic or glutamic acids within their hydrophobic putative transmembrane domains (van den Elsen et al., 1984; Gold et al., 1986; Krissansen et al., 1986). It has been suggested that these charged amino acids may play a role in the association between T3 and Ti. Collectively, these observations support a model of the T cell antigen receptor as a multisubunit complex composed of five to nine chains consisting of T3 and Ti.
2 . Role of the T3lTi Complex in Activation
The T cell antigen receptor must play a role in antigen-induced T cell activation. However, the direct study of the role of the receptor binding to the antigen is hampered by the inherent difficulty in studying cell-cell interactions and the ill-defined structure of antigen associated with M HC molecules. The use of monoclonal antibodies reactive with Ti or T3, which can function as agonists or antagonists, has facilitated the study of the role of the T3/Ti complex in T cell activation. Thus, such antibodies can serve as probes to elucidate the function of the receptor, mimicking the effects of antigen, without the participation of other cell surface molecules which may interact during T cell-APC interactions. Although this approach has limitations, since the effects of agonist mAb may not fully mimic the effects of antigen-antigen receptor interactions, it provides a first approximation toward the :itudy of the function of the antigen receptor.
6
ARTHUR WEISS AND JOHN B . IMBODEN
A large number of studies have revealed that mAb reactive with T3 could function as polyclonal agonists in inducing resting human peripheral blood T cells within peripheral mononuclear cells to secrete the lymphokines IL-2 or interferon-y (IFN-y) (von Wussow et al., 1981; Chang et d., 1982; van Wauwe et al., 1984), to express IL-2 receptors (Meuer et al., 1984a; Schwab et al., 1985; Tsoukas et al., 1985; Ledbetter et al., 1986), or to proliferate (van Wauwe et al., 1980; Chang et al., 1981). These antibodies have also been used to activate T cell clones and tumor lines to produce lymphokines or kill targeted bystander cells (Meuer et al., 1983b; Weiss et al., 1984a; Kranz et al., 1984; Mantzer et al., 1985). Similarly, clonotypic Ti mAb and Ti mAb reactive with nonpolymorphic determinants of Ti, such as mAb WT31, can activate T cells in a manner analogous to T3 mAb (Kappler et al., 1983a; Meuer et al., 1983b; Tax et al., 1983; Kaye and Janeway, 1984; Weiss and Stobo, 1984). It is of interest that under appropriate conditions of antibody immobilization, all anti-T3 or anti-Ti mAb described, with one exception, can function as agonists (Lanier et al., 1986). This implies that, in contrast to the T I 1 molecule (discussed below), perturbation of several distinct sites on the T3/Ti complex can lead to appropriate triggering of the complex. The potency of T3 and Ti mAb suggests that occupancy of relatively few receptors is sufficient to activate T cells (Chang et al., 1982; Kaye and Janeway, 1984). T3 and Ti mAb are also capable of functioning as antagonists, under some circumstances, to block the interactions between T cells and antigen-presenting cells or target cells (Chang et al., 1981; Meuer et al., 1983a; Haskins et al., 1983; Lancki et al., 1983; Samelson et al., 1983). Thus, the use of T3 and Ti mAb has proved to be a powerful tool to examine the role of the T cell antigen receptor in activation. The conditions required for activation of T cells by T3 or Ti mAb are dependent upon the particular manifestation of T cell activation examined. For instance, expression of the IL-2 receptor (IL-2R) has less stringent requirements than T cell proliferation. Hence, IL-2R expression can be induced by T3 or Ti mAb under conditions in which no proliferative response is observed (Schwab et al., 1985; Wakasugi et al., 1985; Tsoukas et al., 1985; Ledbetter et al., 1986). Therefore, production of IL-2 is more stringently regulated than the expression of the IL-2R. Since both the growth factor and its receptor must be produced in order for T cell proliferation to occur, T cell proliferation is primarily limited by the production of IL-2. This view must be qualified by the recent findings that there may be IL-2 independent pathways of T cell proliferation (Moldwin et al., 1986). BSF-1 is produced by T cells and can support the growth of some T cell clones (Smith and Rennick, 1986; Mosman et al., 1986; Yokota et al., 1986; Fernandez-Botran et al., 1986). It is not clear what the requirements are for BSF-1 production or for its role in T cell proliferative responses to antigen.
HUMAN T LYMPHOCYTE ACTIVATION
7
Regardless of the growth factors by which T cell proliferation is mediated, the induction of T cell proliferation by anti-T3 or anti-Ti mAb is dependent upon accessory cells (AC) (Chang et al., 1982; Tax et al., 1983; Landegren et al., 1984) In the case of human PBM, these AC are contained within the adherent cell population (Tax et al., 1983). At least two functions of these AC have been demonstrated. One function is dependent upon an interaction of the Fc portion of the T3 mAb and the Fc receptor on these AC (Tax et al., 1983, 1984; Landegren et al., 1984; Smith et al., 1986; Wakasugi et al., 1985; Ceuppens et al., 1985). This function of AC can be bypassed by immobilization of the T3 mAb onto Sepharose beads or onto the surface of culture dishes (Tax et al., 1984; Ceuppens et al., 1985). This suggests that the formation of a cross-linked matrix of antibody and T3 may be critical in activation requirements. Alternatively, as has been suggested, the immobilization of the T3 mAb may be important in preventing receptor internalization which might result in blunting the stimulatory response (Manger et al., 1985; Ledbetter et al., 1986). Indeed, as reviewed below, transmembrane signaling by soluble and immobilized anti-T3 or anti-Ti has been shown to differ. Whereas neither soluble nor immobilized anti-T3 induce IL-2 production or proliferation of highly purified T cells, only immobilized anti-T3 is able to induce IL-2R expression (Wakasugi et al., 1985; Ledbetter et al., 1986). It is likely that immobilized T3 mAb more closely mimics the Ti/T3 interaction with antigen/MHC on the surface of the T cell and APC. The second function of the AC is revealed by the failure of highly purified resting T cells to proliferate to immobilized anti-T3 or mitogenic lectins (Schwab et al., 1985; Williams et al., 1985; Ledbetter et al., 1986; Manger et al., 1986; Weiss et al., 1986a). The requirements for the activation of purified freshly isolated resting T cells and previously stimulated T cell clones or lines appear to differ (Manger et al., 1985; Meuer and zum Buschenfelde, 1986). Immobilized anti-T3 or anti-Ti alone is sufficient to activate T cell clones to produce IL-2 and to proliferate (Meuer et al., 1983b, 1984a; Manger et al. 1985). Similarly, the T cell leukemic line HUT 78, which phenotypically resembles a previously activated T cell, produces IL-2 in response to immobilized but not soluble T3 mAb (Manger et al., 1985). In contrast, the Jurkat cell line, like resting highly purified T cells, fails to respond to immobilized anti-T3 (Manger et al., 1985; Williams et al., 1985; Ledbetter et al., 1986; Weiss et al., 1986a). These findings suggest that resting T cells require an additional stimulus, provided by AC, which is not observed with T cells previously activated. Relatively small numbers of AC can provide this additional stimulus; hence, the notion that a soluble mediator is involved has emerged. Four ligands that bind to the surface of the T cell can mimic the effect of AC in providing this second function: mAb reactive with T1, T11, or Tp44,
8
ARTHUR WEISS A N D JOHN B . I M H O D E N
as well as IL-1. These will be discussed separately below. In addition, it should be noted that both functions of the AC can be provided by phorbol myristate acetate (PMA), a potent activator of protein kinase C (pkC) (Hara et al., 1985; Ledbetter et al., 1986; Weiss et al., 1986a). The role of PMA and pkC in T cell activation will be discussed at length later in this review. Thus, mAb reactive with T3 or Ti can function as polyclonal activators of T cells in a manner analogous to that of anti-Ig and B cells. However, simple ligand binding to the T cell antigen receptor does not appear to be sufficient for activation. In view of the fact that T cells do not respond to soluble antigen, but, rather, react with cell-bound antigen, the AC dependence of anti-T3 mAb may be quite consistent with the physiologic situation. In addition to its role in antigen-induced T cell activation, the T3/Ti complex appears to be important in the activation of T cells by the T cellspecific mitogenic lectins phytohemagglutinin (PHA) and concanavalin A (Con A). Both lectins bind to large numbers of T cell surface glycoproteins (Henkart and Fisher, 1975; Sitkovsky et al., 1984); however, the cell surface molecules responsible for the ability of these lectins to stimulate T cells have been undefined. Biochemical analyses of solubilized cell surface proteins have demonstrated that Con A can bind to the T3 chains but not Ti, whereas PHA can interact with the Ti heterodimer but not the isolated T3 chains (Kanellopoulos et al., 1985). Thus, among the many cell surface glycoproteins bound by these lectins are component chains of the T3/Ti complex. Indeed, both Con A and PHA can induce cocapping of T3 (Kanellopoulos et al., 1985). Simple demonstration of binding to components of the T3/Ti complex does not establish that the T3/Ti complex mediates the relevant activation signal induced by these mitogens. Evidence supporting the role of the T3/Ti complex in PHA- and Con A-induced T cell activation is the observation that Jurkat mutants which fail to express the T3/Ti complex lose the capacity to produce IL-2 in response to either PHA or Con A (Weiss et al., 1984b; Weiss and Stobo, 1984). Moreover, reconstitution of the T3/Ti expression in one of these mutants by transfection resulted in the restoration of the PHA and Con A responsiveness of this cell (Ohashi et al., 1985; Weiss et al., 198613). These results are in contrast to those suggesting that the T11 (CD2) molecule may function as the relevant PHA receptor. In these studies, anti-T11 mAb were used as antagonists (O’Flynn et al., 1986). The explanation for this discrepancy is not clear but may reflect differences in experimental approach. The evidence that the T3/Ti complex plays a role in mitogenic lectin-induced T cell activation is compelling. B. T11 (CD2, Leu5, LFA-2) The T11 molecule is a 50-kDa glycoprotein on the surface of all T cells and thymocytes (Howard et al., 1981; Kamoun et al., 1981). This molecule func-
HUMAN
'r
LYMPHOCYTE ACTIVATION
9
tions as the sheep erythrocyte receptor on human T cells. As many as six distinct epitopes have been defined by mAb reactive with T11: 9.6/T11,, D66, 35.11, Tll,, T113, and 9.1 (Meuer et al., 1984b; Martin et al., 1983; Brottier ct al., 1985; Yang et al., 1986). Certain epitopes of T11 are not expressed on resting T cells, T11, and 9.1, but can be induced by other mAb reacting with the TS/Ti complex or other epitopes of T11 (Meuer et al., 1984b; Yang et al., 1986). Interest in this molecule has been stimulated by the finding that such mAb can function as agonists or antagonists in inducing T cell activation. Initial studies revealed that an anti-T11 mAb, OKT11, could inhibit lectin and anti-T3-induced lymphokine production T cell proliferation and the lytic activity of cytolytic T cell clones (CTL) (Palacios and Martinez-Maza, 1982; Sanchez-Madrid et al., 1982; Wilkinson and Morris, 1984; Moretta et al., 1985b). This led to the proposal that the T11 molecule might function in immune responses by delivering negative signals (Palacios and MartinezMaza, 1982). Supporting a negative signal role for the T11 molecule is a recent study demonstrating diminished levels of IL-2 transcripts in stimulated T cells preincubated in the presence of mAb 9.6, reactive with T11 (Tadmori et al., 1986). Several studies from independent laboratories have demonstrated that certain combinations of anti-TI1 mAb can activate T cells, as measured by proliferation or IL-2 production (Meuer et al., 1984b; Brottier et al., 1985; Yang et al., 1986). Similarly, non-antigen-specific cytolytic activity of antigen-specific CTL and natural killer (NK) clones can be induced by appropriate anti-T11 mAb (Siliciano et al., 1985).Individual anti-T11 mAb are insuficient in inducing T cell activation. Only certain combinations of appropriate mAb are able to induce the activation of T cells. Whereas mAb reactive with T11, + T l l , or 9.1 + 9.6 can activate T cells in an AC-independent manner (Meuer el al., 1984b; Yang et d., 1986), D66 + 9.6 or D66 + T11, depend upon the presence of Fc receptor-bearing AC (Brottier et al., 1985).Certain T11 mAb can activate resting T cells in the presence of PMA without the addition of a second anti-T11 mAb (Holter et al., 1986). As some combinations of antibodies reactive with different epitopes of T11 do not activate T T l Q , simple cross-linking of molecules does not appear cells (i.e., T11, to account for the ability of certain combinations of mAb to activate. In contrast to the TS/Ti complex, stimulation of T cells via T11 appears to be exquisitely epitope dependent and requires relatively high (probably saturating) amounts of stimulating mAbs (Meuer et al., 198413). The ability to stimulate T cells in the absence of AC with appropriate combinations of antiT11 mAb would appear to exclude the participation of other cell surface molecules in this model of T cell activation. Thus, appropriate triggering of the T11 molecule appears to be able to provide a primary activation signal in
+
10
ARTHUR WEISS A N D JOHN B. IMBODEN
resting T cells which is distinct from that induced by the antigen receptor in view of the AC independence of this pathway. The T11 molecule is nonpolymorphic. Therefore, it is not likely to play a major role in antigen binding. The physiologic function of this molecule is not clear. The activation of T cells via stimulation of the T11 molecule has been termed the alternative pathway of human T cell activation to distinguish it from the antigen-dependent T3/Ti mediated pathway (Meuer et al., 1984b). As it is functional in thymocytes, it has also been proposed to play a role in thymocyte ontogeny (Fox et al., 1985). Recent studies have suggested that LFA-3 may represent the physiologic ligand of T11 (Springer et al., 1987). This 55- to 70-kDa glycoprotein is widely expressed on tissues. Antibodies reactive with LFA-3 inhibit a wide variety of T cell-dependent functions. A role for LFA-3 and T11 interactions in the thymus has also been proposed. Binding studies suggest a direct interaction between LFA-3 and T11. The interaction between the T3/Ti complex and the T11 molecule is of some interest. Stimulation of resting T cells via the T11 pathway does not require interaction with the T3/Ti complex. However, prior modulation of the T3/Ti complex inhibits the ability of T11 mAb to activate T cells (Meuer et al., 1984a; Fox et al., 1986). Conversely, modulation of T11 has little effect upon T3/Ti-induced activation. T11 mAb are able to activate NK cells which do not express T3 molecules (Siliciano et al., 1985). As no physical interaction between T11 and T3/Ti has been demonstrated, the explanation of these findings is not clear. However, one possibility is that in addition to their antigen removal effects, the mAb used in such modulation studies may have physiologic effects upon the cell. Of further interest, however, is a recent report suggesting that anti-T3 and anti-T11 mAb can synergize in inducing proliferation in cultures prepared with highly purified T cells (Yang et al., 1986). Thus, in addition to independent pathways of activation, these two pathways may interact under certain conditions. C. Thy-] Thy-1 is included in this review because of its unique structure and the evidence which had accumulated to support its role as a receptor involved in T cell activation in the murine system. The Thy-1 molecule is a 25- to 30-kDa glycoprotein with two allelic forms expressed on mouse thymocytes, peripheral T cells, fibroblasts, epithelial cells, and neurons (Reif and Allen, 1966a,b, 1984). The cDNAs encoding murine and human Thy-1 genes have been cloned and sequenced and exhibit some sequence homology to immunoglobulin genes (Evans et al., 1984; Seki et al., 1985). The murine Thy-1 gene is located on the ninth chromosome and encodes the 112 amino acid polypeptide chain (Blankenhorn and Douglas, 1972; Cohen et al., 1981). A most interesting structural feature of Thy-1 is the finding that the predicted
H U M A N T LYMPHOCYTE ACTIVATION
11
membrane anchoring region of the molecule does not span the membrane, but is instead truncated and covalently linked to the membrane lipid phosphatidylirtositol (Tse et al., 1985; Low and Kincade, 1985). This feature is of particular interest as Thy-1 can function as a receptor involved in activation, but has no described associated molecule and cannot communicate with intracellular effector molecules via a transmembrane of cytoplasmic domain. Early work with heterosera demonstrated that antibodies reactive with Thy-1 could be mitogenic for murine T cells (Smith et al., 1982). Subsequently, only certain mAb reactive with Thy-1, used individually, could induce ILA-2R,IL-2, or IFN-y production and be mitogenic for murine T cells, whereas other mAb could not (Gunter et al., 1984; MacDonald et al., 1985). However, most anti-Thy-1 mAb failed to induce T cell activation when used alone (Kroczek et al., 1986a). This difference in the agonist effects of these mAb was interpreted to correlate with the distinct epitopes recognized by iigonist versus nonagonist antibodies (Kroczek et al., 1986a). However, T cell proliferation was observed if cross-linking of Thy-l was induced using a rabbit anti-mouse Ig in combination with nonactivating Thy-1 mAb in the presence of PMA (Kroczek et al., 1986a). This response was independent of AC or the epitope of Thy-1 with which the mAb reacted. The requirement for cross-linking Thy-1 is reinforced by the observation that combinations of two Thy-1 mAb, reactive with distinct noncompeting epitopes, were also effective in inducing T cell activation if used in the presence of PMA (Kroczek et al., 1986a). Why some anti-Thy-1 mAb are able to function as agonists in the absence of additional cross-linking antibodies or PMA is not clear. It is of interest that the antigenic epitope recognized by one of these mAb, which by itself can activate T cells, is lost following transfection of Thy-1 into human T cells, murine B cells, or fibroblasts (Kroczek et al., 198613). Thy-1 has been transfected into the human T cell line Jurkat, and Thy-1 mAb can activate this line in the presence of PMA (Gunter et al., 1986). This important study demonstrates that transfer of the Thy-1 molecule alone is sufficient for the active cell surface receptor. Clarification of the mechanisms of Thy-1-induced activation should be possible in such a transfection system. Although the physiologic ligand of Thy-1 has not been identified, the potent effects of anti-Thy-1 mAb would suggest a potential for involvement in murine T cell activation.
D. RECEPTORSWHICHMAYPROVIDEACCESSORY IN T CELLACTIVATION SIGNALS 1 . Tp44
Several reports have implicated the cell surface molecule Tp44 as potentially playing an important role in T cell activation. The only mAb reactive
12
ARTHUR WEISS A N D JOHN B . IMBODEN
with Tp44, 9.3, identifies an 80- to 90-kDa disulfide-linked homodimer composed of 44-kDa subunits which is expressed on the surface of all T4 and -50% of T8 human T cells (Hansen et al., 1980; Yamada et al., 1985). A murine homologue of this molecule may have recently been identified (Nagasawa et al., 1986). Based on modulation studies of normal T cells and studies of mutants of the Jurkat cell line which fail to express the T3/Ti complex, no physical association between Tp44 and T3 exists (Hara et al., 1985; Moretta et al., 1985a; Weiss et al., 1986a). Initial studies demonstrated that 9.3 could inhibit the cytolytic activity of CTL (Fast et al., 1981). However, more evidence has accumulated demonstrating agonist properties of 9.3. The addition of 9.3 mAb to T cell cultures has demonstrated two distinct effects. In the first, 9.3 can play a primary role in inducing T cell proliferation (Hara et al., 1985; Moretta et al., 1985a). In monocyte-depleted cultures, the addition of PMA was required to observe proliferation (Hara et al., 1985). Addition of 9.3 mAb to monocyte-depleted cultures failed to induce IL-2 production or IL-2R expression, whereas abundant IL-2 production and IL-2R expression was observed in cultures containing monocytes or PMA (Hara et al., 1985). Thus, the effects of 9.3 mAb mimic the effects of anti-T3 or anti-Ti mAb, although in one study the kinetics of the response to 9.3 were delayed compared to anti-T3 (Moretta et al., 1985a). The dependency upon the T3/Ti complex for activation by 9.3 has been addressed by modulation of the T3/Ti complex and the study of a Jurkat cell mutant which failed to express the T3/Ti complex (Hara et al., 1985; Moretta et al., 1985a; Weiss et al., 1986a). The modulation experiments performed in different laboratories led to conflicting results regarding this dependency; however, the finding that the Jurkat mutant could still be activated by 9.3 plus PMA supports the notion that activation via the Tp44 molecule is independent of the participation of the T3/Ti complex. Although the ligand of Tp44 is unknown, it is clear that this molecule can be involved in delivering primary activation stimuli. A second accessory function has been demonstrated for Tp44. Addition of 9.3 mAb can substitute for one of the functions of adherent cells in the response to anti-T3, anti-Ti, or T cell mitogenic lectins. If 9.3 is added to cultures of purified T cells in the presence of cross-linked anti-T3, T cell proliferation is observed (Ledbetter et al., 1985; Martin et al., 1986; Weiss et al., 1986a). Thus, 9.3 substitutes for the second function provided by AC, alluded to above, which may involve a soluble factor. In a similar manner, 9.3 can synergize with anti-T3, anti-Ti, or the lectin PHA in inducing Jurkat to produce IL-2 (Martin et al., 1986; Weiss et al., 1986a). Interestingly, 9.3 cannot reconstitute the response to soluble antibody or calcium ionophore by purified T cells or Jurkat (Weiss et al., 1986a). Thus, it does not fully replace the function of AC.
H U M A N T LYMPHOCYTE ACTIVATION
13
The potency of 9.3 in exerting the two effects described is markedly different. Saturating concentrations of 9.3 are required for delivering a primary activation stimulus (Weiss et al., 1986a). The finding that 9.3 can be used at siibsaturating levels of antibody for the accessory function is consistent with the notion that Tp44 is more likely to play an accessory role in T cell activation (Weiss et al., 1986a). The precise role of Tp44 in T cell activation awaits the identification of its relevant physiologic ligand.
2. T1 (CD5, Tp67, L e d ) The T1 antigen is expressed as a 67-kDa protein on all T cells and thymocytes (Reinherz et al., 1979a; Martin et al., 1980). The murine homologue of T1 is Lyt-1 (Ledbetter et al., 1981). The cDNA encoding T1 has recently been isolated (Jones et al., 1986). T1 has a large, 347 amino acid extracellular domain and a 93 amino acid intracytoplasmic domain. Although mAb reactive with T1 have little effect upon T cell activation alone, they appear to have some capacity to deliver accessory stimuli for T cell proliferation. Thus, anti-T1 can augment the proliferative responses and IL-2 production by antiT3-stimulated T cells if immobilized anti-T3 mAb are used (Ledbetter et al., 1985; Ceuppens and Baroja, 1986). The effect of anti-T1 appears to be monocyte independent, since F(ab’)2and Fab are able to provide this accessory function (Ledbetter et al., 1985). Similar effects have been observed in the murine system (Hollander et al., 1981; Logdberg and Shevach, 1985). The accessory function provided by anti-T1 may be distinct from the effect of mAb 9.3 or IL-1, as the effects of these ligands are additive (Ledbetter et al., 1985). The ligand of T1 is unknown.
3. IL-1 Receptor (ZL-1R ) Although the identification of the IL-lR remains somewhat tentative, this receptor 1s included in this discussion because of the numerous studies performed with its ligand, IL-1. Numerous studies in the past have demonstrated that AC function can be, in part, reconstituted by soluble factors in the supernatants of adherent cells (reviewed by Mizel, 1982). One of the most potent of these factors was termed IL-1 (Aarden et al., 1979) and has been purilied to homogeneity (Kronheim et al., 1985). Recently, two cDNAs have been isolated which encode two forms of IL-l,IL-l, and IL-lp(Auron et al., 1984; March et al., 1985). The predicted size of the protein encoded by each of these clones is 31 kDa, though the mature form of the protein is proteolytically processed to 17.5 kDa (Auron et al., 1985). The role of each of these forms of IL-1 remains to be clarified. Interestingly, an activity which would suggest a membrane form of IL-1 has been described (Kurt-Jones et al., 1985). This is of particular interest, since no hydrophobic domain has been identified in the IL-1 sequence (Auron et al., 1984; March et al., 1985).
14
ARTHUR WEISS A N D JOHN B . IMBODEN
Highly purified IL-1, has been used to study the structure of the IL-1R. Radiolabeled IL-lBwas cross-linked to a 75-kDa cell surface protein on the murine T cell leukemic line LBRM-33-1A5 (Dower et al., 1985). This line was used to identify this putative IL-1R because it responds to IL-1 in the presence of suboptimal concentrations of PHA (Gillis and Mizel, 1981). This line was found to express -500 receptors per cell with an affinity of -2 X 10lO/M (Dower et al., 1985). Human T cells were found to express 100 or fewer IL-1R (Dower et al., 1985). Most studies have relied upon the use of purified IL-1, to study the relative function of its receptor. Although IL-1 has little effect upon human T cells by itself, it can substitute for AC if added with appropriate stimuli. IL-1 fails to induce IL-2 production or IL-2R expression when added to cultures of purified T cells or the T cell leukemic line Jurkat. However, it can synergize with PHA or immobilized anti-T3 in the induction of IL-2 production or IL-2R expression (Williams et al., 1985; Manger et al., 1985; Scheurich et al., 1985). Immobilization of the T3 mAb appears to be necessary for the synergistic effects of IL-1 (Williams et al., 1985; Manger et al., 1985). Thus, IL-1 does not completely reconstitute the function of AC, but appears only to substitute for the soluble factors liberated by AC. The phorbol ester PMA can substitute for the role of IL-1 in most systems (deVries et al., 1980; Farrar et al., 1980b). The role of IL-1 in anti-T11 or anti-Thy-l-induced activation remains to be clarified. 4 . Other Accessory Molecules Involved in T Cell Activation Several other T cell surface molecules are thought to play a role in T cell activation. Among these, the most well characterized are LFA-1, expressed on all T cells (Kurzinger et al., 1981), and T4 (CD4, Leu3) and T8 (CD8, Leu2) antigens, which are expressed on the two mutually exclusive major T cell subsets (Reinherz et al., 1979b; Reinherz and Sclossman, 1980). The inhibitory effects of mAb reactive with these antigens upon T cell proliferative responses and cytolytic activity of CTL is the strongest argument for their participation in T cell activation (Davignon et al., 1981; Meuer et al., 1982; Biddison et al., 1982). Whereas the ligand of LFA-1 is unclear, the MHC restriction pattern observed with cells expressing T8 or T4 suggests that these molecules may interact with nonpolymorphic class I or class I1 MHC antigens, respectively. These human cell surface molecules and their murine homologues are generally felt to increase the avidity of the interaction between the T cell and the relevant target cell or APC (Biddison et al., 1982; Swain et al., 1983; MacDonald et al., 1982; Marrack et al., 1983). However, several recent studies have suggested that mAb reactive with the T4 molecule or the murine homologue L3T4 may induce a negative signal apart from diminishing the avidity of the T cell-AC interaction (Bank and
HUMAN
-r LYMPHOCYTE ACTIVATION
15
Chess, 1985; Wassmer et al., 1985). A more detailed discussion of the role of these antigens is beyond the scope of this review and is presented in several recent reviews (Littman, 1987; Springer et al., 1982; Springer et al. ,1987). 111. Synergy between Ca2+ lonophores and Phorbol Esters in T Cell Activation
The cell surface structures that initiate T cell activation must be capable of generating regulatory intracellular signals. One approach to the identification of these signals is to “bypass” the cell surface structures involved in activation by stimulating the cell with pharmacological agents known to activate particular signaling pathways. Application of this approach to T cell activation reveals a remarkable synergy between Ca2 ionophores and phorbol esters. Whereas neither agent alone is mitogenic, the combination of Ca2+ ionophore and phorbol ester activates T cells to produce lymphokines, to express receptors for IL-2, and to proliferate (Weiss et al., 1984a; Truneh et al., 198i3). Of interest, this combination cannot substitute for the growthpromoting effect of IL-2 on T cell lines that are dependent upon exogenous IL-2. In other words, the combination of Ca2+ ionophore and phorbol ester mimics the effect of activation by antigen, but does not bypass the requirement for the IL-%mediated proliferative signal. The implication of these findings is that the early stages of T cell activation involve synergy between at least two discrete intracellular signals, one of which can be supplied by Ca2+ ionolphore and the other by phorbol esters. There is little doubt that the activation signal delivered by Ca2 ionophores is an increase in [Ca2 Ii. There is less certainty as to the identity of the phorbol ester-mediated signal, but because the only known phorbol ester receptor is pkC, the effects of phorbol esters have been attributed to activated pkC. Synergy between increases in [Ca2 Ii and activated pkC appears to regulate cellular activities in a variety of tissues and has been implicated in systems ranging from platelet activation to aldosterone secretion (reviewed by Berridge and Irvine, 1984.; Nishizuka, 1986). +
+
+
+
A. RECEPTOR-MEDIATED INCREASES IN [Ca2+IiA N D ACTIVATIONOF PROTEIN KINASE c: GENERAL CONSIDERATIONS
It is worthwhile considering the general mechanisms by which receptors regulate increases in [Ca2 Ii and activate pkC. A single receptor-mediated event, the hydrolysis of a membrane phospholipid, phosphatidylinositol bisphosphate (PIP,), can stimulate both intracellular pathways. The turnover of PIP, generates two products with second messenger capabilities: inositol 1,4,5-trisphosphate (1,4,5-1P3), which mobilizes intracellular Ca2 , and diacylglycerol (DG), which activates pkC (Berridge, 1983; Berridge and Irvine, 19134). +
+
16
ARTHUR WEISS A N D JOHN B. IMHODEN
1 . IP,-Mediated Increases in [Ca2+Ii 1,4,5-IP3, a water-soluble compound, binds to specific receptors within the endoplasmic reticulum, stimulating an efflux of Ca2+ (Streb et al., 1983; Hirata et al., 1985). In intact cells, release of Ca2+ from intracellular stores can increase [Ca2+Iifrom 100 nM to >SO0 nM (Lew et al., 1984). Increases in [Ca2+Ii that are due solely to intracellular mobilization are invariably transient and are usually of only several minutes duration. Certain receptors, such as T3/Ti and the a-1 adrenergic receptors on hepatocytes, can stimulate extracellular Ca2 uptake as well as mobilize intracellular Ca2 (Imboden and Stobo, 1985; Williamson et aZ., 1985). While intracellular Ca2+ mobilization accounts for the initial response, sustained receptor-mediated increases in [Ca2+Ii require extracellular Ca2+ influx. The mechanism by which these receptors regulate extracellular Ca2+ influx is not understood. The metabolism of 1,4,5-IP3 is complex. Sequential phosphatases can remove phosphates from the inositol ring, eventually converting 1,4,5-IP3 to free inositol which can be recycled into phospholipid (Berridge, 1983). Alternatively, a cytoplasmic kinase can phosphorylate 1,4,5-IP3on the 3 position of the inositol ring, yielding inositol 1,3,4,5-tetrakisphosphate (IP,) (Irvine et al., 1986). The existence of alternative pathways of 1,4,5-IP3 metabolism implies that certain metabolites of 1,4,5-IP3 also have regulatory functions, presumably distinct from those of 1,4,5-IP3. Speculation in this regard focuses on the regulation of extracellular Ca2+ influx and, because the cell invests ATP in its formation, on the possible role of IP,. An additional complexity is imparted to the inositol phosphate system by the observations that these compounds can exist in a 1,2 cyclic form and that cyclic inositol phosphates can be demonstrated following receptor stimulation of intact cells (Dawson et al., 1971; Dixon and Hokin, 1985). Both cyclic and noncyclic 1,4,5-IP3can release Ca2 from permeabilized cells, demonstrating that the cyclic configuration is not required for Ca2+ mobilization (Wilson et al., 1985). Whether the cyclic configuration confers some additional regulatory capability on the inositol phosphates remains to be determined. +
+
+
2 . DG- and Phorhol Ester-Mediated Activation of Protein Kinase C The second signaling pathway linked to PIP, hydrolysis is the activation of pkC. The distinctive feature of pkC, which can phosphorylate a wide range of substrates on serine and threonine residues, is that its activation requires the presence of phospholipid (especially phosphatidylserine) and Ca2 (Takai et al., 1979). In vitro, DG greatly increases the affinity of pkC for phospholipid and Ca2+ and allows pkC activation to occur at Ca2+ concentra+
HUMAN T LYMPHOCYTE ACTIVATION
17
tions that are within the intracellular range. In an insightful series of experiments, Catagna et al. observed that biologically active phorbol esters have effects on pkC similar to those of DG, activating pkC at Ca2+ concentrations in the nanomolar range (Castagna et al., 1982). Subsequent studies demonstrated that pkC activity and phorbol ester binding activity copurify (Kikkawa et al., 1983; Kraft and Anderson, 1983; Nidel et al., 1983). In view of the diverse biological effects of phorbol esters, the identification of pkC as the high-affinity phorbol ester receptor underscores the potential importanoe of pkC as a regulator of cellular activities. Recent cloning studies, however, clearly demonstrate that pkC activity and phorbol ester binding are mediated by a family of closely related, but distinct polypeptides (Knopf et al., 1986; Parker et al., 1986; Coussins et d.,1986). This finding, of course, implies that the biology of pkC is considerably more complex than previously appreciated. It is possible, for example, that there is differential expression of pkC subtypes, that the different species of pkC have distinct substrate specificities, and that there are constraints on the interactions between particular receptors and the different forms of pkC. Since all functional studies to date have treated pkC activity as a single enzyme, we will, of necessity, continue to discuss pkC as if it were a single entity. It is likely that pkC, which has an apparent M , of 80,000, is composed of a regulatory region which binds Ca2+, phospholipid, and DG (or phorbol ester), and a catalytic region (Nishizuka, 1986). The putative catalytic region of pkC has extensive sequence homology with the catalytic subunit of CAMPdependent kinase (Knopf et al., 1986; Parker et al., 1986). In most unstimulated cells, including T lymphocytes, pkC activity is recovered from the cytosol. Activation of pkC, whether mediated by receptor-ligand interaction or phorbol esters, is temporarily associated with a loss of pkC activity from the cytosol and a proportionate recovery of pkC activity in the membrane fraction (Farrar and Anderson, 1985). This translocation of pkC to the membrane is thought to be a critical event in its physiologic activation (Bell, 1986). The binding of pkC to membranes has been studied in a reconstitution system using purified enzyme and inside-out erythrocyte vesicles. As expected, phorbol esters promote the binding of pkC to membranes in this system (Wolf et al., 1985a). Increasing the Ca2+ concentration from 100 nM to 500 nM also promotes the association of pkC with membranes (Wolf et al., 1985a). It is noteworthy that the Ca2 -mediated binding is reversible and occurs at (;a2 concentrations that are within the range of receptor-mediated increases in [Caz+lI(but well below the 5 to 50 F M required for Ca2+ to activate pkC in the absence of DG). Of interest, the effects of Ca2+ and phorbol esters on pkC binding are synergistic. An effect of pkC binding, therefore, may in part explain the widely observed synergy between Ca2+ +
+
18
ARTHUR WEISS A N D JOHN B. IMBODEN
ionophores and phorbol esters. Taken together, these binding studies suggest a model for the physiologic activation of pkC in which receptor-mediated increases in [Ca2+Ii, while probably not of sufficient magnitude to activate pkC directly, “prime” pkC by promoting its binding to plasma membranes. This binding facilitates contact between pkC and DG, leading to the formation of a stable, active complex composed of pkC, phospholipid, Ca2+, and DG. The termination of receptor-mediated pkC activation may be quite complex. Intuitively, a critical factor in the maintenance of an active pkC complex must be continued receptor-mediated generation of DG. Indeed, in reconstitution experiments, a stable membrane-pkC complex requires the continuing presence of phorbol ester. Even in the presence of phorbol ester, however, this complex dissociates when ATP is added (Wolf et d., 1985b). The ATP-induced dissociation is nucleotide specific and requires Mg2+, implying that the release of pkC from membranes is due to phosphorylation of a membrane protein (possibly the autophosphorylation of pkC). Whether a similar ATP-mediated mechanism releases pkC from membranes in vivo is not known. A finding that argues against such an event in vivo is the sustained (>1hour) translocation of pkC that invariably follows the addition of phorbol esters to intact cells. An interesting observation of potential importance for the physiologic regulation of pkC is that pkC can serve as a substrate for a cytosolic proteinase, termed calpain (Inoue et d., 1977). The calpain-mediated proteolysis of pkC releases a 50-kDa fragment which is fully catalytically active in the absence of Ca2 and phospholipid and which is thought to represent the catalytic subunit of pkC freed of its regulatory region (Inoue et al., 1977). Activation of calpain was initially thought to require supraphysiologic concentrations of Ca2+, raising doubts as to whether proteolyltic cleavage of pkC occurs in viuo. Recent studies, however, demonstrate that micromolar concentrations of Ca2+ promote the binding of both pkC and calpain to partially purified plasma membranes (Melloni et al., 1985). Under these conditions, binding is followed by time-dependent proteolysis of pkC. In the presence of membranes, therefore, the in vitro conversion of pkC to its Ca2 and phospholipid-independent form occurs at Ca2 concentrations that approach the intracellular range. These studies suggest an alternative mechanism for the regulation of pkC in which receptor-mediated pkC translocation and increases in [Ca2+Iiresult in the generation of an irreversibly activated fragment of pkC. Implicit in the identification of pkC as a signaling pathway is the assumption that pkC-mediated phosphorylation of particular proteins in some way influences the functions of those proteins. In vitro, a wide range of proteins can serve as substrates for pkC, and a number of cell surface receptors, +
+
+
19
HUMAN T LYMPHOCYTE ACTIVATION
cytoskeletal proteins, and enzymes have been proposed as in vivo substrates for pkC (Nishizuka, 1986). The addition to PMA to intact T lymphocytes leads to the phosphorylation of the IL-2 receptor, T200, the transferrin receptor, and T3 chains of the T3/Ti complex as well as to the hyperphosphorylation of HLA class I antigens (Shackelford and Trowbridge, 1986; Cantrell et al., 1985; Samelson et al., 1985). Compelling data indicate that, at least in the case of the IL-2 receptor, this phosphorylation is directly mediated by pkC. Purified pkC phosphorylates the cytoplasmic domain of immunoprecipitated IL-2 receptor (Shackelford and Trowbridge, 1986). By tryptic peptide analysis, these in uitro phosphorylation sites are identical to those induced in vivo by PMA. The functional consequences of this and other pkC-mediated phosphorylations, however, are far from certain. Studies of the consequences of pkC activation on T lymphocytes and other cells have relied heavily on phorbol esters to stimulate pkC and, to varying degrees, have equated the effects of phorbol esters with pkC activation. The use of phorbol esters for this purpose has several limitations that deserve emphasis. First, phorbol esters bind to more than one species of pkC, but it is not yet known whether there are constraints on receptor interactions with pkC subtypes. Second, although pkC remains the only convincingly demonstrated phorbol ester receptor, it is possible that phorbol esters, in addition to activating pkC, may directly stimulate other signaling pathways. A recent report of the isolation of a phospholipid-dependent, Ca2 -independent kinase that is activated by phorbol esters serves to emphasize this point (Malviya vt al., 1986). Finally, there are marked differences in the metabolism of phorbol esters and DG. Phorbol esters are metabolized slowly, if at all, whereas DG turns over readily. As a result, there can be dramatic differences in the duration and magnitude of pkC activation following the addition of phorbol esters and following receptor stimulation. Phorbol estermediated activation, as reflected by the translocation of pkC to the membrane, is virtually irreversible, whereas receptor-mediated activation of pkC can be transient. +
IV. Receptor-Mediated Signal Transduction during T Cell Activation
The ability of Ca2+ ionophores and phorbol esters to deliver activation signals to T cells implies that T cells express cell surface receptors which regulate increases in [Ca2+Iiand activate pkC. Studies of receptor signaling during physiologic T cell activation are limited to a certain extent by the ambiguities inherent in studying a cell-cell interaction (a situation in which several types of receptors might signal simultaneously). As a result, studies of signal transduction in T cells have relied heavily on the use of mAbs as agonists to stimulate specific receptors. This approach has identified three
20
AKTHUK WEISS A N D J O H N B . IMBODEN
separate T cell surface structures which appear to signal by increasing [Ca2+Ii:T3/Ti, T11, and Thy-1. At least one of these, T3/Ti, also activates pkC. Three additional cell surface structures, Tp44, Tp67, and the IL-1 receptor, can deliver signals during the early stages of T cell activation which are not readily explained by a direct effect on either [Ca2+Iior pkC. We will review recent studies of the mechanisms of signal transduction by these six cell surface structures.
A. T3/Ti COMPLEX Several lines of evidence indicate that an increase in [Ca2+],serves as an intracellular signal for T3/Ti-mediated activation. Soluble T3/Ti mAb and Ca2+ ionophores, such as A23187 and ionomycin, display similar requirements in their abilities to activate T cells. Neither soluble T3/Ti mAb nor Ca2+ ionophores are effective alone, but both can synergize with PMA to elicit lymphokine production (Weiss et al., 198413; Truneh et al., 1985). If PMA is present, Ca2 ionophores activate T3/Ti-negative mutants of Jurkat, demonstrating that ionophore-mediated activation does not require cell surface expression of T3/Ti (Weiss et d . , 198413). The recent development of techniques to monitor [Ca2+Ii in intact, small cells has provided direct evidence that perturbation of T3/Ti increases [Ca2+],.The addition of T3/Ti mAb to T lymphocytes loaded with the Ca2+-sensitive fluor, quin2, induces substantial sustained increases in [Caz+li (Weiss et al., 1984b; Imboden et al., 1985; M. Weiss et al., 1984; Oettgen et a l . , 1985). T3/Ti-mediated increases in [Ca2+], have been demonstrated in peripheral T cells, T cell lines, and T cell clones and are not simply a consequence of the interaction of mAb with the cell surface. Flow cytometric analysis of [Ca2+Ii using the second-generation Ca2 indicator, Indo-1, demonstrates that T3-mediated increases in [Ca2+Ii are not restricted to any T cell subset, but occur in essentially all peripheral T cells exposed to a T3/Ti mAb (June et al., 1986). The nature of the perturbation required for triggering the T3/Ti complex has received limited attention. It is clear that mAb reactive with distinct sites of this complex can induce comparable increases in [Caz+li. Hence, anti-T3 or distinct anti-Ti mAb can all induce greater than $fold increases of [Ca2+Ii (Weiss et al., 1984b; Imboden et al., 1985; Oettgen et al., 1985; O’Flynn et al., 1985; Lanier et al., 1986). These mAb do not appear to trigger T3/Ti-mediated signal transduction by cross-linking receptors as univalent Fab fragments of T3 mAb also induce [Ca2+Iiincreases (Oettgen et al., 1985). This would imply that conformational changes of T3/Ti may be induced by relevant ligand binding events, and this results in receptormediated signal transduction. In support of this notion, one pentameric IgM anti-Ti mAb binds to an epitope of the HPB-ALL Ti effectively, but fails to induce substantial increases in [Ca2+Ii(Lanier et al., 1986). +
+
H U M A N T LYMPHOCYTE ACTIVATION
21
While there is little doubt that perturbation of TS/Ti by mAbs leads to increases in [Ca2+Ii, there are legitimate concerns as to how accurately T3/Ti mAbs mimic physiologic activation of TS/Ti. Antigen-primed AC, however, stimulate increases in [Ca2+Ii when added to quin%loaded, antigen-specific T cell clones (Nisbet-Brown et al., 1985; Shapiro et ul., 1985). When considered together with the studies of T3/Ti mAb, the latter observation strongly supports the notion that physiologic activation of T3/Ti induces an increase in [Ca2+Ii. As is the case with many other receptors that signal via increases in [Ca2 Ii, perturbation of TS/Ti stimulates the turnover of polyphosphoinositides and the generation of inositol phosphates (Imboden and Stobo, 1985). The addition of TS/Ti mAb to Jurkat cells leads to a prompt (detectable in 30 minutes. Similary, antigen recognition by a human T cell clone is associated with a substantial, prolonged increase in IP, (Imboden et ul., 1987). When inositol phosphates are resolved by high-performance liquid chromatography (HPLC), it is clear that the T3/Ti-mediated change in IP, is due in large part to increases in the Ca2+-mobilizing isomer 1,4,5-IP3 (Stewart et al., 1986). A substantial proportion of the 1,4,5-IP, generated by T3/Ti stimulation is converted to IP,, an observation that underscores the possibility of a regulatory role for IP, in T cell activation (Stewart et d., 1986). In all cellular systems studied, receptor-mediated increases in I,4,5-IP3 are associated with the release of Ca2+ from intracellular stores (Berridge and Irvine, 1984). In the presence of extracellular Ca2+, the addition of T3/Ti mAb to Jurkat cells leads to an increase in [Ca2+Ii from 80 nM to a peak of >400 nM within 60 seconds (Imboden et al., 1985). [Ca2+Iithen falls to a plateau of 200-250 nM and remains elevated above basal levels for >30 minutes. When care is taken to minimize the Ca2 -buffering effect of intracellular quin2, depletion of extracellular Ca2 has little effect on the initial peak TS/?’i-mediated increase in [Ca2 Ii, but completely prevents the sustained increase (Imboden and Stobo, 1985). This finding indicates that the initial peak increase in [Ca2+Iifollowing perturbation of T3/Ti Jurkat cells is due to intracellular Ca2 mobilization, while the sustained T3/Ti-mediated increases in [Ca2+Ii require extracellular Ca2+. In addition to mobilizing intracellular Ca2 , therefore, perturbation of T3/Ti must either open a Ca2+ channel in the plasma membrane or regulate the transport of Ca2+ across the plasma membrane. Distinction between these two general mechanisms cannot be made using intact quin2-loaded cells and requires either study of (=a2+ transport in a cell-free system or the application of patch clamping techniques to directly study channel conductance. Initial patch clamping of T lymphocytes failed to demonstrate Ca2+ channels under conditions in which voltage-gated Ca2+ channels, if present, should have been +
+
+
+
+
+
22
ARTHUR WEISS A N D J O H N B . I M B O D E N
identified (Decoursey et al., 1984; Matteson and Deutsch, 1984). Recently, however, Kuno and colleagues have identified a Ca2+ channel in T lymphocytes that is not voltage gated, and its frequency of opening increases following the addition of PHA (Kuno et al., 1986). This channel, whose tissue distribution and gating mechanisms are of considerable interest, is an attractive candidate to explain T3/Ti-mediated extracellular Ca2 influx. One immediate consequence of the T3/Ti-mediated increase in [Ca2 Ii is enhanced activity of the plasma membrane Na+ / H antiporter, leading to a sustained increase in intracellular pH (Rosoff and Cantley, 1985). Activation of the Na+ / H antiporter has been implicated as an important signaling mechanism in the stimulation of quiescent cells by growth factors and in B lymphocyte differentiation. Dimethylamiloride, an inhibitor of the Na+ / H antiporter, inhibits IL-2 production by Jurkat cells stimulated with a T3 mAb and PMA. While the specificity of dimethylamiloride has been questioned, it is of interest that cyclosporine A also blocks the T3/Ti-mediated increase in intracellular pH (Rosoff and Teres, 1986). The link between T3/Ti and polyphosphoinositide turnover suggests that perturbation of T3/Ti, in addition to increasing [Ca2+Ii,also activates pkC. Direct support for this notion stems from the demonstration that T3 mAbs induce the translocation of pkC activity from the cytosol to the membrane fraction. In unstimulated peripheral T cells, pkC activity is recovered almost entirely from the cytosolic fraction. Following the addition of a T3 mAb, there is a >90% decrease in cytosolic pkC activity and a proportionate, concomitant increase in membrane-associated pkC (Farrar and Ruiscetti, 1986). Under these conditions, the TS/Ti-mediated translocation of pkC is maximal at 10 minutes and of 2 hours, demonstrating that the duration of the TS/Ti-mediated translocation can be considerably influenced by the form of the ligand (Manger et al., 1987). Immobilized, but not soluble, T3 mAb can activate certain T cell lines and clones in the absence of additional stimuli. On the other hand, immobilized T3 mAb do not activate resting peripheral T cells and Jurkat cells, suggesting that in these cells PMA may activate an as yet unidentified signaling pathway in addition to pkC. In support of this notion, either IL-1 or 9.3 (an mAb reactive with Tp44) can +
+
+
+
+
H UMAN T LYMPHOCYTE ACTIVATION
23
synergize with immobilized T3/Ti to activate resting peripheral T cells and Jurkat cells, yet neither IL-1 nor 9.3 induce detectable translocations of pkC (Weiss et al., 1986; Manger et al., 1987). At least one function of T3/Ti during signal transduction is to activate the phosphodiesterase that hydrolyzes PIP,. In membranes prepared from several nonlymphoid tissues, the addition of guanosine triphosphate (GTP) increases the activity of the PIP, phosphodiesterase, suggesting that Ca2+mobilizing receptors may be coupled to this enzyme by GTP binding proteins (G proteins) (Cokcroft and Gomperts, 1985). One approach to the identification of these putative G proteins has been to take advantage of the ability of certain bacterial toxins, such as cholera toxin and pertussis toxin, to covalently modify and functionally alter G proteins (Cassel and Pfeuffer, 1978; Katada and Ui, 1982). Exposure of Jurkat cells to cholera toxin completely inhibits TS/Ti-mediated polyphosphoinositide turnover and increases in [Ca2+Ii (Imboden et al., 1986). This effect of cholera toxin on T3/Ti is temporally dissociated from its well-recognized ability to ADPribosylate the stimulating G protein (Gs) of adenylate cyclase and is not mimicked by directly activating adenylate cyclase with forskolin. These observations suggest that a cholera toxin substrate, presumably a G protein other than Gs, regulates signal transduction by T3/Ti. An attractive candidate for this substrate is Go, a G protein that can functionally interact with muscarinic receptors in reconstitution experiments, but whose physiologic role is not known (Florio and Sternweis, 1985). Although originally thought not to be a substrate for cholera toxin, the cloning of Go revealed that it contains a cholera toxin ADP-ribosylation site (Angus et al., 1986). The multimeric complexity of T3/Ti appears to be unique among receptors that are linked to polyphosphoinositide turnover. Even those receptors such as the hepatic a,-adrenergic receptor, which stimulates extracellular Ca2+ influx as well as mobilizes intracellular Ca2+, have no known associated T3-like chains. If we assume that Ti is sufficient for antigen recognition, then the complexity of T3 suggests either that the regulation of signal transduction by T3/Ti is exceptionally complicated or that T3/Ti has a signaling role in addition to activating the hydrolysis of PIP,. Of interest in this regard is the demonstration that T3 polypeptides can be phosphorylated in response to activation by antigen or by treatment with PMA. Antigen recognition by a murine T cell hybridoma results in serine phosphorylation of the T3-6 chain and tyrosine phosphorylation of p21 of the antigen-receptor complex (Samelson et al., 1985, 1986). The kinase(s) involved is (are) not known, but treatment of this T cell hybridoma with PMA also induces serine phosphorylation of the T3-6 chain, raising the possibility that activated pkC phosphorylates this chain during physiologic T cell activation. However, PMA, but not antigen, also stimulates phosphorylation of the
24
ARTHUR WEISS AND J O H N B. I MB O D E N
nonglycosylated 26-kDa E chain of the antigen-receptor complex. Phorbol esters stimulate the phosphorylation of T3 components of the human antigen-receptor complex (Cantrell et al., 1985). Exposure of either human T lymphoblasts or the T cell line, HPB-ALL, to phorbol dibutyrate leads to a rapid (detectable in 50% decrease in the cell surface expression of T3/Ti in human T cells and can render T cell clones unresponsive to antigen (Cantrell et al., 1985; Ando et al., 1985). One possible consequence of phorbol ester-induced phosphorylation of T3, then, is the desensitization and down-regulation of T3/Ti. B. T11 As discussed in Section II,B, perturbation of T11 by the appropriate combination of mAb activates human T lymphocytes. These same combinations of T11 mAb also trigger increases in [Ca2+Iiin T cells, suggesting that an increase in [Ca2+Ii constitutes an intracellular signal for T11-mediated activation (M. Weiss et al., 1984). Perturbation o f T l l also increases [Ca2+Iiin human thymocytes and NK cells that do not express T3, clearly demonstrating that the T11-mediated increase in [Ca2+Ii does not require the cell surface expression of T3/Ti (June et al., 1986; Alcover et al., 1986). It is not known whether perturbation of T11 stimulates the generation of IP,. An indirect argument against a link between T11 and polyphosphoinositide turnover stems from the observation that the T11-mediated increase in [Ca2+], in quin2-loaded T cells is due exclusively to extracellular Ca2+ influx (Alcover et al., 1986). In studies of Indo-1-loaded T cells, however, T11-induced intracellular Ca2+ mobilization has been reported (June et al., 1986).The discrepancy between these two observations may reflect the need for higher intracellular indicator concentrations when using q u i d (all intracellular Ca2+ indicators chelate Ca2+ and, in sufficient concentrations, can prevent increases in [Ca2+Iidue solely to intracellular Ca2+ release) (Lew et al., 1984). Clearly, direct measurements of inositol phosphates following stimulation of T11 are needed. In the absence of other stimuli, perturbation of T11 can activate T lymphocytes, including resting peripheral T cells (Meuer et al., 1984b). Because an increase in [Ca2+Iialone is not sufficient to activate T cells (Weiss et al., 1984a; Truneh et al., 1985), it is very likely that signal generation by T I 1 involves more than an increase in [Ca2+Ii.Moreover, in its ability to provide all the requirements for the activation of resting peripheral T cells, T11 appears to have a signaling capability that is not observed with T3/Ti.
HUMAN T LYMPHOCYTE ACTIVATION
25
Whether these additional signaling effects of T11 are mediated through pkC and/or some other intracellular pathway remains to be determined. C. Thy-1 In the presence of PMA, most Thy-1 mAb activate murine peripheral T lymphocytes when a second-layer anti-Ig antibody is added (Kroczek et al., 1986a). This synergy between the Thy-l-mediated signal and PMA suggests that perturbation of Thy-1 increases [Ca2 Ii. Thy-l-induced increases in [Ca2+Ii have been clearly demonstrated in several cellular systems. Quin2loaded murine T cells and thymocytes that have been pretreated with Thy-1 mAb exhibit an increase in [Ca2+Iiin response to the addition of an anti-Ig antibody (Kroczek et al., 1986a). While most Thy-1 mAb require “crosslinking” with anti-Ig, one Thy-1 mAb, G7, elicits a prompt, sustained increase in [ Ca2+Ii when added directly to quin2-loaded T cell hybridoma cells. G7 activates these hybridomas in the absence of other stimuli. The Thy-l-mediated increase in [Ca2+Ii has been studied further by introducing the Thy-1 gene into cells that do not normally express Thy-1. Following transfection with the Thy-1.2 gene, Jurkat cells and several B cell lines express abundant cell surface Thy-1. The addition of Thy-1 mAb to transfected Jurkat cells, but not to the parental cells, leads to a sustained increase in [Ca2+Ii and, in the presence of PMA, elicits the production of IL-2 (Gunter et al., 1986). Of considerable interest, Thy-l-mediated increases in [Ca2+Iican be demonstrated in two of three B cell transfectants (Kroczek et al., 1986b). These observations establish that Thy-1 can deliver an activation signal to human T cells and that the Thy-l-mediated signal is independent of any other T cell-specific molecule. Thy-1 is anchored to the plasma membrane by a covalent attachment to phosphatidylinositol (or its phosphorylated derivatives) and not by a hydrophobic domain (Low and Kincade, 1985; but see also Seki et al., 1985). This conclusion is particularly intriguing in view of the ability to Thy-1 to transmit an activation signal. The known association between PIP, hydrolysis and Ca2 mobilization suggests that the covalent attachment of Thy-1 to phosphatidylinositol is in some way causally linked to the Thy-l-mediated increase in [Ca2+Ii. It is difficult, however, to develop a mechanistic model along these lines because the inositol ring attached to Thy-1 is extracellular (inositol phosphates are hydrophilic and have only been implicated as intracellular messengers). Alternatively, Thy-l-mediated signaling may require interaction with another plasma membrane protein which in turn triggers an increase in [Ca2+Ii.The transfection studies demonstrate that this putative effector molecule cannot be T cell specific and must be sufficiently well conserved that murine Thy-1 functionally interacts with the human effector. +
+
26
ARTHUR WEISS A N D JOHN B . IMBODEN
D. Tp67, Tp44,
AND THE
IL-1 RECEITOR
There are few studies of signal transduction by Tp67, Tp44, and the IL-1 receptor during T cell activation. These cell surface structures have been variably described as delivering signals that mimic the effects either of Ca2 ionophores or of phorbol esters. Indeed, high concentrations of a Tp44 mAb, 9.3, stimulate IP, generation and increases in [Ca2+Iiwhen added to Jurkat cells (Weiss et al., 1986a). The 9.3-mediated increases in IP,, however, are only a fraction of those observed following stimulation with T3/Ti mAb and are not associated with detectable pkC translocation. A direct effect of Tp44 on either [Ca2+Iior pkC, therefore, does not provide a satisfactory explanation for the ability of low concentrations of 9.3 to synergize with immobilized T3/Ti mAb in the activation of Jurkat and resting peripheral T cells. By a similar argument a direct effect on polyphosphoinositide turnover is not likely to be the only signal transducing mechanism for Tp67 and IL-1 receptor, given the abilities if these structures to synergize only with immobilized T3/Ti mAb in activation. +
V. Role of Intracellular Signals Other Than Ca2+ and pkC
There is indirect evidence for a mitogen-derived intracellular signal other than Ca2 and pkC during T cell activation (Kaibuchi et al., 1985; Gelfand et al., 1985). PHA, for example, can augment the proliferative response of peripheral T cells to the combination of ionomycin and synthetic DG (Kaibuchi et al., 1985). The mechanism of this PHA-mediated effect and the cell surface structure(s) involved is not known. Signaling pathways that have been implicated in T cell activation include changes in cyclic nucleotides (Wedner and Parker, 1976) and the opening of voltage-gated K+ channels (Decoursey et al., 1984). Voltage-gated K + channels, which have been recently reviewed, are the predominant ion channel expressed by T lymphocytes (Decoursey et al., 1984; Matteson and Deutsch, 1984). The gating characteristics of these channels are altered by PHA such that the channels open more frequently and at more negative membrane potentials following the addition of PHA to patch-clamped T cells. K + channel blockers inhibit PHA-mediated mitogenesis (Chandy et al., 1984), but the specificity of these inhibitory effects has been questioned (Gelfand et al., 1986). Specific cell surface structures that regulate the opening of K + channels have not yet been identified. +
VI. Effects of Early T Cell Activation Events upon Gene Regulation
As detailed above, activation of resting T cells is initiated by cell surface ligand-receptor interactions which result in second-messenger generation.
H U M A N T LYMPHOCYTE ACTIVATION
27
Subsequently, poorly understood events lead to the transcriptional activation of a certain set of genes which are responsible for the early manifestations of T cell activation. It is not clear whether the activation of the relevant genes is in response to the second messengers described above or, more likely, through an as yet undefined cascade of events which leads to the appearance of elements that can bind to targeted sequences and regulate this set of responsive genes. It is likely that following the initial activation of a certain set of targeted genes, a cascade of later gene activation events is initiated by the secondary effects of the products of these primary target genes. Indeed, the activation of IL-2 and its receptor can result in a second wave of gene activation, leading to mitogenesis (Stern and Smith, 1986). The regulation of T cell activation would appear to be most tightly regulated at the level of the activation of this initial set of responsive genes and will thus be the focus of further discussion. Individual T cells appear to respond to activating stimuli with the expression of different overlapping menus of responsive gene products. Hence, heterogeneity among the different lymphokines secreted by individual T cell clones is well documented (Kelso et al., 1982; Prystowsky et al., 1982). The basis for this response heterogeneity of target genes is not clear. Among the many genes that are activated during the initial phase of T cell activation are the oncogenes c-myc and c-fos, the IL-2 receptor, and a variety of lymphokines, including IL-2 and IFN-y.
A. c-myc AND c-fos The protooncogenes c-myc and c-fos encode nucleoproteins (Curran et al., 1984; Eisenman et al., 1985). They have attracted particular interest because they are among the earliest genes to become transcriptionally active in stimulated T cells (Persson et al., 1984; Kaczmarek et al., 1985; Reed et al., 1985). Thus, c-fos and c-myc transcripts can be observed within 10 minutes following the stimulation of resting PBL with PHA and persist for 2 or 48 hours, respectively (Reed et al., 1986). Based on nuclear runoff technology in which the synthesis of nascent RNA chains is examined, this appears to represent true transcriptional activation of c-myc (Kronke et al., 1985). The requirements for the expression of c-myc and c-fos have not been intensively studied. Whereas it is clear that the lectin PHA, a calcium ionophore, or PMA alone can induce c-myc expression (Reed et al., 1985), a synergistic effect is observed if PMA is used in combination with either PHA or Ca2+ ionophore (Yamamoto et al., 1985; Granelli-Piperno et al., 1986). The presence of adherent cells can also increase the inductive effects of PHA upon c-myc expression (Kern et al., 1986). These results support the notion
28
ARTHUR WEISS A N D JOHN B . IMRODEN
that activation of c-myc and c-fos involves the synergistic effects of an increase in [Ca2+],and the activation of pkC, as discussed above. As the appearance of c-fos and c-myc transcripts precedes the accumulation of IL-2 and IL-SR, it was tempting to speculate that these nucleoproteins might influence the expression of IL-2 and IL-R (Kronke et al., 1985; Reed et al., 1985). However, the addition of a protein synthesis inhibitor, cycloheximide, failed to affect the expression of c-myc, c-fos, IL-2, or IL-2R transcripts (Kronke et al., 1985; Reed et al., 1985). This result has recently been challenged (Weiss et al., 1987). Thus, the role of c-myc and c-fos expression in T cell activation is not clear.
B. INDUCTION OF IL-2
AND
IFN-y
The cloning of the IL-2 and IFN-y genes has greatly facilitated the study of their regulation (Taniguchi et al., 1983; Holbrook et al., 1984; Gray and Goeddel, 1982). The appearance of IL-2 and IFN transcripts correlates well with the production of the protein products of these genes (Efrat et al., 1982). This involves the transcriptional activation of both these genes (Efrat and Kaempfer, 1984; Kronke et al., 1984; Weiss et al., 1986b). Posttranscriptional control is suggested by the ability of cycloheximide to superinduce IL-2 mRNA (Efrat and Kaempfer, 1984). The appearance of both transcripts requires simultaneous stimulation of purified resting T cells or Jurkat cells with PMA together with ligands which increase [Ca2+Ii,such as lectins PHA or Con A, mAb reactive with the T3/Ti complex, or calcium ionophores, although small amounts of IL-2 and IFN-y transcripts may be seen with the lectins alone (Weiss et al., 1984a; Wiskocil et al., 1985; Granelli-Piperno et al., 1986; Kern et al., 1986; Yamamoto et al., 1985). Exceptions to this general model have been well documented, as in the case of the murine EL-4 and the human HUT-78 lines in which PMA alone may be sufficient for IL-2 activation (Farrar et al., 1980a). The basis for this discrepancy is not clear, but, at least in the case of HUT-78, may reflect the more activated phenotype of the cell (Manger et al., 1985). Hence, the synergistic effects of the activation of pKC and increases in [Ca2+Iiappear to be involved in the activation of IL-2 and IFN-y genes in resting T cells. It takes 2-6 hours for IL-2 or IFN-y transcripts to accumulate following stimulation (Wiskocil et al., 1985; Kronke et al., 1985; Reed et al., 1986). It is not clear what events are required to occur during this time period. The failure of cycloheximide to inhibit the appearance of IL-2 or IFN-y transcripts would argue against the participation of another protein whose synthesis is induced earlier (Kronke et al., 1985; Reed et al., 1986). However, conflicting findings have recently been obtained (Weiss et al., 1987). The ability of cyclosporin A to inhibit the induction of IL-2 and IFN-y transcrip-
HUMAN T LYMPHOCYTE ACTIVATION
29
tion may provide some clues regarding these earlier events (Wiskocil et al., 1985; Kronke et al., 1984; Reed et al., 1985). Since cyclosporin A does not interfere with the increase in [Ca2+Ii or activation of pkC in stimulated Jurkat cells, it must interfere with an event more closely linked to the transcriptional activating event (Wiskocil et al., 1985; Manger et al., 1986). Recent reports suggest that cyclosporin A binds to and inhibits calmodulin (Colombani et al., 1985). Another set of studies has identified yet another cyclosporin binding protein, cyclophyllin (Handschumacher et al., 1984). The role of these two proteins in the activation of the lymphokine genes is not yet clear. Apparent coordinate regulation of IL-2 and IFN-y has been observed. Thus, the stimuli required for their activation, their kinetics of appearance, and their inhibition by cyclosporin are sirnilar (Wiskocil et al., 1985). Whether other lymphokines are similarly coordinately regulated is not clear. In contrast, as discussed below, IL-2R is not coordinately regulated with its ligand, IL-2 (Kronke et al., 1984, 1985). The unique conformation of regulatory regions of genes in intact nuclei can render these regions unusually sensitive to digestion by DNase I. The mapping of DNase I hypersensitive sites is one approach to the identification of DNA sequences which are potential sites of transcriptional regulation. The application of this approach to the human IFN-y gene reveals a prominent hypersensitive site within the first intron in nuclei obtained from Jurkat cells (Hardy et al., 1985). This site is far less sensitive to DNase I in a spontaneous variant of Jurkat that has lost the ability to produce IFN-y and is not present in B cells and nonlymphoid cells. The presence of this hypersensitivity, therefore, correlates with the capability of a cell to produce IFN-y. Computer scanning of this intronic hypersensitive region reveals a 25 bp sequence, with 83% homology to a sequence located 300 bp upstream from the promoter of the human IL-2 gene and may be relevant for their coordinate regulation. A similar analysis has been applied to the IL-2 gene (Siebenlist et d.,1986). In order to identify functional DNA sequences of the human IL-2 gene, Fujita et al. (1986) fused potential regulatory regions of the IL-2 gene to the chloramphenicol acetyltransferase (CAT) structural gene. The expression of these recombinants was then studied in resting and activated T cell 1'ines as well as in several non-T cell lines. In an elegant series of experiments, Fujita et al. identified a 200 bp segment in the 5' flanking region of the IL-2 gene that can mediate inducible T cell-specific gene expression. This sequence functioned in an orientation-independent fashion, suggesting this region of the IL-2 gene is a regulatory enhancer whose function is restricted to activated T lymphocytes. Similar results have recently been obtained by Durand et al. (1986).
30
ARTHUR WEISS A N D JOHN B. IMBODEN
C. REGULATIONOF
THE
IL-2 RECEPTOR
The IL-2 receptor (IL-2R) is expressed on activated T cells, but more recently also has been identified on other cells (reviewed in Waldman, 1986). From the partial protein sequence, the cDNA encoding the IL-2R has been isolated (Leonard et al., 1984; Cosman et al., 1984; Nikaido et al., 1984). Its expression is transcriptionally regulated (Leonard et al., 1985b; Kronke et al., 1985). At least three distinct initiation sites can be used and appear to be differentially utilized, depending on how the gene is induced (Leonard et al., 1985a). Its regulation in T cell activation is distinct from the regulation of its ligand, IL-2. As discussed above, its expression can be induced by PMA only or by other ligands which effectively activate pkC (Leonard et al., 1985b). Moreover, unlike IL-2 or IFN-7, its activation is not inhibited by cyclosporin A (Kronke et al., 1984). In addition to the effects of the primary activation stimuli, it is up-regulated by its ligand, IL-2 (Leonard et al., 1985b; Hemler et al., 1984; Smith and Cantrell, 1985; Reem et al., 1985). This may explain some of the synergistic effects observed with stimuli used together with PMA, which induce IL-2. Thus, the effect may not be upon the transcription of the IL-2R directly, but may be an indirect effect mediated by the IL-2 induced. A more detailed review of the regulation of the expression of the IL-2R has recently been published (Waldman, 1986). VII. Summary
The physiologic activation of human T cells by antigen involves events that occur between ligands and receptors at the interface of the T cell and antigen-presenting cell (or target cell). These events have been examined by identifying the cell surface receptors involved in such interactions using mAb. Whereas the T3/T cell antigen receptor plays a central role in such interactions, other T cell receptors have been identified which may also contribute to T cell activation in providing primary activation signals or by functioning as accessory molecules. Although the ligands of these other receptors are currently unknown or ill defined, it is likely that this will provide a fruitful area of investigation. The use of mAb as probes to mimic these putative ligands has facilitated the study of the requirements for activation and the biochemical events initiated by the receptors involved. The T cell receptor, a multisubunit complex, has been most intensively studied. Ligands that bind to T3/Ti cannot initiate activation by themselves and require the participation of accessory molecules. Stimulation of T3/Ti results in the formation of at least two potent intracellular second messengers, IP3 and DG, through the hydrolysis of PIP,. These second messengers, in turn, induce an increase in
HUMAN T LYMPHOCYTE ACTIVATION
31
[Ca2+Iiand the activation of pkC. These two events appear to be essential in the transcriptional activation of certain targeted genes through ill-defined pathways leading to the manifestations of T cell activation.
ACKNOWLEDGMENTS We would like to thank Mr. Michael Armanini and Ms. Denise Go for their excellent secretarial assistance in the preparation of the manuscript. This work was supported in part by a grant from the Arthritis Foundation to A.W. J.I. is a Pfizer Scholar.
REFERENCE s Aarden, L. A., Burnner, T. K., Cerottini, J. C . , Dayer, J. M., de Weck, A. L., Dinarello, C. A., Di Sabato, G., Farrar, J. J., Gery, I., Gillis, S., Handschumacher, R. E., Henney, C. S., Hoffman, M. K., Koopman, W. J., Karane, S. M . , Lachman, L. B., Lekowits, I., Mishell, R. I., Mizel, S. B., Oppenheim, J. J., Paetkau, V., Plate, J., Rollinghoff, M . , Rosenstreich, D., Rosenthal, A. S., Rosenwasser, L. J., Schimpl, A , , Shin, H. S., Simon, P. L., Smith, K. A . , Wagner, H., Watson, J. D., Weckner, E., and Wood, D. D. (1979).J. Zmmunol. 123, 2928. Acuto, O., Hussey, R. E., Fitzgerald, K. A., Protentis, J. P., Meuer, S. C., Schlossman, S. F., and Reinherz, E. L. (1983). Cell 34, 717. Alcover, A., Weiss, M., Daley, J., and Reinherz, E. (1986). Proc. Natl. Acad. Sci. U.S.A. 83,
2614. Allison, J. P., and Lanier, L. L. (1985). Nature (London) 314, 107. Allison, J. P., McIntyre, B. W., and Bloch, D. (1982).J . Zrnmunol. 129, 2293. Ando, I., Hariri, G., Wallace, D., and Beverley, P. (1985). Eur. J. Zmmunol. 15, 196. Angus, C., Van Meurs, K., Tsai, S., Adamik, P., Medel, M., Pan, Y., Kung, H . , Moss, J., and Vaughn, M. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 15813. Auron, P. E., Webb, A. C . , Rosenwesser, L. J., Mucci, S. F., Rich, A,, Wolff, S. M . , and Dinarello, C. A. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 7907. Auron, P. E., Rossenwasser, L. J., Matsushima, K., Copeland, T., Dinarello, C. A,, Oppenheim, J. J., and Webb, A. C. (1985).J. Mol. Cell. Zmmunol. 2, 169. Bank, I., and Chess, L. (1985).J. E x p . Med. 162, 1294. Bank, I., DePinho, A., Brenner, M. B., Cassimeris, J., Alt, F. W., and Chess, L. (1986). Nature (London) 322, 179. Bell, R. (19815).Cell 45, 631. Berridge, M. J. (1983). Biochem. J. 212, 849. Berridge, M. J., and Irvine, R. F. (1984). Nature (London) 312, 315. Biddison, W. F . , Rao, P. E., Talle, M. A . , Goldstein, G., and Shaw, S. (1982).J. E x p . Med. 156, 1065. Blankenhorn. E., and Douglas, T. (1972).J . Zfered. 63, 259. Borst, J., Prendiville, M. A , , and Terhorst, C. (1982).J. Zrnmunol. 128, 1560. Borst, J., Prendiville, M. A , , and Terhorst, C. (1983a). Eur. J. Zmmunol. 13, 576. Borst, J., Alexander, S., Elder, J . , and Terhorst, C. (1983b).J. Biol. Chem. 258, 5135. Brenner, M. B., Trowbridge, I. S., and Strominger, J. L. (1985). Cell 40, 183. Brenner, M. B., McLean, J . , Dialynas, D. P., Strominger, J. L., Smith, J. A , , Owen, F. L., Seidman, J. G., Ip, S., Rosen, F., and Krangel, M. S. (1986). Nature (London) 322, 145. Brottier, P., Boumsell, L., Gelin, C., and Bernard, A. (1985).J. Zmmunol. 135, 1624. Campbell, D., Gagnan, J., Reid, K., and Williams, A. (1981). Biochem. J. 195, 15.
32
ARTHUR WEISS AND JOHN B . IMBODEN
Cantrell, D., Davies, A,, and Crumpton, M. (1985). Proc. Natl. Acad. Sci. U . S . A . 82, 8158. Cassel, D., and Pfeuffer, T. (1978). Proc. Natl. Acad. Sci. U . S . A . 75, 2669. Catagna, M., Takai, Y . , Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. (1982).J. B i d . Chem. 257, 7847. Ceuppens, J. L., and Baroja, M. L. (1986).J. Zmmunol. 137, 1816. Ceuppens, J. L., Bloemmen, F. J., and Van Wauwe, J. P. (1985). J. Zmmunol. 135, 3882. Chandy, K., DeCoursey, T. E., Cahalan, M., McLaughlin, C., and Gupta, S. (1984).J. E x p . Med. 160, 369. Chang, T. W., Kung, P. C . , Gingras, S. P., and Goldstein, G. (1981). Proc. Natl. Acad. Sci. U . S . A . 78, 1805. Chang, T.-W., Testa, D., Kung, P. C., Perry, L., Dreskin, H. J., and Goldstein, G . (1982).J. Zmmunol. 128, 585. Cockcroft, S . , and Gomperts, B. (1985). Nature (London) 314, 534. Cohen, F., Novotny, J., Steinberg, M., Campbell, D., and Williams, A. (1981). Biochem. J. 195, 31. Collins, M. K. L., Tanigawa, G . , Kissonerghis, A.-M., Ritter, M., Price, K. M., Tonegawa, S., and Owen, M. J. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 4503. Colombani, P. M., Robb, A , , and Hess, A. D. (1985). Science 228, 337. Cosman, D., Cerretti, D. P., Larsen, A , , Park, L., March, C., Dower, S.,Gillis, S., and Urdal, D. (1984). Nature (London) 312, 768. Coussens, L., Parker, P., Rhee, L., Yang-Feng, T., Chen, E., Waterfield, M., Francke, U . , and Ullrich, A. (1986). Science 233, 859. Curran, T., Miller, A. D., Zokas, L., and Verma, J. M. (1984). Cell 36, 259. Davignon, D., Martz, E., Reynolds, T., Kurzinger, K., and Springer, T. A. (1981).J. Zmmunol. 127, 590. Dawson, R., Frienkel, M., Jungalwala, F., and Clark, N. (1971). Biochem. J. 122, 605. DeCoursey, T. C., Chandy, K. G., Gupta, S., and Cahalan, M. D. (1984). Nature (London) 307, 465. Dembic, Z . , Haas, W., Weiss, S., McCubrey, J., Kiefer, H., von Boehmer, H., and Steinmetz, M . (1986). Nature (London) 320, 232. DeVries, J. E., Vyth-Dreese, F. A., van der Hulst, R., Sminia, P., Figdor, C. G., Bont, W. S., and Spits, H. (1982). Zmmunobiology 162, 103. Dixon, J., and Hokin, L. (1985).J. B i d . Chem. 260, 16068. Dower, S. K., Kronheim, S. R., March, C. J., Conlon, P. J., Hopp, T. P., Gillis, S.,and Urdal, D. L. (1985).J. E x p . Med. 162, 1985. Durand, D. B., Bush, M. R., Morgan, J. G., Weiss, A , , and Crabtree, G. R. (1987).J. E x p . Med., 165, 395. Efrat, S., and Kaempfer, R. (1984). Proc. Natl. Acad. Sci. U . S . A . 81, 2601. Efrat, S., Pilo, S., and Kaempfer, R. (1982). Nature (London) 297, 1982. Eisenman, R. N . , Tachibana, C. Y., Advams, H. D., and Hann, S. R. (1985).Mol. Cell. Biol. 5, 114. Evans, G. A . , Ingraham, H . A., Lewis, K., Cunningham, Seki, T., Moriuchi, T., Chang, H. C., Silver, J., and Hyman, R. (1984). Zmmunology 81, 5532. Farrar, W., and Anderson, W. (1985). Nature (London) 315, 233. Farrar, W. L., and Ruscetti, F. W. (1986).J. Zmmunol. 136, 1266. Farrar, J., Mizel, S.,Fuller-Farrar, J , , Farrar, W., and Hilfiker, M. (1980a).J . Zmmunol. 125, 793. Farrar, J. J., Fuller-Farrar, J., Simon, P. L . , Hilfiker, M. L., Stadler, B. M., and Farrar, W. L. (1980b).J . Zmmunol. 125, 2555. Fast, L. D., Hansen, J. A,, and Newman, W. (1981).J. Zmmunol. 127, 448.
H U M A N T LYMPHOCYTE ACTIVATION
33
Fernandez-Botran, R . , Krammer, P. H., Diamantstein, T., Uhr, J. W., and Vitetta, E. S. (1986).J. E x p . Med. 164, 580. Fink, P., Matis, L. A , , McElligott, D. L., Bookman, J., and Hedrick, S. M. (1986). Nature (London) 321, 219. Florio, V., and Sternweiss, P. (1985).J. Biol. Chern. 260, 3477. Fox, D. A,, Ilussey, R. E. Fitzgerald, K. A,, Bensussan, A,, Daley, J. F., Scholossman, S. F., and Reinherz, E. L. (1985).J . Znirnunol. 134, 330. Fox, D. A,, Schlossman, S . F., and Reinherz, E. L. (1986).J. Zrnrnunol. 136, 1945. Fujita, T., Shibuya, H., Ohashi, T., Yamanishi, K., and Taniguchi, T. (1986). Cell 46, 401. Garman, R. D., Doherty, P. J., and Raulet, D. H. (1986).Cell 45, 733. Gelfand, E., Cheung, R., Mills, G . , and Grnnstein, S. (1985). Nature (London) 315, 419. Gelfand, E., Cheung, R . , and Grunstein, S. (1986).J. B i d . Chern. 261, 11520. Gery, I., Gershon, R. K., and Waksman, B. H. (1972).J. E x p . Med. 136, 128. Gillis, S., and Mizel, S. B. (1981).Proc. Natl. Acad. Sci. U.S.A. 78, 1133. Gold, D. P., Puck, J. M., Pettey, C. L., Cho, M., Coligan, J., Woody, J. N . , andTerhorst, C. (1986). Nature (London) 321, 431. Gonwa, T. A , , Peterlin, B. M., and Stobo, J. D. (1983).Ado. Zrnrnunol. 34, 71. Granelli-Piperno, A., Andrus, L., and Steinman, R. M. (1986).J. E x p . Med. 163, 922. Gray, P. W.: and Goeddel, D. V. (1982). Nature (London) 298, 859. Gunter, K. C., Malek, T. R . , and Shevach, E. M. (1984).J. E x p . Med. 159, 716. Gunter, K. C., Kroczek, R. A,, Shevach, E. M . , and Germain, R. N. (1986).J. E x p . Med. 163, 285. Haars, R . , Kronenberg, M., Gallatin, W. M., Weissman, I. L., Owen, F. L., and Hood, L. (1986).J. E x p . Med. 64, 1. Handschumacher, R. E., Harding, M. W., Rice, J., and Drugge, R. J. (1984).Science226,544. Hansen, J. A , , Martin, P. J . , and Nowinski, R. C. (1980). Irnrnunogenetics 10, 247. Hara, T., and Fu, S. M. (1985).J. E x p . Med. 161, 641. Hara, T., Fu, S. M., and Hansen, J. A. (1985).J. E x p . Med. 161, 1513. Hardy, K., Peterlin, M., Atchison, R . , and Stobo, J. (1985).Proc. Natl. Acad. Sci. U.S.A. 82, 8173. Haskins, K., Kubo, R . , White, J., Pigeon, M., Kappler, J . , and Marrack, P. (1983).J. E x p . Med. 157, 1149. Hayday, A. C., Saito, H., Gillies, S. D., Kranz, D. M . , Tanigawa, G., Eisen, H. N . , and Tonegawa, S. (1985). Cell 40, 259. Hemler, M. E., Brenner, M . B., McLean, J. M., and Strominger, J. L. (1984). Proc. Natl. Acad. S c i . U . S . A . 81, 2172. Henkart, P. A , , and Fisher, R. I. (1975).J. Zrnmunol. 114, 710. Hirata, M., Sasaguri, T., Hamachi, T., Hashimoto, T., Knkita, M., and Koga, T. (1985).Nature (London) 317, 723. Holbrook, N . J., Smith, K. A,, Fornace, A. J., Comeau, C. M., Wiskocil, R. L., and Crabtree, 6. R. (1984). Proc. Natl. Acad. Sci. U . S . A . 81, 1634. Hollander, K.,Pillemer, E., and Weissman, I. (1981).Proc. Natl. Acad. Sci. U.S.A. 78, 1148. Holter, W., Fischer, 6. F., Majdic, O., Stockinger, H., and Knapp, W. (1986).J. E x p . Med. 163, 654. Hood, L., Kronenberg, M . , and Hnnkapiller, T. (1985). Cell 40, 225. Howard, F. I]., Ledbetter, J. A,, Wong, J., Bieber, C. P., Stinson, E. B . , and Herzenberg, L. A. (1981).J. Irnrnunol. 126, 2117. Imboden, J. B., and Stobo, J. D. (1985).J. E x p . Med. 161, 446. Imboden, J., Weiss, A , , and Stoho, J. D. (1985).J. Zrnrnunol. 134. 663. Imboden, J., Shoback, D., Pattison, G., and Stobo, J. (1986).Proc. Natl. Acad. Sci. U . S . A . 83, 5673.
34
ARTHUR WEISS A N D J OHN B. I M B O D E N
Imboden, J., Weyand, C., and Goronzy, J. (1987).J. Zmmunol. 138, 1322. Inoue, M., Kishimoto, A., Takai, Y., and Nishizuka, Y. (1977).J . Biol. Chem. 252, 7610. Irvine, R. F., Letcher, A. J., Heslop, J. P., and Berrideg, M . J. (1986). Nature (London) 320, 631. Jones, N. H., Clabby, M. L., Dialynas, D. P., Huang, H.-J. S., Herzenberg, L. A., and Strominger, J. L. (1986). Nature (London) 323, 346. June, C., Ledbetter, J,, Rabinovitch, P., Martin, P., Beatty, P., and Hansen, J. (1986).J . Clin. Invest. 77, 1224. Kaczmarek, L., Calabretta, B., and Baserga, R. (1985). Proc. Natl. Acad. Sci. U.S.A.82, 5375. Kaibuchi, K., Takai, Y., and Nishihzuka, Y. (1985).J. Biol. Chem. 260, 1366. 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. Kanellopoulos, J. M., Wigglesworth, N. M . , Owen, M. J., and Crumpton, M. J. (1983).EMBO J. 2, 1807. Kanellopoulos, J. M., De Petris, S., Leca, G., and Crumpton, M. J. (1985). Eur. J. Zmmunol. 15, 479. Kappler, J., Kubo, R., Haskins, K., White, J., and Marrack, P. (1983a). Cell 34, 727. Kappler, J., Kubo, R., Haskins, K., Hannum, C., Marrack, P., Pigeon, M., McIntyre, B., Allison, J., and Trowbridge, I. (1983b). Cell 35, 295. Katada, T., and Ui, M. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 3129. Kaye, J., and Janeway, C. A. (1984).J . E x p . Med. 159, 1397. Kelso, A,, Glasebrook, A. L., Kanagawa, O., and Brunner, K. T. (1982).J . Zmmunol. 129,550. Kern, J. A., Reed, J. C., Daniele, R. P., and Nowell, P. C. (1986).J. Zmmunol. 137, 764. Kikkawa, U., Takai, Y., Tanaka, Y., Miyake, R., and Nishizuka, Y. (1983).J . Biol. Chem. 258, 11442. Knopf, J., Lee, M., Sultzman, L., Kruz, R., Loomis, C., Hewick, R., and Bell, R. (1986). Cell 46, 491. Kraft, A., and Anderson, W. (1983). Nature (London) 301, 621. Kranz, D. M., Tonegawa, S., and Eisen, H. N. (1984).Proc. Natl. Acad. Sci. U . S . A . 81, 7922. Kranz, D. M., Saito, H . , Heller, M., Takagaki, Y., Haas, W., Eisen, H. N., andTonegawa, S. (1985). Nature (London) 313, 752. Krissansen, G. W., Owen, M. J., Verbi, W., and Crumpton, M. J. (1986). EMBOJ. 5, 1799. Kroczek, R. A., Gunter, K. C., Seligmann, B., and Shevach, E. M. (1986a).J . Zmmunol. 136, 4379. Kroczek, R. A., Gunter, K. C., Germain, R. N., and Shevach, E. M. (1986b).Nature (London) 322, 181. Kronheim, S. R., March, C. J., Erb, S. K., Conlon, P. J., Mochizuki, D. Y., and Hopp, T. P. (1985).J. Exp. Med. 161, 490. Kronke, M., Leonard, W. J., Depper, J. M., Arya, S. K., Wong-Staal, F., Gallo, R. C., Waldmann, T. A., and Greene, W. C. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 5214. Kronke, M., Leonard, W. J., Depper, J. M., and Greene, W. C. (1985).J . Exp. Med. 161, 1593. Kuno, M., Goronzy, J., Weyland, C., and Gardner, P. (1986). Nature (London) 323, 269. Kurt-Jones, E. A,, Beller, D. I., Mizel, S. B., and Unanue, E. R. (1985).Proc. Natl. Acad. Sci. U.S.A. 82, 1204. Kurzinger, K., Reynolds, T., Germain, R. N., Davignon, D., Martz, E., and Springer, T. A. (1981).J. Zmmunol. 127, 596. Lancki, D. W., Lorber, M. I., Loken, M. R., and Fitch, F. W. (1983).J . Exp. Med. 157, 921. Landegren, U., Anderson, J., and Wigzell, H. (1984). Eur. J. Zmmunol. 14, 325. Lanier, L. L., and Weiss, A. (1986). Nature (London), 324, 268.
HUMAN T LYMPHOCYTE ACTIVATION
35
Lanier, L. L . , Ruitenberg, J. J., Allison, J. P., and Weiss, A. (1986).J. Zmmunol. 137, 2286. Ledbetter, J. A., Evans, R. L., Lipinski, M., Cunningham-Rnndles, C., Good, R. A,, and Herzenberg, L. A. (1981).J. E x p . Med. 153, 310. Ledbetter, J. A., Martin, P. J., Spooner, C. E., Wofsy, D., Tsu, T. T., Beatty, P. G., and Gladstone, P. (1985).J . Zmmunol. 135, 2331. Ledbetter, J. A,, June, C. H., Martin, P. J., Spooner, C. E., Hansen, J. A,, and Meier, K. E. (1986).J , Zmmunol. 136, 3945. LeFranc, M.-P., Forster, A , , Baer, R . , Stinson, M. A,, and Rabbitts, T. H. (1986).Cell 45,237. Leonard, W. J., Depper, J. M., Crabtree, G . R., Rudikoff, S., Pumphrey, J . , Robb, R. J., Kronke, M., Svetlik, P. B., Peffer, N. J.. Waldmann, T. A., and Greene, W. C. (1984). Nature (London) 311, 627. Leonard, W. J., Depper, J. M . , Kanehisa, M., Kronke, M., Peffer, N . J., Svetlik, P. R., Sullivan, M., and Greene, W. C. (1985a). Science 230, 633. Leonard, W. J., Kronke, M., Peffer, N. J., Depper, J . M., and Greene, W. C. (1985b). Proc. Natl. Acczd. Sci. U . S . A . 82, 6281. Lew, P. D., Wollhaim, C. B., Waldvogel, F. A., and Pozzan, T. (1984).J. Cell B i d . 99, 1212. Littman, D. (1987). Annu. Reo. Zmmunol. 5, 561. Logdberg, L., and Shevach, E. M. (1985). Eur. J. Zmmunol. 15, 1007. Low, M. G . , and Kincade, P. W. (1985). Nature (London) 318, 62. MacDonald, H . R . , Glasebrook, A. L., and Cerottini, J.-C. (1982).J. E x p . Med. 156, 1711. MacDonald, H . R . , Bron, C., Rousseaux, M . , Horvath, C., and Cerottini, J.-C. (1985). Eur. J. Zmmunol. 15, 495. McIntyre, B. W., and Allison, J. P. (1984). Cell 38, 659. Malvija, A., Louis, J., and Zwiller, J. (1986). FEBS Lett. 199, 213. Manger, B., Weiss, A,, Weyand, C., Goronzy, J., and Stobo, J. D. (1985).J. Zmmunol. 135, 3669. Manger, B., Weiss, A , , Imboden, J.. and Stobo, J. D. (1987).J. Zmmunol., in press. March, C. J., Mosley, B., Larsen, A,, Cerretti, D. P., Braedt, G., Price, V., Gillis, S., Henney, C. S., Kronheim, S. R., Grabstein, K., Conlon, P. J., Hopp, T. P., and Cosman, D. (1985). Nature (London) 315, 641. Marrack, P., Endrews, R., Shimonkevitz, R., Zlotnik, A., Dialynas, D., Fitch, F., and Kappler, J. (1983).J. E x p . Med. 158, 1077. Martin, P. J., Hansen, J. A , , Nowinski, R. C., and Brown, M. A. (1980). Zmmunogenetics 11, 429. Martin, P. J . , Longton, G., Ledbetter, J . A , , Newman, W., Braun, M. P., Beatty, P. G., and Hansen, J. A. (1983).J. Zmmunol. 131, 180. Martin, P., Ledbetter, J., Morishita, Y., June, C., Beatty, P., and Hansen, J. (1986). J. Zmmunol. 136, 3282. Matteson, D. R., and Deutsch, C. (1984). Nature (London) 307, 468. Mellonni, E . , Pontremoli, S., Michetti, M., Sacco, D., Sparatore, B., Salammo, F., and Horecker, B. (1985). Proc. Natl. Acad. Sci. U . S . A . 82, 6435. Mentzer, S. J , Barbosa, J. A , , and Burakoff, S. J. (1985).J. Zmmunol. 135, 34. Meuer, S. C., and Meyer zum Buschenfelde, K.-H. (1986).J. Zmmunol. 136, 4106. Meuer, S. C . , Schlossman, S. F., and Reinherz, E. L. (1982). Zmmunology 79, 4395. Meuer, S. C . , Fitzgerald, K. A., Hussey, R. E . , Hodgdon, J. C., Schlossman, S. F ., and Reinherz, E. L. (1983a).J. E r p . Med. 157, 705. Meuer, S. C., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1983b).J. Exp. Med. 158, 988. Meuer, S. C., Hussey, R. E . , Cantrell, D. A,, Hodgdon, J. C., Schlossman, S. F., Smith, K. A., and Reinherz, E. L. (1984a). Proc. Natl. Acad. Sci. U . S . A . 81, 1509.
36
ARTHUR WEISS A N D J O H N B. IMBODEN
Meuer, S. C., Hussey, R. E., Fabbi, M., Fox, I)., Acuto, O., Fitzgerald, K. A,, Hodgdon, J. C., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1984b). Cell 36, 897. Mizel, S. B. (1982). Zmmunol. Reo. 63, 51. Moldwin, R. L., Lancki, D. W., Herold, K. C., and Fitch, F. W. (1986).J. E x p . Med. 163, 1566. Moretta, A,, Pantaleo, G., Lopez-Botet, M., and Moretta, L. (1985a).J. E x p . Med. 162, 823. Moretta, A , , Pantaleo, G., Lopez-Botet, M., Mingari, M.-C., Carrel, S., and Moretta, L. (1985b). E u r . J. Zmmunol. 15, 841. Mosmann, T. R., Bond, M. W., Coffman, R. L., Ohara, J., and Paul, W. E. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 5654. Nagasawa, R., Gross, J., Kanagawa, O., Townsend, K., Lanier, L., Chiller, J., and Allison, J. (1987).J. Zmmunol., 138, 815. Nidel, J.. Kuhn, L. J., and Vanderbark, G. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 36. Nikaido, T., Shimizu, A,, Ishida, N., Sabe, H., Teshigawara, K., Maeda, M., Uchiyzma, T., Yodoi, J., and Honjo, T. (1984). Nature (London) 311, 631. Nisbet-Brown, E., Cheung, R. K., Lee, J. W. W., and Gelfand, E. W. (1985).Nature (London) 316, 545. Nishizuka, Y. (1986). Science 233, 305. Oettgen, H., Terhorst, C., Cantley, L., and Rosoff, P. (1985). Cell 40, 583.10. Oettgen, H. C., Pettey, C. L., Maloy, W. L., andTerhorst, C. (1986). Nature(London)320, 272. O’Flynn, K., Krensky, A. M., Beverley, P. C. L., Burakoff, S. J., and Linch, D. C. (1986). Nature (London) 313, 686. Ohashi, P., Mak, T. W., Van den Elsen, P., Yanagi, Y., Yoshikai, Y., Calman, A. F., Terhorst, C., Stobo, J. D., and Weiss, A. (1985). Nature (London) 316, 602. Palacios, R., and Martinez-Maza, 0. (1982). J . Zmmunol. 129, 2479. Parker, P., Loussens, L., Totty, N., Rhee, L., Young, S., Chen, E., Stabel, S., Waterfeld, M., and Ullrich, A. (1986). Science 233, 853. Persson, H., Henninghausen, L., Taub, R., DuGrode, W., and Leder, P. (1984). Science 225, 687. Prystowsky, M. B., Ely, J. M., Beller, D. I., Eisenberg, L., Goldman, J., Goldman, M., Goldwasser, E., Ihle, J., Quintans, J., Remold, H., Vogel, S. N., and Fitch, F. W. (1982). J. Zmmunol. 129, 2337. Raulet, D. H., Garman, R. D., Saito, H., and Tonegawa, S. (1985). Nature (London) 314, 103. Reed, J. C., Nowell, P. C., and Hoover, R. G. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 4221. Reed, J. C., Alpers, J. D., Nowell, P. C., and Hoover, R. G. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 3982. Reem, G. H., Yeh, N.-H., Urdal, D. L., Kilian, P. L., and Farrar, J. J. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 8663. Reif, A. E., and Allen, J. M. (1966a). Nature (London) 209, 521. Reif, A. E., and Allen, J. M. (1966b). Nature (London) 209, 523. Reif, A. E., and Allen, J. M. (1984).J. E x p . Med. 120, 413. Reinherz, E. L., and Schlossman, S. F. (1980). Cell 19, 821. Reinherz, E. L., Kung, P. C., Goldstein, G., and Schlossman, S. (1979a).J . Zmmunol. 123, 1312. Reinherz, E. L., Kung, P. C., Goldstein, G., and Schlossman, S. F. (1979b).Proc. Natl. Acad. Sci. U.S.A. 76, 4061. Reinherz, E. L., Meuer, S. C., Fitzgerald, K. A., Hussey, R. E., Hodgdon, J. C., Acuto, O., and Schlossman, S. F. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 4104. Rosoff, P., and Cantley, L. (1985).J. B i d . Chem. 260, 14053. Rosoff, P., and Teres, 6. (1986). Cell. Biol. 103, 457.
HUMAN T LYMPHOCYTE ACTIVATION
37
Royer, H. D., Acuto, O., Fabbi, M., Tizard, R., Ramachandran, K., Smart, J. E., and Reinherz, E. L. (1984). Cell 39, 261. Royer, H. D., Ramarli, D., Acuto, O., Campen, T. J., and Reinherz, E. L. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 5510. Saito, H . , Kranz, D. M., Takagaki, Y . , Hayday, A. C., Eisen, H. N . , andTonegawa, S. (1984). Nature (London) 309, 757. Samelson, L E . , Germain, R. N . , and Schwartz, R. H. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 6972. Samelson, L. E., Harford, J., and Klausner, R. D. (1985). Cell 43, 223. Samelson, L. E., Patel, M. D., Weissman, A. M., Harford, J. B., and Klausner, R. D. (1986). Cell, 46, 1083. Sanchez-Madrid, F., Krensky, A. M., Ware, C. F., Hobbins, E., Strominger, J. L., Burakoff, S. J., and Springer, T. A. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 7489. Scheurich, P., Ucer, R., Wrann, M., and Pfizenmaier, K. (1985). Eur. J. lmmunol. 15, 1091. Schwab, R., Crow, M. K., Russo, C . , and Weksler, M. E. (1985). J. Zmmunol. 135, 1714. Seki, T., Chang, H., Moriuchi, T., Denome, R., Pleogh, H., and Silver, J. (1985). Science 227, 649. Shackelford, D., and Trowbridge, I. (1986).J. B i d . Chem. 261, 8334. Shapiro, D., Adams, B., and Niederhbuber, J. (1985).J. Zmmunol. 135, 2256. Siebenlist, V . , Durand, D. B., Bressler, P., Holbrook, N. J., Norris, C. A., Kamoun, M . , Kant, J. A., and Crabtree, G. R. (1986). Mol. Cell. Biol., 6, 3042. Siliciano, R. F., Pratt, J. C., Schmidt, R. E . , Ritz, J., and Reinherz, E. L. (1985). Nature (London) 317, 428. Sim, G. K., Yague, J., Nelson, Marrack, P . , Palmer, E . , Augustin, A , , and Kappler, J. (1984). Nature (London)312, 771. Sitkovsky, M. V . , Pasternack, M. S., Lugo, J. P., Klein, J. R., and Eisen, H. V. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 1519. Siu, G . , Clark, S. P., Yoshikai, Y., Malissen, M., Yanagi, Y . , Strauss, E., Mak, T. W., and Hood, L,. (1984). Cell 37, 393. Smith, C. A , , and Rennick, D. M. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 1857. Smith, K. A , , and Cantrell, D. A. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 864. Smith, K. G. C., Austyn, J. M., Kairiri, C . , Beverley, P. C. L., and Morris, P. J. (1986).Eur. J. Zmmunol. 16, 478. Smith, R. T.. Shimizu, S., Norcross, M. A , , and Maino, V. C. (1982). Clin. Zmmunol. Zmmunopathoi. 23, 286. Springer, T. ,4., Davignon, D., Ho, M. K., Kurzinger, K., Martz, E., and Sanchez-Madrid, F. (1982). lmmunol. Reo. 68, 111. Springer, T. A , , Dustin, M. L., Kishimoto, T. K., and Marlin, S. D. (1987). Annu. Reu Zmmunol. 5, 223. Stern, J. B., and Smith, K. A. (1986). Science 233, 203. Stewart, S. J . , Prpic, V., Powers, F. S., Bocckino, S. B., Isaacks, R. E., and Exton, J. H. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 6098. Streb, H., Irvine, R., Berridge, M., and Schulz, I. (1983). Nature (London) 306, 67. Swain, S . L., Dutton, R. W., Schwab, R., and Yamamoto, J. (1983).J. E x p . Med. 157, 720. Tadmori, W . , Kant, J. A,, and Kamoun, M. (1986).J. Zmmunol. 136, 1155. Takai, Y . , Kishirnoto, A , , Iwasa, Y., Kawahara, Y . , Mori, T., and Nishizuka, Y. (1979).J. B i d . Chem. 254, 3692. Taniguchi, T., Matsui, H., Fugita, T., Takaoka, C . , Kashima, N., Yoshimoto, R. and Hamuro, J. (1983). Nature (London) 302, 305. Tax, W. J. M , Willems, H. W. Reekers, P. P. M., Capel, P. J. A , , and Koene, R. A. P. (1983). Nature (London) 304, 445.
38
ARTHUR WEISS A N D JOHN B. I M B O D E N
Tax, W. J. M., Hermes, F. F. M., Willems, R. W., Capel, P. J. A., and Koene, R. A. P. (1984) J . Zmmunol. 133, 1185. Truneh, A., Albert, F., Colstein, P., and Schmitt-Verhulst, A. (1985). Nature (London) 313, 318. Tse, A. G . D., Barclay, A. N., Watts, A., and Williams, A. F. (1985). Science 230, 1003. Tsoukas, C. D., Landgraf, B., Bentin, J., Valentine, M., Lotz, M., Vaughan, J. H., andcarson, D. A. (1985). J. Zmmunol. 135, 1719. Van den Elsen, P., Shepley, B.-A,, Borst, J., Coligan, J. E., Markham, A. F., Orkin, S., and Terhorst, C. (1984). Nature (London) 312, 413. Van Wauwe, J. P., De Mey, J. R., and Goossens, J. G. (1980).J. Zmmunol. 124, 2708. Van Wauwe, J. P., Goossens, J. G., and Beverley, P. C. L. (1984). J . Zmmunol. 133, 129. Von Wussow, P., Platsoucas, C. D., Wiranowska-Stewart, M., and Stewart, W. E. (1981).J. Zmmunol. 127, 1197. Wakasugi, H., Bertoglio, J., Tursz, T., and Fradelizi, D. (1985).J . Zmmunol. 135, 321. Waldmann, T. A. (1986). Science 232, 727. Wedner, H., and Parker, C. (1976). Prog. Allergy 20, 195. Weiss, A., and Stobo, J. D. (1984).J . E x p . Med. 160, 1284. Weiss, A., Wiskocil, R. L., and Stobo, J. D. (1984a).J . Zmmunol. 133, 123. Weiss, A., Imboden, J., Shoback, D., and Stobo, J . (1984b). Proc. Natl. Acad. Sci. U.S.A. 81, 4169. Weiss, A., Manger, B., and Imboden, J. (1986a).J. Zmmunol. 137, 819. Weiss, A., Imboden, J., Hardy, K., Manger, B., Terhorst, C., and Stobo, J. (1986b). Annu. Reu. Zmmunol. 4, 593. Weiss, A., Newton, M., and Crommie, D. (1986~).Proc. Natl. Acad. Sci. U.S.A. 83, 6998. Weiss, A., Shields, R., Newton, M., Manger, B., and Imboden, J. (1987). J . Zmmunol. 138, 2169. Weiss, M., Daley, J., Hodgdon, J . , and Reinherz, E. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 6836. Wessmer, P., Chan, C., Logdberg, L., and Shervach, E. M. (1985).J. Zmmunol. 135, 2237. Wilkinson, M., and Morris, A. (1984). Eur. J. Zmmunol. 14, 708. Williams, J. M., Deloria, D., Hansen, J. A,, Dinarello, C. A,, Loertscher, R., Shapiro, H. M., and Strom, T. B. (1985). J. Zmmunol. 135, 2249. Williamson, J., Cooper, R., Joseph, S., and Thomas A. (1985). Am. J. Physiol. 248, C203. Wilson, D. B., Connolly, T. M., Bross, T. E., Majerus, P. W., Sherman, W. R., Tyler, A. N., Rubin, L. J., and Brown, J. E. (1985). J . Biol. Chem. 260, 13496. Wiskocil, R . , Weiss, A., Imboden, J., Kamin-Lewis, R., and Stobo, J. (1985).J . Zmmunol. 134, 1599. Wolf, M., LeVine, H., May, S . , Cuatrecasas, P., and Sahyoun, N. (1985a). Nature (London) 317, 546. Wolf, M., Cuatrecasas, P., and Sahyoun, N. (1985b). J. Biol. Chem. 260, 15718. Yamada, H., Martin, P. J., Bean, M. A., Braun, M. P., Beatty, P. G., Sadamoto, K., and Hansen, J. A. (1985). Eur. J . Zmmunol. 15, 1164. Yamamoto, Y., Ohmura, T., Fujimoto, K., and Onoue, K. (1985). Eur. J . Zmmunol. 15, 1204. Yanagi, Y., Yoshikai, Y., Leggett, K., Clark, S. P., Aleksander, I., and Mak, T. W. (1984). Nature (London) 308, 145. Yang, S. Y., Chouaib, S . , and Dupont, B. (1986).J . Zmmunol. 137, 1097. Yokota, T., Otsuka, T., Mosmann, T., Banchereau, J., DeFrance, T., Blanchard, D., De Vries, J. E., Lee, F., and Arai, K.-I. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 5894. Yoshikai, Y., Yanagi, Y., Suciu-Foca, N., and Mak, T. W. (1984). Nature (London) 310, 506. Yoshikai, Y., Clark, S. P., Taylor, S., Sohn, U., Wilson, B. I., Minden, M . D., and Mak, T. W. (1985). Nature (London) 316, 837. Zauderer, M., Iwamoto, A., and Mak, T. W. (1986).J . E x p . Med. 163, 1314.
ADVANCES IN IMMUNOLOGY, VOL. 41
Function and Specificity of T Cell Subsets in the Mouse JONATHAN SPRENT AND SUSAN R. WEBB Department of Immunology, IMM4A, Research Institute of Scripps Clinic, La Jollo, California 92037
1. Introduction
Specific antibody molecules play a vital role in counteracting infectious organisms in the extracellular milieu, but antibodies and antibody progenitors, B cells, are poorly equipped to react against organisms harbored inside cells. To deal with intracellular organisms, the immune system has evolved a quite different set of immunocompetent cells, T cells. These cells are imbued with a number of interesting properties (1).First, in contrast to B cells and antibody molecules, T cells generally do not manifest specificity for free antigen, despite the fact that T cells are highly antigen specific. Second, unlike B cells, T cells do not secrete their antigen-specific receptors. Third, under physiological conditions, T cells only respond to antigen displayed on the surface of living cells. Fourth, T cells show the intriguing requirement that, to be immunogenic, antigen has to be aligned on the cell surface in association with gene products of the major histocompatibility complex (MHC),’ the H-2 complex in mice. With their disregard for free antigen, T cells are thus programmed to concentrate their attentions on parasitized cells, these cells being flagged by breakdown products of the pathogen linked to surface H-2 molecules. Especially in the case of viral infections., parasitized cells recognized by T cells are destroyed by cytotoxic T lymphocytes (CTL), a subset of T cells with cytolytic properties. 1 Abbreviations: Ab, antibody; AC, accessory cells; APC, antigen-presenting cells; ATS, antithymocyte serum; “B” mice, thymectomized, irradiated mice reconstituted with T-depleted marrow cell:;; BUdR, bromodeoxyuridine; CAS, supernatant from concanavalin A-stimulated lymphoid cells; CFA, complete Freund’s adjuvant; CML, cell-mediated lympholysis; Con A, concanavalin A; cOVA, chicken ovalbumin; CTL, cytotoxic T lymphocytes; DC, dendritic cells; dCuo, deoxy guanosine; DTH, delayed-type hypersensitivity; GVHD, graft-versus-host disease; HA, histocompatibility antigens; HAN, hemagglutinin; HEV, high endothelial venules; HRC, horse red blood cells; Ig, immunoglobulin; IL, interleukin; KLH, heyhole limpet hemocyanin; mAb, monoclonal antibody; MHC, major histocompatibility complex; MLR, mixed-lymphocyte reaction; M+, macrophage; PHA, phytohemagglutinin; PMA, phorbol myristate acetate; self + X, foreign non-MHC antigen seen in association with self MHC determinants; SRC, sheep red blood cells; TCR, T cell receptor; TNP, trinitrophenol.
39 Copyright 0 1987 by Academic Press, Inc. All rights of‘reproduction in any form reserved.
40
JONATHAN SPRENT A N D SlJSAN R . W E B B
In addition to destroying cells harboring pathogenic organisms, T cells also play a major role in controlling the quantity and quality of specific antibody made by B cells. This function of T cells is controlled by a different subset of T cells, T helper cells. Whereas T killer cells provide a negative (cytotoxic) signal to antigen-bearing cells, T helper cells provide a positive signal. Thus, when a T helper cell encounters a foreign antigen linked to H-2 molecules on the surface of a B cell, the T cell delivers a helper signal which enables the B cell to differentiate into an antibody-forming cell. Another interesting property o f T cells is that, although both T killer and T helper cells generally ignore self components, including self H-2 molecules per se, T cells show a marked propensity to respond to foreign H-2 molecules-the phenomenon of alloreactivity. The issue of how T cells discriminate between self and nonself H-2 determinants while remaining reactive to antigen is one of the main themes of this review. The principal aim of this article is to outline how the specificity, function, and induction of T cells and T cell subsets are under the strict control of H-2 molecules. The scope of the subject matter covered in this article is quite large: It should be emphasized that our underlying intention is not to discuss each topic in exhaustive detail, but rather to give an overview of a highly complex field. II. Cell Surface Molecules Controlling T Cell Specificity and Function
Formulating concepts of how and why T cells display H-2-restricted specificity inevitably hinges on understanding the various receptor-ligand interactions which take place during T cell recognition. Three types of molecules are of obvious importance: antigen-specific T cell receptors (TCR), H-2 molecules, and so-called accessory molecules on T cells. A brief description of these molecules and their genes is given below.
A. THE T CELLRECEPTOR Although the basic structure of antibody molecules has been known for many years, the nature of the TCR remained a near-total mystery until only a short time ago. During the 197Os, a considerable body of evidence from a number of influential investigators led to the view that Ig genes encoded the TCR (reviewed in Ref. 2). Although some of this evidence looked highly convincing on paper, a sizable cross section of immunologists remained skeptical as typified by the extensive review on T cells by Jensenius and Williams (2) in 1982. The first direct information on TCR molecules came from studies in 19821983 with monoclonal antibodies (mAb) reacting with clonotype-specific structures on a T lymphoma (3) and on antigen-specific T cell clones and
T CELL SUBSETS IN T H E M O U S E
41
hybridomas (4-7). These studies revealed that the TCR consists of a disulfide-linked, glycosylated heterodimer consisting of two chains, a and p. Each chain is a transmembrane protein and, in mice, has an MW of 40,00043,000, the MW of the heterodimer being about 82,000. Like certain other surface proteins, both chains of the TCR show considerable homology with immunoglobulin (Ig) and are thus viewed as forming part of the “Ig supergene family” (reviewed in Ref. 8). By isoelectric focusing (9,10), peptide mapping (10-12), polypeptide sequencing (13- 15), and gene sequencing (see below), TCR a and f3 chains show extensive polymorphism. As for Ig molecules, polymorphism is largely limited to the variable (V) NH,-terminal domain of each chain, the constant (C) COOH-terminal domain anchored to the cell membrane showing little variability. In a brief %year period from 1984 to 1986, recombinant DNA technology has provided a detailed picture of the genetic organization of the genes encoding the TCR. Using the technique of subtractive cDNA-mRNA hybridization, initial studies on mouse (16) and human (17) T cells led to the isolation of cDNA clones for segments of DNA which rearranged exclusively in T cells. The rearranged genes were shown to encode the V and C regions of the TCR p chain. Subsequent work has revealed the following information on TCR p genes (18-34). The p gene complex in mice is situated on chromosome 6 (19,20) and contains a cluster of about 30 V gene segments (27,29);in contrast to Ig genes, the family of V genes is divided into a large number of subfamilies (19,27,29),each containing relatively few members, usually 1-3 (19,27,29). Downstream from the V genes, there are two consecutive clusters of D (diversity), J (joining), and C region segments, the two C regions being almost identical (21-25); each cluster has 1 D region and 6 J regions. Recently, a third C region, CPO, has been mapped between Jp1 and C p l (30). A single Vp gene has been mapped 3’ to Cp2 (28); curiously, this gene is in inverted transcriptional polarity relative to all the other known genes in the p complex. Soon after the initial description of p chain genes, subtractive cDNA hybridization led to the isolation of cDNA clones for both mouse (35,36) and human (3‘7) TCR a chain genes. The a complex is situated on chromosome 14 in mice (38) and shows close similarities with the p complex, although there are some distinct differences (38-43). Thus, in contrast to the p complex, there are large numbers of V a genes, perhaps 100 (33). Only one C a gene segment has been identified and, to date, no D a regions have been found. A unique aspect of the a complex is that there are probably of the order of 50 Jasegments which are dispersed over at least 60 kb of DNA 5’ to
Ca (39,40). TCR genes show the same type of ordered rearrangement observed for Ig genes (8,34). Thus, through flanking sequences analogous to those for Ig
42
JONATHAN SPRENT AND SUSAN R. WEBB
genes, D-J joining of p genes is followed by V-D-J rearrangement to form a single exon. VQ gene segments can rearrange to either of the two Cp segments, since each C p segment has its own set of upstream D and J segments. At present, there is no evidence that individual T cells express more than one type of a-p heterodimer on their surface. Although both chromosomes can show rearrangement of (Y and Q genes, productive rearrangements seem to be restricted to one chromosome, rearrangements on the other chromosome being incomplete or abortive (26). A number of different mechanisms shape the diversity of the TCR (22,27,31,33):(1) combinatorial rearrangement of different V, D, and J gene segments, (2) junctional diversity, i. e., imprecise V-D- J joining plus V-D joining in any of three translational reading frames, and (3) N-region diversity, i.e., addition of extra nucleotides at the V-D-J joint. It has been calculated (33) that these mechanisms in toto can generate in the order of 8 X lo6 combinatorial associations of expressed a,p genes. This is quite close to the estimate of 2 X lo7 different associations for Ig (V,-V,) molecules. An interesting difference between T and B cell receptors is that somatic hypermutation seems to be very rare in T cells (22,26,27,44). Thus, whereas primed B cells and myeloma cells show considerable hypermutation (which leads to increased affinity for antigen) (reviewed in Ref. 45), there is currently only one example of such mutation occurring in T cells, and these changes were observed in a T hybridoma (46). T cell clones do occasionally change their specificity over prolonged periods in vitro (47-49), but so far these alterations in specificity have not been shown to involve point mutations. In a recent report, the change in specificity found in a long-term cytotoxic clone was attributed to secondary rearrangement and expression of Q chain gene segments (48). Although the available evidence is clearly consistent with TCR being intrinsically resistant to somatic hypermutation (perhaps to guard against the emergence of self-reactive cells), the evidence is still too fragmentary to make firm statements on this important issue. In addition to TCR (Y and p chains, considerable attention is now being focused on a third chain, y (49-61). Originally mistaken for the a gene, the y complex of genes is situated on chromosome 13 in mice (50) and contains a fairly small number of V gene segments, of which at least four are functional (59). There are currently four known Cy segments (one silent), but each Cy gene is flanked 5’ by only a single J region. The potential diversity of y gene products thus appears to be fairly limited. Although both T cell subsets can express y mRNA (53,59,60), many of these transcripts might be nonfunctional. Indeed, until very recently (see below), the protein product of the y complex has been a topic of mere speculation. Although models have been proposed in which the y chain forms heterodimers with (Y or p chains (62), the function of at least some typical T cells clearly does not depend on
T CELL SUBSETS I N THE M O U S E
43
functional y transcripts (52,55,56,63,64). The possible role of y in thymic differentiation will be considered in Section V,A. The discovery of y gene products hinged on the fact that classic a-P heterodimers are linked to T3, a heterogeneous complex of polypeptides containing, in mice, at least 3 monomeric glycoproteins and a family of homo- and heterodimers (65-69). Under defined conditions, mAb to human T3 molecules precipitate the a-f3 dimer from nearly all T cells. However, a small proportion of T cells (1-3%) in the thymus and blood are T 3 + , but lack a-p TCR heterodimers. Studies on cell lines of this type isolated from blood showed that anti-T3 mAb coprecipitated two non-disulfide-linked chains of MW 55,000 and 40,000 (70). Significantly, the 55,000 MW protein was precipitated by antisera raised against synthetic peptides corresponding to nucleotide sequences of Cy and Vy genes. Since the cell lines expressed potentially functional y mRNA transcripts, the data strongly suggest that the 55,000 MW is the y product. The 40,000 MW protein, provisionally termed 6, was not precipitated by anti-y antisera, and its relationship to the 55,000 MW protein is unknown. In parallel studies by another group (71), an a-pT3+ cell line was isolated from human thymus. Anti-T3 antibodies co-precipitated two chains from the cell line, with MWs of 62,000 and 44,000. Only the 44,000 MW protein was precipitated with anti-y antibodies. More recently, 7-6 heterodimers have been found on mouse T cells (D. Pardoll, personal communication); y-6 heterodimers are disulfide linked in mice and, as in man, appear to be expressed largely and perhaps solely on the minority population of a- p- T3+ cells. At this time, the functions of a-6 heterodimers on T cells is totally obscure, although it is intriguing that ybearing cells triggered by anti-T3 antibodies develop cytolytic activity (71).
B. H-2 MOLECULES Early studies on histocompatibility loci controlling skin allograft rejection in mice showed that histocompatibility differences encoded by one particular locus, the Histocompatibility-2 (H-2)locus (now known to be a complex of genes), led to a particularly rapid rate of graft rejection (72-74). It soon became clear that all mammals examined and many lower species have a major histocompatibility complex (MHC) and that the MHC controls all forms of “strong” histocompatibility reactions, including allograft rejection and graft-versus-host disease (GVHD), as well as the in vitro counterparts of these reactions, such as the mixed-lymphocyte reaction (MLR) and assays for cell-mediated lympholysis (CML) (reviewed in Ref. 74). At a time when “cellular” immunology is rapidly becoming subcellular, it is instructive to reflect that contemporary T cell immunobiology had its humble beginnings in studies on transplantation reactions. Evidence from these early studies led to a number of key conclusions. First, the studies of
44
JONATHAN SPRENT A N D SUSAN R . WEBB
Owen (75) on tolerance in chimeric cattle twins led Burnet and Fenner (76) to propose that self tolerance is an immunological process and is acquired in neonatal life; proof for this theory was obtained by Billingham et al. (77) and HaSek (78). Second, Simonsen (79,80),working with chickens, demonstrated that a very high proportion of normal lymphoid cells express MHC alloreactivity. Third, Gowans (reviewed in Ref. 81) proved that alloreactivity is mediated by a population of long-lived small lymphocytes which recirculate continuously from one lymphoid organ to another; this was a decisive breakthrough because small lymphocytes at that time were regarded as ignominious cells with a limited life span whose main function was to act as trephocytes, i.e., to die quietly and make their DNA available for reutilization by other more “important” cells (82). Fourth, the studies of Miller (83) on the effects of neonatal thymectomy in mice demonstrated that alloreactivity is controlled by thymus-derived (T) lymphocytes. Fifth, groups headed by Benacerraf (84) and McDevitt (85) showed that the inability of certain members of a species to respond to particular foreign antigens was controlled by “immune response” (Ir) genes; these genes mapped within the MHC. Finally, a number of workers in the early 1970s (reviewed in Refs. 1,74) demonstrated that, in addition to serving as alloantigens, self MHC molecules guide T cell reactivity to typical foreign (non-MHC) antigens. Following these pioneering studies, the genetic makeup of the MHC and the gene products of this complex have come under close scrutiny. Although the MHC encodes a variety of different cell surface molecules, two types of MHC molecules, termed class I and class 11, are of particular relevance to T cell function. These molecules differ from TCR and Ig molecules in always being expressed codominantly. An overview of the H - 2 complex is given below; for detailed information, see Refs. 74, 86-92. Class I molecules are expressed on nearly all cells in the body and consist of two glycoprotein chains: (1) a heavy 45,000 highly polymorphic a chain encoded by the H - 2 complex on chromosome 17, and (2) a light (12,000) nonpolymorphic p chain, &-microglobulin, encoded on chromosome 2. The two chains are noncovalently associated and only the heavy chain is anchored to the cell membrane. Class I a chains consist of three external “domains” (yet to be confirmed by X-ray crystallography), termed a1 (NH,-terminal), a2,and a3; intrachain disulfide bonds exist in the a2 and a3 domains, but not in al. The a3 domain shows relatively little polymorphism and associates with &-microglobulin. The a1 and a 2 domains, by contrast, show extensive polymorphism and probably express all of the sites (epitopes) recognized by T cells. Polymorphic class I a chains are the products of two regions of the H - 2 complex, K and D. The K region encodes at least two different types of class I molecules (K, K’), although it is not yet clear whether these molecules are products of different loci or, conversely, reflect
T C E L L SUBSETS IN THE MOUSE
45
differential processing of a single species of mRNA (93). The D region encodes a varying number of molecules, ranging from a single molecule, in H-2‘’ mice to as many as 5 molecules in H-2“ mice (Dd, L“, R“, M”, and L2d) (9495). Precisely how many loci encode these molecules is still unclear, although protein sequencing analysis has shown that H-24 mice have at least three structurally distinct molecules, Dq, Lq, and Rq (96). As mentioned above, typical K and D molecules show extensive polymorphism, with possibly as many as 100 alleles for each locus (97). K and D molecules are not separable structurally, i. e., the alleles do not display obvious “K-ness” or “D-ness.” Thus, although it seems highly likely that the K and D regions developed by gene duplication from an ancestral gene, this process must have occurred quite early in evolution and been followed by rapid diversification (89). Class I gene segments consist of 8 exons (88).The first exon encodes the hydrophobic leader peptide which guides insertion of the class I molecule through the cell membrane. The second, third, and fourth exons encode the three external domains of the molecule, while the remaining exons encode the transmembrane and cytoplasmic regions; as expected from protein sequencing, extensive polymorphism is apparent only in the second and third exons. Although there are only a limited number of known class I gene products (mice of the H-2b haplotype express only two molecules, K” and D’)), there are many class I genes--26 in the B10 (H-2”) mouse (98) and 36 in the BALBIc (Ei-2”)mouse (99,100). Most of these genes, however, are situated in the Qa and Tla regions, which are telomeric to the K and D regions. The function of’these genes is still unclear, although it has long been known that the QaITZa region encodes a series of class I-like molecules, collectively termed Qa molecules. These molecules differ from typical class I molecules in several respects (87). First, the expression of Qa molecules appears to be restricted to lymphohematopoietic cells; it was initially thought that Qa molecules were expressed only on subsets of T cells, but more recent evidence suggests that all T cells and at least some B cells are Qa+ (87).Second, Qa molecules display only limited polymorphism, with only 2-6 alleles for each molecule (87). Third, although Qa molecules can serve as targets for allo (foreign) CML responses and can be detected serologically, there is currently 110 evidence that these molecules can serve as “restriction elements”: in other words, unlike typical class I (and class 11) molecules, T cells do not appear to respond to antigen in the context of self Qa molecules. Fourth, in contrast to typical class I molecules, certain Qa molecules, e.g., Qa-2 and (210, are secreted in a soluble form (101,102). Recently, it has been suggested that QaITla genes might act as a reservoir of diversity for class I KID genes (103-105). This evidence stems from ex-
46
JONATHAN SPRENT AND SUSAN H. WEBB
periments on spontaneous class I mutant strains of mice, i.e., mice whose class I H - 2 K molecules differ at 1-3 amino acids relative to the “wild-type” strain of origin (106). Examination of one of these mutants, Kbml, at the DNA level indicated that a 13 nucleotide region of the H-2K region had 7 nucleotide changes relative to the wild-type Kb gene, these changes leading to 3 amino acid substitutions at the protein level (104,105). This finding was difficult to explain by simple, random point mutations. Stimulated by recent studies on globin genes (107), two groups (104,105) took the view that the changes at the DNA level might reflect a gene conversion-like event in which a segment of the K b gene had been replaced by a segment from some other gene. In support of this idea, it was found that one of the QalTla genes, QlO, had a stretch of DNA that was identical to the “new” segment of DNA in the KbnL1gene (106).On the basis of this finding, the authors suggest that gene conversion might be the main mechanism responsible for generating H-2 polymorphism. Although proof for gene conversion, as defined in lower eukaryotes, is not yet available, it is striking that for many of the “bg” series of mutants (of which Kbml is the prototype), potential donor genes have now been found in the QalTla region (106). Unlike class I molecules, the expression of class I1 molecules is limited to certain cell types, especially B cells and dendritic cells (a class of cells involved in presenting antigen to T cells) (see Section II1,C) (74). Various other cell types, such as macrophages, endothelial cells, epithelial cells, and fibroblasts, express class I1 molecules, but only when induced by interferon-? (IFN-7) (108,109). Whereas there appears to be a wide variety of different class I1 molecules in man (11O), mice have only two sets of molecules, termed I-A and I-E (88,97,111).Each of these molecules (often collectively termed Ia molecules) is composed of an a and p chain. These chains, which are not covalently linked, are glycoproteins of similar size (-35 kDa for the a chain and 29 kDa for the p chain). Both chains are transmembrane proteins and show sequence homology with TCR and Ig molecules, and are thus members of the Ig supergene family (88). The p chain has two intrachain disulfide bonds and is presumed to form two domains, p l and p2; the a chain also probably has two domains, a1 and 012, although only a2 (closest to the cell membrane) has an intrachain disulfide bond. Before being expressed on the cell surface, the a and p chains are noncovalently associated with a nonpolymorphic chain, termed the invariant chain (111,112). This chain is encoded on a different chromosome and has little or no homology with the a and p chains (or with the other members of the Ig supergene family) (113). The main function of the invariant chain is presumably to regulate the intracellular transport andlor association of the a and p chains (113-115). Curiously, the invariant chain is reported to appear on the cell
T CELL SUBSETS IN THE MOUSE
47
surface after dissociating from the a dnd P chains (116). Since the invariant chain lacks a signal sequence, it has been suggested that it lies “upside down” in the cell membrane, the carboxy-terminus being extracellular (114,115). The biological significance of cell-surface invariant chain is obscure. At the DNA level there are at least 8 class I1 gene segments in mice, three AP, three E P , one A a , and one Ea gene (117-120). These genes reside in the 1 region of the H - 2 complex between the K and D loci, the order of the genes (and gene clusters) being K , AP, A a , E P , Ea, D (except for one EP gene which maps between Ea and D). Interestingly, the a and P genes are transcribed off opposite strands of DNA, i.e., the genes are in tail-to-tail orientation. As yet, it is not clear how many of the P genes are expressed, although at least one of these genes (AP3) is known to be a pseudogene (120). Strain variation in the expression of Ea and EP genes will be considered later. The arrangement of the exons and introns of Ia a and P chain genes is slightly different. For a typical a chain, a signal sequence exon encodes the leader peptide and the first few amino acids of the mature protein. The second and third exons encode the a 1 and a 2 domains, and the fourth exon encodes the transmembrane region and the cytoplasmic tail as well as part of the 3’ untranslated sequence; the remainder of the untranslated sequence is encoded by a fifth exon. The arrangement of the genes for the P chain is essentially the same, except that the transmembrane and cytoplasmic regions are encoded by two separate exons. Based on nucleotide sequencing, the P chains for both I-A and I-E molecules show extensive polymorphism in their N-terminal domains. Polymorphism is less marked in the I-A a chain and is virtually absent in the I-E a chain. Although the a chains (especially Ea) play a comparatively minor role in contributing to polymorphism of class I1 molecules, a chains are essential for the cell surface expression of the P chains. In this respect, it should be mentioned that a considerable number of mouse strains, e.g., mice of the s, b, f a n d y H - 2 haplotypes, do not express an I-E molecule. In these mice, the E a chain is not expressed. Defective E a expression in these mice involves at least three different mechanisms (121): (1) a deletion in the Ea gene for the b and s haplotypes; (2) an E a mRNA of aberrant size for thef haplotype; and (3) a defect in mRNA processing and/or mRNA instability for the y haplotype. In the absence of the E a chain, the EP chain can be synthesized in the cytoplasm, but fails to be expressed on the cell surface. This situation occurs in mice of b and s haplotypes. If these mice are crossed with mice that do express an E a chain, transchain association can occur. Thus, the E a chain of one parent associates with the EP chain of the other
48
JONATHAN SPHENT A N D SUSAN R. WEBB
parent to form a functional heterodimer which is then expressed on the cell surface. These same heterodimers can be expressed by cis-chain association in appropriate H - 2 recombinant mice, e.g., in recombinants carrying the E; allele and a functional Ea allele. In contrast to the b and s haplotypes, mice of thefand q haplotypes do not express an EP product. These mice are thus incapable of synthesizing an I-E molecule even when crossed with an E,+ strain. Because of transchain association, H-2 heterozygous mice can express unique “F, hybrid” Ia molecules lacking in the two parental strains (122). Thus, in the case of F, hybrids between I-Aa (A;-A$ and I-A“ (At-Abp), one finds four I-A heterodimers: the A:-A; and At-A; heterodimers of the two parental strains plus 2 sets of F1-unique heterodimers, Ai-A; and At-A;. Because of the nonpolymorphism of the E a chain, trans-chain association can create only one F,-unique I-E molecule (with the proviso that one of the parents is EL). H-2 heterozygous mice can therefore express a total of 6 Ia molecules, 4 I-A molecules and 2 I-E molecules. It should be pointed out that, in theory, two additional mechanisms might generate further sets of Ia heterodimers. First, mixed heterodimers might exist between I-A and I-E a and (3 chains, thus creating Ea-AP and/or Aa-EP molecules. Although heterodimers of this type have yet to be observed under normal physiological conditions, gene transfection experiments suggest that Aa-EP heterodimers do form in certain situations (123). Second, it is by no means clear that the clusters of three EP and three AP genes each contain only one functional gene. If several of these genes were functional, the potential diversity of Ia molecules would obviously be considerable. By classical genetic methods involving studies on a variety of H - 2 recornbinant mice, it was originally concluded that the I region of the H - 2 complex was divided into 5 subregions: A, B , J , E , C (reviewed in Ref. 87). It is now generally accepted that the A subregion (a term no longer used) encodes the A@,A a , and EP genes and the E subregion encodes the E a chain. The B, 1, and C subregions, however, appear to be nonexistent at the DNA level (117-120). In view of this disturbing discrepancy, the A, B , J , E , C terminology has fallen into rapid disfavor. The phenomenology that led to the postulated existence of the B , J , and C subregions is outside the scope of this review (see Ref. 87). In addition to the genes for class I and class I1 molecules, the H - 2 complex also contains a number of “passenger” genes, e.g., genes for neuraminidase-1 and the C4 component of complement (87);these genes map between Ea and the D locus. Why these genes reside in the H - 2 complex is obscure. As suggested by Klein et al. (87), the simplest explanation is that these genes were entrapped accidentally during evolution.
T CELL SUBSETS IN THE MOUSE
49
C. ACCESSORYMOLECULES ON T CELLS Like other cells, T cells display a multiplicity of different cell surface molecules, and antibodies to some of these molecules provide invaluable tools for isolating T cells and separating these cells into functionally distinct subsets (124). Defining the physiological role of the various molecules expressed on T cells is still in its infancy, but two sets of molecules have aroused particular interest, Lyt-2/3 (125, 126) and L3T4 (127). These molecules are of importance for two reasons, one practical and one theoretical. First, the expression of L3T4 and Lyt-2/3 molecules on extrathymic T cells is niutually exclusive. Thus, with the exception of rare “doublepositive” ‘r cell clones maintained in uitro (128,129), peripheral T cells are either L3r4 , Lyt-2/3 -, or L3T4 -, Lyt-213 (127). Antibodies to these markers are therefore extremely useful for isolating phenotypically distinct T cell subsets. As discussed in more detail in a later section, the second reason for the high interest in L3T4 and Lyt-213 molecules is that the expression of these molecules correlates quite closely with the class of H-2 molecules recognized by T cells. Moreover, there is reason to believe that L3T4 and Lyt-2/3 molecules might actually bind to H-2 molecules. Before discussing these topical issues, it is usefd to consider recent information on the structure and genetic organization of L3T4 and Lyt-2/3 molecules and genes and their homologues in other species, e.g., CD4 and CD8 molecules in man (130,131) and W3/25 and OX8 molecules in the rat (132). Except for a subset of natural killer (NK) cells (133), the expression of Lyt-2/3 appears to be restricted to T cells (124). The Lyt-2/3 molecule is a 70-kDa heterodimer of two covalently linked chains, an a chain (35-38 kDaj expressing epitopes detected by anti-Lyt-2 antibodies, and a p chain (30-34 kDa) expressing Lyt-3 epitopes (124, 134-137). The Lyt-2 chain is the homologue of the CD8 chain in man, the CD8 molecule consisting of homodimers and multimers of a single chain; there is no apparent counterpart of the Lyt-3 chain in man. The Lyt-2 and Lyt-3 chains are both transmembrane glycoproteins and are encoded by two closely linked genes on chromosome 6. The Lyt-2 gene segment contains 5 exons (138, 139). The first exon encodes the signal peptide and an amino terminal domain showing close homology with Ig V, light-chain regions, including cysteines for an intrachain disulfide bond. Exons 2-5 encode, in order, the spacer region, the transmembrane region, and two cytoplasmic regions, C 1 and C2. Interestingly, although there is only a single Lyt-2 gene segment, there are two types of Lyt-2 chains, a and a’,of slightly different size (124,134,140143). The two chains are identical except that the a‘ chain lacks the C1 and +
+
+
50
IONATHAN SPRENT A N D SUSAN R . WEBB
C2 intracytoplasmic regions of the a chain (138,139). The chains arise from separate species of mRNA and are presumed to reflect alternative modes of mRNA splicing (138,139). The tissue distribution of a and a’chains is somewhat different (124,134). Thymocytes express both chains, whereas lymph node cells express only the a chain (despite the presence of a’mRNA in the cytoplasm). Interest in this differential expression of the a‘ chain is muted by reports that there is no counterpart of the a’chain in human T cells (see Ref. 139). Likewise, the absence of the Lyt-3 chain in human T cells suggests that the biological function of the Lyt-213 heterodimer is largely determined by the Lyt-2 chain. For simplicity, we shall henceforth refer to “Lyt-2 molecules” rather than Lyt-213 molecules. The structure and genetic organization of L3T4 molecules have only recently come under close scrutiny. Whereas Lyt-2 molecules are heterodimers, L3T4 and their homologues in other species are monomeric glycoproteins and are somewhat smaller (55K)than Lyt-2 (127,143-147). Genes for the human (CD4) molecule have now been cloned and sequenced (144) and reveal a single polypeptide chain with three extracellular domains: an amino-terminal Ig V-region-like domain, a joining (J)-like region, and a third extracellular domain. There is also a membrane-spanning domain analogous to class I1 MHC Q chains and a highly charged cytoplasmic domain; the intron-exon organization of the genes is still unclear. Sequencing at both the cDNA (144) and protein (147) levels suggests that the three external domains each have intrachain disulfide bonds. The V-like domain shows significant homology with v k and also with the amino-terminal domain of the CD8 (Lyt-2) molecule. However, the degree of homology between the amino-terminal domains of CD4 and CD8 is fairly weak [28% for CD8 vs CD4 (144)l; in fact, the domains of these two molecules are less homologous to each other than each is to v k (144). Like CD8 molecules, CD4 (L3T4) molecules show considerable divergence from their homologues in other species, suggesting that these molecules have undergone rapid evolution to maintain complementarity with their respective “ligands” (? MHC molecules-see below) (147). In addition to Lyt-2 and L3T4 molecules, T cells also express several other types of “accessory” molecules that might play a role in T cell recognition of antigen, e.g., LFA-1, LFA-2, and LFA-3 (148). Moreover, in the case of human T cells, a variety of mAb are being used to separate T cells into a bewildering complexity of subsets displaying different functions (149,150). We have elected not to discuss this phenomenology, largely because, as yet, there is little or no evidence that murine T cells-the main subject of this review-are divided into more than two phenotypically and functionally distinct subsets.
T CELL SUBSETS IN THE MOUSE
51
111. H-2-Restricted Recognition of Antigen by Mature T Cells
After the discovery of H-2-linked Ir genes (85), the first evidence that T cell function is controlled by H-2 gene products came from studies on T-B collaboration. In the late 1960s, experiments of Claman et al. (151) and Davies et al. (152) followed by the definitive experiments of Mitchell and Miller (153)showed that antibody responses to “T-dependent” antigens involve cooperation between thymus-derived (T) helper cells and bone marrow-derived antibody-forming cell precursors (B cells). In 1972-1973, Kindred and Shreffler (154) and Katz et al. (155) provided convincing evidence that this interaction between T and B cells exhibits H-2 restriction, i.e., T-B collaboration fails to occur unless the two cell types share H-2 determinants; the restricting determinants were shown to map to the 1 region of the H-2 complex. Similar restriction was observed by Rosenthal and Shevach (156) in an in uitro system involving proliferative responses of T cells to antigen presented by syngeneic versus allogeneic macrophages. In 1974, the studies of Zinkernagel and Doherty (157) on antiviral responses and Shearer (158)on responses to the hapten TNP indicated that CTL lyse H-2-compatible but not H-2-incompatible target cells; lysis requires a sharing of class I (K or D) H-2 determinants. Similar restriction to self H-2 determinants was found for CML to minor histocompatibility antigens (HA) by Bevan (159) and Gordon et al. (160). Although Zinkernagel and Doherty were not the first to discover the phenomenon of H-2 restriction, the speculations of these two workers on the physiological significance of H-2 restriction (161,162) made the phenomenon “accessible” to the immunological community at large. Doherty and Zinkernagel put forward two models to account for H-2 restriction: altered self and dual recognition. Both models assert that T cells have joint specificity for self H-2 plus foreign antigen. According to the altered-self model, T cells express a single recognition unit (receptor) with specificity directed neither to self H-2 determinants per se nor to antigen, but to new antigenic determinants (NADs) created by an association of the two ligands. The opposing model, dual recognition, argues that T cells express two linked recognition units, one specific for self H-2 and the other for antigen, the two recognition units either being expressed on two different (though linked) receptors or on a single polypeptide chain. The exposition of these two models was enormously influential, and even now there is no direct proof for either model, although the “two receptor” variant of the dual recognition model would seem to be ruled out. Although the three-dimensional structure of the TCR is still uncertain, the conservative view is that the TCR binding site will show close similarities
52
JONATHAN SPRENT A N D SUSAN R. WEBB
with the binding site of Ig molecules. Thus, despite the accumulating evidence that T cells have specificity for small peptides (see below) whereas antibody (Ab) molecules recognize three-dimensional configurations on native molecules, it is not unreasonable to suppose that the combining sites of TCR and Ab molecules both simply bind complementary “shapes” of best fit. Although it is relatively easy to study antibody-antigen interactions, attempting to establish how T cells recognize antigen is fraught with two imposing difficulties. First, the fact that TCR are not secreted makes it very difficult to prepare these molecules in pure form. Second, we still have only a vague idea of precisely “what” T cells recognize. Before dealing with the complex issue of how T cells recognize antigen, one apparent difference between the V regions of TCR and Ab molecules should be mentioned. In the case of Ab, it has long been known from protein sequencing studies that the V region displays three discrete areas of hypervariability and that these complementarity-determining regions (CDRs) converge in the three-dimensional structure of the combining site to form the contact residues for antigen (163). Whether this also applies for TCR molecules is uncertain, since hypervariability is seen throughout the TCR V region (18, 164). Therefore, it is possible that a wide area of the TCR V region can act as a combining site (or sites). The physiological response of a virgin T cell to antigen can be envisaged as having three components: (1)an induction phase in which the T cell recognizes an immunogenic form of the antigen and is induced to blast transformation, (2) a proliferative phase where the induced T cell undergoes clonal expansion, and (3) a stage of differentiation where the proliferating T cells acquire some type of effector function, e.g., the capacity to lyse appropriate target cells or provide help for B cells. These three components of the immune response need to be addressed separately. Examining the induction phase of the T cell response is complicated by the fact that activation of unprimed T cells depends not only on recognition of antigen plus H-2, but also on additional signals from other cells. Primed T cells are less dependent on these other signals, and most of the information on antigen recognition by T cells has come from studies with preactivated T cells, especially class 11-restricted antigen-specific T helper clones and hybridomas.
A. TRIGGERING OF ACTIVATEDT CELLSAND HYBRIDOMAS In trying to establish how T cells are triggered under physiological conditions, one has to work backward from the simplest system currently available: the capacity of T hybridomas to synthesize the lymphokine interleukin-2 (IL-2)after contact with specific antigen or other ligands (165). With this system one can address a very basic question: Is TCR-ligand interaction
T CELL SUBSETS I N T H E M O U S E
53
alone sufficient to deliver a signal to the T cell? Studies with anti-TCR antibodies-a very simple type of ligand-suggest that this is indeed the case. Thus, aggregated anti-TCR mAb, e.g., mAb attached to beads, is highly efficient at inducing IL-2 production by T cell hybridomas and clones (6,9,166,167). This finding suggests that cross-linking of TCR is all that is required to deliver at least one type of positive signal to a T cell. As discussed earlier, the physiological ligand for T cells is presumed to be an association or juxtaposition of antigen and self H-2 molecules. Until recently it has generally been assumed that T cell triggering requires that antigen plus self H-2 be displayed on the surface of viable “antigen-presenting cells” i(APC)(168). In the case of Ia-restricted T helper cells, the APCs have to express class I1 molecules, i.e., be I a + . For complex antigens, it has long been argued that native antigens have to be broken down (“processed”) by the APC into small immunogenic peptide fragments which then align themselves with surface Ia molecules (168-171). Indirect support for this notion stemmed from findings that (1)peptides cleaved from native antigens are strongly immunogenic (168,170-175) and (2) that, after feeding native antigens to APCs, effective presentation of antigen to T cells requires a lag period of about 1 hour at 37°C (176-178). The first direct evidence for antigen processing by APCs came from the experiments of Shimonkevitz et ul. (179,180) on the capacity of T hybridomas to respond to chicken ovalbumin (cOVA) presented by glutaraldehyde-fixed APCs. The key finding was that although fixed APCs were not able to present native cOVA to the T hybridomas, fixed APCs were fully capable of presenting enzymatically degraded or chemically disrupted fragments of cOVA. This observation, since confirmed by other groups (181-183), strongly suggests that antigen processing simply involves partial proteolysis or unfolding of native antigen. The issue of how the TCR recognizes peptide fragments aligned with Ia molecules was first addressed in depth by the group of Heber-Katz, Hansburg, Schwartz et uZ. (170,184-188). Using T cell lines and hybridomas specific for fragments of cytochrome c and two sets of APCs expressing slightly different Ia molecules (E);-Ef: vs E$-E$), these workers have assembled impressive evidence that T cell recognition of antigen involves the formation of a trimolecular complex between the TCR, the immunogenic peptide, arid Ia molecules on the APC. This group envisages that the immunogenic peptide has two distinct contact points, an “epitope” recognized by the TCR and an “agretope” that binds to the Ia molecule. Likewise, the polymorphic part of the Ia molecule also has two contact points, a “histotope” recognized by the TCH and a “desotope” that binds the peptide. As proposed earlier by other workers (189; see below), it is argued that association of the peptide with the Ia molecule results in “determinant selection.” In other words, the desotope of the la molecule orients the peptide in an
54
JONATHAN SPRENT AND SUSAN R . WEBB
“immunogenic” position such that the epitope can be seen by the T cell, the type of orientation being unique for each Ia allele. An accumulating body of evidence (190-192) supports this model. The trimolecular model for T cell recognition of antigen raises two questions. First, is there any physicochemical difference between the peptide epitope recognized by the TCR and the Ia-binding agretope? An interesting suggestion here is that immunogenic peptides are amphipathic and have an a-helical structure with hydrophobic and hydrophilic polarities (170,193195); the hydrophobic aspect of the peptide makes contact with the Ia molecule, whereas the hydrophilic portion is recognized by the TCR. [Note that if immunogenic peptides do indeed have an a-helical structure, one would be able to discard the popular view that, unlike B cells, T cells see linear sequences of amino acids (169); T cell recognition of peptides in the context of Ia molecules would then be closely analogous to recognition of conformational (“discontinuous”) determinants by antibody molecules (see also Ref. 196).] The second, and crucial, question is whether the immunogenic peptide and the Ia molecule do actually enter into a physical association. This question-which is obviously central to the altered-self versus dual-recognition controversy-has been a topic of speculation €or several years. Although a number of groups have reported the existence of complexes of Ia and antigen released from APCs (197-199), most investigators were not able to reproduce these findings. Recently, several groups have reinvestigated the question of antigen-Ia association by examining whether complexes occur when immunogenic peptides are incubated with purified Ia molecules in a cell-free system. Before considering the results of these experiments, the concept of Ir genes needs to be discussed (for a comprehensive review, see Ref. 170). It has long been known that many antigens, especially simple synthetic antigens and viruses, are under H-2-linked Ir gene control (85,170). Such antigens are strongly immunogenic in some strains (high responders), but not in others (low or nonresponder strains). The nature of Ir genes gradually became apparent when it was discovered that T cells from a high responder strain recognized an Ir gene-controlled antigen in the context of one self H-2 molecule, but not another. For example, T killer cells from a high responder strain, e.g., H-2k, were found to mount CTL responses to a particular virus only in the context of the Kk molecule, but not the Dk molecule (1).Likewise, T helper cells recognized a particular protein only in the context of the I-E molecule and not the I-A molecule (170,200). These and other findings led to the now well-established view that H-2 molecules and Ir gene products are one and the same, and that only certain class I or class II alleles are “permissive” for a particular antigen. Two main theories have been advanced to explain T cell unresponsiveness to antigens under Ir gene control:
T CELL SUBSETS IN THE MOUSE
55
(1)creation of “holes” in the cell repertoire, and (2) a defect at the level of antigen presentation by APC (156,169-171,188,189,201,202). The essence of this second possibility-the determinant selection theory-is that unresponsiveness reflects a failure of antigen to enter into an immunogenic association or alignment with H-2 molecules. Thus, in the words of Rosenthal et al. (189), the original proponents of determinant selection, Ir gene products “focus or orient distinct regions of the antigen for presentation to the T cell.” In assessing this theory, one is forced back to the central issue of whether antigen and Ia can form stable associations in the absence of T cells. If such associations do occur normally, a peptide under Zr gene control might be expected to show significant binding affinity for high responder Ia molecules, but not for low responder molecules. Babbitt et al. (203,206) recently addressed this issue with the aid of an equilibrium dialysis system in which 16 amino acid fragments of hen egg lysozyme, HEL (46-61), were incubated with purified I-Ak (high responder) versus I-A“ (low responder) molecules. There were two important findings. First, the HEL (46-61) peptide did form quite a strong association with high responder I-Ak molecules, with an apparent equilibrium constant of 2 x 10-6 M . Second, the HEL (46-61) peptide did not show measurable binding to low responder I-Ad molecules. These findings thus demonstrate that antigen and Ia can indeed physically associate with one another (see also Refs. 204-206). The data also suggest that a failure of antigen-Ia association accounts for (or at least correlates with) Zr gene-controlled unresponsiveness. Striking confirmation of the data of Babbitt et al. (203) has since been reported by Buus et al. (207). Using an essentially similar approach, these workers repeated the observation of Babbitt et al. that H E L (46-61) binds to I-Ak, but not to I-Ad molecules; in addition, binding failed to occur with a second low responder molecule, I-Ek. With another antigen, a peptide of ovalbumin, Buus et al. (207) found a quite different binding pattern, i.e., binding to I-A“ (high responder), but not to I-Ak or I-Ek (both low responders). Collectively, these two studies provide powerful support for the view that the immunogenicity of peptides depends upon physical association with Ia molecules. The data also strongly support the view that l r genes act at the level of antigen presentation. Although the above studies are highly convincing, Watts et al. (192) have recently reported somewhat different findings. On the basis of examining resonance-energy transfer from donor peptides to acceptor I-A molecules on a solid matrix, these workers concluded that significant peptide-Ia association only occurs when specific T cells are present. It is not clear how this finding can be reconciled with the above evidence that antigen and Ia do associate in the absence of T cells. An obvious possibility is that the system used by Watts et al. is less sensitive than equilibrium dialysis. Whatever the
56
JONATHAN SPRENT AND SUSAN R .
wEnn
explanation for this discrepancy, the important finding of Watts et al. is that the interaction between antigen and Ia molecules is stabilized by the TCR. The data thus provide further strong support for a trimolecular complex of antigen, Ia, and TCR molecules. It was mentioned earlier that cross-linking of TCRs with anti-TCR antibodies is sufficient to trigger IL-2 production by T hybridomas in the absence of APCs. Similar findings might therefore be expected for antigenspecific T cells exposed to a cross-linked association of antigen plus self Ia molecules. Watts et al. (195,208) have verified this prediction with the APCfree system described above, i.e., purified Ia molecules plus peptide antigens supported on glass coverslips. Antigen-specific hybridomas placed on the coverslips responded with vigorous production of IL-2. These data provide formal support for the view that, after processing of antigen, the role of APCs in presenting antigen plus self H-2 to activated T cells is simply to display these two ligands in cross-linked form. Until recently, the evidence that antigen processing is a prerequisite for T cell recognition of native antigens has come almost entirely from studies with Ia-restricted T cells. In the case of class I-restricted T killer cells, it has usually been assumed that CTL recognize membrane proteins, such as viral envelope glycoproteins or minor HA, which associate in some way with cellsurface class I molecules. As pointed out by Germain (209), however, the evidence that the particular antigenic epitopes recognized by CTL do exist as intact transmembrane proteins is fairly sparse. Interest in this issue has been kindled by the finding that a sizable proportion of influenza virusspecific CTL are reactive to viral nucleoprotein, i. e., to a component that does not have a recognizable leader sequence and is therefore incapable of being inserted in the cell membrane (210-213). The explanation of Townsend et al. (214) for this paradox is that nonmembrane viral proteins are processed in the cytoplasm: The proteins are degraded to small peptides which then somehow reach the surface and associate with class I molecules. In support of this idea, Townsend et al. (214) demonstrated that nucleoprotein-specific CTL were able to lyse target cells incubated with short (14 amino acid) synthetic peptides derived from the nucleoprotein sequence; lysis was antigen specific as well as H-2 (Db)specific. Since a wide variety of virus-infected cells can act as targets for CTL under physiological conditions, the authors suggest that processing of viral antigens might be a property of a multiplicity of different cell types. An interesting feature of processing of viral nucleoprotein is that treatment of virus-infected cells with the lysosomotropic agent, chloroquine, does not interfere with the display of the surface epitopes recognized by nucleoprotein-specific CTL (214). This is surprising because chloroquine is highly effective at inhibiting processing of antigens recognized by Ia-re-
T CELL SUBSETS IN T H E MOUSE
57
stricted T cells, including Ia-restricted CTL (178,215-219). Further information on this issue has come from studies by Morrison et al. (219). Using T cell clones specific for influenza hemagglutinin (HAN), these workers have assembled impressive evidence that the intracellular handling of HAN for class I- and class 11-restricted CTL is quite different. In the case of HAN detected by class I-restricted CTL, the expression of HAN was chloroquine resistant and depended on endogenous synthesis of the new HAN antigens, either by infectious virus or by transfected HAN genes; as found by Townsend et al. (214) with whole (nondegraded) nucleoprotein, exposing the target cells to exogenous HAN failed to elicit lysis. In marked contrast, expression of HAN detected by class 11-restricted CTL was chloroquine sensitive ancl depended solely on exogenous uptake of HAN. To accommodate the findings of the above two groups, Germain (209) proposes that the differential processing of viral antigens for class I- versus class 11-restricted T cells involves two quite separate mechanisms. For class I-restricted T cells, he suggests that endogenous viral proteins are degraded to simple peptides in a portion of the Golgi apparatus; these peptides then associate with class I molecules, either locally or after migration to the cell surface. For class 11-restricted CTL, the viral proteins have to make their way to enclosomes before being processed into small peptides; these peptides then associate with class I1 molecules, either after export to the cell surface, or possibly within the endosomes (see Ref. 220). A key aspect of this scheme is that, except for artificially degraded proteins or synthetic peptides, proteins introduced into cells exogenously associate only with class I1 and not class I molecules. A critical assessment of this attractive theory will obviously require a great deal of additional information on the intracellular handling of viral antigens. It will also be essential to obtain comparable information on other antigens seen by class I-restricted CTL, especially minor HA. Though excellent targets for CML (159,160), minor HA are very poorly characterized, largely because these antigens are extremely difficult to detect serologically (221). By analogy with CTL responses to viruses, one can now toy with the possibility that minor HA are not integral membrane proteins, but instead represent breakdown (processed) products of certain endogenous intracellular proteins, the display of these products on the cell surface being dependent on their capacity to associate with class I molecules. Although the results discussed above are still fragmentary, it seems quite possible that all typical protein antigens seen by H-2-restricted T cells, whether class I or class 11 restricted, are handled in much the same way: Native antigens are degraded into simple peptides, which then associate immunogeriically with permissive H-2 molecules. This pathway is unnecessary if the antigens are already in the form of “pre-processed” simple pep-
58
JONATHAN SPRENT AND SWSAN R . WEBB
tides (or if H-2 molecules are directly modified with a hapten). Of course, we are still left with the fundamental enigma of how foreign antigens make contact with H-2 molecules in the presence of a sea of selfproteins (222,223). This paradox is unlikely to be resolved until we have detailed information on the capacity of autologous proteins and peptides to associate with H-2 molecules (see Ref. 224). Despite the increasing evidence that class I and class I1 molecules have binding sites for processed antigens, the relationship of these sites (desotopes) to the histotopes recognized by the TCR is still unclear. Since polymorphism in H-2 molecules is largely limited to the membrane-distal domains, i.e., the a1 and 012 domains of class I molecules and the a1 and p l domains of class I1 molecules, one would obviously expect both recognition sites to be displayed on these domains. Studies with recombinant H-2 molecules prepared by transfection with “exon-shuffled’ genes imply that this is indeed the case (reviewed in Ref. 225). H-2 gene transfection experiments and studies with mutant H-2 molecules also suggest that the membranedistal domains interact closely with one another to form a single quaternary structure, both for class I and class I1 molecules (106,123,226-228). It seems likely, therefore, that desotopes and histotopes are both combinatorial in nature. But, at present, this seems to be the limit of our knowledge. Although the basic concept of the trimolecular complex is now gaining general acceptance, it should be borne in mind that so far there is no compelling evidence which sheds light on the altered-self versus dual-recognition controversy. Despite widespread evidence to the contrary, there are now several reports that under certain circumstances H-2-restricted T cells can bind and respond to free antigen, e.g., haptens conjugated to polymers on protein carriers (229-234) or (undenatured) proteins held in liposomes (235). Although these data are easier to interpret in terms of dual recognition than altered self, the possibility that the data reflect covert association of antigen with H-2 molecules, e.g., from the responding T cells (and/or contaminating APCs), has not been totally ruled out; alternatively, T cells might indeed be able to bind free antigen, but with much lower affinity than antigen plus H-2.2 2 Given the enormous diversity of TCRs, it is quite possible that some unprimed T cells do have the capacity to bind free antigen, e.g., native ovalbumin, with high affinity. But how could one isolate these particular T cells? As discussed later (Section III,C), it appears that inductive signals provided by APCs are essential for the activation of resting T cells. Hence, the induction of a resting T cell specific for native ovalbumin would require that this antigen be displayed on a living APC without being degraded. Such presentation would be most unlikely under physiological conditions. Under normal conditions, the antigen would be broken down and presented in association with H-2 molecules. This complex of processed antigen plus self H-2 would then be immunogenic for a different set of T cells. These cells, at best, would have only low binding
T CELL SUBSETS IN THE MOUSE
59
It would be interesting to repeat the above experiments on T cell binding of free antigen with a cell-free system using soluble TCRs prepared by exonshuffling techniques (236). Another important issue that could be addressed with a cell-free system is whether the TCR has any specific binding affinity for self H-2 molecules in the absence of antigen. This question is central to the problem of how T cells “learn” restriction to self H-2 (see Section V). Finally, cell-free systems will undoubtedly be the key to establishing the nature of the binding site(s) of the TCR and how the TCR a and p chains contribute to this site. It should be mentioned that gene transfection experiments have shown that, at least in one situation, the TCR binding site for class I-restricted CTL is assembled solely from the products of the TCR a and p genes (64). It therefore seems unlikely (though not proved) that other chains, e.g., y chains, contribute to the binding site of typical H-2-restricted T cells. It also seems clear that the pattern of TCR V a and Vp genes used by class I- and class 11-restricted T cells is very similar (237-239). What is not clear is how the various polymorphic a and p gene elements assemble a binding site that has dual specificity, i. e., specificity for antigen and self H-2. The obvious question here is whether the H-2-restricted specificity of the TCR correlates with the usage of particular gene segments (e.g., certain Ja genes), whereas specificity for antigen is controlled by other genes. The available data suggest that this is unlikely (see Ref. 239), although some groups do report provocative correlations in some instances (e.g., Ref. 240).
B. FUNCTION OF T ACCESSORYMOLECULES Although the specificity of T cells is probably controlled solely by the TCR, there is increasing evidence that interaction with antigen also involves other molecules on T cells, especially L3T4 and Lyt-2 molecules. As discussed earlier, extrathymic T cells generally express either L3T4 or Lyt-2 molecules, but not both. It was mentioned that occasional clones express both markers (128,129), but this phenotype is very rare outside the thymus. Double-positive T cells are virtually undetectable in the peripheral lym-
&nity for the native antigen. The point to emphasize is that T cells with specificity for native antigen would not be triggered under physiological circumstances and so would go undetected. In theory, T cells specific for a native antigen could be isolated by exposing rinprimed T cells to cross-linked native antigen plus a soluble source of the inductive signals provided by APCs. The stumbling blcick here is that the nature of the signals provided by APCs is still unclear (see Section III,C), so the experiment is unfeasible at present. It may be noted that the capacity of primed T cells/hybridomas to bind and respond to anti-TCR mAb could be viewed as an example of T cells binding a native antigen. It is also worth pointing out that T cells generally recognize H-2 alloantigens in native form (Section IV) and that the strong immunogenicity of these particular antigens for unprimed T cells is in part a reflection of the fact that H-2 molecules reside on the cell surface of APCs in intact (nondenatured) form.
60
JONATHAN SPRENT AND SUSAN R . WEHE
phoid organs of mice, although 3% of human peripheral blood T cells are reported to be C D 4 + , CD8+ (241). Interestingly, quite a high proportion of mitogen-activated human T cells can show transient expression of both markers (241);for obscure reasons, double-positive cells are also common in rats treated with cyclosporine (242). There are also reports of a phenotype switch for both human (243) and mouse (244) T cell subsets. The general finding, however, is that the expression of accessory molecules by mature T cells is mutually exclusive and remarkably stable (245, 246). Studies on a wide variety of T cell clones and hybridomas have shown a close correlation between (1) Lyt-2 expression and restriction by class I molecules and (2) L3T4 expression and class I1 restriction (247-254). (Exceptions to this correlation will be discussed later in Section IV.) To explain this finding, the prevailing view is that accessory molecules function by binding to mononiorphic sites on H-2 molecules and thus strengthen the avidity of binding via the TCR (247-254). A wide body of evidence supports this idea. First, antibodies to T accessory molecules are generally highly efficient at inhibiting T cell responses to antigen plus H-2, T cell clones with apparently low binding avidity being more easily inhibited than high avidity clones (247,251,255-257). Second, one group has reported that T cell triggering via anti-TCK antibodies is not inhibited by antibodies to accessory molecules (258). Third, T hybridomas which show spontaneous loss of accessory molecules in vitro often show a reduction in their capacity to respond to antigen, consistent with a lowering of binding avidity (S. Webb, unpublished data). Fourth, an unusual L3T4+ class I-restricted T hybridoma is inhibited by anti-L3T4 antibody if the APCs express both class I and class I1 molecules, but is not inhibited if the APCs are Ia- (253). Collectively, these data would seem to make a strong case that L3T4 and Lyt-2 molecules do bind selectively to H-2 determinants and that the blocking effects observed with anti-L3T4 and anti-Lyt-2 antibodies cannot be attributed simply to down-regulation of T cells. Surprisingly, however, there is accumulating evidence that, under certain conditions, antibodies to accessory molecules can down-regulate T cells. First, the capacity of L3T4+ T cells to respond to mitogens such as phytohemagglutinin (PHA) can be inhibited by anti-L3T4 antibody in the absence of Ia+ cells (259-262); this effect is most prominent if limiting doses of mitogen are used. Second, it has recently been reported that the proliferative response of resting T cells to anti-TCR antibodies can be inhibited by anti-L3T4 antibodies in the absence of Ia+ APC (263); likewise, stimulation of human T cells with anti-T3 antibody can be inhibited with anti-CD4 antibody (264). Third, the capacity of Lyt-2+ CTL to induce lectin-mediated lysis of class I-negative target cells (the lectin “glues” the CTL and target cells together) can be inhibited by anti-Lyt-2 antibody (265). To account for these discrepant findings, Tite et
T CELL S UBS E T S I N THE M O U S E
61
al. (262) have suggested that interaction between T cells and APCs involves two phase:;: (1)an antigen-independent phase of binding between T accessory m o l e d e s and monomorphic H-2 determinants on APCs, followed by (2) formation of a trimolecular complex between the TCR and Ia plus antigen on APCs. Tite et al. argue that if the phase of binding between T accessory molecules and H-2 is not followed by TCR interaction with antigen, e.g., if the APC 1.acks antigen, the T cell receives a down-regulation signal, which enables th'e T cell to break free from the APC and wander away. TCR contact with antigen counteracts this negative signal and the T cell is triggered. Though fitting most of the available facts, this theory raises certain questions. In particular, one has to explain how interaction of T accessory molecules with H-2 determinants (or with antibodies to T accessory molecules) transmits a negative signal to the T cell whereas TCR contact with antigen plus H-2 molecules results in a positive signal. Fazekas de St. Groth et al. (129)offer a somewhat different explanation for the blocking effects of antibodies to accessory molecules. These workers isolated two rare class 11-restricted T cell clones that express both L3T4 and Lyt-2 molecules. The interesting finding was that only anti-L3T4 and not anti-Lyt-2 antibodies blocked the response of the clone to antigen (hapten) coupled to Ia+ APCs. Since the APCs expressed both class I and class I1 determinants, the selective inhibition observed with anti-L3T4 antibody is clearly difficult to explain in terms of simple competitive inhibition. The provocative suggestion of the authors is that the L3T4 molecules on the clones form some type of inultiinolecular complex with TCRs during interaction with antigen, Lyt-2 molecules being excluded from this complex; antiL3T4 antibodies inhibit the formation of this complex and thereby impair the capacity of the TCR to recognize (or respond to) antigen." None of the above evidence is inconsistent with the notion that T accessory molecules do bind to H-2 molecules, but it should be emphasized that this hypothesis is based almost entirely on circumstantial evidence. There are two main problems with this theory. First, one has to argue that all class I and class 11 molecules express class-unique monomorphic regions accessible to one set of accessory molecules, but not the other. This has yet to be proved. Second, an ingenious experiment by Golding et al. (266) suggests that if H-2 molecules do have unique binding sites for accessory molecules, In the case of Lyt-2+ cells, recent studies of N . Crispe and M . Bevan (personal coinmunication) have shown that attachment of anti-Lyt-2 inAb to a solid matrix next to an artificial antigen (anti-TCR rn.4b) can lead to enhanced responses of Lyt-2 + cells to the antigen; conipara1)le findings have been observed for L3T4 cells with iniiinobilized anti-LX1'4 mAl) (K. Eichlnann, unpublished data). Although these findings are open to various interpretations, the data are in line with the suggestion of Fazekas de St. Grot11 rt ul. (129) that T cell triggering involves the formation of a complex between the TCR and appropriate T accessory molecules. +
62
JONATHAN S PR E NT A N D SUSAN R . WEBB
these H-2 sites must be represented on the polymorphic domains. By exon shuffling and gene transfection techniques, these workers fashioned a recombinant H-2 molecule consisting of the NH,-terminal (polymorphic) domain of a class I1 p chain (A:,) covalently attached to the a3, transmembrane, and intracytoplasmic portions of a class I molecule. An L3T4+ T cell line raised against allogeneic I-Ak (A~,,,--A~~,,) determinants proved capable of lysing target cells expressing the hybrid molecule, implying that at least some of the T cells were specific for Abl epitopes. The key finding was that lysis directed to these epitopes was totally inhibited by anti-L3T4 antibody. Although the possibility that T accessory molecules bind to polymorphic epitopes on H-2 molecules still remains open, the results of Golding et al. are clearly much easier to explain in terms of the model of Fazekas de St. Groth et al. (see above). What is clearly needed is direct evidence on whether T accessory molecules bind to H-2 molecules. Before leaving this issue, one should mention that other accessory molecules on mouse T cells, such as LFA-1, also might help to stabilize T-APC interactions (259,267270) (see also Section IV,C). As for L3T4 and Lyt-2 molecules, the function of these other accessory molecules is still poorly understood. C. TRIGGERING OF UNPRIMEDA N D RESTINGT CELLS As discussed earlier, simple cross-linking of TCRs through contact with anti-TCR antibodies or antigen plus H-2 is sufficient to cause T hybridoma cells to synthesize IL-2. Triggering of normal T cells, especially resting T cells, is much more complex. Addressing this issue necessitates defining what stimuli are required to induce a small resting T cell in Go to enter cell cycle and initiate synthesis of growth-promoting lymphokines such as IL-2. The ideal system for approaching this question would be to prepare purified populations of unprimed antigen-specific T cells and then define what particular signals are required to induce these cells to respond to purified antigen plus H-2 on a solid matrix. There are two insuperable problems with this type of approach. First, there are no known techniques for isolating antigen-specific unprimed T cells from other T cells. Second, the precursor frequency of unprimed T cells for antigen plus self H-2 is extremely low. Indeed, with one exception (201), no one to our knowledge has been able to reproducibly demonstrate proliferative responses to protein antigens in vitro with unprimed T cells as responders. In view of these problems, investigators have had to resort to artificial systems for studying the induction of unprimed T cells. Most of the information on this issue has come from studying the response of normal T cells to three sets of stimuli which trigger a high proportion of unprimed cells:
T CELL SUBSETS IN THE MOUSE
63
(1) H-2 alloantigens (dealt with in Section IV), (2) T cell mitogens such as concanavalin A (Con A) or PHA, and (3)antibodies directed to the TCR/T3 complex of molecules. Recent information gained from the latter two systems can be summarized as follows. Purified small resting T cells can proliferate and synthesize IL-2 when cultured with anti-T3 antibodies (tested only for human T cells) or mitogens in the presence of accessory cells (macrophage/monocytes or dendritic cells) (271-275). 'There is general agreement that, in the abs'ence of added factors, T cell stimiilation by these ligands fails to occur if the T cells are rigorously depleted of accessory cells (AC). Precisely how AC control T cell responses to mitogens, however, is far from clear. One possibility is that AC merely provide a source of H-2 molecules, stimulation of T cells by mitogens being dependent on corecognition of H-2 molecules, especially Ia molecules. This idea seems unlikely in view of reports that Ia- cells can act as AC, even for L3T4+ cells (261,274,275). A more plausible possibility is that AC function simply by cross-linking the ligand. Although AC undoubtedly do play an important role in cross-linking ligand, there is increasing evidence that a predominant function of AC is to display or release certain activation signals required by resting T cells. The nature of these signals is a source of continuing controversy. For both mouse and human cells, it is well accepted that activation of resting T cells leads initially to synthesis and surface expression of receptors for IL-2 (IL-2R) (276-278). In order for the T cells to proceed to the stage of cell division and proliferation, the cells have to make contact with IL-2, either exogenous IL-2 or IL-2 made by the cells themselves. Contact with IL-2 through IL-2R then activates the cells to enter cell cycle. With regard to the initial events in T cell induction, the bulk of the evidence from studies with mitogens and anti-TCR/T3 antibodies suggests that cross-linking of the TCR/T3 complex, e.g., by using anti-T3 antibodies on Sepharose beads or mitogens, is sufficient to trigger at least a proportion of T cells to produce IL-2R (271.274). The subsequent signals required to induce the cell to secrete IL-2 (and thus enter cell cycle), however, are still poorly understood. For human T cells, a convincing case has been made that IL-2 synthesis is under the control of IL-1, a lymphokine synthesized by typical AC, but not T cells (279). The evidence implicating IL-1 in the activation of unprimed T cells is as follows (271,280,281): When AC-depleted resting T cells are exposed to anti-T3 antibodies cross-linked on Sepharose beads or attached to glass dishes, the cells express IL-2R, but do not proliferate. If the cells are supplemented with IL-1 (or IL-2), however, vigorous proliferation occurs; addition of IL-1 to normal (unsensitized) T cells causes no proliferation, resting T cells being insensitive to both IL-1 and IL-2 (and a mixture of IL-1
64
JONATHAN SPHENT AND SUSAN R . WEBB
and IL-2). Although the precise mechanism of action of IL-1 is not known (271,280-283), the prevailing view is that IL-1 somehow induces the cells to begin endogenous synthesis of IL-2. The main problem with this scheme is that although IL-1 is highly effective in the hands of some groups, other workers, especially those working with murine T cells, have had considerable difficulty in finding a convincing role for IL-1 in T cell induction (263,274,275,284,285) [although there is at least one murine T cell clone that responds dramatically to anti-TCR antibodies supplemented with IL-1 (283)l. Most of the skepticism concerning IL-1 has come from studies on the differential triggering requirements of purified populations of unprimed Lyt-2 and L3T4 cells responding to mitogens or anti-TCR antibodies. In the case of mitogens, most groups find that purified resting Lyt-2+ cells respond well to mitogens such as Con A providing AC are present (274,275,284,285-287). If AC are removed, responses to mitogens are abolished or considerably reduced, but can be restored to high levels by addition of crude supernatants of Con A-activated spleen cells (CAS). Since CAS contains a wide variety of different lymphokines, including IL-2, much effort has been devoted to defining which particular lymphokines can overcome the unresponsiveness of AC-depleted Lyt-2+ cells to Con A. Several groups find that IL-2 alone is sufficient to restore the response provided that high doses of IL-2 are used; one group (275) has the extra proviso that responsiveness of Lyt-2+ cells to Con A plus IL-2 in the absence of AC depends on reducing the net negative charge on the T cells, e.g., by pretreating the cells with neuraminidase. If Lyt-2+ cells are cultured with Con A plus low doses of IL-2, some groups argue that additional factors, e.g., certain factors present in CAS, are required to cause optimal proliferation and differentiation into CTL (285,287); these factors, some of which are still not well characterized, include IL-2R-inducing factor (RIF)4(285), CTL differentiation factor (CTDF) (285,288), and interferon-y (287). The important point to emphasize is that IL-1 has virtually no effect in this system. Thus, even in high doses, IL-1 added to AC-depleted Lyt-2+ cells fails to induce these cells to proliferate in response to Con A (274,284,285). In marked contrast to Lyt-2+ cells, the response of L3T4+ cells to mitogens depends heavily on the presence of AC (274,284,285). Thus, addition of even very high doses of IL-2 fails to allow purifed AC-depleted +
+
Recent studies of Wagner’s group (H. Wagner, personal communication) suggest that RIF is synthesized by dendritic cells and that AC-depleted, small, resting, high-density Lyt-2+ cells cannot respond to Con A plus high concentrations of IL-2 unless the cells (used in small numbers) are supplemented with RIF. This finding suggests that at least two signals, mitogen and RIF, are required for IL-2R expression by Lyt-2+ cells.
65
T CELL SUBSETS IN THE MOUSE
L3T4 cells to respond to mitogens such as Con A; adding a mixture of IL-2 and IL-1 is similarly ineffective. Interestingly, however, AC-depleted L3T4+ cells can be induced to respond to mitogens by supplementing the cells with a mixture of IL-2 and the synthetic compound, phorbol myristic acetate (PMA), an activator of protein kinase C (274). Even better stimulation occurs with a mixture of IL-2, PMA, and the calcium ionophore, ionomycin (284). The above data suggest that in response to mitogens such as Con A the requirement for AC is weak for Lyt-2+ cells, but strong for L3T4+ cells. Very similar findings have been reported for activation of resting T cells by an anti-TCR antibody, F23.1; this antibody reacts with about 25% of Lyt-2+ and L3T4+ T cells from normal mice and is specific for TCR P-chain determinants (VP8) (289). When the F23.1 antibody is coupled to Sepharose beads, bulk cultures of AC-depleted purified T cells (a mixture of Lyt-2+ and L3T4+ cells) give high proliferative and CTL responses provided that the cells are supplemented with CAS (290); IL-2 (used only at low doses) is less effective than CAS, and IL-1 is totally ineffective. In the presence of CAS, the vast majority of the T cells stimulated by F23. l-beads are Lyt-2+ cells, ey, P. J., Sharrow, S. O., Kohno, Y . , Berzofsky, J. A , , and Singer, A. (1985). Transplantation 40, 68. Kimoto, M., and Kishimoto, S. (1986). Eur. J. Zmmunol. 16, 835. Villartay, J.-P., Griscelli, C., and Fischer, A. (1986). Eur. J. Zmmunol. 16, 117. Glazier, A . , Tutschka, P. J.. Farmer, E. R., and Santos, G. W. (1983). J. E x p . Med. 158, 1. Cheney, R . T., and Sprent, J. (1985). Transplant. Proc. 17, 528. Nossal, 6. J, V. (1983). Annu. Rec. Zinmuno/. 1, 33. Miller, R. 6. (1980). Nature (London) 287, 544. Crispe, I. N., and Owens, T. (1985). Zmmunol. Today 6, 40. Rammensee, H.-G., Bevan, M. J., and Fink, P. J. (1985). Zmmunol. Today 6, 41.
This Page Intentionally Left Blank
ADVANCES IN IMMUNOLOGY, VOL 41
Determinants on Major Histocompatibility Complex Class I Molecules Recognized by Cytotoxic T Lymphocytes JAMES FORMAN Department of Microbiology, University of Texas Health Science Center at Dallas, Dallas, Texas 75235
1. Introduction
Cytotoxic T lymphocytes (CTL) are effector cells that play an active role in the immune response against virus infection, tumors, and allografts (see reviews by Zinkernagel and Doherty, 1979; Doherty et al., 1984). The antigen specificity of these cells is a result of the recognition of class I molecules on target cells that are encoded by genes in the major histocompatibility complex (MHC). These molecules consist of a 40,000-45,000 Da heavy chain noncovalently bound to a 12,000-Da light chain, P,-microgIobulin (Pz-M) (Kimball and Coligan, 1983) (Fig. 1). The heavy chain is an integral membrane protein composed of three extracellular domains (al-a3) of -90 amino acids each, a transmembrane (TM) domain, and a cytoplasmic region that varies in length between 30 and 39 amino acids (Steinmetz et al., 198la) (Table I). These molecules are glycosylated through N-linked oligosaccharides (Nathenson and Cullen, 1974) and have serine, tyrosine, and possibly threonine residues in their cytoplasmic region that can be phosphorylated (Pober et td., 1978; Lalanne et al., 1982; Guild et al., 1983; Guild and Strominger, 1984a,b). The molecule can be fatty acylated to palmitic acid via a cysteine in the TM region (Kaufman et al., 1984). Although class I molecules expressed by an individual are invariant, the alleles of each gene which encode class I molecules are highly polymorphic such that most species possess a large number of antigenically diverse class I alloantigens (Duncan et al., 1979; Klein and Figueroa, 1981). The polymorphism found in class I molecules is not randomly distributed, but rather localized to the a1 and a 2 domains (K.imbal1 and Coligan, 1983). In contrast to the class I heavy chain, &-M is invariant within a species, although in the mouse limited polymorphism has been detected (see Section V). In the mouse, class I genes number between 23 and 37 (Steinmetz et al., 1982; Weis,s et al., 1984; Fisher et al., 1985). They can be divided into two groups: (1)H-2K, D, and L, and (2) genes in the QalTla region including Iimt (Fischer-Lindahl et al., 1983). H-2K, -D, and -L are target molecules for both alloreactive and antigen-specific H-%restricted CTL. These mole135
Copyright 6 I987 by Academic Press, Inc All nghts of reproduction In any form reserved
136
JAMES FORMAN
cules are expressed on most cells in varying amounts, the highest density usually found on cells of lymphoid origin where as many as 5 X 105 molecules/cell of each antigen can be detected (Liberti et al., 1979). Qa-1, Qa-2, and Mta antigens serve as target structures for alloreactive CTL, but have not been demonstrated to restrict antigen-specific CTL (Kastner et al., 1979; Fischer-Lindahl et al., 1980; Forman et al., 1982). These latter molecules have a lowered level of cell membrane expression and a more limited tissue distribution than H-2K, D, or L (Flaherty, 1981), which may explain this discrepancy. Although class I molecules within a species are -80-90% homologous, the fact that T cells readily distinguish between class I alloantigens indicates that the immune system focuses on the polymorphic features of these molecules. In order to accomplish this, CTL utilize a and p chains of their T cell receptor to recognize antigen in the context of class I molecules (see reviews by Reinherz et al., 1984; Marrack and Kappler, 1986). Other molecules also play a role in CTL interactions with target cells including Lyt-2/CD8 (MacDonald et al., 1981; Swain, 1981; Meuer et al., 1982) and LFA molecules (Krensky et al., 1983, 1984). Helper T cells have been demonstrated to specifically recognize antigenic peptides in the context of class I1 molecules (Shimonkevitz et al., 1983). Townsend et al. (1986a) has shown that peptides of influenze virus nucleoprotein interact with the membrane of target cells to create determinants recognized by MHC-restricted anti-influenza CTL. Recently, it has TABLE I DOMAIN CIIAHACTEHISTICS OF CLASSI MOLECULES= Domain Characteristic Approximate size Homology with P2-M Homology with immunoglobulin Association with P2-M Phosphorylation site N-linked oligosaccharide H-2Kb mutation sites HLA CTL subtype variant sites CTL epitopes Serological epitopes
a1
a2
a3
TM
CYTO
90
90
90
24
30-39
+?
+?
+?
+b
+b
+ + + + +
+ + + +
+ +
+
-?
+
~~
See text for description and references. In man, only the a1 domain is glycosylated; in mouse, the a2 and in some molecules the a3 domains are also glycosylated. 0
MHC CLASS I MOLECULES RECOGNIZED BY CTL
137
3 are FIG.1. Schematic diagram of a class I MHC molecule, including P2-M. a l - 1 ~domains denoted by different patterns. Potential attachment sites for N-linked oligosaccharides are indicated. POi indicates region for phosphorylation, P indicates site for attachment to palmitic acid. Figure is drawn to indicate that most epitopes in the u l and a2 domains are controlled by the interaction of these domains with each other.
been reported that there is a weak interaction between specific peptides and class I1 molecules (Babbitt et al., 1985). Presumably, this peptide-class I1 molecular complex is recognized by the T cell receptor on helper T cells. Since CTL utilize the same receptors as helper T cells (Rupp et al.,1985; see review, Marrack and Kappler, 1986), it is likely that polymorphic regions of class I molecules not only intereact with CTL receptors, but that these molecules also interact with specific viral peptides. This review will examine the various approaches used to characterize determinants on class I molecules that are recognized by both alloreactive and antigen-specific class I-restricted CTL. Studies attempting to examine the nature of the determinants recognized by CTL have taken advantage of
138
J A M E S FORMAN
variants of normal class I molecules. These variants have arisen by (1)germ line mutation, (2) somatic cell mutation usually induced or detected in uitro, (3) site-specific mutagenesis of class I genes, and (4) exon shuffling of class I genes which results in the expression of domain-shuffled molecules in transfected cells. These topics as well as the role of carbohydrates and P2-M in controlling epitopes recognized by CTL will be discussed. II. Exon Shuffling to Produce Domain-Shuffled Molecules
A. TRANSFECTION OF CLASSI GENES cDNA clones have been isolated that encode class I MHC genes (Sood et al., 1981; Ploegh et al., 1981; Kvist et al., 1981; Steinmetz et al., 1981b). These cDNA clones were used to identify genomic clones, including H-2Kb (Mellor et al., 1982), H-2Ld (Evans et al., 1982a; Goodenow et al., 1982a), S L A d (Singer et al., 1982), and HLA-A2 and -B7 (Barbosa et al., 1983). To test the functional properties of these genes, genomic clones were introduced into murine L cells. The products of these genes were detectable on the L cell plasma membrane (Mellor et al., 1982; Barbosa et al., 1983; Evans et al., 1982b; Singer et al., 1982; Goodenow et al., 1982a). One assumption in using transfectants for functional assays is that the cell membrane product of the transfected gene is identical to the membrane product encoded by the endogenous gene. In general, most evidence supports this, especially when considering the transfection of murine class I genes into murine L cells. Thus, the size and charge heterogeneity of H-2Ld on L cells appear identical to H-2Ld on spleen cells (Goodenow et al., 1982a), the size of H-2Kb is similar on L cells as on EL-4 tumor cells (Mellor et al., 1982), and H-2Kd from spleen cells is similar to H-2Kd expressed on L cell transformants (Goodenow et al., 1982b). The results of serological studies are also consistent with the notion that class I-transfected molecules expressed as a result of transfection are similar to the same molecule expressed endogenously. Thus, transfected class I molecules only react with expected serological reagents (Evans et al., 1982b; Ozato et al., 1983a,b; Allen et al., 1984; Arnold et al., 1985). One exception to these similarities is the difference in glycosylation of the class I QlO molecules secreted by L cells transfected with the 010 gene compared to QlO found in serum (Devlin et al., 1985). However, it cannot be ascertained whether L cells glycosylate this molecule in a unique fashion as opposed to the possibility that serum QlO is altered while in the circulation. Transfection of human class I genes into murine cells could yield an abnormal molecule, since the P2-M is of murine origin. In some studies, human P2-M and human leukocyte antigen (HLA) have been introduced into mouse
MHC CLASS I MOLECULES RECOGNIZED BY CTL
139
L cells (van de Rijn et al., 1984; Bernabeu et al., 1984a; Barbosa et al., 1984). It should also be noted that there is a rapid exchange between the &-M in the medium and endogenously synthesized P2-M (see Section V). Nevertheless, the size and charge heterogeneity of HLA heavy chain in transfected L cells or human cells is similar to the endogenous products (van de Rijn et al., 1984; Bernabeu et al., 1983). While the antigen-specific T cell receptor recognizes products of class I genes, other molecules also play a role in T cell recognition, including Lyt-2/CD8 (MacDonald et al., 1981; Swain, 1981; Meuer et al., 1982) and LFA molecules (Krensky et al., 1983, 1984). It is likely that murine L cells, which are frequently used for transfection studies, do not express all the appropriate ligands for accessory molecules found on commonly used targets in CTL assays (tumor cells or lymphoblasts) (Golde et al., 1985; Naquet et al., 1985), nor is the efficiency of accessory molecule interactions between cells from different species understood (Spits et al., 1986). Therefore, caution must be taken in the interpretation of unexpected data (see Section 111). Orn et al. (1982), Mellor et al. (1982), Forman et al. (1983), Levy et al. (1983), Herman et al. (1983), Margulies et al. (1983), and Ozato et al. (1983a) demonstrated that L cells transfected with exogenous class I genes and expressing the corresponding gene product could be recognized by alloreactive and H-%restricted antigen-specific CTL. Epitopes recognized on the transfected class I molecules by CTL appeared to be similar to those recognized on normal cells, since the transfected antigen could competitively inhibit 1ySi:jdirected against the endogenously expressed antigen (Forman et al., 1983). The ability to exon shuffle genes allowed for the construction of hybrid molecules containing domains from different class I genes. Several approaches have been used, including (1)shuffling the 5’ end of one gene by splicing within the large intervening sequence between exons 3 and 4 with the remaining 3’ end of a second gene using conserved XbaI sites (Evans et al., 198213; Allen et al., 1984), (2) utilization of the conserved BamHI sites in the large intervening sequence and the 3’ end of the gene (Stroynowski et al., 1984), and (3)generation of partial deletions of class I genes followed by ligation of subclones of these truncated genes in the same vector (Arnold et al., 1984). A schematic diagram indicating some of the exon-shuffled class I genes that have been constructed is given in Fig. 2.
B.
EXON-SHUFFLED
GENESBETWEEN H-2Ld A N D H-2Dd 1 . Serology
Evans et al. (1982b) and Ozato et al. (1983a) exon shuffled H-2Ld and H-2Dd between the a 2 and a3 encoding exons so that the domain-shuffled
140
JAMES FORMAN
A. - '5
Es-
1
2
3
4
5 6 7 0 H
KHH,3'+HzNi
HHHJ-U-3'+HzN
a1 a2 a3TMCy I w n w COOH
r
m
FIG. 2. Exon shuffling of class I genes to produce domain-shuffled molecules. The left portion of the figure schematically indicates the exon-intron structure of a class 1 gene while the right portion depicts the protein structure of a class I molecule. Exon 1 encodes the leader; exons 2-4 encode the al-a3 domains, respectively; exon 5 encodes the transmembrane (TM) segment; and exons 6-8 encode the cytoplasmic (Cy)portion of the molecule. A and B represent two class I genes that have been spliced to produce constructs C-E (see text). The resultant molecules have homologous a1 and a2 domains and heterologous a3 domains (C and D) or heterologous a1 and a2 domains and homologous a2 and a3 domains (E and F). See Section I1 and Tables II-IV for explanation of results.
molecules expressed on transfected L cells consisted of Ld/Ld/Dd or Dd/Dd/Ld(referring to the origin of the a l / a 2 / a 3 domains). All monoclonal antibodies (mAbs) specific for H-2Ld or H-2Dd reacted with the domainshuffled antigens. Most mAbs appear to react with epitopes controlled by the a l / a 2 domains of these molecules, while only two mAbs, 28-14-8 and 34-2-12, reacted with the a3 domains of H-2Ld and H-2Dd, respectively. The fact that all mAbs tested retained activity against these domain-shuffled molecules suggests that most serological determinants are not controlled by the interaction of a3 with the a V a 2 domains. Further studies have involved exon shuffling of H-2Ld, H-2Dd, and Q7d (the 27.1 gene described by Steinmetz et al., 1981a) to produce nonhomologous a l - a 2 domain-shuffled molecules. Murre et al. (1984a) constructed a Dd/Ld/Ld molecule. Although the reciprocal Ld/Dd/Dd molecule was not available, they provisionally mapped mAb sites to the a1 and a 2 domains. For example, mAb 34-5-8 reacted with Dd/Dd/Ld, but poorly reacted with Dd/Ld/Ld, suggesting that this defines an epitope in either the a 2 domain of H-2Dd or a conformational determinant controlled by a l / a 2 . Other mAbs were provisionally mapped to the a1 or a 2 domains of H-2Ld or H-2Dd. McCluskey et al. (1986a)constructed the reciprocal domain-shuffled molecule (Ld/Dd/Dd)and was able to further localize monoclonal binding epitopes. Six monoclonal antibodies bound to a 2 of H-2Dd, 5 to a1 of -Dd,
MHC CLASS 1 MOLECULES RECOGNIZED BY CTL
141
and 4 to a1 of -Ld. The binding of 34-5-8 to these domain-shuffled molecules further illustrates the problem in definitively assigning epitopes to particular domains. 34-5-8 binds to Dd/Dd/Dd and Dd/Dd/Ld, but not Ld/Ld/Dd, mapping its reactivity to a determinant in the d / a 2 and not a3 domain of H-2Dd (Evans et al., 1982b). The antibody binds about half as well to Dd/Ld/Ld and Ld/Dd/Dd, suggesting that it interacts with a site controlled by both the a l / a 2 domains or similar sites in a1 and a 2 of H-2Dd. However, the Dd/Q7d/Dd molecule does not bind this antibody (Stroynowski et al., 1985a). Thus, the a 2 domain of H-2Ld, but not Q7d, permits the expression of the 34-5-8 epitope. A similar observation was made by Darsley et al. (1987), who noted that several mAb which were assigned a1 domain reactivity with H-2Dd based on H-2Dd-H-2Ld domain shuffled molecules (Ozato et al., 1985) did not react with the a1 domain of -Dd when this molecule was shuffled with H-2Dp. Therefore, many mAb-defined epitopes currently mapped to a particular domain may later be found to be indirectly controlled by another domain as further domain-shuffled molecules are tested. McCluskey et al. (1986a) noted that molecules shuffled between the a 2 and a3 domains bound mAb at roughly equivalent amounts to the homologous molecules. This was not the case when the a l / a 2 domains were exchanged. Thus, Dd/Ld/Ld molecules bound mAb somewhat less than Dd/Dd/Dd or Ld/Ld/Ld and Ld/Dd/Dd molecules bound mAbs much less than Dd/Dd/Dd or Ld/Ld/Ld. Therefore, constructing molecules from nonhomologous a l / a 2 domains may grossly alter the epitope structure of these antigens.
2 . Cytolytic T Lymphocyte Recognition Structural data indicate that most of the polymorphism found in class I molecules is in the a1 and a 2 domains. Further, studies with H-2 mutants and HLA subtypes (see Section VI1,A) have revealed that point mutation or minor variation in either the a1 or a 2 domain leads to strong alloreactivity. Therefore, it might be expected that the a1 and a2, but not a3 domains control determinants recognized by CTL. Ozato et al. (1983a) used alloreactive anti-H-2Ld CTL to demonstrate that both Ld/Lcl/Ldand Ld/Ld/Dd molecules were recognized effectively. Similar results were seen with anti-H-2Dd CTL and Dd/Dd/Dd and Dd/Dd/Ld molecules. Further evidence indicating that the a3 domain does not control polymorphic determinants recognized by alloreactive CTL was shown by the finding that domain-shuffled molecules expressed on unlabeled cells could block the 1ysis directed against labeled targets expressing the homologous molecules (Ozato et al., 1983a; Stroynowski et al., 1984). McCluskey et al.
142
J A M E S FORMAN
(198613) also noted that cells expressing the a3 domain only of H-2Ld or H-2Dd were not recognized by alloreactive CTL (see Table 11). Antigen-specific CTL restricted by H-2Dd or -Ld were able to interact with domain-shuffled molecules. Thus, antihapten (Ozato et al., 198310) and antiviral CTL (Reiss et al., 1983; Stroynowski et al., 1984) restricted to either H-2Dd or -Ld interacted with Dd/Dd/Ld or Ld/Ld/Dd, respectively, but not the reverse. Unlabeled cells expressing domain-shuffled molecules could block the lysis by antivirus-specific CTL of infected target cells expressing the homologous molecule, demonstrating that few, if any, polymorphic restricting determinants are controlled by the a3 domain (Stroynowski et al., 1984). Levy et al. (1985) demonstrated that both alloreactive and haptenrestricted CTL clones recognize epitopes controlled by the a l / a 2 domains irrespective of whether the a3 domain was from H-2Ld or -Dd. Engelhard et al. (1985) exon shuffled HLA-B7 with H-2Ld and -Dd and showed that the presence of a murine-derived a3 domain had no effect on the recognition of polymorphic HLA determinants in the domains. However, Maziarz et al. (1985) observed that both human ani-HLA-A2 and murine anti-H-2Kb CTL were less efficient at lysing human/mouse and mouse/human hybrid molecules, respectively, than the appropriate homologous molecule. Since the a3 domain of human and mouse class I molecules is -70% homologous in amino acid sequence, it is possible that this domain can affect the conformation of d / a 2 epitopes, or alternatively, the a3 domain could serve as a ligand for an accessory molecule that functions inefficiently between species. Since certain point mutations or minor variations in class I molecules induce strong alloreactive responses (Pimsler and Forman, 1980), it might be expected that class I molecules shuffled between the a1 and a 2 domains would affect most of the CTL epitopes. However, alloreactive bulk CTL and CTL clones specific for H-2Dd and H-2Ld recognize a Dd/Ld/Ld molecule (Murre et al., 1984a; Reiss et al., 1986). H-2Dd- and H-2Ld-restricted antiinfluenza virus-specific CTL also showed lytic activity against this target molecule, but anti-vesicular stomatitis virus (VSV) CTL, which are H-2Ld restricted, did not interact with Dd/Ld/Ld VSV-infected targets (Murre et al., 1984a). In contrast to alloreactive and H-2-restricted recognition of Dd/Ld/Ld, McCluskey et al. (1986a) tested the reciprocal exon-shuffled gene (Ld/Dd/Dd)and found that both alloreactive and antigen-specific anti-VSV CTL failed to react with this molecule. While some anti-H-2Ld and -Dd CTL clones reacted with Dd/Ld/Ld, none reacted with Ld/Dd/Dd. Mouse strain B10. D2-H-2dm1 is a meiotic H-2Dd/Ld mutant resulting from the fusion of the H-2Dd and -Ld genes (Sun et al., 1985). Thus, this fusion molecule represents a “natural” domain-shuffled molecule. The fusion occurred between residues 122 and 155, resulting in the a1 and most of the a 2 domain derived from H-2Dd, while the C-terminal amino acids of the
TABLE I1 EFFECTOF SHUFFLING THE al-a3 DOMAINS OF H-2Dd Derivation of domains
AND
H-2Ld
Alloreactive CTL
Constructh
a1
a2
a3
Anti-H-2Dd
Anti-H-2L"
2C/D 2C/D 2E/F 2EIF
Dd Ld Dd Ld Dd Dd
Dd Ld L'I Dd Q7d Dd/Ld
Ld Dd Ld D" Dd LdY
+'
-
Fig. Fig. Fig. Fig.
-
+
+ +
ON
CTL RECOGNITION^
H-%restricted CTL H-2Dd
H-ZLd
Referencef
+ +
-
1 2 3 4 5 6
+ +
-
-
-
-d
-
+
+
ND
-
-
ND
-
a CTL were generated against H-2Dd, H-2Ld, or antigen (hapten or virus) restricted by H-2Dd or H-2Ld and tested against the domain-shuffled molecules expressed on L cells. See schematic diagram in Fig. 2 for construction of exon-shuffled genes. c indicates cross-reaction of CTL on domain-shuffled molecule, - indicates lack of cross-reaction. CTL were generated against Q7 determinants rather than H-2Ld. p This construct is the H-2dm' molecule (see text for explanation). f References: (l),Reiss et al. (1983), Ozato et al. (1983a,b), Levy et al. (1985), (2), Ozato et al. (1983a,b), Levy e t a / . (1985), Stroynowski et al. (1984), (3), McCluskey et al. (1986a), Reiss et a / . (1986), (4), McCluskey et al. (1986a), (5), Stroynowski et a / . (1985), (6), Burnside et al. (1984), Forman and Klein (1975a), Ciavarra and Forman (1982), Forman (unpublished data).
+
144
J A M E S F ORMAN
a 2 domain are derived from -Ld. Previous studies (Forman and Klein, 1975a; Burnside et al., 1984; Forman, unpublished data) have noted that anti-H-2Dd and -Ld CTL cross-react on H-2dm1target cells and therefore are consistent with the studies of Murre et al. (1984a) showing that H-2Dd/Ld molecules retain alloreactive determinants characteristic of both parents. HLA-Aw69 is another natural domain-shuffled molecule consisting of HLA-Aw68 sequences in the a1 domain and HLA-A2 sequences in the a 2 and a3 domains (Holmes and Parham, 1985). Most CTL clones generated against HLA-AS, -Aw68, and -Aw69 fail to recognize determinants localized to either the a1 or a 2 domain only, but rather recognize epitopes requiring the presence of both a1 and a 2 (Clayberger et al., 1985). However, Wallace et al. (1986) found that anti-Epstein-Barr virus (EBV) -specific HLA-Aw69restricted CTL cross-react on HLA-AS. This finding is similar to that of the Dd/Ld/Ld construct which retained H-2Dd-restricting epitopes for anti-influenza CTL. Stroynowski et al. (1985a) produced a domain-shuffled molecule that contained H-2Dd in the a1 domain, Q7d in the a 2 domain, and -Dd in the a3 domain (Dd/Q7d/Dd).This molecule is not recognized by either anti-H-2Dd, anti-Q7d, or antigen-specific H-2Dd-restricted CTL. Thus, in contrast to the study of Murre et al. (1984a), changing the a 2 domain of the H-2Dd molecule to Q7d, as opposed to -Ld, completely alters the alloreactive and restricting specificity of this molecule. C. EXON-SHUFFLED GENESBETWEEN H-2Kb AND H - 2 D b Allen et al. (1984) constructed hybrid class I genes between H-2Kb and H - 2 D b . These hybrids consisted of Db/Kb/Kb, Db/Db/Kb, Kb/Db/Db, and Kb/Kb/Db. Of 17 antibodies tested, only one (28-14-8) reacted with the a3 domain of H-2Db. This result is not surprising, since 28-14-8 reacts with the a3 domain of H-2Ld, a molecule that has an identical sequence to H-2Db in the a3 domain (Kimball and Coligan, 1983). This antibody bound to H-2Db expressed on transfected L cells to the same extent as the Kb/Db/Db construct, indicating that molecules with nonhomologous a l / a 2 domains (between H-2Kb and -Db) do not necessarily have a reduced cell membrane expression. Although most antibodies were shown to interact with determinants which map to the a1 and a2 domains, their binding was greatly reduced compared to the binding to homologous molecules. Five antibodies lost reactivity with the a U a 2 domain-shuffled molecules. Similar findings were reported by Bluestone et al. (1985). Alloreactive anti-H-2Kb CTL recognized Kb/Kb/Kb and Kb/Kb/Db constructs, but not K ~ / D ~ or / D Db/Kb/Kb. ~ Analogous results were noted with anti-H-2Db CTL (see Table 111). H-2Kb- and -Db-restricted influenza virus-
145
MHC CLASS I MOLECULES RECOGNIZED BY CTL
Derivation of domains Construct”
a1
a2
Alloreactive CTL a3
AntiH-2KI)
AntiH-ZDI)
H-2-resticted CTI,
H-2Ki)
H-2D1>
Data taken from Allen et u1. (1984)and Bluestone et a2. (1985).CTL were generated against H-2K‘) or Hl-2DlJ, or antigen (virus) restricted I)y H-ZKI’ or H-211’) and tested against the domain-shuffled molecules expressed on L cells. b.r See Ta.ble I1 for explanation.
specific C‘TL and CTL clones showed similar patterns of reactivity (Allen et al., 1984). Thus, these results are similar to those noted with the Ld/Dd/Dd and Dd/Q7d/Dd constructs (McCluskey et a l . , 1986a; Stroynowski et al., 1985a) where nonhomologous a l / a 2 domain-shuffled molecules lost all detectable CTL-controlled epitopes. Mouse strains bearing a point mutation in either the a1 or a 2 domains of H-2Kh (H-2Kbmmutant strains, see Section VII,A) were used to generate CTL against H-2Kb. Since an H-2Khm mutant strain differs with respect to H-2Kb in either the a1 or a 2 domain only, these alloreactive CTL were tested to determine if the epitopes they recognized could be localized to that particular domain. Such CTL failed to recognize any molecules shuffled between the a1 and a2 domains of H-2Kh, indicating that both a1 and a2 domains control epitopes recognized by CTL. In a secondary response, (C3H x C57BL)F, mice generated CTL that were Dh/Kh/Ktl specific. However, it was not ruled out that these CTL recognize this hybrid molecule in an H-2-restricted rather than in an unrestricted fashion. D. EXON-SHUFFLED GENESBETWEEN H-2Kd, H-2Kk, A N D H-2Kb Arnold et al. (1984, 1985) constructed hybrid genes using the H-2K alleles,
k, d, and b. These hybrid genes were transfected into IT22-6 fibroblasts or the IC9 fibrosarcoma. Two antibodies reacted with the a3 domain of H-2Kk and one with the a3 domain of -Kd. Most antibodies appeared to be specific for determinants requiring homologous a l / a 2 domains. While some mAbs showed reactivity with the a1 domain on H-2Kk or -Kd, none was reported
146
JAMES FORMAN
TABLE IV EFFECTOF SHUFFLINGTHE al-a3 DOMAINS OF H - 2 K d RECOGNITION" Derivation of domains
cxl
a2
Fig. 2C/D Fig. 2E/F Fig, 2C/D Fig. E/F NAe NA NA
Kd Kd Kk Kk Kk Kk Kd
Kd Kk Kk Kd Kk-Kdf Kk-Kb Kd-Kk
H-2Kk
ON
CTL
Alloreactive CTL Anti-
Constructb
AND
a3
Anti-
H-2Kd
H-2Kk
Kk Kk
+'
-
?(5/41)d
*(1/46)
Kd
-
Kd K" or Kk Kk K d o r Kk
-C(6/30)
-
+ -(0/30) + +
-
-
+
H-%restrictedCTL
H-2Kd
+ -
ND
ND ND ND ND
0 Data taken from Arnold et al. (1984, 1985) and Schder et al. (1986). CTL were generated against H-2Kk or H-2Kd or virus restricted by H-ZKd and tested against the domain-shuffled molecules expressed on fibroblasts. b,c See table I1 for explanation. d Number of CTL clones reactive/total number tested. e Not applicable, f Molecule shuffled at residue 142 in a2 domain (Scheller et al., 1986).
to react with the a 2 domain of either of these molecules. Antibodies that reacted with the a1 domain of H-2Kd or -Kk on domain-shuffled molecules showed decreased binding compared to binding to the homologous antigen. Bulk cultured anti-H-2Kk and -Kd CTL recognized Kk/Kk/Kd and Kd/ Kd/Kk molecules, respectively (see Table IV). No recognition was detected when the antigens were shuffled between the a1 and a 2 domains (Arnold et al., 1985). The authors extended these studies by examining CTL clones generated in limiting dilution. Of 76 anti-H-2Kk CTL clones, one was reported to recognize Kd/Kk/Kk. Of 71 anti-H-2Kd CTL clones, 11 recognized either Kd/Kk/Kkor Kk/Kd/Kd.Thus, a small minority of CTL clones generated in alloreactions can recognize determinants expressed on nonhomologous a l / a 2 molecules. H-2Kd-restricted anti-influenza CTL recognized Kd/Kd/Kk, but not Kd/Kk/Kk(Arnold et al., 1984). Domain-shuffled molecules with nonhomologous TM or cytoplasmic domains were efficiently recognized by CTL specific for the a l l a 2 domains. Scheller et al. (1986) shuffled H-2Kd, -Kk,and -Kb genes within the a 2 domain encoding exon to produce molecules with heterologous sequences between positions 142 and 182 of the a 2 domain. Bulk cultured anti-H-2Kk and anti-H-2Kd CTL readily reacted with Kk/Kk-Kd/Kk and Kk/Kk-Kd/Kd (refers to al/a2[91-141]a2[ 142-182]/a3) domain shuffled molecules but not Kd/Kd-Kk/Kd or
MHC CLASS I MOLECULES RECOGNIZED BY CTL
147
Kd/Kd-Kk/Kk. The data above (Arnold et al., 1985)indicate that most antiH-2Kd clones do not react with Kk/Kd/Kd.Thus, it is likely that the carboxyterminal end of the a 2 domain of H-2Kd, when shuffled with H-2Kk, can retain a conformation allowing for the expression of H-2Kd CTL epitopes while the complete a 2 domain of H-2Kd cannot. Since most a l / a 2 domain-shuffled molecules have lost most of the epitopes recognized by CTL, it may be expected that these molecules would acquire new epitopes unique to their structure. Support for this was provided by Horstmann et d . (1986) who studied Kd/Kk/Kk and Kk/Kd/Kd constructs and found that both were specifically recognized in secondary responses by anti-Kd/Kk/Kkand anti-Kk/Kd/KdCTL, respectively. The lysis was H-2 unrestricted. Further, these CTL did not cross-react on either native H-2:Kd or H-2Kk.
E. EXON-SHUFFLED GENESBETWEEN ZAPk AND H-2Dd OR H-2Ld McCluskey et al. (1985) exon shuffled the leader (L) and p l exons of IAkp with the a3, TM, and cytoplasmic exons of H-2Dd and transfected the hybrid gene into 1, cells. A domain-shuffled molecule was expressed on the cell membrane which associated weakly with P2-M microglobulin (McCluskey et al., 1985). This molecule was recognized by anti-IAk CTL lines generated by multiple rounds of in uitro stimulation (Golding et al., 1985). The lysis was partially blocked with the 10-2-16 mAb (anti-IAg,), further substantiating that at least a portion of anti-IAk9 CTL activity is directed against epitopes in the p l domain of IAk. An additional line that was L3T4+ had anti-IAkp, specificity, and the lytic activity of this line was blocked with anti-L3T4 antibody. These data demonstrate that at least some of the epitopes in IA molecules that are recognized by CTL do not require the presence of the a chain or the p2 domain of the p chain. Further, these data suggest that if the L3T4 molecule has a ligand, it need not be in the p2 domain of IAb or the a chain. McCluskey et al. (1986b) deleted the p l exon from the above-described exon-shuffled gene to derive a construct with the L exon from IApkjoined to the a3, TM, and cytoplasmic exons of H-2Dd. A similar construct was also prepared with the 013, TM, and cytoplasmic exons derived from H-2Ld. Both of these shuffled genes would be expected to have their first 4 amino acids contributed by the L exon of IAb. A third mutant gene was also constructed containing the L exon of H-2Dd together with the a3, TM, and cytoplasmic exons from the same genes (the a1 and a 2 exons were excised). Alloreactive anti-H-2Ld or -Dd CTL did not recognize L cells expressing these gene products, consistent with previous findings demonstrating the specificity of alloreactive CTL for the a1 and a 2 domains (see Sections 11,B-11,D). On the other hand, these constructs were recognized if mice were previously
148
JAM E S F OR M AN
primed in vivo with the L / a 3 constructs. These a3 domain-specific CTL did not recognize the a3 domain on intact H-2Ld or H-2Dd. It is possible that the recognition of the a3 domain is H-2 restricted by H-2k molecules on the L cells. The lysis did not appear to depend on the amino terminal amino acids controlled by the L exon. F. EXON-SHUFFLED GENESWITH ALTEREDCYTOPLASMIC REGIONS Class I molecules have been constructed with altered cytoplasmic domains. Zuniga et al. (1983) generated two mutant H-2Ld molecules, one containing 7 cytoplasmic amino acids and the other with 25 amino acids, including 19 random amino acids at the C-terminus. Murre et al. (198413) substituted the 3 cytoplasmic domain-encoding exons of H-2Ld for the second cytoplasmic exon of Z-Apd. The resulting protein contained 10 instead of the 31 cytoplasmic amino acids found in the normal H-2Ld product. In both studies, the truncated molecules were expressed efficiently on the surface of transfected L cells. Zuniga and Hood (1986) generated several additional mutants of class I H-2Ld molecules, including some with no cytoplasmic tail, a 4 amino acid tail, and one with a shortened TM domain and a 2 amino acid tail. All of these latter mutants had a reduced cell membrane expression. The constructs produced by both laboratories were recognized by both alloreactive and H-2-restricted CTL. However, Murre et al. (1984b) noted that VSV-specific CTL lysed cells expressing truncated H-2Ld less efficiently than the wild-type molecule. No difference was seen with anti-influenza CTL. In contrast, Zuniga et al. (1983) and Zuniga and Forman (unpublished observations) noted no difference in the lytic activity of anti-VSV CTL directed against any of the truncated H-2Ld versus normal -Ld molecules. Arnold et al. (1984) exon shuffled H-2Kd, -Kk and -Kb genes between the TM and first cytoplasmic exons. No effect on CTL lytic activity was noted. The cytoplasmic domains of class I molecules contain a highly conserved tyrosine at residue 321 and serine at 335, which can be phosphorylated (Pober et al., 1978; Guild et al., 1983). Some of these molecules are also covalently linked to palmitic acid via a cysteine encoded by the TM exon (Kaufman et al., 1984). Analysis of cDNA clones of H-2Kq (Kress et al., 1983) and H-2Dd (Brickell et al., 1983, 1985)indicate that these genes may undergo differential mRNA processing at the 3’ end. McCluskey et al. (198613) noted that an H-2Dd gene construct could give rise to two proteins, one of which represented a deletion of the peptide encoded by exon 7 . Taken together, these findings suggest that this region of the molecule plays an important functional role. However Kranz et al. (1984) and Staerz et al. (1985) have shown that class I molecules may merely play a passive role with respect to recognition and lysis by CTL effector cells. Thus, H-2- target cells bearing an anti-idiotype mAb coupled to their membrane can trigger
MHC CLASS I MOLECULES RECOGNIZED BY CTL
149
CTL clones with the corresponding idiotype to cause their specific lysis. Therefore, the presence of a class I molecule on the target cell membrane is not required for lysis, but rather activates the CTL through binding to its receptor molecules. If the cytoplasmic domain does play a functional role for CTL activity, then it must be at some point earlier in the generation of CTL from their precursors. Ill. Recognition of HLA Class I Antigens on Transfected Cells
LeBoutdler et al. (1982), Lemonnier et al. (1982, 1983a), and Barbosa et al. (1983)lransfected HLA class I genes into murine L cells. The expression was stable for several months. LeBouteiller et al. (1983, 1985) estimated that there were -4 x lo5 HLA molecules/cell, similar to the expression of endogenoiis H-2Kk molecules. One of the unexpected results from studies of HLA expressed on L cells is that in many cases these molecules are not recognized by specific anti-HLA CTL. Herman et al. (1983) transfected L cells with HLA-B7 and -A2 and used these cells as targets for specific mouse anti-HLA-B7 or -A2 CTL clones. Although some CTL clones lysed the L cell transformants, other clones had no activity. Bernabeau et al. (1983) noted that L cells expressing HLA-A2 or -B7 were not lysed using either bulk cultured or cloned human anti-HLA CTL. The L cells did not block specific lysis in a cold target inhibition assay. The ability to lyse HLA appears to be related to the species from which the cells used for transfection are derived. Thus, Barbosa et al. (1984) trarisfected HLA-A2 and -B7 into human cells; murine L, embryonic liver, and B lymphoma cells; and monkey kidney cells. Human anti-HLA CTL clones lysed human transformants, a minority of clones lysed monkey cells but to a lesser extent, and no lysis was observed against transformed mouse cells. These results were not due to a low level of expression of HLA on these cells. Although HLA on mouse cells is not readily recognized by anti-HLA CTL, H-2 antigens are readily detected on human cells by antiH-2 CTL (Maziarz et al., 1985). The failure of CTL to recognize HLA expressed on the surface of transfected murine cells could be due to a number of different factors, including (1)quantiiy of cell membrane HLA molecules, (2) qualitative character of HLA or P2-Mon cultured murine cells, or (3) the interaction of accessory molecules on CTL with ligands on the target cell. Quantitative expression of HLA on murine cells does not account for the lack of lysis, since there is relatively high expression of these molecules on L cells (LeBouteiller et al., 1983, 1985), and human cells transfected with the same HLA gene as L cells and expressing equivalent amounts of cell membrane antigen are readily lysed (Barbosa et al., 1984; Maziarz et al., 1985;
150
JAMES FORMAN
Maryanski et al., 1986). There is no serological evidence to indicate that determinants on HLA molecules are altered relative to their endogenous expression on human cells, since mAb typing reagents react with both transfected and endogenous HLA (LeBouteiller et al., 1982, 1983; Bernabeu et al., 1983; Herman et al., 1983; Layet et al., 1984). However, since serological epitopes do not define CTL specificities (see Sections VII and VIII), this type of analysis may not be definitive. Since HLA molecules expressed on murine L cells would not be associated with human &-M, L cells expressing high levels of human P2-M were transfected with HLA (van de Rijn et al., 1984). However, these target cells were not susceptible to lysis either. The role of P2-M was further addressed by Bernabeu et al. (1984a)who examined L cells transfected with both HLA and human P2-M or L cells cultured in human serum. In the latter case, it was shown that serum P,-M rapidly interacts with membrane HLA through an exchange reaction (Bernabeu et al., 1984b). In neither instance were L cells lysed by anti-HLA CTL. Engelhard et al. (1985) and Maziarz et al. (1985) constructed hybrid molecules consisting of HLA in the 01l/a2 domains and H-2 in the 013 and carboxy-terminal domains. This did not alter the inability of anti-HLA CTL to lyse L cells transfected with HLA. No lysis was induced by the addition of lectin to mixtures of human effector cells and transformed L cells (Bernabeu et al., 1983, 1984a), although lysis was reported to have occurred when mouse effector cells were used. Since some accessory molecules may be required for lectin-induced lysis (Springer et al., 1982), it is possible that an interaction between intraspecies accessory molecules and their ligands is required for optimal recognition/lysis of anti-HLA effector cells and their targets. This is consistent with the findings of van de Rijn et al. (1984), who noted that an anti-HLA-A2 CTL clone that lysed human, but not mouse fibroblasts transfected with HLA-A2 was inhibited by antibodies against LFA-1 and CD8. Spits et al. (1986) showed that two human CTL clones could not form conjugates with murine L cells, but did with murine-derived P815 cells, suggesting that L cells lack some accessory molecules that permit effector-target cell interactions. Golde et al. (1985)and Naquet et al. (1985) found that T hybridoma recognition of antigen on class I1 gene-transfected L cells were not inhibited by anti-LFA-1 mAb. Cowan et al. (1985) transfected an HLA-A3 subtype gene from the E l donor (see Section VI1,B) into L cells and showed that human secondary and tertiary anti-HLA CTL lysed these target cells, even in the absence of human P2-M. Antibodies to CD3, CD8, and LFA-1 blocked lysis of both L cell transformants and human PHA blasts expressing the HLA-A3 variant molecule. These data suggest that if CD8 and LFA-1 interact with ligands, they can do so across the species barrier and thus would not account for the
MHC CLASS I MOLECULES RECOGNIZED BY CTL
151
lack of lysis observed in the other studies described above. Greenstein et al. (1986) also showed that Lyt-2+ mouse anti-human CTL could be blocked from lysing specific target cells with anti-LFA-1 antibodies, although the same cells could not be blocked with anti-Lyt-2 antibody. Barbosa et al. (1986) showed that murine L cells lack LFA-3, but that acquisition of this molecule through somatic cell hybridization did not restore recognition by CTL clones incapable of lysing HLA-transfected target cells. Mentzer et al. (1986) noted that most the human anti-HLA CTL clones did not recognize HLA expressed on murine cells, although 2 CTL clones could specifically lyse L cells transfected with HLA. Based on the results of inhibition of target cell lysis using anti-HLA mAb, these two clones appear to have a higher avidity for HLA compared to CTL clones unable to kill mouse transfectants. While these clones could lyse HLA expressing human and mouse cell targets, anti-LFA-1 mAb inhibited lysis to a greater extent when mouse rather than human targets were used. mAb to LFA-2 and -3 inhibited lysis against human, but not mouse targets. These data suggest that either murine cells do not use LFA-2 and -3 accessory molecules or that these receptor-ligand interactions do not occur between human and mouse cells, and as a result LFA-1 interactions become relatively more important for conjugate formation. It should be noted that the lack of lysis observed in the studies of Herman ef al. (1983), Bernabeu et al. (1984a), and Engelhard et al. (1985) involved anti-HLA CTL that were of murine origin where accessory molecules on niurine effector cells should be able to interact with ligands on murine L cell targets. However, some of these ligands may be lacking on murine fibroblasts (Shimonkevitz et al., 1985) and the xenogeneic anti-HLA CTL may be of low avidity and require these ligands for recognition. It is also possible that class I molecules interact with species-specific molecules, both of which control polymorphic epitopes recognized by alloreactive CTL (Matzinger and Bevan, 1977). In addition to the finding that some anti-HLA CTL clones can lyse HLAtransfected L cell targets (Herman et al., 1983; Maryanski et al., 1986), Maryanski et al. (1985) transfected HLA-Aw24 into murine P815 cells and showed that bulk cultured anti-Aw24 CTL readily lysed these cells. Achour et al. (1986) noted that anti-HLA CTL could be generated that recognized transfected L cells, albeit weakly, but that lysis directed against P815-transfected target was highly efficient. Gomard et al. (1986) transfected HLA-A2 into P815 cells and showed that anti-influenza-specific CTL restricted by HLA-A2 lysed infected targets. Since Spits et al. (1986) showed that P815 cells form conjugates with human CTL clones, it would appear that the appropriate ligands are expressed on these cells which can interact with human accessory molecules. Thus, present data suggest that most human anti-HLA CTL require and
152
JAMES FORMAN
participation of accessory molecules, possibly including LFA-2 (CD2) and LFA-3, for efficient recognition of HLA by the T cell receptor. Whether murine fibroblast targets express all of these molecules or their ligands is not known, nor is it known whether accessory molecule interactions occur between species. Further, there may be heterogeneity between L cell lines in regard to the expression of accessory molecules, since Mentzer et al. (1986) could block recognition of antigen on L cells with the anti-LFA-1, whereas Golde et al. (1985) could not. Xenogeneic murine anti-HLA CTL responses may be of low affinity and require an accessory molecule interaction that may not occur with fibroblast targets. Achour et al. (1986), Maryanski et al. (1986a), and Holterman and Engelhard (1986) noted that mice primed in uiuo with murine cells transfected with HLA allowed for the generation in uitro of H-%restricted anti-HLA CTL. The results are similar to those for CTL generated against minor H antigens, although in this case the minor H antigen is HLA. Thus, it appears that presentation of xenogeneic class I molecules on xenogeneic cells gives rise to an MHC-unrestricted anti-class I response, while presentation of the same molecules on cells from the same species elicits an MHC-restricted antigen-specific response. Further, in the latter case the recognition of HLA is similar to that described for the nucleoprotein of influenza in that an HLA peptide from HLA-CW3 spanning amino acid residues 171-186 sensitized H-2d target cells to the lytic effects of H-2Kd-restricted DBAIB anti-HLA CTL (Maryanski et al., 1986b). IV. Role of Carbohydrate Moieties in Determining CTL Recognition of Class I Molecules
The heavy chain of class I MHC molecules contains N-linked oligosaccharides attached to the al, the a1 and a2, or the al, a2, and a3 domains. Doberstein et al. (1979) showed that glycosylation of class I MHC molecules is similar to that described for other viral and cell membrane glycoproteins. Class I molecules from all species have a conserved asparagine at residue 86 to which a glycosyl group is attached. In the mouse, an asparagine at 176 also serves as an attachment site for glycans, and in H-2Ld, H-2Db, and H-2Kd molecules there is a third asparagine acting as an attachment site at residue 256 (Kimball and Coligan, 1983). Oligosaccharides have been shown to play a role in the transport of class I molecules to the cell membrane (Ploegh et al., 1981; Landolfi et al., 1985; Miyazaki et al., 1986). For class I molecules, glycans could also (1)comprise a portion of an epitope recognized by CTL, (2) control the conformation of class I epitopes, or (3)affect the interaction of viral peptides with class I molecules. These issues have been approached with drugs that inhibit glycosylation of
MHC CLASS I MOLECULES RECOGNIZED BY CTL
153
proteins. Drugs frequently used are tunicamycin (Tm), which inhibits the attachment of N-acetylglucosamine to dolichol phosphates and thus prevents N-linked oligosaccharide linkage (Tkacz and Lampen, 1975), and 2-deoxy-~glucose (2-DOG), which at low concentrations inhibits glycosylation (reviewed in Scholtissek, 1975). The effect of Tm on the expression of class I molecules varies. Swaminathan and Gooding (1983) removed H-2Kk molecules from C3H cells with papain and showed that their reappearance was prevented by the presence of Tm. Colombo et al. (1983) showed that H-2b antigens were reduced on cells cultured in Tm, although H-2Kb was reduced less than -Db. They also provided evidence that nonglycosylated H-2Kb molecules were present on the membrane of Tm-cultured cells. Carter et al. (1981) did not quantitate H-2 on Tm or 2-DOG cultured cells, but noted that subconfluent SV40transformed C57BL/6 embryo fibroblasts had a reduced sensitivity to lysis mediated hy anti-H-2b CTL. Black et al. (1981) found that P815 cells cultured with Tm for 18 hours expressed normal amounts of cell membrane H-2Dd, but in a nonglycosylated form. Miyazaki et al. (1986) used sitespecific mutagenesis to change the consensus sequences of Asn-X-Ser/Thr by altering the asparagines at position 86, 86 and 176, or 86, 176, or 256. The result was the production of mutant H-2Ld genes which, when transfected into L cells, encoded proteins devoid of oligosaccharides in the al,a1 and a2, or al- a3 domains, respectively. While mutant molecules lacking a glycan in the a1 or a1 and a 2 domains expressed equivalent amounts of H-2 on the cell membrane as native H-2Ld, mutant molecules lacking carbohydrate in all 3 domains were poorly expressed at the plasma membrane. The low level of surface expression of this latter mutant was not due to accelerated degradation or increased shedding of the molecule, but rather to a decrease in intracellular transport. Ploegh et al. (1981) cultured JY cells with Tm and showed that nonglycosylated HLA-A and -B antigens reached the surface of these cells at rates indistinguishable from normal HLA antigens, although the amount of HLA synthesized in such cells was decreased. In general, CTL functional data parallel the effect of Tm on the cell membrane expression of class I molecules. Thus, antiviral-specific CTL directed against SV40 (Swaminathan and Gooding, 1983), herpes simplex virus (HSV) type 1 (Carter et al., 1981), VSV (Black et al., 1981), and Moloney sarcoma-leukemia virus (Colombo et al., 1983; Watson and Bach, 1980) all showed reduced amounts of lysis when the target cells were cultured with Tm or 2-DOG. In contrast, the extent of lysis directed against H-2 by alloreactive T cells was not altered (Watson and Bach, 1980; Colombo et al., 1983) except in one study (Swaminathan and Gooding, 1983).Thus, the lack of lysis in the antigen-specific systems has been suggested to be due to alteration of viral proteins such that these unglycosylated molecules are
154
J A M E S FORMAN
either not expressed on the cell membrane or not recognized by CTL. Black et al. (1981) demonstrated that unglycosylated VSV G protein is expressed on the cell membrane of Tm-cultured P815 cells. Recent data have indicated that at least a portion of antiviral CTL reactivity is directed against nonglycosylated viral proteins (Townsend et al., 1984; Abastado et al., 1985). Therefore the effect of Tm in decreasing antigen-specific CTL-mediated lysis could be explained by altering the expression of ligands on target cells that bind accessory molecules on cytolytic T cells; e.g., Lyt-2, LFA molecules. Pimlott and Miller (1986) noted that anti-H-2 conjugate formation was specifically inhibited by glycopeptides extracted from tumor cells. However, these glycopeptides weakly inhibited target cell lysis, and it is not known if these molecules are derived from class I antigens. Jenkins et al. (1985) showed that some of the epitopes on the Qa-1 molecule that are recognized by alloreactive anti-Qa-1 CTL clones are lost with Tm treatment. Since Tm and 2-DOG can have pleiotropic effects (MacDonald and Cerottini, 1979), clearer studies have been performed by Miyazaki et al. (1986) who employed site-specific mutagenesis of class I genes, and Goldstein and Mescher (1986) who prepared liposomes containing deglycosylated H-2Kk molecules. Goldstein and Mescher (1986) showed that the liposomes containing deglycosylated H-2Kk were able to induce alloreactive CTL to the same extent as liposomes with native H-2Kk. Miyazaki et al. (1986) showed that mutant H-2Ld molecules that lacked carbohydrate attachment sites in the a l , a1 and 012 or al-a3 domains, when expressed on the surface of L cells, functioned as target molecules for both alloreactive and VSV-specific CTL. In addition, five out of five alloreactive anti-H-2Ld CTL clones all retained reactivity against the mutant molecules. Thus, it is unlikely that carbohydrates on class I molecules play any more than a minor role in contributing to antigenic determinants recognized by alloreactive or H-%restricted CTL. V. Role of &-Microglobulin in T Cell Recognition of Class I Molecules
P2-M noncovalently associates with the heavy chain of class I molecules (Nakamuro et al., 1973; Natori et al., 1974; Rask et al.,1974; Silver and Hood, 1974;Vitetta et al., 1975). This molecule has a molecular weight of 12,000 and is usually invariant within a species, with the exception of mice where 7 allelic forms have been detected (Michaelson et al., 1980; Gates et al., 1981; Robinson et al., 1984; Gasser et al., 1985; K. Fischer-Lindahl, personal communication). The amino acid sequence homology between human and murine P2-M is -70%, between murine and bovine 68%, and between human and bovine 76% (Gates et al., 1981). Yokoyama and Nathenson (1983) digested H-2 with glycosidases and proteolytic enzymes and found that
MHC CLASS I MOLECULES RECOGNIZED BY CTL
155
fragments associated with the P2-M subunit were derived from the a3 domain, suggesting that the a3 domain predominantly interacts with P,-M. However, the binding of the W6/32 mAb, which reacts with a determinant localized to the a l / a 2 domain of HLA (Maziarz et al., 1985), is greatly reduced if the P2-M is of murine rather than human origin (Ferrier et al., 1985). Further, the a l / a 2 domains of HLA-AS, but not -Cw3, can alter epitopes 011 murine P2-M (Jordan et al., 1983). This indicates that P2-M can interact or affect the conformation of d / a 2 domains of class I molecules. Similarly, McCluskey et al. (1986a) noted that the association of P2-M with H-2 was affected by the a1 domain ofclass I molecules, further supporting the idea that P2-M interacts with heavy chains at more than just the a3 domain. The interaction of @,-M with the heavy chain of class I molecules results in an alteration of the conformation of P2-M. Lemonnier et al. (1983b) and Agthoven et al. (1984) noted that a murine anti-human P2-M mAb reacted with murine P2-M associated with HLA. Apparently the murine P2-M requires a conformational determinant that resembles a human p,-M epitope when the murine P2-M interacts with a human class I molecule. P2-M from serum can readily exchange with endogenous P,-M on the cell membrane. Bernabeu et al. (198413) noted that human cells grown in fetal bovine serum (FBS) had both human and bovine P2-M associated with T6 and HLA-MB. H-2KID antigens exchange murine for bovine @,-M with a t,,, of -2 hours. HLA-B7 expressed on JY cells cultured in FBS associates predominantly with human P2-M, whereas HLA-B7 expressed on murine L cells associates predominantly with bovine P2-M. Thus, the ability of one species of pz-M to exchange with another appears to depend on the source of the endogenous &-M. Kubota (1984) noted that P2-M from FBS exchanged with murine P,-M on L cells and that the acceptor molecules for the exchange were the heavy chain of class I molecules. The association of P,-M with class I molecules could have an effect on the structure of class I molecules themselves. Thus, most anti-HLA antisera do not react with isolated heavy chains (Krangel et al., 1979) and HLA loses some of its @-pleatedsheet structure in the absence of P2-M (Lancet et al., 1979). P,-M may also be required for the transport of class I molecules to the cell membrane. Thus, P,-M-negative Daudi cells make cytoplasmic HLA, but do not transport it to the cell membrane (Ploegh et al., 1979; d e Preval and Mach, 1983). Similar findings have been described for the P,-M- Rl(TL-) mutant cell line (Hyman and Stallings, 1977) which also fails to express membrane H-2. There is little information as to whether P2-M plays a role in controlling epitopes recognized by alloreactive CTL. Although there is only -70% homology of P,-M sequences between species (Gates et al., 1981), Cowan et
156
JAMES FORMAN
al. (1985)noted that anti-HLA CTL could lyse murine cells expressing transfected HLA in the absence of a source of human P2-M. It is not known whether a large panel of CTL clones would be able to distinguish class I molecules associated with different species of P2-M microglobulin. An mAb directed against P2-M blocks about 40% of alloreactive CTL activity, while blocking of H-2Db-restricted anti-H-Y-specific CTL was only inhibited -20% (Tomonari et al., 1982). The blocking could be due to steric factors or reflect poor inhibition by a low-affinity antibody. Langhorne and FischerLindahl (1982) failed to detect any alloreactive CTL clones that distinguish between the two murine alleles of P2-M. However, Rammensee et al. (1986) generated a CTL line across the H-3 barrier which includes the P2-M locus. This CTL line was specific for the b allelic form of P2-M and was restricted by H-2Kb. Target cells from &-Ma strains incubated with purified P2-Mb acquired the target antigen. Thus, in this case specific reactivity against a minor H antigen was shown to be directed against a determinant dependent on a P2-M alloantigen. VI. Use of Monoclonal Antibodies to Block CTL Recognition of Class I Molecules
Fischer-Lindahl and Lemke (1979) demonstrated that anti-H-2Kb mAbs could block alloreactive and H-2-Kb-restricted CTL. One antibody, 27R9, was the least effective at blocking, but the other two mAbs blocked lysis almost completely. Epstein et al. (1980) tested eight anti-H-2k mAbs, all of which were able to block CTL activity. In general, the affinity of the mAb for the H-2 antigen correlated with its blocking activity. The use of mAbs directed against particular domains of class I molecules may allow an assignment of CTL reactivity to epitopes localized to one of the three external domains. Hammerling et al. (1982) described a series of mAbs directed against H-2Kb and further defined their specificity based on their reactivity with H-2Kbm mutant strains. Two antigenic clusters were identified, one localized to the a 2 domain based on the fact that the antibodies do not react with the H-2Kbm1 and Kbm4 molecules, which are a 2 domain mutants (see Section VI1,A). The other cluster presumably reacts with the a1 domain, since these antibodies do not react with the a1 mutant, H-2Kbm3. Several of these antibodies were also later described by Allen et al. (1984)and Bluestone et al. (1985) using the technique of exon shuffling. The cluster of antibodies defined by Hammerling et al. (1982) as reacting with the a 2 domain was shown to have reactivity with the a 2 domain of H-2Kb on a domain-shuffled molecule, although it was much weaker when compared to binding to the homologous molecule (Allen et al., 1984; Bluestone et al., 1985). Three antibodies defining the a1 domain were shown to react with the a1domain of H-2Kb only or al/012 using domain-shuffled molecules. Thus, an
MHC CLASS I MOLECULES RECOGNIZED BY CTL
157
antibody (e.g., 10-56) that lost reactivity with an al-domain mutant (H-2Kbm3)and is presumably specific for an epitope in the a1 domain does not necessarily react with a molecule containing H-2Kb in only the a1 domain, supporting the concept that many serological epitopes require interaction of both a1 and a 2 domains. Lemke et al. (1979) described six mAbs that reacted with H-2Kk; three reacted with one region of the molecule (cluster A) while the other three reacted with a second region (cluster B), as defined by inhibition of binding. Antibodies to either cluster blocked a portion of alloreactive anti-H-2Kk CTL activity as well as -Kk-restricted anti-TNP activity (Weyand et al., 1981a). When all six mAbs were used in a mixture, the lysis was completely blocked. Blanden et id. (1979)noted that H-2Dk-restricted anti-influenza, but not antiBebaru virus CTL were inhibited by a -DkmAb. Stringfellow et al. (1983) noted that the ability of mAbs directed against H-2Kk to inhibit influenza virus-specific CTL varied with the mouse strain, suggesting that different selfrestricting epitopes were selected by mice of different background genes. Pircher et al. (1984) described H-2Kb-restricted CTL clones specific for the hapten I-AED [N-iodoacetyl-N-(5-sulfonic-naphthyl)ethylenediamine]. One set of clone:; (type A) did not recognize H-2Kbm mutants that had amino acid substitutions in their a1 domain (H-2Kbm3,8,11),while a second set (type B) did not recognize mutants that had amino acid substitutions in the “loop” region of the a 2 domain (H-2Kbm536,9).Type B CTL clones were inhibited from lysing H-2Kb AED coupled targets by three mAbs specific for the a 2 domain, while type A was not inhibited by the same mAbs. Thus, these type B clones could be predominantly a 2 domain specific. Further support for the existence of domain-specific CTL based on mAb inhibition of CTL activity was produced by Weyand et al. (198lb), who showed in limiting dilution that anti-H-2Kk CTL consist of three populations with different precursor frequencies. rnAbs directed against cluster A block -70% of clones with the highest precursor frequency and only 25% of clones of the lowest precursor frequency. In contrast, antibodies to cluster B do no block high-frequency CTL, but do block 50-70% of clones of the lowest frequency. There are numerous reports indicating a discordance between epitopes on class I molecules defined by mAbs versus CTL. Davignon et al. (1983) generated TNP-specific H-2K“-restricted CTL clones and used mAbs to inhibit their ability to lyse TNP-derivatized targets. Two mAbs, which do not react with the H-2Kbm3mutant molecules, were able to block the reactivity of CTL clones on H-2Kb-TNP targets. These clones also lysed H-2Kbm3-TNP targets, indicating that the antibody specificity does not correlate with the CTL epitope. Rusch et al. (1983) tested CTL generated in bulk culture between H-2Kbm mutant strains and C57BL/6 (H-2Kb). H-2Kbm3anti-B6 (putatively a1 domain specific) and H-2Kbm1anti-B6 (putatively a 2 domain
158
J A M E S FORMAN
specific)were tested with domain-directed antibodies (termed epitope groups A, B, and C) for inhibition of CTL activity. No correlation was observed between the epitope groups that the mAb was directed against (epitope group A corresponds partly to the a1 domain and epitope group C corresponds to the a 2 domain) and the putative domain specificity of the CTL. Similarly, Bluestone et al. (1984a) noted that an anti-H-2Ld CTL clone which did not react against an mAb-selected H-2Ld variant cell could be blocked from killing a cell expressing wild-type H-2Ld using the same mAb. Wraith et al. (1983) showed that an anti-influenza-specific H-2Kk-restricted CTL clone lysed a target cell that expressed an H-2Kk variant molecule which lacked reactivity with an anti-H-2Kk mAb, yet could be blocked from lysing target cells expressing wild-type H-2Kk with the same mAb. Thus, these studies demonstrate that determinants defined by mAbs are different from those recognized by CTL. Since exon-shuffling studies have shown that the a3 domains do not control polymorphic determinants recognized by CTL (see Section 11),it might be expected that mAbs directed against a3 would not block CTL recognition of class- I molecules. Thus, McCluskey et al. (1986b) did not block antiH-2Ld CTL activity with the 28-14-8 mAb directed against the a3 domain, but did note inhibition of a3 domain-specific CTL with the same mAb directed against a domain-shuffled molecule missing the a1 and &2 domains, but expressing the a3 domain on the cell membrane. However, there are several reports indicating that a3 domain mAbs do block both alloreactive and H-2-restricted CTL (Ciavarra and Forman, 1982; Orn et al., 1982; Forman et al., 1983; Levy et al., 1983; Ozato et al., 1983a; Stroynowski et al., 1985b). While these data would support the concept that blocking of CTL activity with mAb directed against the a3 comain is steric, it is possible that the a3 domain contributes a constant portion of a polymorphic epitope or that the a3 domain serves as a ligand for accessory molecules on CTL that assist in binding. Further evidence consistent with this is the findings of Maziarz et al. (1985) who noted that hybrid HLA molecules containing H-2 in the a3 domain and remaining carboxy-terminus were recognized less efficiently by anti-HLA CTL than the homologous molecule. Analogous results were also observed for reciprocal H-2/HLA constructs. VII. Class I Heavy Chains Bearing Defined Amino Acid Changes: Effect on Polymorphic Determinants Recognized by CTL
A. H-2 MUTANTSTRAINS Bailey and Kohn (1965), Egorov and Blandova (1968), Kohn and Melvold (1974), and Melvold and Kohn (1976) skin grafted syngeneic mice to screen
MHC CLASS I MOLECULES RECOGNIZED BY CTL
159
for histoconipatibility mutants. Several mutant mice were identified with a mutation in one of their M H C genes, and coisogeneic strains were derived from these mutant animals (for review, see Klein, 1978). Most of the mutants were detected in (BALB/c x C57BL/6)Fl animals, and of these, most involved the (257BL/G-derived class I H - 2 K b gene. Other strains bearing mutant M H C genes have also been described, including the H-2dm1 and H-2dm2mulants involving H-2D and H-2L molecules encoded by the D end of the H-2d haplotype (Egorov, 1967; Melvold and Kohn, 1976; Sun et al., 1985; see Section 11,B). Nathenson and his colleagues (see Nathenson et al., 1986)have analyzed the structure of 12 of the H-2Kb in vivo-detected mutant genes. These mutations have occurred in either the a1 or a 2 domain, are complex, involving changes in several bases, and are probably the result of gene conversion mediated by other donor genes in the H-2b haplotype (Pease et a/'., 1983). All of these mutant molecules have an alteration in H-2Kb epitopes detected by alloreactive anti-H-2Kb CTL (Forman and Klein, 197513; Melief et al., 1980; Albert et al., 1982; Sherman, 1980) and H-2Kb-restricted CTL (Blanden et al., 1976; Zinkernagel, 1976; Whitmore and Gooding, 1981; Pan et al., 1982; Wettstein, 1982; de Waal et al., 1983b). In contrast, most of these molecules are serologically very similar to native H-2Kb. Since these mutant strains are detected by their ability to elicit skin graft rejection (Bailey and Kohn, 1965), it is expected that all these mutant molecules should elicit T cell responses by altering polymorphic epitopes on H-2Kb. An extensive analysis of these mutants has been reviewed by McKenzie et al. (1977), Klein (1978), and Nathenson et al. (1986). The 12 H-2Kbm mutant molecules that have been characterized either at the protein or nucleotide level are shown in Tables V and VI. Although the number of mutants is limited, several seem to involve the same amino acid changes, e.g., H-2bm5,16and H-2bm6,7,9.There is a clustering of mutations involving certain amino acids. Thus, mutants H-2bm3,11,23have an Asp * Ser interchange at residue 77. This residue is also one of the more variable positions between H-2 alloantigens (Kimball and Coligan, 1983). H-2Kbm8 has a Glu -* Ser interchange at residue 24, which represents another variable amino acid. H-2Kbm1, the mutant least related to H-2Kb at the T cell level (Melief et al., 1980; Sherman, 1980; Clark and Forman, 1983), has alterations at residues 152, 155, and 156, and residues 155and 156 are highly polymorphic within the murine species. HLA-A2 and -A3 subtype molecules also have amino acid changes involving residues 152 and 156 (Krangel et al., 1982, 1983; van Schravendijk et al., 1985). There is a hierarchy of relatedness of H-2 mutants to H-2Kb (Melief et al., 1980; Sherman, 1980; Clark and Forman, 1983), as defined by CTL reactivity, with H-2Kbm1being the least related, while H-2Kbm5,6,7,9,16 are the most similar to H-2Kb. This latter group of mutants has a Tyr * Phe in-
TABLE V AMINO ACID ALTERATIONS ON CLASS I MOLECULES AFFECTING T CELLREACTIVITY~ Amino acid residue a1 domain 1 11 21 31 41 51 61 71 81 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890
c
8
Variable residuesb Mutants H-2Kb KbmE Kbm3 Kbmll Kbm2.3 R8.313 HLA subtypes HLA-A2 M7;DRl
0
0
x
y
0
x
0
0
0
0
0
x
X
X
x x x x X
X
a Data taken from Nathenson et al. (1986), G. Pfaenbach and S. G. Nathenson (personal communication), Nakagawa et al. (1986), Krangel et al. (1982, 1983), Taketani et al. (1984), and van Schravendijk et al. (1985). Residues of murine class I molecules with a variability score >6 (Kimhall and Coligan, 1983) calculated according to Wu and Kabat (1970).
TABLE VI AMINOACID ALTEHATIONS ON CL.\SSI MOI.ECUIXS AFFECTINGT CELLR E A C T I V I ~ Amino Acid Residue a 2 Domain
151 161 171 181 91 101 111 121 131 141 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 1234567890 12 Variable residues" Mutants H-2K"
0 0
n
00
Kbml
r
x x x
Kbrn5.16
X
Kbrn6.7.9
X
X
Klim4
B
xx x x x
KbrnlO
HLA subtypes HLA-A2 M7;DKl DK1 HLA-A3 El HLA-B7 CF a
See Table V for explanation and references. Indicated region of mutation(s), site not defined
h
. . . .
x
X
x
x
x
x
xx
xx
162
JAMES FORMAN
terchange at residue 116 [mutants H-2Kbm6,7,9have an additional Cys + Arg interchange at 121, but are indistinguishable from mutants H-2Kbm5,16 at the T cell level (Melvold et al., 1982)l. Shiroishi et al. (1984) changed the tyrosine at position 116 of H-2Ld to phenylalanine. Thus, it was to be expected that alteration of this residue would alter this molecule in a fashion similar to what happens with H-2Kbm5,l6.However, alloreactive anti-H-2Ld CTL lysed these target cells like normal -Ld-bearing targets. Since mutants H-2bm5,16are closely related to H-2Kb, as defined by CTL reactivity, a more detailed analysis with anti-H-2Ld CTL clones may be required to reveal differences. C57BL/6 (H-2b) anti-H-2Kbm1 CTL cross-react on H-2Ld expressing targets (Nabholz et al., 1975), and this molecule bears the three mutant amino acids of H-2Kbm1,i.e., Ala, Tyr, Tyr at positions 152, 155, 156, respectively, suggesting that these amino acids are recognized in a linear fashion by antiH-2Kbm1 CTL. Hunt and Sears (1983) demonstrated that H-2b anti-H-2a CTL lysed H-2Kbm1 target cells, and a portion of this activity could be blocked by H-2d, but not H-2dm2 inhibitors, the latter of which lack expression of H-2Ld. Similarly, Mann et al. (1987; and unpublished observations) noted that CTL directed against the QlO molecule, which also bears the three mutant amino acids of H-2Kbm1(Mellor et al., 1983), react against target cells expressing H-2Ld. De Waal et al. (1983a), however, noted that most H-2Kb anti-H-2Kbm1 CTL activity was not cross-reactive on H-2Ld, but was reactive against other unrelated H-2 molecules. Further, H-2dm2 anti-H-2d CTL (anti-H-2Ld) do not distinguish H-2b from H-2bm1target cells (Hunt and Sears, 1983). Thus, definitive evidence is lacking for recognition of linear determinants on class I molecules by alloreactive CTL. It should also be noted that although the assignment of H-2Kbm1cross-reactivity with H-2Ld relies on studies using the H-2dm2mutant strain, H-2dm2mice have deleted three other class I H-2d genes in addition to H-2Ld (Stephan et al., 1986). A recent report by Parham et al. (1987) indicates that anti-HLA-A2 CTL can be specifically blocked from recognizing HLA-A2 targets by a peptide spanning amino acids 94-112 of HLA-A2, albeit at high concentrations. Whether this peptide directly interacts with the receptors on CTL or interacts with class I molecules on the target cell has not yet been ascertained. Most data derived from studies with H-2 mutant mice suggest that CTL recognize determinants controlled by both the a1 and a 2 domains. Thus, B6 anti-H-2Kbm" CTL clones which might presumably be specific for an epitope in the a1 domain (the H-2Kbm11 mutation has altered amino acids at positions 77 and 80) recognize other H-2Kb mutant molecules carrying the wild-type residues at positions 77 and 80, but vary elsewhere in the molecule (Sherman, 1982). Similarly, both a1 domain mutants H-2Kbm3,11and
M H C CLASS I MOLECULES RECOGNIZED BY CTL
163
a 2 domain mutant H-2Kbm1have lost most of their H-2Kb-restricting determinants for Sendai virus and VSV (de Waal et al., 1983; Clark and Forman, 1983). Furthermore, (H-2Kbm3 x H-2Kbm1)F, mice which express H-2Kb molecules with mutations in the a1 and a 2 domains, respectively, are able to generate anti-H-2Kb responses (Apt et al., 1975; Melief et al., 1980). If allodeterminants were solely domain specific, then the F, animals would be expected to be unresponsive to H-2Kb. While the above data suggest that alloantigenic and H-2-restricting epitopes result from the interaction of the a1 and a 2 domains, it is not clear whether the epitopes are localized to one domain as opposed to being formed by the interaction of amino acids from both domains. The description of restricting activity by class I mutant molecules generally correlates with alloantigenic activity. Thus, the H-2Kbm1molecule is the least related to H-2Kb as defined by alloreactive CTL and does not crossreact with H-2Kb with respect to restricting determinants (Table VII). H-2Kbm5,6,’3 molecules are closely related to H-2Kb, as defined by alloreactive CTL, and are highly cross-reactive with H-2Kb-restricting epitopes. Most other class I mutants show a varied retention of allo- and restricting determinants compared to H-2Kb, suggesting that epitopes on class I molecules can be recognized by T cell receptors in many different ways. Shirioshi et al. (1984) used site-specific mutagenesis to change the cysteine at position 101 of H-2Ld to serine so as to interrupt the disulfide bridge in the a 2 domain. This mutant gene was transfected into L cells where marked changes in the binding of H-2Ld-specific mAbs were noted. Not unexpectedly, alloreactive CTL failed to recognize the mutant molecule. Jordan et al. (1983)noted that an HLA gene lacking a cysteine at residue 164 (Malissen et al., 1982), which also should not be able to form a disulfide bridge in the a 2 domain, could not be successfully transfected into L cells. B. HLA SUBTYPESDEFINEDBY CYTOTOXIC T LYMPHOCYTES Many serologically defined HLA antigens can be further subdivided into subtypes using alloreactive or HLA-restricted CTL (Biddison et al., 1982a; Breuning et‘ al., 1982; Horai et al., 1982; Spits et al., 1982; van der Poel et al., 1983a,h; Molders et al., 1983; Goulmy et al., 1984). Thus, these subtypes bear resemblance to the H-2Kb mutant antigens described above. HLA-A2 can be divided into one major and three minor subtypes using both CTL and biochemical analyses (van der Poel et al., 198313). Approximately 11% of Caucasian populations express one of the minor HLA-A2 subtypes. The HLA-A2 minor subtype from donor DK1 is not recognized by anti-influenza HLA-A2-restricted CTL (Biddison et al., 1982a) or an alloreactive anti HI,A-A2 CTL line (Ware et al., 1983). This HLA-A2 molecule has three amino acid changes at residues 149, 152, and 156 (Krangel et al., 1983)
TABLE VII O N C u s s I MUTANT OR SUBTYPE MOLECULES THATCROSS-REACT WITH NATIVE DETERMINANTS~ CTL DETERMINANTS Antigen Class I molecule Parent H-2Kb
Mutantlsubtype
(az)~
~-2~bm5 H-2Kbms.’ (a2) ~ - 2 ~ b m 4(a2) H-2Kbm1 (4 H-2Kbms (al) ~ - 2 ~ b m 3(4 H-2Kbmll (4 HLA-A2 M7, DR1 (a1 and a2) DK1 (4 HLA-A3 E l (a2) (4 HLA-B7 C F
l b 2 TNP VSV
+ +
+ +
+
-
-
*
* +
*
+ -
-
3 4 5 6 SV ECTRO VAC LCM
+ + +
-
+ -
-
+ +
+ + -
-
+
-
-
-
+
+ +
7 SV40
8 HSV
+ +
+ +
9 10 MoLV FLU
11 12 13 14 15 16 EBV H-Y H-1 H-3 H-4 ALLO
+ +
+
-
+ + +
-
-
-
-
-
+
-
-
+ + +
-
* - -
-
+ +
+ + +
-
-
-
+ + 2
-
’’
-
-
-
a Abbreviations: TNP, Trinitrophenylated cells; SV, Sendai virus; ECTRO, ectromelia virus; VAC, vaccinia virus; LCM, lymphocytic choriomeningitis virus; SV40, simian virus 40; HSV, herpes simplex virus; MoLV, Moloney leukemia virus; ALLO, alloreactive. Because of different assay systems used, some variation in the extent of cross-reactivity exists. Therefore, the results are meant to indicate the relative extent of cross-reactivity only. + indicates moderate-to-high cross-reactivity of antigen-specific class I restricted CTL on mutant or subtype targets; 2 indicates weak cross-reactivity; - indicates little or no cross-reactivity. Reference key: 1, Forman (1979), Levy and Shearer (1982), Davignon et al. (1983), 2, Clark et al. (1985), 3, de Waal et al. (1983b), 4, Blanden et al. (1976), Pang et al. (1977), 5, Zinkernagel(1976), Zinkemagel and Klein (1977), Forman (1979), 6, Blanden et al. (1976), Zinkernagel(1976), Zinkernagel and Klein (1977), Byrne et al. (1984), 7, Whitmore and Gooding (1981), Pan et al. (1982), 8, Jennings et al. (1984), 9, Stukart et al. (1984), 10, Townsend et al. (1983), Biddison et al. (1980a,b, 1981, 1983), 11, Gaston et al. (1983), 12, Goulmy et al. (1982), 13, Chiang and Klein (1978), 14, Chiang and Klein (1978), Wettstein (1982), 15, Wettstein (1982), 16, Melief et al. (1980), Sherman (1982), Clark and Forman (1983). c Refers to domain bearing mutation.
MHC CLASS I MOLECULES RECOGNIZED BY CTL
165
(see Table YI). HLA-A2 subtypes from donors M7 and DR1 have a Gln + Arg substitution at residue 43, a change in the 147-157 peptide of the molecule, and could also have an alteration in their carbohydrate (Krangel et al., 1982). M7 is not recognized by either alloreactive, influenza-specific, EBV-specific, or H-Y HLA-A2-restricted CTL (Biddison et al., 1980a,b, 1982; Goulmy et al., 1982; Ware et al., 1983; Gaston et al., 1983; Molders et al., 1983). HLA-A3 subtype from donor E l has an alteration at residues 152 (Glu + Val) and 156 (Leu + Gln) (van Schravendijk et al., 1985). This HLAA 3 subtype can induce an alloreactive CTL response with HLA-A3 cells (Monos et al., 1984). Further, influenza virus-specific HLA-A3-restricted CTL do not interact with virus-infected cells from this -A3 subtype (Biddison et al., 1981). HLA-B7 subtype from donor C F has a change in residue 116 from Tyr to an unidentified amino acid (Taketani et al., 1984) and is not recognized by alloreactive anti-HLA-B7 CTL (Spits et al., 1982). The parallel between the H-2Kb mutants and HLA subtypes is striking, both with respect to the site of alteration and to changes in CTL-defined determinants. The HLA-A2 and -A3 subtypes have an alteration in the molecule which includes the same region as the H-2Kbm1mutant. Clearly, mutation in this region is not readily detected by serological reactivity whereas it is by the T cell system. Similarly, the HLA-B7 subtype from donor C F is similar to H-2Kb1n5,16in that there is a change in residue 116. VIII. CTL Recognition of Monoclonal Antibody-Selected Somatic Cell Class I Variants
The H-2 mutant molecules described above represent variants that have been selected by their capacity to elicit skin graft rejection in uiuo. Subtypes of HLA molecules defined by CTL also represent class I variants that arose in uiuo and are subject to selective pressures. Therefore, alternative approaches have been attempted to characterize CTL-defined epitopes using somatic cell variants selected by mAbs. Rajan (1977, 1980) used anti-H-2 antibodies to select stable heterozygous H-2 variant cell lines. Potter et al. (1983) used this technology to isolate somatic cell variants in (H-2d x H-2b)F, cell lines with mAb 34-2-12, specific for the a3 domain of H-2Dd. Two variants were described that retained reactivity with other H-2Dd mAb, but lacked reactivity with 34-2-12. These variants (selected with anti-a3 mAb) were poorly recognized by anti-H-2Ddreactive CTL. Although this finding appears to contradict studies with exonshuffled mctlecules whereby it was found that the a3 domain does not control polymorphic epitopes (see Section 11), a gross change in molecular weight was noted in one of the mutant molecules which was not due to an alteration in carbohydrate structure. It is possible that the ability of the molecule to interact with P2-M or other proteins could have been affected by this structural change.
166
JAMES FORMAN
Bluestone et al. (1984b) and Geier et al. (1986) used technology similar to that described by Potter et al. (1983) to derive heterozygous somatic cell H-2Kb variants. The variants were selected with anti-H-2Kb mAb specific for epitopes controlled by the a1 or a 2 domains. CTL clones generated in the strain combination H-2Kbm8anti-H-2Kb (H-2Kbm8is a mutant strain bearing mutant residues at amino acid positions 22-24, 30 (G. Pfaffenbach and S. G. Nathenson, personal communication) of the a1 domain of H-2Kb) recognized H-2Kb, but lost reactivity with H-2Kb variants selected for a loss of either al- or &defined serological epitopes. H-2Kbm10anti-H-2Kb (H-2Kbm10is a mutant strain bearing mutant residues at amino acid positions 165 and 173 of the a 2 domain of H-2Kb) -derived CTL clones lost reactivity against H-2Kb variant molecules selected for loss of either al- or a2-controlled serological epitopes. Cell line R8.313, which has a variant H-2Kb molecule with a single amino acid change at residue 82 from a leucine to phenylalanine (Nakagawa et al., 1986), lost reactivity with most anti-H-2Kb CTL clones generated in H-2Kbm strains (Bluestone et al., 1986). These data are consistent with a contribution of both a1 and a 2 domains in controlling epitopes recognized by monoclonal CTL. Clones that were inhibited from recognizing H-2Kb by mAbs could react with loss variants selected using the same mAb, further indicating that mAb-defined epitopes do not necessarily correlate with CTL epitopes. Bluestone et al. (1984a) selected variants of (H-2k x H-2d)F, cell lines with anti-H-2Ld mAbs directed against the a 2 or a3 domains. Of 36 antiH-2Ld CTL clones tested, only 3 did not react with the variants selected with anti-a2 mAbs, and all reacted with those selected with anti-& Similar to the findings with H-2Kb variants (Bluestone et al., 1984b), mAb defined epitopes did not correlate with CTL epitopes. Vohr et al. (1983) selected (H-2k x H-2d)F1 variants with two mAbs directed against H-2Kk. Approximately 40% of H-2Kk-TNP CTL clones generated in a limiting dilution assay lost reactivity with cell lines that were selected with either of the two antibodies. However, these serological epitopes probably do not define CTL specificities, since variant cell lines expressing H-2Kk molecules that lost reactivity with both mAbs retained reactivity with all H-2Kk-TNP CTL clones. H-2Kk-restricted CTL specific for fluorescein isothiocyanate (FITC), Newcastle disease virus (NDV), and influenza virus did not distinguish wildtype H-2Kk from the variants. Although the antibodies used to select the variants block lysis of influenza virus-specific H-2Kk-restricted CTL clones (Wraith et al., 1983), H-2Kk variants lacking these epitopes still were recognized by anti-influenza CTL. Vohr et al. (1983) suggest that since -40% of H-2Kk-restricted anti-TNP CTL clones lose reactivity with the variants, the number of restricting epitopes on an H-2 molecule is limited. While it may be argued that the variants represent gross alterations in H-2Kk, the fact that
MHC CLASS I MOLECULES RECOGNIZED BY CTL
167
all NDV, influenza, and FITC clones retain reactivity with the same variants mitigate against this. The authors suggest that since TNP covalently couples to H-2, which probably accounts for the specificity of TNP-H-2Kk CTL (Forman and Vitetta, 1978; Handa and Herrmann, 1985), these variants may not allow relevant covalent coupling of this hapten. Pious et al. (1982) selected HLA-A2 somatic cell variants from the T5-1 cell line with the BB7.2 anti-HLA-A2 mAb. Cell line 8.6.1 was described as having a change at amino acid 161 from Glu + Lys, while variants 8.18.1 and 8.21.1 have an alteration in the region from 98 to 108 (Taketani et al., 1983). These variants have been tested for their ability to react with both alloreactive and HLA-restricted CTL. Ware et al. (1983) and Brenner et al. (1985) found that these variants react with anti-HLA-A2 CTL and Gaston et al. (1984) noted that these variant molecules are recognized by HLA-A2-restricted EIW-specific CTL. Similar to studies with variant H-2 molecules selected by mAbs, deletion of serologically defined HLA epitopes does not lead to readily detectable changes in CTL determinants. IX. Concluding Remarks
MHC molecules may play at least three roles in determining the specificity of T lymphocyte recognition. For class I1 molecules, these roles are characterized by interactions with (1)the a / p chains of the T cell receptor, (2) antigenic peptides, and possibly (3) L3T4 or CD4 receptors from the mouse or human, respectively. Since CTL use the same aY/Pchains to recognize antigen in the context of class I molecules, it is worth considering whether similar interactions might govern class I molecules in determining the specificity of CTL. First, it is clear that CTL use the a and p chains of the T cell receptor heterodimer to recognize antigen (Dembic et al., 1986). These genes are uniquely rearranged in different CTL clones, and antibody directed against clonotypic determinants on these molecules can either block or activate CTL clones, depending on the assay in which they are used (Meuer et al., 1983; Reinherz et al., 1984). T helper cells and CTL clones have also been described that use the same V, and J, segments in their receptor genes (Rupp et al., 1985). The second issue concerns the interaction of MHC molecules with antigen. Recent evidence by Babbitt et al. (1985, 1986) indicates that a peptide of hen egg lysozyme interacts with Ia molecules, albeit weakly. Further, H-2k strains are able to generate an immune response to the peptide and show binding of the peptide to IAk, while H-2d is a nonresponder strain and the peptide does not demonstrate detectable binding to IAd. Similar findings have been reported by Buus et al. (1986)using ovalbumin peptides. Watts et
168
J A M E S FORMAN
al. (1986) found an interaction between an ovalbumin peptide and Ia using resonance energy transfer, although it was reported that this interaction required the presence of specific T cells. Antigen presentation involving class 11 molecules appears to use a pathway whereby molecules from the exterior are internalized, interact with the lysosomal compartment, and then are reexpressed on the cell membrane in a “processed” state (reviewed in Unanue, 1984). In contrast, inhibitors of lysosomal activity do not inhibit the ability of target cells to present antigens restricted to class I molecules to CTL (Morrison et al., 1986). Apparently, internally synthesized polypeptides which do not need to be expressed on the cell membrane as intact molecules are processed and transported to the cell membrane through an unknown mechanism where they can be recognized by CTL (Townsend et al., 198613). Examples include nonenvelope proteins of influenza, VSV, and Friend virus (Townsend et al., 1984; Holt et al., 1986; Yewdell et al., 1986). Since peptides of the influenza virus nucleoprotein and HLA-CW3 can be added to target cells to create antigens recognized by CTL (Townsend et al., 1986a; Maryanski et al., 1986b), it would appear likely that the same type of interaction described between peptides and Ia would also apply to peptides and class I molecules. This would be consistent with class I and I1 molecules having a similar structure and the fact that they interact with the same a/p chains of the T cell receptor. Therefore, the ability of class I molecules to act as restricting molecules for antigen-specific responses may be controlled in part by the ability of these molecules to interact with various antigenic components of internal antigens, e.g. virus and minor H antigens. The third aspect to consider is whether accessory molecules interact with portions of class I molecules. Such interactions have been postulated to account for the functional activity of T cell subsets (Forman, 1984). Lyt-2/CD8 and L3T4/CD4 are nonpolymorphic immunoglobulin-like molecules which could interact with nonpolymorphic portions of MHC molecules (Littman et al., 1985; Maddon et al., 1985; Sukhatme et al., 1985). Lyt-2/CD8 is expressed on cells that interact with class I molecules, while L3T4/CD4 cells interact with class I1 molecules (Swain, 1981; Meuer et al., 1982), and this observation has led to the suggestion that these molecules aid in the binding of T cells to specific antigen controlled by class I or I1 molecules, respectively (Biddison et al., 1982b; Marrack et al., 1983; Dialynas et al., 1981, 1983; Reinherz et d., 1983; Greenstein et al., 1984). L3T4+ T hybridoma cells react to low amounts of antigen more readily than L3T4cells and the presence of L3T4 enhances the ability of cells to become activated when specific antigen is limiting (Marrack et al., 1983).The ligand(s) for L3T4 is not known. Greenstein et al. (1984, 1985) noted that an antiH-2Dd reactive hybridoma that expressed L3T4 was inhibited by anti-L3T4 mAb recognizing H-2Dd on an Ia+ L cell only if the expression of H-2Dd was
MHC CLASS I MOLECULES RECOGNIZED BY CTL
169
relatively low. Watts et al. (1984) and Gay et al. (1986) showed that hybridomas sensitized against planar membranes or glass beads containing lipid, Ia, and peptide could be inhibited from responding by anti-UT4 mAb. Golding et al. (1985)showed that an L3T4+ anti-class I1 CTL line could be inhibited fxom recognizing a class II/class I hybrid molecule with anti-L3T4 mAb. The molecule lacked the a chain and the (32 domain of the (3 chain of Ia (see Section 11,E). Therefore, while evidence is suggestive that class 11 molecules may be a ligand for L3T4, other molecules may also serve as ligands. Alternatively, L3T4 could interact directly with the T cell receptor or play a role in binding after initial antigen recognition occurs. It should also be noted that mitogen-activated T cell responses directed against class I1 or class I negative cells can be inhibited by anti-L3T4 or Lyt-2 mAbs, respectively (Hunig, 1984; Malek et al., 1985; Tite et al., 1986). I have discussed data demonstrating that most of the polymorphic epitopes recognized by CTL are controlled by both the a1 and a 2 domains. These data were first obtained by studies using H-2 mutant strains where mutation in either the a1 or a 2 domain resulted in a loss of most of the alloreactive and restricting epitopes recognized by CTL (Section VII). These findings have been extended more recently using the techniques of exon shuffling and site-specific mutagenesis. Most class I molecules that bear a1 and a 2 domains derived from different class I genes or alleles lack most of their alloantigenic epitopes recognized by CTL generated against the native molecules (Section 11). A few exceptions exist, e.g., Dd/Ld/Ld and the H-2dm1 mutant, the latter of which represents a natural domain-shuffled molecule similar in structure to Dd/Ld/Ld (Section 11,B). Both of these molecules retain epitopes recognized by alloreactive and H-2-restricted CTL reactive with both H-2Dd and H-2Ld. While it is tempting to conclude that epitopes can be confined to both the a1 and a 2 domains, it is more likely that substituting the a2 domain of H-2Dd with H-2Ld allows much of the native epitope structure of the H-2Dd molecule to remain intact. This interpretation is supported by the finding that changing the a 2 domain of H-2Dd to Q7d completely alters the alloantigenic and restricting H-2Dd epitopes recognized by CTL. Further, the reciprocally shuffled molecule, LJd/Dd/Dd.is devoid of detectable H-2Ld and H-2Dd CTL-defined determinants. Studies with other molecules show that altering either a1 or a2 domains results in a loss of the vast majority of CTL-defined epitopes. In contrast to the role of the a1 and a 2 domains, a change in the a3 domain appears to have no detectable effect in CTL specificity. This has been noted even when the a3 domain is derived from a different species, although in one report it was found that such a molecule was recognized less efficiently. Although current data indicate that the a3 domain does not contribute to polymorphic determinants recognized by CTL, this does not preclude the
170
JAMES FORMAN
possibility that a conserved segment of the a3 domain either constitutes a portion of CTL-defined determinants or interacts with accessory molecules on the effector cell. A role for P2-M in controlling polymorphic determinants recognized by alloreactive T cells is also lacking; however, an allele of P2-M has been shown to be recognized by MHC-restricted CTL (Section V). Current evidence also indicates that carbohydrate does not play a major role in controlling epitopes recognized by either alloreactive or H-2-restricted CTL (Section IV). The H-2Kbm1 mutant strain as well as several serologically silent CTLdefined HLA subtypes have an alteration in amino acids involving residues 149 to 156. The substitutions frequently involved a change in a charged amino acid, and this type of change appears to be a characteristic of H-2 mutant and HLA subtype molecules (Monos et al., 1984). This region of the molecule has been postulated to comprise a turn of an a helical structure (Vega et al., 1984) and may define a part of a class I antigen important for T cell recognition. In the mouse, CTL generated against these molecules bearing mutant amino acids appear to cross-react on other H-2 molecules bearing the same “mutant” amino acids (Mann and Forman, 1987). However, most data would suggest that this cross-reaction is not directed specifically at these mutant amino acids, but rather at a conformation that they induce elsewhere in the molecule. It would also appear that many mutations have little effect on CTL epitopes. Thus, most mutant class I molecules which have been selected by mAb treatment of somatic cells show no alteration in their CTL epitopes (Sections VI and VIII), and this finding is consistent with a lack of correlation between serological and CTL-defined determinants. As the three-dimensional structure of class I molecules will soon be available, this puzzle may readily be solved (Bjorkman et al., 1985). At present, however, this finding suggests that CTL epitopes may be partially cryptic and revealed following interaction with a T cell. Since conjugate formation precedes antigen recognition (Spits et al., 1986), it is possible that as a result of conjugate formation alterations occur in the conformation of class I molecules revealing previously undetectable alloantigenic sites to the T cell receptor. Changes in the interaction of one mAb directed against a class I molecule have been reported to occur subsequent to the binding of a second mAb (Lemonnier et aZ., 1984; Diamond et al., 1984). This type of observation could be a model for exposure of cryptic CTL sites.
ACKNOWLEDGMENTS This work was supported by NIH Grants AI13111, CA41009, and AI11851. I wish to thank Ms. Betty Jo Washington for her excellent secretarial help and Drs. K. Fischer-Lindahl and B. Loveland for their careful reading of the manuscript together with their valuable suggestions. I
MHC CLASS I MOLECULES RECOGNIZED BY CTL
171
am grateful for the many discussions I have had with Dr. Stephen Clark and Mr. Don Mann regarding this work.
REFERENCES Abastado, J.-P., Plata, F., Morello, D., Daniel-Vedele, F., and Kourilsky, P. (1985).J. Zmmunol. 135, 3512. Achour, A., Begue, B., Gomard, E., Paul, P., Sayagh, B., van Pel, A., and Levy, J.-P. (1986). Eur. J. Ztnmunol. 16, 604. Agthoven, A. J . V., Lemonnier, F. A., Kourilsky, F. M., and Jordan, B. R. (1984). Mol. Zmmunol. 21, 175. Albert, F., Buferne, M., Boyer, C., and Schmitt-Verhulst, A. M. (1982). Zmmunogenetics 16, 533. Allen, H., Wraith, D., Pala, P., Askonas, B., and Flavell, R. A. (1984). Nature (London) 309, 279. Apt, A. S., Blandova, Z., Dishkant, I., Shumova, T., Vedernikov, A. A , , and Egorov, I. K. (1975). Inamunogenetics 1, 444. Arnold, B., Burgert, H.-G., Hamann, U., Hammerling, G., Kees, U., and Kvist, S. (1984). Cell 38, 79. Arnold, B., Horstmann, U., Kuon, W., Burgert, H.-G., Hammerling, 6. J., and Kvist, S. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 7030. Babbitt, B. P , Allen, P. M., Masueda, G., Haber, E., and Unanue, E. R. (1985). Nature (London) 317, 359. Babbitt, B. P . , Matsueda, G., Haber, E . , Unanue, E. R., and Allen, P. M . (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 4509. Bailey, D. W., and Kohn, H. I. (1965). Genet. Res. 6, 330. Barbosa, J. A. Kamarck, M. E . , Biro, P. A,, Weissman, S. M., and Ruddle, F. H. (1983). Proc. Natl. Acad. Sci. U . S . A . 79, 6327. Barbosa, J. A.. Mentzer, S. J., Minowada, G., Strominger, J. L., Burakoff, S. J., and Biro, P. A. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 7549. Barbosa, J . A., Mentzer, S. J . , Kamarck, M. E., Hart, J., Biro, P. A,, Strominger, J. L., and Burakoff, S. J . (1986).J. Zmmunol. 136, 3085. Bernabeu, C., Finlay, D., van de Rijn, M., Maziarz, R. T., Biro, P. A., Spits, H., de Vries, J., and Terhorst, C. P. (1983). J. Zmmunol. 131, 2032. Bernabeu, C., Maziarz, R., Spits, H., de Vries, J., Burakoff, S. J., and Terhorst, C. (1984a).J. Zmmunol. 133, 3188. Bernabeu, C. van de Rijn, M., Lerch, P. G., and Terhorst, C. P. (1984b). Nature (London) 308, 642. Biddison, W. E., Ward, F. E., Shearer, 6. M., and Shaw, S. (1980a).J. Zmmunol. 124, 548. Biddison, W. E., Krangel, M. S . , Strominger, J. L., Ward, F. E., Shearer, 6 . M., and Shaw, S. (1980b). Hum. Zmmunol. 3, 225. Biddison, W. E., Shearer, G. M., and Shaw, S. (1981). J. Zmmunol. 127, 2231. Biddison, W. E., Kostyu, D. D., Strominger, J. L., and Krangel, M. S. (1982a).J. Zmmunol. 129, 730. Biddison, W. E . , Rao, P. E., Talle, M. A , , Goldstein, G., and Shaw, S. (1982h). J. Erp. Med. 156, 1065. Bjorkman, P. J., Strominger, J . L., and Wiley, D. C. (1985). J. Mol. Biol. 186, 205. Black, P. L., Vitetta, E. S., Forman, J.. Kang, C.-Y., May, R. D., and Uhr, J. W. (1981). Eur. J. Zmmun.ol. 11, 48. Blanden, R. !I., Dunlop, M. B. C., Doherty, P. C., Kohn, H. I., and McKenzie, I. F. C. (1976). Zrnmunogenetics 3, 541.
172
JAMES FORMAN
Blanden, R. V., Mullbacher, A,, and Ashman, R. B. (1979).J . Exp. Med. 150, 166. Bluestone, J. A., Potter, T. A., Chattrjee-Hasrouni, S., and Rajan, T. V. (1984a).J . Zmmunol. 133, 1168. Bluestone, J. A., Palman, C., Foo, M., Geier, S. S., and Nathenson, S. G. (1984b).In “Regulation of the Immune System” (H. Cantor, L. Chess, and E. E. Sercarz, eds.), p. 89. Liss, New York. Bluestone, J. A., Foo, M., Allen, H., Segal, D., and Flavell, R. A. (1985).J . Exp. Med. 162, 268. Bluestone, J. A., Langlet, C., Geier, S. S . , Nathenson, S. G., Foo, M., and Schmitt-Verhulst, A.-M. (1986).J . Zmmunol. 137, 1244. Brenner, M. B., McLean, J., Yang, S. Y., van der Poel, J., Pious, D., and Strominger, J . L. (1985). J. Immunol. 135, 384. Breuning, M. H., Lucas, C. J., Breur, B. S., Engelsma, M. Y., de Lange, G. G., Dekker, A. J., Biddison, W. E., and Ivanyi, P. (1982). Hum. Zmmunol. 5, 259. Brickell, P. M., Latchman, D. S., Murphy, D., Willison, K., and Rigby, P. W. J. (1983). Nature (London) 306, 756. Brickell, P. M., Latchman, D. S., Murphy, D., Willison, K., and Rigby, P. W. J. (1985). Nature (London) 316, 162. Burnside, S. S., Hunt, P., Ozato, K., and Sears, D. W. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 5204. Buus, S . , Colon, S., Smith, C., Freed, J. H., Miles, C., and Grey, H. M. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 3968. Byrne, J. A., Ahmed, R., and Oldstone, M. B. A. (1984).J. Zmmunol. 133, 433. Carter, V. C., Schaffer, P. A., and Tevethia, S. S. (1981).J. Zmmunol. 126, 1655. Chiang, C. L., and Klein, J. (1978). Zmmunogenetics 6, 333. Ciavarra, R . , and Forman, J. (1982).J . E x p . Med. 156, 778. Clark, S. S . , and Forman, J. (1983). Transplant. Proc. 15, 2090. Clark, S. S., Geier, S., Nathenson, S. G., and Forman, J. (1985). Zmmunogenetics 22, 391. Clayberger, C., Holmes, N., Wang, P. L., Koller, T. D., Parham, P., and Krensky, A. M. (1985).J. Exp. Med. 162, 1709. Colombo, M. P., Pierotti, M. A., Ballinari, D., and Parmiani, G. (1983). Zmmunobiology 165, 186. Cowan, E. P., Coligan, J. E., and Biddison, W. E. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 4490. Darsley, M. J., Takahashi, H., Macchi, M. J . , Frelinger, J. A., Ozato, K., and Apella, E. (1987).J . Exp. Med. 165, 211, 1987. Davignon, J.-L., Guimezanes, A,, and Schmitt-Verhulst, A.-M. (1983).J . Zmmunol. 131, 1073. Dembic, Z., Haas, W., Weiss, S., McCubrey, J., Kiefer, H., von Boehmer, H., and Steinmetz, M. (1986). Nature (London) 320, 232. de Preval, C., and Mach, B. (1983). Zmmunogenetics 17, 133. Devlin, J. J., Lew, A. M., Flavell, R. A., and Coligan, J. E. (1985). EMBOJ. 4, 369. de Waal, L. P., Nathenson, S. G., and Melief, C. J. M. (1983a).J . E x p . Med. 158, 1720. de Waal, L. P., Fast, W. M., Melvold, R. W., and Melief, C. J. M. (198317).J . Zmmunol. 130, 1090. Dialynas, D. P., Loken, M. R., Sarmiento, M., and Fitch, F. W. (1981).J. Exp. Med. 153,595. Dialynas, D. P., Wilde, D. B., Marrack, P., Pierres, A., Wall, K. A., Havran, W., Otten, G., Loken, M. R., Pierres, M., Kappler, J., and Fitch, F. W. (1983). Zmmunol. Rew. 74, 29. Diamond, A. G., Butcher, G. W., and Howard, J. C. (1984).J . Zmmunol. 132, 1169. Doberstein, G., Garoff, H., and Warren, 6. (1979). Cell 17, 759. Doberty, P. C., Knowles, B. B., and Wettstein, P. J. (1984). Adw. Cancer Res. 42, 1. Duncan, W. R . , Wakeland, E., and Klein, J. (1979). Zmmunogenetics 9, 261.
MHC CLASS I MOLECULES RECOGNIZED BY CTL
173
Egorov, I. K. (1967). Genetika (Moskow) 9, 136. Egorov, I. K . , and Blandova, Z. K. (1968). Genetika (Moskow) 9, 136. Engelhard, V. H., Yannelli, J. R., Evans, 6. A,, Walk, S. F., and Holterman, M . J. (1985).J. Zmmunol. 134, 4218. Epstein, S. L , Ozato, K., and’ Sachs, D. H. (1980).J . Zrnrnunol. 125, 129. Evans, 6. A., Margulies, D. H., Camerini-Otero, R. D., Ozato, K., and Seidman, J . 6. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79, 1994. Evans, G. A., Margulies, D. H., Shykind, B., Seidman, J. G., and Ozato, K. (1982b).Nature (London) 300, 755. Ferrier, P., Layet, C., Caillol, D. H., Jordan, B. R., and Lemonnier, F. A. (1985).J. Zmmunol. 135, ,1281. Fisher, D. A.. Hunt, S. W., and Hood, L. (1985).J. E x p . Med. 162, 528. Fischer-Lindahl, K. F., and Lemke, H. (1979). E u r . J. Zmmunol. 9, 526. Fischer-Lindahl, K., Bocchieri, M., and Riblet, R. (1980).J . E x p . Med. 152, 1583. Fischer-Lindahl, K., Hausmann, B., and Chapman, V. M . (1983). Nature (London) 306, 383. Flaherty, L. (1981). The “Tla” Region Antigens. In “The Role of the Major Histocompatibility Complex in Immunobiology” (M. Dorf, ed.), p. 33. Garland, New York. Forman, J. (1!379).Zmmunogenetics 8, 487. Forman, J. (1!384). Znimunol. Reo. 81, 203. Forman, J . , and Klein, K. (1975a).J. Zmmunol. 115, 711.60, 23. Forman, J., and Klein, J. (197513). Immunogenetics 1, 469. Forman, J., and Klein, J. (1977). Zmmunogenetics 4, 183. Forman, J., and Vitetta, E. S. (1978).J. Zmmunol. 121, 1102. Forman, J., Trial, J.-A,, Tonkonogy, S., and Flaherty, L. (1982).J. E x p . Med. 155, 767. Forman, J., Goodenow, R. S., Hood, L., and Ciavarra, R. (1983). J . E x p . Med. 157, 1261. Gasser, D. L., Klein, K. A., Choi, E . , and Seidman, J. 6. (1985). Zmmunogenetics 22, 413. Gaston, J. S. H., Rickinson, A. B., and Epstein, M. A. (1983).J . Exp. Med. 158, 280. Gaston, J. S. H . , Wallace, L. E . , Rickinson, A. B., Epstein, M. A,, and Pious, D. (1984). Zmmunogenetics 19, 475. Gates, 111, F. T., Coligan, J. E., and Kindy, T. J. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 554. Gay, D., Coeshott, C., Golde, W., Kappler, J., and Marrack, P. J. (1986).J. Zmmunol. 136, 2026. Geier, S. S., :Zeff, R. A,, McGovern, D. M., Rajan, T. V., and Nathenson, S. 6. (1986). J. Zmmunol. 137, 1239. Golde, W. T., Kappler, J. W., Greenstein, J., Malissen, B., Hood, L., and Marrack, P. (1985). J . E x p . M’ed. 161, 635. Golding, H., McCluskey, J., Munitz, T. I., Germain, R. N . , Margulies, D. H., and Singer, A. (1985). Nature (London) 317, 425. Goldstein, S. A. N . , and Mescher, M. F. (1986).J. E x p . Med. 162, 1381. Gornard, E., Begue, B., Sodoyer, S., Maryanski, J. L., Jordan, B. R., and Levy, J . P. (1986). Nature (L,ondon)319, 153. Goodenow, R. S . , McMillan, M., Orn, A , , Nicolson, M., Davidson, N., Frelinger, J. A., and Hood, L. (1982a). Science 215, 677. Goodenow, R. S., Mc Millan, M., Nicolson, M., Sher, B. T., Eakle, K., Davidson, N., and Hood, L. (1982b). Nature (London) 300, 231. Goulmy, E., van Leeuwen, A,, Blokland, E., van Rood, J. J., and Biddison, W. E. (1982).J. E x p . Med. 155, 1567. Goulmy, E., van der Poel, J , , Giphart, M . , and van Rood, J. J. (1984). Zmmunogenetics 20, 13. Greenstein, J. L., Kappler, J., Marrack, P., and Burakoff, S. J. (1984).J. E x p . Med. 159, 1213. Greenstein, J. L., Malissen, B., and Burakoff, S. J. (1985).J . E x p . Med. 162, 369.
174
JAMES FORMAN
Greenstein, J. L., Foran, J. A,, Gorga, J. C., and Burakoff, S. J. (1986).J. Zmmunol. 136,2358. Guild, B. C., and Strominger, J. L. (1984a).J. Biol. Chem. 259, 13504. Guild, B. C., and Strominger, J. L. (1984b).J. B i d . Chem. 259, 9235. Guild, B. C., Erikson, R. L., and Strominger, J. L. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2894. Hammerling, G., Riisch, E., Tada, N., Kimura, S., and Hammerling, U. (1982). Proc. Natl. Acad. Sci. U .S.A. 79, 4737. Handa, K., and Hermann, S. (1985).J . Zmmunol. 135, 1564. Herman, A., Parham, P., Weissman, S. M., and Engelhard, V. H. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 5056. Holmes, N., and Parham, P. (1985). E M B O J . 4, 2849. Holt, C. A., Osorio, K., and Lilly, F. (1986).J. E x p . Med. 164, 211. Holterman, M. J . , and Engelhard, V. J. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 9699. Horai, S . , van der Poel, J. J., and Glulmy, E. (1982). Zmmunogenetics 16, 1982. Horstmann, U . , Arnold, B., and Hammerling, G. J. (1986). Eur. J . Zmmunol. 16, 863. Hiinig, T. (1984).J . Exp. Med. 159, 551. Hunt, P., and Sears, D. W. (1983).J. Zmmunol. 130, 1439. Hyman, R., and Stallings, V. (1977). Zmmunogenetics 4, 171. Jenkins, R. N., Aldrich, C. J., Lopez, L. A., and Rich, R. R. (1985). J. Zmmunol. 134, 3218. Jennings, S. R., Rice, P. L., Pan, S., Knowles, B. B., and Tevethia, S. S. (1984).J. Zmmunol. 132, 475. Jordan, B. R., Lemonnier, F. A., Caillol, D. H., and Trucy, J. (1983).Zmmunogenetics 18, 165. Kastner, D. L., Rich, R. R., and Chu, L. (1979).J. Zmmunol. 123, 1239. Kaufman, J. F., Krangel, M. S., and Strominger, J. L. (1984).J. Biol. Chem. 259, 7230. Kimball, E. S., and Coligan, J. E. (1983). Contemp. Top. Mol. Zmmunol. 9, 1. Klein, J. (1978). Ado. Zmmunol. 26, 55. Klein, J . , and Figueroa, F. (1981). Zmmunol. Reo. 60, 23. Kohn, H. I., and Melvold, R. W. (1974). Mutat. Res. 24, 163. Krangel, M. S . , Orr, H. T., and Strominger, J. L. (1979). Cell 18, 979. Krangel, M. S., Taketani, S., Biddison, W. E., Strong, D. M., and Strominger, J. L. (1982). Biochemistry 21, 6313. Krangel, M. S., Biddison, W. E., and Strominger, J. L. (1983).J. Zmmunol. 130, 1856. Kranz, D. M . , Tonegawa, S., and Eisen, H. N. (1984).Proc. Natl. Acad. Sci. U . S . A . 81,7922. Krensky, A. M., Sanchez-Madrid, F., Robbins, E., Nagy, J. A., Springer, T. A,, and Burakoff, S. J. (1983).J. Zmmunol. 131, 611. Krensky, A. M., Robbins, E., Springer, T. A., and Burakoff, S. J. (1984).J. Zmmunol. 132, 2180. Kress, M., Glaros, D., Khoury, G., and Jay, G. (1983). Nature (London) 306, 602. Kubota, K. (1984).J. Immunol. 133, 3203. Kvist, S . , Bregegere, F., Rask, L., Cami, B., Garoff, H., Daniel, F., Wiman, K., Larhammar, D., Abastado, J. P., Gachelin, G., Peterson, P. A , , Dobberstein, B., and Kourilsky, P. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 2772. Lalanne, J. L., Bregegere, F., Delarbre, C., Abastado, J. P., Gachelin, G., and Kourilsky, P. (1982). Nucleic Acids Res. 10, 1039. Lancet, D., Parham, P., and Srominger, J. L. (1979). Proc. Natl. Acad. Sci. U.S.A.76, 3844. Landolfi, N. F., Rich, R. R., and Cook, R. G. (1985).J. Zmmunol. 134, 423. Langhorne, J., and Fischer-Lindahl, K. F. (1982). Eur. J. Zmmunol. 12, 102. Layet, C., Le Bouteiller, P. P., N'Guyen, C., Mercier, P., Rosa, F., Fellous, M., Caillol, D. H., Jordan, B. R., and Lemonnier, F. A. (1984). Hum. Zmmunol. 11, 31. Le Bouteiller, P. L., Foa, C., Malissen, M., Golstein, P., Galindo, J. R., Mishal, Z., Caillol, D., and Lemonnier, F. A. (1982). E x p . Cell Res. 141, 473.
MHC CLASS I MOLECULES RECOGNIZED BY CTL
175
Le Bouteiller, P. P. L., Lemonnier, F. A., and Mishal, Z. (1983). Erp. Cell Res. 149, 587. Le Bouteiller, P. P., N'Guyen, C., Lemonnier, F. A , , Jordan, B. R., Caillol, D., and Mishal, Z. (1985). C!ytometry 6, 414. Lemke, H., Hammerling, G. J., and Hammerling, U. (1979). Zmmunol. Rew. 47, 175. Lemonnier, F. A., Malissen, M., Golstein, P., Le Bouteiller, P., Rebai, N., Damotte, M., Birnbaum, D., Caillol, D., Trucy, J., and Jordan, B. R. (1982). Zmmunogenetics 16, 355. Lernonnier, F. A., Dubreuil, P. C., Layet, C., Malissen, M., Bourel, D., Mercier, P., Jakobsen, B. K., Caillol, D. H., Svejgaard, A , , Kourilsky, F. M., and Jordan, B. R. (1983a). Zmmunoganetics 18, 65. Lemonnier, F. A., Le Bouteiller, P. P., Malissen, B., Golstein, P., Malissen, M., Mishal, Z., Caillol, D'. H., Jordan, B. R., and Kourilsky, F. M. (1983b).J . Zmmunol. 130, 1432. Lemonnier, F. A., LeBouteiller, P. P., Olive, D., Malissen, B., Mishal, Z., Layet, C., Dubreuil, J., Caillol, D. H., Kourilsky, F. M., and Jordan, B. R. (1984).J . Zmmunol. 132, 1176. Levy, R. B., and Shearer, G. M. (1982).J. Zmmunol. 129. Levy, R. B., Richardson, J. C., Margulies, D. H., Evans, 6. A., Seidman, J. G., and Ozato, K. (1983).J . Zmmunol. 130, 2514. Levy, R. B., Ozato, K., Richardson, J. C., and Bluestone, J. A. (1985).J. Zmmunol. 134, 677. Liberti, P. A,, Hackett, C. J., and Askonas, B. A. (1979). Eur. J . Immunol. 9, 751. Littrnan, D.R., Thomas, Y., Maddon, P. J., Chess, L., and Axel, R. (1985). Cell 40, 237. McCluskey, J., Germain, R. N., and Margulies, D. H. (1985). Cell 40, 247. McCluskey, J . , Boyd, L., Foo, M., Forman, J., Margulies, D. H., and Bluestone, J. A. (1986a). J. Zmmunol.. in press. McCluskey, J . , Bluestone, J. A., Coligan, J. E., Maloy, W. L., and Margulies, D. H. (1986b).J . Zmmunol. 136, 1472. MacDonald, H. R., and Cerottini, J.-C. (1979). Eur. J. Zmmunol. 9, 466. MacDonald, €1. R . , Thiernesse, N . , and Cerottini, J. C. (1981).J. Zmmunol. 126, 1671. McKenzie, I. F., Pang, T., and Blanden, R. V. (1977). Zmmunol. Rew. 35, 179. Maddon, P. J . , Littrnan, D. R., Godfrey, M., Maddon, D. E., Chess, L., a n d h e l , R. (1985). Cell 42, 93. Malek, T. R., Chan, C., Glirncher, L. H., Germain, R. N . , and Shevach, E. M. (1985). J . Zmmunol. 135, 1826. Malissen, M., Malissen, B., and Jordan, B. R. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 893. Mann, D. W.. Stroynowski, I., Hood, L., and Forman, J. (1987).J . Zmmunol. 138, 240. Mann, D. W.: and Forman, J. (1987). In preparation. Margulies, D. H., Evans, G. A., Ozato, K., Camerini-Otero, R. D., Tanaka, K., Appella, E . , and Seidman, J. G. (1983).J . Zmmunol. 130, 463. Marrack, P., and Kappler, J. (1986). Adu. Zmmunol. 38, 1. Marrack, P., Ehdres, R., Shimonkevitz, R., Zlotnik, A,, Dialynas, D., Fitch, F., and Kappler, J. (1983). 1. Erp. Med. 158, 1077. Maryanski, J. L., Moreta, A,, Jordan, B., De Plaen, E., Van Pel, A , , Boon, T., and Cerottini, J.-C. (1985). Eur. J . Zmmunol. 15, 1111. Maryanski, J. L., Accolla, R. S., and Jordan, B. (1986a).J . Zmmunol. 136, 4340. Maryanski, J. L., Pala, P., Corradin, G . , Jordan, B. R., and Cerottini, J.-C. (1986b). Nature (London) 324, 578. Matzinger, P., and Bevan, M. J. (1977). Cell. Zmmunol. 29, 1. Maziarz, R., Allen, H., Strominger, J. L., Flavell, R., Biro, P. A,, and Burakoff, S. J. (1985). Proc. Natl. Acad. Sci. U.S.A. 82, 6276. Melief, C.J. kl., de Waal, L. P., van der Meulen, M. Y., Melvold, R. W., and Kohn, H. I. (1980).J. E r p . Med. 151, 993. Mellor, A. L., Golden, L., Weiss, E . , Bullman, H., Hurst, J., Simpson, E . , James, R. F. L.,
176
JAMES FORMAN
Townsend, A. R. M., Taylor, P. M., Schmidt, W., Ferluga, J., Leben, L., Santamaria, M., Atfield, G., Festenstein, H., and Flavell, R. A. (1982). Nature (London) 298, 529. Mellor, A. L., Weiss, E. H., Ramachandran, K., and Flavell, R. A. (1983). Nature (London) 306, 792. Melvold, R. W., and Kohn, H. I. (1976). Zmmunogenetics 3, 241. Melvold, R. W . , Kohn, H . I., and Dunn, G. R. (1982). Zmmunogenetics 15, 177. Mentzer, S. J., Barbosa, J. A,, Strominger, J. L., Biro, P. A,, and Burakoff, S. J. (1986).J . Zmmunol. 137, 408. Meuer, S. C., Hussey, R. E., Hodgdon, J. C., Hercend, T., Schlossman, S. F., and Reinherz, E. L. (1982). Science 218, 471. Meuer, S. G., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1983).J. E x p . Med. 158, 988. Michaelson, J., Rothenberg, E., and Boyse, E. A. (1980). Zmmunogenetics 11, 93. Miyazaki, J.-I., Appella, E., Zhao, H., Forman, J., and Ozato, K. (1986).J . E x p . Med. 163, 856. Molders, H. M., Brewing, M. H., Ivanyi, P., and Ploegh, H. L. (1983). Hum. Immunol. 6, 111. Monos, D. S., Tekolf, W. A., Shaw, S., and Cooper, H. L. (1984).J . Zmmunol. 132, 1379. Morrison, L. A., Lukacher, A. E., Braciale, V. L., Van, D. P., and Braciale, T. J. (1986).J . E x p . Med. 163, 903. Murre, C., Choi, E., Weis, J., Seidman, J. G., Ozato, K., Liu, L., Burakoff, S. J., andReiss, C. S. (1984a).J . E x p . Med. 160, 167. Murre, C., Reiss, C. S., Bernabeu, C., Chen, L. B., Burakoff, S. J.. and Seidman, J. 6. (1984b). Nature (London) 307, 432. Nabholz, M., Young, H., Meo, T., Miggiano, V., Rijnbeek, A,, and Schreffler, D. F. (1975). Zmmunogenetics 1, 457. Nakagawa, M., Zeff, R. A., Geier, S. S., Bluestone, J. A., and Nathenson, S. 6. (1986). Zmmunogenetics 24, 381. Nakamuro, K., Tanigaki, N., and Pressman, D. (1973).Proc. Natl. Acad. Sci. U . S . A . 70, 2863. Naquet, P., Malissen, B., Bekkhoucha, F., Pont, S., Pierres, A,, Hood, L., and Pierres, M. (1985). Zmmunogenetics 22, 247. Nathenson, S. G . , and Cullen, S. E. (1974). Biochim. Biophys. Acta 344, 1. Nathenson, S. G., Geliebter, J., PfafTenbach, G. M., and Zeff, R. A. (1986). Annu. Rev. Zmmunol. 4, 471. Natori, T., Katagiri, M., Tanigaki, N., and Pressman, D. (1974). Transplantation 18, 550. Orn, A., Goodenow, R. S . , Hood, L., Brayton, P. R., Woodward, J. G., Harmon, R. C., and Frelinger, J. A. (1982). Nature (London) 297, 415. Ozato, K., Evans, G. A., Shykind, B., Margulies, D. H., and Seidman, J. 6. (1983a). Proc. Natl. Acad. Sci. U.S.A. 80, 2040. Ozato, K., Evans, G. A., Margulies, D. H., Seidman, J. G . , and Levy, R. B. (1983b). Transplant Proc. 15, 2074. Ozato, K., Takahashi, H., Appella, E., Sears, D. W., Murre, C . , Seidman, J. G., Kimura, S., and Tada, N. (1985).J . Zmmunol. 123, 1749. Pan, S.-H., Wettstein, P. J., and Knowles, B. R. (1982).J. Zmmunol. 128, 243. Pang, T., Andrew, M. E . , Melvold, R. W., and Blanden, R. V. (1977).Aust. J . E x p . B i d . Med. 55, 39. Parham, P., Clayberger, C., Zorn, S. L., Ludwig, D. S., Schoolnik, G. K., and Krensky, A. M. (1987). Nature (London), in press. Pease, L. R., Schulze, D. H., PfafTenbach, G. M., and Nathenson, S. 6. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 242. Pimlott, N. J. G., and Miller, R. 6 . (1986).J . Zmmunol. 136, 6.
MHC CLASS I MOLECULES RECOGNIZED BY CTL
177
Pimsler, M., ;md Forman, J. (1980). Zmmunogenetics 11, 111. Pious, D., Krztngel, M. S., Dixon, L. L., Parham, P., and Strominger, J. L. (1982).Proc. Natl. Acad. Sci. U.S.A. 79, 7832. Pircher, H., Hammerling, G., and Hengartner, H. (1984). Eur. J . Immunol. 14, 144. Ploegh, H. L., Cannon, L. E., and Strominger, J. L. (1979). Proc. Natl. Acad. Sci. U . S . A . 76, 2273. Ploegh, H. L., Orr, H. T., and Strominger, J. L. (1981).J . Immunol. 126, 270. Pober, J. S . , Guild, B. C., and Strominger, J. L. (1978).Proc. Natl. Acad. Sci. U . S . A . 75,6002. Potter, T. A,, Palladino, M. A,, Wilson, D. B., and Rajan, T. V. (1983).J . Exp. Med. 158, 1061. Rajan, T. V. (1977). Zmmunogenetics 4, 105. Rajan, T. V. (1980). Zmmunogenetics 10, 423. Rammensee, H.-G., Robinson, P. J., Crisanti, A , , and Bevan, M . J . (1986). Nature (London) 319, 502. Rask, L., Lindblom, J. B., and Peterson, P. A. (1974). Nature (London) 249, 833. Reinherz, E. L., Meuer, S. C., and Schlossman, S. F. (1983). Zmmunol. Today 4, 5. Reinherz, E. L,Acuto, O., Fabbi, M., Bensussan, A., Malanese, C., Royer, H. D., Metier, S. C., and Schlossman, S. F. (1984). Zmmunol. Reo. 81, 95. Reiss, C. S., EIvans, 6. A., Margulies, D. H., Seidman, J. G., andBurakoff, S. J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2709. Reiss, C. S., Greenstein, J. L., Crimmins, M. A. V., Liu, L. L. M., Rao, A., Maziaraz, R. T., Murre, C . , and Burakoff, S. J. (1986).J . Zmmunol. 136, 2191. Robinson, P. J . , Steinmetz, M., Moriwaki, K., and Fischer-Lindahl, K. (1984).Zmmunogenetics 20, 655. Rupp, F., Acha-Orbea, H., Hengartner, H., Zinkernagel, R., and Joho, R. (1985). Nature (London) 315, 425. Riisch, E., Knon, W., and Hammerling, G. J. (1983). Transplant. Proc. 15, 2093. Scheller, J.. Shimonkevitz, R., MacDonald, H. R., and Kvist, S. (1986). J . E x p . Med. 164, 1823. Scholtissek, C. (1975). Curr. Top. Microbiol. Immunol. 70, 101. Sherman, L. A. (1980).J. E x p . Med. 151, 1386. Sherman, L. A. (1982). Nature (London) 297, 511. Shimonkevitz, R . , Kappler, J., Marrack, P., and Grey, H. (1983).J. E x p . Med. 158, 303. Shimonkevitz, R . , Cerottini, 1.-C., and MacDonald, H. R. (1985). J . Zmmunol. 135, 1555. Shiroishi, T., Evans, 6. A., Appella, E., and Ozatto, K. (1984). Proc. Natl. Acad. Sci. U . S . A . 81, 7544. Silver, J., and Hood, L. (1974). Nature (London) 249, 764. Singer, D. S., Camerini-Otero, R. D., Satz, M. L., Osborne, B., Sachs, D., and Rudikoff, S. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 1403. Sood, A. K., I’ereira, D., and Weissman, S. M . (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 616. Spits, H., Breuning, M., Ivanyi, P., Russo, C., and de Vries, J. E. (1982).Zmmunogenetics 16, 503. Spits, H., Schooten, W. v., Keizer, H., Seventer, G. v., van de Rijn, M., Terhorst, C., and De Vries, J. E. (1986). Science 232, 403. Springer, T. A . , Davignon, D., Ho, M. K., Kurzinger, K., Martz, E., and Sanchez-Madrid, F. (1982). Znamunol. Rev. 68, 171. Staerz, U. D., Kanagawa, 0.. and Bevan, M. J. (1985). Nature (London) 314, 628. Steinmetz, M , Moore, K. W., Frelinger, J. G., Sher, B. T., Shen, F.-W., Boyse, E. A., and Hood, L. (1981a). Cell 25, 683. Steinmetz, M , Frelinger, J. G., Fisher, D., Hunkapiller, T., Pereira, D., Weissman, S. M. Uehara, JI., Nathenson, S., and Hood, L. (1981b). Cell 24, 125. Steinmetz, M , Winoto, A,, Minard, K., and Hood, L. (1982). Cell 28, 489.
178
J A M E S FORMAN
Stephan, D., Sun, H., Lindahl, K. F., Meyer, E., Hammerling, G., Hood, L., and Steinmetz, M. (1986).J. E x p . Med. 163, 1227. Stringfellow, M., Wraith, D. C., and Askonas, B. A. (1983). Zmmunogenetics 18, 177. Stroynowski, I., Orn, A., Goodenow, R. S . , McMillan, M., Forman, J., Brayton, P. R., Frelinger, J., and Hood, L. (1984). Zmmunogetics 20, 141. Stroynowski, I., Clark, S., Henderson, L. A,, Hood, L., McMillan, M., and Forman, J. (1985a).J. Zmmunol. 135, 2160. Stroynowski, I., Forman, J., Goodenow, R. S., Schiffer, S. G., McMillan, M., Sharrow, S. O., Sachs, D. H., and Hood, L. (1985b). J . Exp. Med. 161, 935. Stukart, M. J., Boes, J., and Melief, C. J. M. (1984).J. Zmmunol. 133, 24. Sukhatme, V. P., Sizer, K. C., Vollmer, A. C., Hunkapiller, T., and Parnes, J. R. (1985). Cell 40, 591. Sun, Y. H., Goodenow, R. S., and Hood, L. (1985).J. E r p . Med. 162, 1588. Swain, S. L. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 7101. Swaminathan, S., and Gooding, L. R. (1983). Eur. J. Zmmunol. 13, 335. Taketani, S . , Krangel, M. S . , Pious, D., and Strominger, J. L. (1983).J. Zmmunol. 131, 2935. Taketani, S., Krangel, M. S., Spits, H., de Vries, J., and Strominger, J. L. (1984).J. Zmmunol. 133, 816. Tite, J. P., Sloan, A., and Janeway, C. A. (1986).J. Mol. Cell. Zmmunol. 2, 179. Tkacz, J. S . , and Lampan, 0. J. (1975). Biochem. Biophys. Res. Commun. 65, 248. Tomonari, K . , Nobuhiko, T., Kimura, S . , Hammerling, U., and Weksler, M. E. (1982).Zmmunogenetics 15, 605. Townsend, A. R. M., Taylor, P. M., Mellor, A. L., and Askonas, B. A. (1983).Zmmunogenetics 17, 283. Townsend, A. R. M., McMichael, A. J,, Carter, N. P., Huddleston, J, A,, and Brownlee, G. G. (1984). Cell 39, 13. Townsend, A. R. M., Rothhard, J., Gotch, F. M., Bahadur, G., Wraith, D., and McMichael, A. J. (1986a). Cell 44, 959. Townsend, A. R. M., Bastin, J., Gould, K., and Brownless, G. G. (1986b). Nature (London) 324, 575. Unanue, E. (1984). Annu. Reo. Zmmunol. 2, 395. van de Rijn, M., Bernabeu, C., and Royer-Pokora, B. (1984). Science 226, 1083. van der Poel, J. J., Pooly, J., Goulmy, E., and van Rood, J. J. (1983a).Zmmunogenetics 17,599. van der Poel, J. J., Molders, H., Thompson, A., and Ploegh, H. L. (198313). Zmmunogenetics 17, 609. van Schravendijk, M. R., Biddison, W. E . , Berger, A. E., and Coligan, J. E. (1985). J . Zmmunol. 134, 410. Vega, M. A., Bragado, R., Ezquerra, A,, and Lopez, de Castro, J. A. (1984). Biochemistry 23, 823. Vitetta, E. S., Uhr, J. W., and Boyse, E. A. (1975).J. Zmmunol. 114, 252. Vohr, H.-W., Holtkamp, B., and Rajewsky, K. (1983). Eur. J. Zmmunol. 13, 846. Wallace, L. E., Kennedy, L. J., Landon, C., Bodmer, J. G., and Rickinson, A. B. (1986).Tissue Antigens 27, 298. Ware, C. F., Krangel, M. S . , Pious, D., Burakoff, S. J., and Strominger, J. L. (1983). J. Zmmunol. 131, 1312. Watson, A., and Bach, F. H. (1980). Znt. J. Cancer 26, 483. Watts, T. H., Brian, A. A,, Kapler, J. W., Marrack, P., and McConnell, H. M. (1984). Proc. Natl. Acad. Sci. U.S.A: 81, 7564. Watts, T. H., Gaub, H. E . , and McConnell, H. M. (1986). Nature (London) 320, 179. Weiss, E. H., Golden, L., Fahrner, K., Mellor, A. L., Devlin, J. J., Bullman, H., Tiddens, H., Bud, H., and Flavell, R. A. (1984). Nature (London) 310, 650.
MHC CLASS I MOLECULES RECOGNIZED BY CTL
179
Wettstein, P J. (1982).J. Zmmunol. 128, 2629. Weyand, C., Hammerling, G. J., and Goronzy, J. (1981a). Nature (London) 292, 627. Weyand, C., Goronzy, J., and Hammerling, G. J. (L98lb). J. E r p . Med. 54, 1717. Whitmore, A. C., and Gooding, L. R. (1981). J. Zrnmuno2. 127, 1207. Wraith, D. C:., Holtkamp, B., and Askonas, B. A. (1983). Eur. J. Zmmunol. 13, 762. Wu, T. T., and Kabat, E. A. (1970).J. E x p . Med. 132, 211. Yewdell, J. W.,Bennink, J. R., Mackett, M., LeFrancois, L., Lyles, D. S., and Moss, B. (1986). J E x p . Med. 163, 1529. Yokoyama, K . , and Nathenson, S. 6. (1983).J. Zmmunol. 130, 1419. Zinkernagel, R . M. (1976). J. Erp. Med. 143, 437. Zinkernagel, R. M., and Doherty, P. C. (1979). Ado. Zmmunol. 27, 51. Zinkernagel, R. M . , and Klein, J. (1977). Zmmunogenetics 4, 581. Zuniga, M. C., and Hood, L. (1986).J. Cell Biol. 102, 1. Zuniga, M. C:. , Malissen, B . , McMillan, M., Brayton, P. R . , Clark, S. S., Forman, J., and Hood, L. (1983). Cell 34, 535.
This Page Intentionally Left Blank
ADVANCES IN IMMUNOLOGY, VOL 41
Experimental Models for Understanding 6 lymphocyte Formation PAUL W. KINCADE !minunobiology Laboratory, Oklahomo Medical Reseorch Foundation, Oklahoma City, Oklahoma 73104
1. An Introductory Overview
While many points that can be made in an introduction are arguable and await further experimentation, the outlines are beginning to emerge of a fascinating process which can produce hundreds of billions of B lymphocytes daily and for the lifetime of an individual. This is achieved in a highly regulated manner, with requisite stem cells occasionally being drawn from a common, but very small pool. Increasingly less reversible changes commit them to a differentiation pathway as a series of divisions expand cell numbers. This is coordinated with simultaneous and parallel production of at least eight other lineages of blood cells which will coexist and be functionally interrelated when mature. The name B lymphocyte denotes the origin of these cells in birds (bursa of Fabricius) and mammals (bone marrow), and this review will emphasize events taking place in the latter. Progress in this field is being achieved by rapidly converging technologies and experimental approaches as the same general questions attract molecular, cellular, and developmental biologists. The objective of this chapter will be to highlight some recent developments, with a focus on issues that have been of interest to our laboratory. Studies involving normal and genetically defective experimental animals, monoclonal antibodies, inducible cell lines, soluble mediators, and innovations in long-term culture will receive particular emphasis. Early in gestation, before blood circulation begins, cells destined to become hemopoietic stem cells migrate from the neural crest area to extraembryonic tissues (yolk sac) as well as to the midline of the e m b q o . Both populations are potentially capable of making lymphocytes, but only those within the embryo which are incorporated into the developing liver and spleen norrnally do so (78, 272). Precursor cells sharing some of the distinctive properties of B lymphocyte lineage cells in adult marrow are found within embryonic tissues, and their differentiation potential can be demonstrated with a number of functional assays (review: 186). However, there are a number of notable differences in the initial emergence of lymphocytes in embryonic liver and the steady-state production of B cells within adult bone marrow (192, 197, 223, 326). In addition, early cells make preferential use of 181 Copyright Q 1987 by Academic Press, Inc. All rights of reproduction in any farm reserved.
182
PAUL W. KINCADE
certain families of variable region genes to make immunoglobulin heavy chains and assemble antibody molecules which will not be common in adult life (338, 436, 499, 502). At around the time of birth, hemopoiesis and the production of lymphocytes move from liver and spleen to the bone marrow (272, 457). The models and arguments used in this chapter are mainly based on the adult steady state and may not be completely applicable to the emergence of lymphocytes during ontogeny. While some multipotential stem cells can be found in the circulation, these may have limited proliferative potential, and it has not been proved that they are capable of becoming B lymphocytes (277). More likely, a relatively small pool of stem cells is retained within specialized “niches” within bone marrow and maintained by minimal self-renewal (442). They have finite proliferative potential which can be exhausted by serial transplantation, but aging is not appreciable during a normal individual’s life span (139). Such early stem cells can be considerably enriched experimentally by selective cytotoxic drugs, use of monoclonal antibodies, and electronic cell sorting strategies (17, 122, 151, 462). When injected into lethally irradiated recipient mice, they probably must mature briefly within bone marrow before being detectable as spleen colony-forming cells (CFU-s) (251, 455). While many of the latter are also multipotential, some are committed progenitors of erythroid, myeloid, and megakaryocytic lineages. Few, if any, CFU-s have the option of becoming B cells (187, 328; Fig. 1). The multilineage potential of stem cells was originally demonstrated in mice by use of cytogenetic markers in cloned populations(1). Premalignant expansion of genetically marked hemopoietic stem cells provided evidence for their existence in humans (101, 102). More recently, retrovirus vectors are being used to introduce selectable markers into hemopoietic stem cells (77, 169, 183, 488). This approach should pave the way for eventual gene replacement therapy (14) and provide a more practical means of studying stem cell population dynamics (235). Monoclonal antibodies can be used to identify normal cells which are committed to becoming B lymphocytes, and this has opened many investigative possbilities (43, 195). For example, we know the morphology, size, and frequency of B lineage precursors in late embryonic, neonatal, and adult life as well as in some genetically defective animals (see Section 111). It has also been possible to enrich for them and follow their short-term fate in a number of culture systems. Transformed counterparts of normal B cell precursors can undergo spontaneous or induced changes in culture (Section VII). These cell lines have been particularly useful in learning that functional immunoglobulin (Ig) genes are made by first reconfiguring heavy chain (D-J and then V-D-J) and then light chain (V-J) gene segments. There are several opportunities for errors as Ig genes are rearranged, and this contributes to
B LYMPHOCYTE FORMATION
183
multipotential stem cell (pre-CFU - s )
large small newly early B lineage pre - B pre-B formed cell B cell precursor cell
FIG.1. Lineage relationships between hemopoietic stem cells and B cell precursors. Areas of particular uncertainty are indicated by dotted lines.
potential antibody diversity. Some unknown fraction of lymphocytes may reach a “dead end” because a functional gene is not assembled. The fate of these defective cells is not known, but a mechanism has recently been found through which some may be functionally rescued (Section IV). The first DNA rearrangement event probably occurs soon after a distinctive surface marker [ Ly-5(220)]is expressed on large bone marrow lymphocytes which have basophilic cytoplasms containing numerous polyribosomes (45, 223). Detailed kinetic studies suggest that these cells probably initiate synthesis of the heavy chains of IgM while still large and that light chains will not be made for several more days (Section V). During this interval, replicative activity of the B cell precursors declines, and there is a marked reduction in cell size (Section 111). During an approximately 2day postreplicative period, additional cell surface molecules that are typical of mature I) cells are acquired, and functional receptors for certain soluble mediators can be demonstrated (Section IX). B lineage cells are not influenced by specific antigen or idiotypic networks before surface Ig appears (201, 202, 296). However, their production is affected by nonspecific environmental stimulation (110, 114, 311, 343). An interesting period of hypersensitivity exists just as receptors for antigen emerge on membranes of newly formed small B cells (274, 344, 350). Through an energy-dependent mechanism, cells exposed to small concentra-
184
PAUL W. KINCADE
tions of specific antigen are functionally inactivated, and this “clonal abortion” mechanism contributes to immunological tolerance (362, 437); i. e., many clones which have high-affinity receptors for abundant self-antigens are eliminated. It has been estimated that 5 X lo7 small lymphocytes are made daily in the bone marrow of a mouse, and many of them probably migrate via the circulation to the spleen (310; Section V). Depletion of mature B cells does not change the rate of B cell production (112), and we have little idea how a steady mind is maintained. Most of this chapter will deal with recent studies aimed at learning the nature and function of cells and molecules which regulate this process. Soluble mediators that influence several cell lineages have been found to augment the maturation of normal B cell precursors in culture (Section IX). These are now made by recombinant DNA technology and are available in highly purified form, which will permit molecular studies of receptor-factor interactions. Other factors, which seemed like better candidates for local regulation of bone marrow lymphocytes, were identified through studies of genetically defective animals and man (177, 225). However, the origins of these materials are not known, and thus far, only small quantities have been obtained in partially purified form. Lymphocytes in bone marrow are crowded among immature cells of all of the other blood lineages in what might seem a random array. However, they are probably closely associated with specialized cells of the bone marrow stroma, which provide a favorable microenvironment (oxygen, tension, nutrients, soluble mediators, inductive stimuli, etc.) for their proliferation and maturation (444). One or more adhesion molecules are probably involved in precursor cell-stromal cell recognition (Section VI), but this probably does not result in a permanent fusion. Maturing B cell precursors must gradually be carried along stromal cell surfaces to eventually be favorably located for discharge into venous sinuses. Essentially nothing is known about this process, but it presumably involves modulation in the expression of particular adhesion molecules and energy-dependent contractile elements. Replication of early precursors is essential to maintain population sizes within adult marrow, yet it has been difficult to find concentrated collections of lymphocytes. This suggests that a type of “conveyor belt” mechanism must be operating and that daughter calls from a mitosis are moved away from each other even as they mature. Long-term bone marrow culture approaches now make it possible to ask what types of microenvironmental cells interact with B lymphocyte precursors (74, 483). Different variations on the culture system permit long-term propagation of stem cells, B lineage precursors, or intermediate cell types in
€3 LYMPHOCYTE FORMATION
185
absolute dependence on adherent stromal cells (Section VI). Some possible sources of lab-to-lab and culture dish-to-culture dish variation are considered below, but it is already clear that this is a powerful methodology which can be manipulated in many ways. Our laboratory has characterized the large stronial cells to which lymphocytes are bound in culture and hopes that this will lead to an identification of critical adhesion molecules. Other have reported in workshops that it is possible to establish cloned lines which can replace some of the functions of adherent stromal cells in long-term cultures. These should yield large quantities of at least some of the critical regulators of precursor replication and maturation. Antagonists of this process are probably selected against by culture strategies, and we think that their characterization will also provide an exciting area of investigation. This is important not only for understanding how homeostasis is achieved, but also because of their potential in leukemia therapy. Genetic defects have long been informative about precursor-product and functional relationships between cells (Section VIII). We find it useful to return to the same models again and again because, as our methodology steadily improves, it is possible to appreciate ever more subtle connections and realize new opportunities to exploit particular mutations. A summary of our recent studies of genetic defects which influence B lymphocyte formation will comprise one section of this chapter. Finally, after considering individual cell types, mediators, and approaches, we hope to encourage a more catholic appreciation of relationships between different blood cell lineages. This will certainly be required in order to understand how the balance between massive production, maintenance, and death of cells of the humoral immune system is maintained. II. Organization of Lymphohemopoietic Tissues
The microanatomy of normal adult, embryonic, and regenerating hemopoietic tissues has been described in detail (64, 239, 240, 477-479, 482). However, until recently, it was only known that large collections of lymphocytes were not conspicuous. The precise relationships of differentiating B lymphocyte lineage cells to other cells in mammalian tissues have now become an object of investigation, and this occurs just as in uitro models are becoming better understood. A simplified interpretation of the anatomical studies will hopefully provide a basis for comparison with the culture work described below. Formal descriptions can be found in the many excellent reviews on this subject (240, 479, 478). The volume of hemopoietic tissue is determined by space not occupied by bone cortex, bone trabeculae, and specialized fat cells, while the latter may
186
PAUL W. KINCADE
be able to contract in situations of hemopoietic demand. Blood-forming areas of marrow are situated in cords between two types of circulation. The subendosteal area has a rich capillary plexus, and the core of the marrow space is traversed by interconnecting venous sinuses. The latter provides an exit route for newly formed blood cells and are comprised of a minimal basement membrane and overlapping endothelial cells which lack tight junctions. The abluminal surface of these sinuses is largely covered by fibroblastic “adventitial reticular” cells and occasional macrophages. Radiating out from these are processes of reticular cells and fibrils which form the basic spongy scaffolding of the hemopoietic cords. It is generally assumed that bone marrow “stromal” cells provide inductive signals as well as support to hemopoietic progenitors, and we use the word interchangeably with “microenvironment” or “microenvironmental elements.” Two particularly significant types of cellular associations have been noted by experimental hematologists. Maturing erythroid cells can be found clustered around a central macrophage in what has been termed an erythroblastic island (18). Immature granulocytes are physically associated with nonphagocytic adventitial reticular cells, which can express alkaline phosphatase (482).The same relationship has been appreciated between myeloid progenitors and adherent cells in long-term cultures, but in that circumstance, their well-spread appearance has led to such names as blanket cells (4, 5). This is particularly appropriate in the sense that progenitor cells can crawl underneath and proliferate. An analogy will be drawn between this phenomenon in lymphoid cultures and the behavior previously noted for tumor cells, termed pseudoemperipolesis (Section VI). Established bone marrow stromal cell lines can at least briefly support myelopoiesis, and some have the potential of turning into fat cells (preadipocytes) (207, 409, 505, 506). Large flattened cells to which lymphocytes are intimately associated in long-term bone marrow cultures have many similarities to adventitial reticular cells, and it is possible that functionally specialized microenvironmental elements would not be distinct morphologically (85; Section VI). This would be as dificult as resolving B and T cells on the basis of appearance alone. Alternatively, stromal cells capable of supporting myelopoiesis might under different circumstances be converted to components of a lymphoid microenvironment. Precursors of B cells have been identified in sections of fetal and neonatal liver (180). Before birth, they were found in scattered arrays or “star bursts.” As hemopoiesis is shifted from liver to bone marrow during the first days of life, small collections of hemopoietic cells are left. These were shown to be clonal and to contain foci of pre-B cells (368). It is not clear why such collections of lymphocytes can only be found in this special circumstance, but one could speculate that physical limitations are imposed by the rapidly
B LYMPHOCYTE FORMATION
187
developing nonhemopoietic areas of liver. In other tissues, such as bone marrow, the progeny of replicating precursors are permitted to move away from each other. However, within an individual bone, it is still possible to identify clonally related precursors (201). Finding multiple pre-B cells together has prompted speculation about their self-renewal at this stage of differentiation (Section V). Osmond and Batten (319)perfused murine femurs with radiolabeled antiIg antibodies and were subsequently able to identify B cells by autoradiography of marrow cross sections. Some mature, but presumably recently formed B cells were found adhering to venous sinuses. Other B cells were found scattered throughout the hemopoietic cords or in collections of less than four cells. The same laboratory has now utilized this approach to identify earlier cells with a monoclonal antibody to Ly-5(220) (D. G. Osmond, personal communication). Again, large lymphoid foci were not conspicuous, but there was a tendency for a gradient to be seen; i. e., there may have been more immature cells just beneath the bone capsule. Dissection of cells from that region of viable bone marrow and phenotyping also indicated that precursors may be enriched in the well-vascularized endosteal area (218). Earlier studies had revealed that active proliferation and immature myeloid cells are most prominent in this region, suggesting that as cells become closer to the stage of discharge into the bloodstream, they are moved toward the collecting sinuses in the center of the bone (241, 394). Parallel studies are being done with rat bone marrow where a new monoclonal antibody is particularly useful for identifying the pre-B cells (312; D. Opstelten, personal communication). One can expect that distinctive markers will soon be identified on subsets of bone marrow stromal cells, and this should make possible their localization in situ in relation to B lineage lymphocytes. The avian bursa of Fabricius should be used as a frame of reference of mammalian B lymphopoiesis because this represents a centralized and highly specialized site of lymphocyte formation (48, 120, 121). Some myelopoiesis can occur within the bursal mesenchyme during embryonic life, but it is essentially a lymphoid tissue (232). At a discrete stage of development, small numbers of stem cells colonize the bursa and move into epithelial buds. These proliferate and quickly differentiate to form small follicles of B cells, and it is notable that in birds, there is little delay between expression of heavy and light chain immunoglobulin genes (49).Much remains to be learned about the bursal microenvironment, but an interesting type of “secretory cell” has been identified in ultrastructural studies which might ultimately be associated with the production of critical soluble mediators (307). As we move toward an understanding of the nonlymphoid cells which induce andl regulate B lymphocyte formation, the question of their developmental origin recurs. There are clonal assays and other experimental ap-
188
PAUL W. KINCADE
proaches for studying the progenitors of fibroblasts, which share at least some properties with marrow stromal cells. These cells are noncycling, more radioresistant than hemopoietic progenitors, and not usually transplanted with marrow grafts. In addition, they can lack, the Philadelphia chromosome which marks all other lineages of cells in chronic myeloid leukemia (review: 490). However, in at least one situation, a clonal relationship was found between cultured adherent cells and premalignant hemopoietic cells (403). It is controversial whether adherent stromal cells are donor type in recipients of bone marrow transplants, and resolution of this question may await better definition of the cell types involved (70). 111. Resolution of B Cell Precursors
Terminology is certainly a problem for newcomers to immunology, and attempts will be made here to be consistent with a previous contribution to this series (186). Cells with potential for extensive self-renewal as well as differentiation into all eight blood cell lineages will be called multipotential hemopoietic stem cells. Cells bearing Ig chains which they made will be referred to as B cells. The immediate precursors of functional B cells in mammals lack surface Ig, but probably contain sufficient quantities of the p. heavy chain of IgM to be detected by immunofluorescence (351). Although the term was originally coined to describe a cell with slightly different characteristics (216), we refer to cells with this particular staining pattern as pre-
B cells (47). There are practical limitations to characterization of cells on the basis of a test which requires their fixation. Also, it is not clear how many cells successfully rearrange and express p. chain genes, but do not make light chains and, consequently, never become B lymphocytes. Many monoclonal antibodies and flow cytometers are now available, and it has been tempting to refer to lymphocytes which display one B lineage marker, but not sIgM, as pre-B cells. However, it is not clear that a single set of surface characteristics defines all murine pre-B cells (see below). The lesson has emerged from studies of tumor lines that cells which have rearranged one or more Ig genes and which display one or more B lineage markers may still retain the option of becoming something other than a B cell. Similarly, although we are learning a great deal about cells that are slightly earlier in the differentiation series, no set of characteristics unambiguously defines them. We refer to lymphocytes which may be irrevocably committed to becoming pre-B cells as early B lineage precursors. Such definitions may have only temporary usefulness as we gain more insight into molecular aspects of differentiation, and it is unfortunately the case that different investigators use the same words to describe different or incompletely overlapping cell populations. An
B LYMPHOCYTE FORMATION
189
attempt will be made in the following sections to summarize the many techniques. antibodies, and functional assays now being brought to bear on these questions.
A. AN OVERVIEW OF CELLSURFACE MARKERS It is easy to appreciate the value of monoclonal antibodies in resolving and manipulating lymphocyte populations. The surface markers most commonly used to study murine B lineage precursor cells are listed in Table I. No known antigen is exclusively expressed on C F + , sIg- pre-B cells. This may mean that pre-B cells represent a transient differentiation compartment rather than a discrete stage, and the extent to which these lymphocytes can self-renew under different circumstances is discussed below. The assumption is generally made that lymphocyte antigens must be receptors with unknown, but presumably important functions. Pre-B cells are unresponsive to specific foreign antigen because they have not yet acquired surface Ig. However, most membrane molecules required for regulation of pre-B cell activities must be shared with cells earlier and/or later in this pathway as well as with cells in other lineages. Indeed, our studies reveal that multiple soluble mediators can influence pre-B cells in culture, and thus far, none has been found which exclusively influences cells at this stage of differentiation (Section IX). Formal demonstration that a marker is on the precursors of B cells requires that it be used to deplete and/or enrich for them before testing their ability to become B cells in a functional assay. This has been shown for only the most widely utilized reagents (Section 111,H). However, a wealth of information has been obtained from studies of established pre-B cell lines and lymphocytes in long-term cultures (Table I). While no single antibody is stage specific, some fairly restricted markers, such as one detected by the recently described BP-1 antibody (50), can be used together with other reagents to define the immediate precursors of B cells. B. Ly-5 FAMILY OF GLYCOPROTEINS Macromolecular antigens encoded by the Ly-5 gene are very immunogenic in rats and many monoclonal antibodies are available to them. These have been extremely valuable as experimental tools, and extensive study of this antigen family has revealed a most interesting pattern of tissue-specific gene regulation. Trowbridge and colleagues (446) first demonstrated macromolecules that were uniquely expressed on T cells (T ZOO), and related antigens of slightly different sizes were subsequently found on all blood cells with the exception of erythrocytes (276, 309, 382, 399, 445). B lineage cells preferentially express the largest member of this family (220 kDa) and intermediate-sized (205 kDa) molecules have be.en demonstrated on myeloid
190
PAUL W. KINCADE
TABLE I SOMEMARKERS EXPRESSED BY MURINE B CELLPRECURSORS Demonstrated on some Markera
Pre-B tumors
LTBM cultures
c p + , sIgcells
Functional precursors
Lyb-2 Lyb-8 BP-1 GF-1.2
Yes Yes Yes Yes
Yes
Yes
Yes
?
p
Yes Negative
Yes Yes
? ? ?
Ly-5 (220)
Yes
Yes
Yes
Yes
ThB Ly-1 IL-2R Me1 14 TdT Lym-19
t
Yes t
Yes
Yes ?
2
r + Yes Yes
? Negative Yes
p
p ? ? Neg. Yes
b,c,138,192
? ?
Yes t
AA4.1 Mac-1 (CR-3) LFA-1 Ly-17 (FcR)
Yes Yes Yes Yes
Yes t t Negative
Yes
Yes
6C3 PNA MU75 M 1/69 (HSA) JllD LGPl00
Yes ? t Yes Yes Yes
? ? Yes Yes Yes t
43,44,59,69 198,223,289 57,59,92 1,59,158,228 c
t Negative
Yes -
57,59,195 158,430 50,492 17,262,263
? ? ? ?
2
Qa-2 la
Selected references
b,C
129,265,335,482 c,431 c,57,59,158, 497 b,262,263,330 59,155
? ?
p
b,c
Yes
Yes
b,',59,155 157,158,501 b,158,441,345 318,320,321 bsC,492 b,c , 492 262,263,492 153,228,230,492
?
?
Yes
Yes
?
? ? ? ?
Yes Yes
?
Markers are grouped according to probable restriction to B lineage cells, B and T lineages, or broader representation on hemopoietic cells. In many cases, only a subset of cells was clearly positive for the marker, and in others, the results are equivocal or ambiguous (t). All pre-B cells also express H-2 as well as the Ly-5 common leukocyte antigen, and some stain with a monoclonal antibody to the transferrin receptor (G. Lee et al., Ref. 509). b P. L. Witte (unpublished observations). 6. Lee (unpublished observations). (I
cells (383,445,448).Although the molecules are glycosylated, the major size differences have been attributed to varying lengths of the polypeptide chains (354, 473). The Ly-5 alloantigen system appears to identify virtually (380) all of this group of molecules in mice, regardless of their cellular origin, and we refer to them collectively as the Ly-5 family.
B LYMPHOCYTE FORMATION
191
A single structural gene has been cloned from rats and mice which corresponds to this family of macromolecules (395, 439). The nucleotide sequence predicts a transmembrane protein which has a very large intracellular domain. Northern blot analysis of RNA from B and T cells revealed differentsized messages, and recent studies suggest that RNA processing determines which type of molecule will be synthesized (374, 439). The largest message corresponds to the 220-kDa protein, whereas the shortest message results in the 200 kDa molecule preferentially expressed by T cells. Coffman and Weissman (44) found that one of their monoclonal rat antimouse B lineage antibodies (RA3-3A1)precipitated a 220-kDa antigen. Subsequent studies revealed that five of our independently developed antibodies, additional antibodies prepared by Coffman and Weissman, and a similar one to a human lymphocyte antigen all recognized molecules of this approximate size (42, 56, 383). We know from typing lymphocytes of different species that at least several epitopes are recognized by our antibodies. For example, only two of the five bind to human cells (222). Therefore, the N-terminal portion of the largest Ly-5(220) molecules must carry several unique antigenic determinants which have not yet been characterized in molecular terms. Essentially all peripheral B cells, all precursor cells that can quickly become B cells in culture, and a majority that do so within 10 days of transplantation express the Ly-5(220) molecule (43, 195). Most of the cells enriched from fetal liver and adult bone marrow with anti-Ly-5(220) antibodies have rearranged at least one allele of Ig heavy chain genes (45; Section IV). At the other end of the differentiation pathway, only some of the antibody-secreting cells in a primary immune response or plasma cell tumors are Ly-5(220)+ (42, 195). The WEHI-3 myelomonocytic leukemia cell line is recognized by our antibodies, but myeloid progenitor cells (CFU-c) and multipotential hemopoietic stem cells (CFU-s) are uniformly negative (195). Similarly, when we have used these reagents to enrich cells from fetal, neonatal, and adult hemopoietic tissues as well as from human bone marrow, they have been virtually all lymphocytes (222, 223). In one study, electronically sorted cells included some granulocytes, especially when the suspensions were obtained from neonatal liver (457). We know that autofluorescent nonlymphoid cells are problematic in that circumstance (unpublished observations). However, given the result with WEHI-3 cells, it remains possible that a very small number of myeloid lineage cells express the Ly-5(220) marker. Significant binding to platelets and macrophages as well as partial inhibition of CFU-s was reported with one, but not all antibodies to this antigen family (353). While th ymocytes and thymus-seeking cells in bone marrow are Ly-5(220) negative, the marker is detectable on a subset of peripheral T cells (195, 236). These are most conspicuous in lymph nodes of mice, where they
192
PAUL W. KINCADE
normally are also positive for Lyt-2 (195, 211, 383). Our two human reactive antibodies bind to a subset of T cells in peripheral blood, as do antibodies made against human and rat antigens of this size by others (55, 56, 222, 231, 412). The size of the molecule recognized by these antibodies on the normal peripheral T cells of mice has not been determined. However, this has been studied by several groups in the case of MRL/Lpr mice, where a lymphoproliferative disorder causes expansion of T cells with a very unusual phenotype (60, 88, 286, 383). In that circumstance, Lyt-2-, Thy-l+ lymphocytes make the 220 form of Ly-5 antigen. We know of only one situation where monoclonal antibodies to the Ly-5(220) molecule do not recognize at least small numbers of peripheral T cells. This is said to be the case for the RA3-6B2 antibody (42). If at least some T cells can transcribe the entire Ly-5 gene, special RNA processing or postsynthetic modification mechanisms would have to be envoked to explain how the 6B2 epitope is not expressed on those T cell Ly-5 molecules. For example, alternate patterns of glycosylation could yield different epitopes on these macromolecules (299). By analogy to other markers that were initially thought to be lineage restricted, the term B-220 was coined (44). However, until it is formally shown that any unique molecule of this size is exclusively made by B cells, we prefer the designation Ly-5(220). This is used throughout this chapter to refer to antigens recognized by our monoclonal 14.8 antibody and other antibodies with related specificities (195). While the Ly-5 family antigens are relatively abundant on leukocyte surfaces (445), their function remains unclear. Hemopoietic stem cells must synthesize the smallest member of this family of “common leukocyte antigens” and at some point early in the B lymphocyte differentiation lineage, a decision to change the pattern of transcription must be made. Early B precursors, pre-B cells, B cells, and a subset of peripheral T cells all make the largest form of these antigens. The Ly-5(220) molecule bears several unique epitopes that presumably correspond to functions most required by cells of the humoral immune system. It will be interesting indeed to learn what those functions might be. Antibodies to Ly-5 inhibit antibody responses to particular antigens in vitro, possibly by interfering with macrophage-B cell communication (495). It has been shown that antibodies to this family or antigens of similar size on human cells can depress T cell and NK cellmediated cytotoxicity (234, 298, 299, 380). Mitogen-driven proliferation of human T cells was augmented by trace amounts of antibody to a 220-kDa glycoprotein (231). It is important to stress that cultured or transformed B lineage lymphocytes do not always express the 220 form of Ly-5 antigens. This was first obvious in the case of established B lymphoma cell lines such as WEHI-231, which is Ly-5+, but lacks epitopes associated with Ly-5(220), (195, 228). All
B LYMPHOCYTE FORMATION
193
of a series of human pre-B leukemias were negative for Ly-5(220), and we have found that this is often the case with pre-B cells in long-term bone marrow cultures (198, 219, 231, 493; Section IV). Although from 16 days of gestation, all pre-B cells are Ly-5(220) , pre-B cells and functional precursors of B cells in earlier embryos can be negative (195, 223, 265, 457). Therefore, while expression of Ly-5(220) probably precedes rearrangement and expression of Ig genes in normal circumstances, this is not an obligatory sequence of events. Use of this or any other marker alone to detect B lineage cells could be problematic, and especially so when working with early embryos, cultured cells, or transformed cells. +
C. MARKERSMERITING MOREATTENTION
The BP-1 marker is interesting because it represents a new approach to preparing monoclonal antibodies as well as a unique cellular representation (50).Reasoning that many genetic polymorphisms might not be appreciated by inbred strains of mice, Cooper and colleagues immunized wild mice with a pre-B cell tumor. One of the resulting antibodies detects a 140-kDa disulfide-linked heterodimer which is expressed on pre-B and some B cells in bone marrow, but not on any lymphocytes in peripheral tissues. Studies in our laboratory have shown that the BP-1 antigen is expressed on long-term cultured lymphocytes and that its density changes with culture age (492). It has also been found that the density of this marker on an inducible pre-B tumor can be regulated up or down by coculture with particular factors or residence in uivo (342; G . Lee, unpublished observations). The receptor for interleukin 2 (T cell growth factor) was initially thought to be restricted to T lymphocytes, but it is also present on activated B cells (171, 248, 352, 313, 465). Large, cycling pre-B cells might be considered to be in an “activated” state, and it therefore seemed possible for them to be IL-2 receptor positive (509). Pre-B cell lines were indeed stained with a monoclonal anti-IL-2 receptor antibody, and it will be interesting to learn if these receptors mediate functionally significant signals on normal pre-B cells. Similarly, all dividing cells express receptors for transferrin (237), and this could also be demonstrated on at least some normal and transformed pre-B cells (G. Lee et al., unpublished observations). Evidence has recently been presented which suggests that a close functional relationship may exist between the receptor for B cell stimulating factor-1 (BSF-1) and the murine Lyb-2 alloantigen (497). This marker appears to be exclusively expressed on B lineage cells, including pre-B cells in adult mice, but it was difficult to demonstrate by cytotoxicity on functional B cell precurs,ors in fetal liver (195, 193; see also below). It also appears to be most easily detectable on pre-B cells which are small (509). A recently described antibody to the ci chain of the LFA-1 antigen has
194
PAUL W. KINCADE
similar effects to BSF-1 on mature B cells (281); i.e., it causes increased Ia antigen expression, augments proliferative responses, and influences isotype switching in culture. LFA-1 is a member of a very interesting antigen family that has been implicated in leukocyte adhesion and function (215, 414, 416). Patients who are unable to make the common p chain of this family do not express surface LFA-1, Mac-1, or p155, 95 and have recurrent infections (367, 415). Some studies suggest that the function of LFA-1 is to stabilize weak recognition bonds between cells of the immune system (407). At least small amounts of LFA-1 are detectable on murine pre-B cell lines (509). Recent evidence that pre-B cells can receive signals via the BSF-1 receptor will be discussed below. Previous studies suggested that newly formed B cells in bone marrow must acquire MHC class I1 (Ia) antigens at around the same time as surface IgM (195, 217). This question has recently been reexamined by direct staining of sIg- ,Ly-5(220) cells, and approximately one-third of them had detectable Ia. Moreover, it was possible to influence the density of both class I (H-2) and Ia antigens on established pre-B cell lines (349, 509; Section IX). An association has been found between a lymphocyte alloantigen and one type of receptor for the Fc portion of Ig molecules (157). Normal pre-B cells, as well as pre-B cell lines stain with monoclonal antibodies to this molecule, but long-term cultured lymphocytes usually do not (59, 155, 156, 158, 509; P. L. Witte, unpublished observations). Several types of Fc receptors have been described and at least three different ones can be simultaneously expressed by macrophages (104, 452, 458). The monoclonal Mac-1 antibody detects a C 3 receptor-related molecule on pre-B cell tumors (13, 155; see below). However, it is usually very low on long-term cultured lymphocytes, and there is no evidence that normal pre-B cells express it (492). The nuclear enzyme terminal deoxynucleotidyltransferase (TdT) has been frequently utilized as a marker for differentiating lymphocytes. There are indications that TdT might help to diversify antibody and T cell receptor diversity by introducing non-germ line-encoded nucleotides into the functionally configured genes (11, 68, 212, 400). Much of the recent information regarding TdT expression has come from studies of rats and humans, and in those circumstances, it appears that the marker is associated with early B lineage precursors in bone marrow (29, 146, 166, 312). Some murine lymphocytes in a long-term bone marrow culture system simultaneously expressed Ly-5(220) and TdT (129, 266). A detailed kinetic study of lymphocytes in murine marrow has just been completed and it included an assessment of TdT expression (335). Approximately one-half of the TdT+ cells were medium sized and also positive for Ly-5(220). Such cells would appear to be good candidates for the immediate precursors of large pre-B cells. Many other markers are listed in Table I which have not received detailed +
B LYMPHOCYTE FORMATION
195
study or which have relatively broad tissue representation. However, most have unknown functions and might be invoked to explain any of the soluble mediator responses, adhesion, and cellular recognition processes needed by B lineage precursors.
D. TUMORCELLLINESAND LINEAGEFIDELITY Spontaneous pre-B cell tumors are not common in mice, but they can easily be derived with Abelson and other transforming retroviruses (59, 156, 158, 470, 471). Established in uitro lines of these tumors have been extremely important for understanding the process of Ig gene rearrangement (7). Some of the differentiation events associated with normal B cells occur spontaneously or following stimulation of murine pre-B cells, and this is providing insight into receptor-factor interactions (Sections VII and IX). By typing large numbers of B lineage cell lines with the available monoclonal antibodies to lymphocyte antigens and correlating this with Ig gene rearrangements, a possible order of normal differentiation events can be inferred (59). Schemes have been similarly derived from characterizing large numbers of human tumor cells (9). However, it remains to be formally demonstrated that such sequences of marker acquisition occur with untransformed cells. A variety of experimental evidence has been used to conclude that progenitors of ‘rand B lymphocytes do not descend directly from cells which can become granulocytes and macrophages (186, 187). However, some studies with transformed cell lines have revealed an unexpected kinship between pre-B cells .and macrophages. A pre-B cell line mutagenized with 5-azacytidine gave rise to cells with macrophage characteristics, and there is precedent for a similar transition with human tumor cells (23, 98, 405, 426). Holmes and colleagues (12, 155, 158) recently provided definitive evidence that rearrangement of one Ig heavy chain allele and expression of some B lineage antigens does not preclude transformed cells from taking on macrophage properties. We need to know if this is an anomalous process which only occurs in tumors, and it will be interesting to learn if Ig gene rearrangements can be found in normal macrophages. However, together with the Ly-5(220) findings discussed above, these observations suggest that caution should be exercised in extrapolation from cell lines to the normal situation in uiuo. They also demonstrate that no single criterion can be used to assign cells to and position them within the B lymphocyte lineage. Ly-1 is a particularly interesting marker inasmuch as it is preferentially expressed by a B cell subset with a unique pattern of tissue localization and autoreactivity (144, 145). A series of transplantation experiments indicate that in adult mice, Ly-1+-bearing B cells may derive from a self-renewing pool of stem cells which are independent of bone marrow (136). It has been
196
PAUL W. K I N C A D E
difficult to demonstrate Ly-1 on significant numbers of lymphocytes in longterm bone marrow cultures or by direct staining of pre-B cells (492, 509). However, pre-B tumor lines often express this marker (137). Ly-1 B cells are most frequently among mature populations in the peritoneal cavity (145). However, sublines of a pre-B tumor which expressed increased amounts of Ly-1 after LPS stimulation did not have a notably different pattern of tissue localization when injected intravenously; i.e., many of them were still recovered in the bone marrow (342). It remains unclear if Ly-l+ tumors represent transformed equivalents of a particular lineage of functionally specialized B cells. E. TECHNICAL CONSIDERATIONS Most antigens are expressed in variable densities on murine B cell precursors and often at the limit of detectability. The sensitivity of fluorescence microscopes can be optimized with frequent bulb changes, proper alignment, and use of high numerical aperture optics. Even then, it is usually possible to detect more positive cells by flow cytometry. However, the difference in sensitivity could be more apparent than real. With the microscope, phase contrast is used to select “nucleated” cells for scoring, whereas low-angle forward and 90” light scatter is used in flow cytometry. It is not certain that identical populations of hemopoietic cells are evaluated with the two methods. We are very concerned about experimental protocols that employ hypotonic shock to prepare cell suspensions. Recovery of murine bone marrow nucleated cells is often 50% or less with many NH,Cl lysis procedures, and occasionally, the ratio of small-to-large cells appears to be skewed. Our best results have been obtained with Gey’s solution (280), and we use it only when absolutely necessary (such as with particular sorting protocols). It might appear that clear-cut staining results would always be obtained with recently cloned lymphoid tumor cells. This is the case with some antigens and cell lines. However, with some others, the cultures include both positive and negative cells. The density of some antigens is markedly influenced by growing the cells in serum-free medium or different concentrations of fetal calf serum ( G . Lee, unpublished observations). Variant subsets might grow out under the different conditions or this might also reflect the influence of hormones or other soluble mediators in serum. However, it points out the difficulty in obtaining consistent results from laboratory to laboratory, especially when dealing with normal, heterogeneous cell populations from different strains of mice. The relative abundance of the different sets of B lineage precursor cells is reasonably predictable in any given mouse strain maintained under the same conditions. However, it should be noted that the kinetics of production of
0 LYMPHOCYTE F ORMAT I O N
197
these cells is sensitive to nonspecific environmental stimulation (110, 114, 311, 343). Animals from a given conventional breeding room at times give higher or lower than average expected values. It is easy to imagine that some of the systemic, non-lineage-specific mediators such as IL-1, IFN gamma, and BSF-1 participate in such fluctuations (see below). The effect of corticosteroids on the bone marrow merits further investigation because preliminary studies in this laboratory suggest that stress can have major effects on precursor population sizes (P. L. Witte, unpublished observations). One should also be aware that in addition to polymorphism of alloantigens, amounts of markers such as Qa and ThB vary greatly on different pedigrees. In these two examples, the regulation has been mapped to an MHC-linked gene and the Ly-6 locus, respectively (92, 275). The aim of these comments is not to diminish confidence in published descriptions of antigens on murine B lineage cells or to suggest that our approaches are superior to those used by others. Rather, it is hoped that the reader will appreciate the difficulty in categorizing low densities of antigens on small subsets of hemopoietic cells and be aware of some causes for differences between studies. In our experience, it has been difficult to “nest” all defined populations into a single lineage scheme. This may reflect branching avenues of differentiation or the limits of our present technology.
F.
PHOSI’HATIDYLINOSITOL-LINKED
LYMPHOCYTE ANTIGENS
Most of the known lymphocyte antigens are probably anchored in the membrane by means of a stretch of hydrophobic amino acids embedded in the lipid bilayer. However, it has recently been appreciated that a novel form of membrane attachment is used by a subset of lymphocyte markers, and this may be important for understanding the function of these molecules. Tissue-specific control over this attachment mechanism could permit utilization of a single structural gene to produce a secreted protein, a releasable membrane protein, or a permanently anchored surface molecule. Investigation of this question has also provided a practical means of distinguishing and independently studying molecules which are similar in other respects. Alkaline phosphatase was the first molecule shown to be attached to cell membranes via phosphatidylinositol (PI) (245). A similar linkage mechanism was subsequently discovered for 5’-nucleotidase, the VSG antigen of trypanosomes, a form of acetylcholinesterase, decay-accelerating factor, and putative second messengers associated with the insulin receptor (62, 99, 115, 244, 376, 377). Where it has been examined in detail, the PI is attached via a short carbohydrate spacer and ethanolamine to the C-terminal amino acid of the protein (52, 99, 247; review: 243). The lymphocyte Thy-1 antigen usually utilizes this form of membrane attachment (246, 447). While the nucleotide sequence for this molecule can
198
PAUL W . KINCADE
encode hydrophobic C-terminal amino acid residues, none were found by direct sequencing of the purified protein (31, 391). A staphylococcus aureus-derived phospholipase is specific for the PI linkage, and this selectively removed Thy-1 from viable thymocytes or thymoma cells (242, 246). An extension of those studies revealed that ThB and some of the Qa-2-related antigens, but not TL, Qa-1, µglobulin, or other class I or class I1 molecules are released from lymphocytes by this enzyme (424). Spleen cell responses to LPS were only slightly affected, whereas those to concanavalin A (Con A), phytohemagglutinin (PHA), and pokeweed mitogen (PWM) were nearly abolished by this treatment. These results demonstrate the selectivity of the enzyme and indicate that some PI-linked molecules may be important in mature lymphocyte responses. Activated T cells have been shown to release intact Qa-2 molecules (406). It is not known if these were derived from PI-linked cell surface molecules, but we previously speculated that endogenous phospholipases might allow molecules that are attached in this way to be shed under appropriate circumstances (246). Some cell types, exemplified by L cells, seem not to be able to express surface Qa-2 when transfected with the unmodified gene, and this is the case even when RNA transcription is achieved (123, 427; G. Waneck, personal communication). It seems possible that such cells might make and secrete soluble forms of potential membrane molecules, as has been found for a mutant T cell line that secretes Thy-1 antigen (97). The antigens on some lymphocytes out of a population seem unaffected by treatment with PI-PLC (424; unpublished observations with G. Waneck). While there is precedent for the PI linkage being cryptic on some cells (62, 115, 363), it could also be that those lymphocytes synthesized molecules with conventional hydrophobic C-terminal residues. Further study may show that sensitivity of an antigen to PI-PLC reflects the maturation state or functional capability of the cell type on which it is expressed. It would then be important to understand the RNA processing and/or postbiosynthetic modification mechanisms responsible for such selective use of a PI anchor. As with the Ly-5 family of common leukocyte antigens discussed above, it provides considerable flexibility in use of a single structural gene. The functional significance of this for B lineage precursors and other lymphocytes remains unknown. However, preliminary results indicate that a PI-linked molecule may contribute to the adhesion of maturing pre-B cells to bone marrow stromal cells (Section VI,C).
G. OTHERDISTINCTIONS In addition to monoclonal antibodies and surface markers, there are other approaches to identifying and characterizing B lymphocyte lineage precursors. Forman and colleagues (103) described an IgH-linked alloantigen sys-
€3 LYMPHOCYTE FORMATION
199
tem (H-40) which is defined only by cytotoxic T cells and whose expression on B cells seemed to correlate with their initial acquisition of surface Ig. O’Toole and Bosma (323) discovered a broader range of specificities encoded by genes of this general region (Lm-1) and also recognized by T cells, but expressed on cells other than B cells. They noted several important implications for such alloreactivity in bone marrow transplantation as well as experimentation involving allotype congenic animals. Storkus and Dawson (425) found that B lineage cells become sensitive to recognition and lysis by natural killer (NK) cells at a particular stage of differentiation. Earlier studies showed that some bone marrow cells and subsets of thymocytes are potential NK-sensitive targets (133, 134, 304). Molecules recognized in these circumstances could play a role in normal B lineage differentiation and function. Subtractive RNA hybridization techniques have been extensively used to isolate genes which are expressed in a lineage-restricted manner (61). For example, this approach was used to clone the first T cell receptor gene and an X chromosome-linked gene family associated with murine immunodeficiency (46, 147). A series of genes have recently been identified which are utilized only in B lineage cells and which may correspond to different stages of maturation (148, 374b, 374c). On the basis of size and partial sequences, some of these have been assigned to previously known B cell products (Ig light and heavy chains, etc.). Others appear to encode molecules with unknown characteristics and functions (148). Studies now under way should reveal if these are inducible with any of the known lymphokines and B cell stimuli and if they are aberrantly expressed in mutant animals. While protooncogenes can be expressed in many embryonic and dividing cells, there is tissue specificity, and certain ones of these might be used as an additional criterion for resolving B lineage cells which otherwise appear to be similar (503). It is also interesting that Myc expression in transgenic mice causes preferential expansion of a population of B lineage precursors in prelymphomatous animals (2, 226).
H. RELATIONSHIPOF MARKERS TO FUNCTIONAL ASSAYS Until recently, the transition of uncommitted stem cells to functional B cells could only be studied by transplantation to irradiated or immunodeficient animals in a maturation process which required 6 to 8 weeks (186, 187, 195). Fortunately, a variety of systems are now available for demonstrating that murine cells expressing a particular set of surface markers can give rise to B cells. For example, the immediate progenitors of B cells mature quickly and have traditionally been characterized with a variety of short-term culture procedures. Furthermore, innovations in long-term culture methodology may not only facilitate a detailed study of the characteristics of B lineage precursors, but allow their emergence from multipotential stem cells
200
PAUL W. KINCADE
to be followed in culture (Section VI). Soluble mediators now known to augment differentiation steps in culture will be discussed later (Section IX). Representative studies of early and late B lineage precursors will be summarized here. A semisolid agar cloning system has been useful for monitoring the emergence of functional B cells in a variety of experimental circumstances. Cells bearing surface IgM proliferate in response to mitogens intrinsic to laboratory agar, and colonies are scored after 6-7 days of culture (190, 273). Inclusion of anti-p antibodies in the medium of semisolid agar cultures completely prevents colony formation (193).Therefore, cells must mature before being capable of replication in this system, and the maturation of sIg- cells after being dispersed in agar is inefficient. However, this transition can be augmented by hyperactive regulatory cells and soluble mediators (Sections VII1,D and IX). Partially immunodeficient CBA/N mice (Xid mutation) totally lack B cells that can respond in this assay (185). This is not because CBA/N B cells are mitogen unresponsive, but seems linked to their inability to be diluted in culture. For example, Xid B cells divided in crowded liquid cultures in response to agar-associated mitogens or LPS, but the blasts quickly died when plated in agar. Normal histocompatible hemopoietic cells almost completely reconstitute this functional deficiency, and this facilitated many studies of the characteristics of B lineage precursors (review: 186). Distinctions between early stem cells and late progenitors have been made on the basis of size, surface marker expression, and the time required to yield clonable B cells after transfer to CBA/N mice. For example, cytotoxic elimination of cell suspensions with the 19B5 monoclonal antibody greatly diminished precursors which could generate functional B cells within 6 weeks after engraftment into irradiated CBA/N recipients (195, 328). Particularly significant was the finding that multipotential cells (CFU-s) capable of spleen colony formation were unaffected by 19B5. This provided one of the first indications that early B lineage precursors are not closely related to myeloid stem cells (Fig. 1). Numbers of clonable B cells found per recipient spleen 10 days after transplantation are directly related to numbers of injected precursors. At this interval, -70% of the precursors could be depleted with monoclonal antibodies to Ly-5(220) (195). In addition, cells bearing this marker, but lacking surface Ig, were enriched by positive selection with the same antibodies and shown to give rise to functional B cells (223). Cells that spontaneously matured during 2-4 days of conventional culture to express sIgM and clone in agar were found to be Ly-5(220)+ (43, 195). Similar findings were recently made when cells with these characteristics were placed on preestablished layers of stromal adherent cells (289).
B LYMPHOCYTE FORMATION
20 1
Paige and colleagues developed conditions which permit relatively early cells to be cloned in agar and followed to an Ig-secreting stage (119, 324, 331b). Using an adherent underlayer of fetal liver cells, they found a linear relationship between numbers of precursors plated and immunoglobulinsecreting clones which developed. At 12 days of gestation, the precursors did not bear B lineage-associated antigens (330, 331). One day later, AA4.1 was acquired, and still later, Ly-5(220) was expressed. Finally, by 15-16 days of gestation, it became possible to enrich for the precursors with monoclonal antibody to Lyb-2. Precursors taken from late-stage embryos required the least amount of time in culture before giving rise to plaque-forming colonies. Osmond and Owen found that the lectin peanut agglutinin (PNA) preferentially bound to c p + , sIg- pre-B cells in adult murine bone marrow (320). The lectin was then used to enrich for large and small pre-B cells by electronic cell sorting (321). Both populations seemed to give rise to sIgM B cells and clonable B cells when held in liquid culture, and this was most convincing for the large fraction, which initially contained very few B cells. Many noa-lineage-specific receptors are acquired by differentiating B cell precursors (Table I). For example, the PI-linked Qa-2 family of antigens is demonstrable on pre-B cells and other hemopoietic cells in adult marrow, and developmental age rather than position in the lineage determines their expression i(192).Yang et al. (501) used a double-labeling procedure to associate the expression of sIg, receptors for the F c portion of Ig, and complement receptors with the postmitotic age of bone marrow lymphocytes. The results would be consistent with sequential acquisition of sIg, Fc, and C’ receptors by some cells within bone marrow, and when newly formed, radiolabeled, marrow lymphocytes were transferred to recipient mice, expression of these markers increased. However, it was pointed out that subsets of maturing B cells could acquire Fc and C’ receptors at different times. This may explain in part why numbers of B lineage cells bearing the Fc receptor detected by monoclonal 2.4G2 are particularly variable (509). A splenic fragment cloning assay developed by Klinman has been invaluable for studying some aspects of B lymphocyte lineage differentiation (202). Cells to be assayed are allowed to home to spleens of carrier-primed, irradiated recipient mice. The spleens are removed, diced, and cultured with antigen, and antibody secretion is subsequently measured. It is clear that sIg - precursors are capable of quickly maturing and responding to antigen in this system, and this has formed the basis for several studies of “prereceptor” B cell characteristics (201). Small marrow lymphocytes which are phenotypically Ly-5(220) , c p , sIg - comprise a majority of the cells which give rise to functional B cells under these conditions (296). Although other markers and physical characteristics of prereceptor B cells h a w not been +
+
+
202
PAUL W. KINCADE
described in detail, several interesting aspects of their potential antibody specificities have been reported. As a population, they already reflect certain biases found among the repertoire of mature splenic B cells, and it seems clear that they have not been specifically influenced by idiotypic networks or environmental antigens. I. RELATIONSHIP OF MARKERS TO CELLSIZE CHANGES Transformed cell lines consist of very large blast cells which are continuously replicating. This is true regardless of whether the lymphomas are classified as representative of mature B cells, pre-B cells, or earlier B lineage lymphocytes. However, small size has been repeatedly stressed as a distinguishing feature of the immediate precursors of B cells, newly formed B cells, and unstimulated B cells in peripheral lymphoid tissues. Kinetic studies of bone marrow lymphocyte populations are consistent with large pre-B cells giving rise to small pre-B cells (Section V). The significance of this change is not clear because lymphocytes in long-term bone marrow cultures are relatively homogeneous and small to medium in size despite the fact that they are cycling (493). Low fonvard-angle light scatter as determined by flow cytometry is being frequently used as a convenient way to discriminate small and large lymphocyte populations, but there is no agreed upon criterion for distinguishing the two. The upper size threshold for small cells in smears has traditionally been 8 p, and for cytocentrifuged lymphocytes 10 p has been used (221, 310, 316). These are experimentally convenient criteria for monitoring what must be a gradual shrinking process. It has been proposed that the ThB antigen is preferentially associated with small bone marrow lymphocytes (42). However, recent studies in this laboratory suggest that while this is true for unfractionated (B6 x DBA)F, bone marrow, most of the ThB-bearing cells are actually small B cells, and the marker is certainly expressed by some large sIg- lymphocytes (G. Lee, manuscript in preparation). The magnitude of ThB expression is governed by genes linked to the Ly-6locus, and the ThB molecule is linked to the lymphocyte surface via PI (92, 424). It is possible that differences in strains and/or husbandry of animals will provide explanations for discrepancies between results obtained by different laboratories. It is clear that cells destined to become B lymphocytes already express a variety of distinct surface markers. Although a few of the available monoclonal antibodies have been used in conjunction with functional assays, no definitive picture of the order of acquisition or significance of these putative receptors has emerged. Particularly useful would be any correlations that could be found between cell surface marker expression and changes in proliferative activity, rearrangement of immunoglobulin genes, association with
B LYMPHOCYTE FORMATION
203
particular marrow stromal cells, and immigration from this central lymphoid tissue. IV. Rearrangement and Utilization of Immunoglobulin Genes
Antibody molecules represent the most distinctive and important product of the B lymphocyte lineage. Their synthesis and display in membranebound and secreted forms provide a most complex and interesting system of coordinate gene expression during development. The NH,-terminus of an Ig heavy chain is encoded by just one of at least 100 variable region gene segments which, in the mouse, are grouped into seven families on chromosome 12. Selection of this gene segment is done during the second of a twostep rearrangement process. The components of an Ig gene are separated (in germ line configuration) in nonlymphoid cells and early embryonic tissue. An initial event involves alignment of 1 of 4 J segments with 1 of 12 D region segments and deletion of the intervening DNA. This is done via nonamer and heptamer recognition sequences which flank the J and D segments and which are separated from each other by 12 or 23 base pairs. A second and presumably similar process involves deletion of the DNA between a V region gene and the D-J complex. Use of this fully assembled gene to synthesize heavy chains then precedes rearrangement of gene segments for K light chains on chromosome 2. If this results in a functional Ig molecule, the process is usually terminated. However, if both alleles of K light chains are abortively rearranged, rearrangement of h light chains is attempted (chromosome 22). All of these events have been described in detail (7, 26, 90, 159). Only recent developments and particular aspects that are relevant to B cell precursors will be considered here. The alignment of V and D segments during rearrangement commonly does not result in a correct reading frame (91, 132). In addition, extra nucleotides (non-germ line elements) can be randomly inserted at this junction through the action of TdT (11, 68, 212). Since the D segment and V-D junction encode portions of the third hypervariable region of antibody molecules, errors contribute to the diversity of antibody-combining sites (90, 91). The process requires that as many as half of all B cell precursors generate nonfunctional heavy chain genes on both chromosomes (7). Until recently, it was assumed that all such precursors would represent abortive or “dead end” clones (132). However, a mechanism has been found through which some of these may be rescued. Studies with two transformed cell lines indicate that V-D-J rearrangement, regardless of whether it yields a functional Ig gene, can be followed by a second rearrangement which essentially exchanges a second upstream V region for the first (6, 200, 360). This seems to be possible because a heptamer, which is part of most V region genes, can
204
PAUL W. KINCADE
be used as a recognition sequence for further rearrangement. The net result is that most of an initially selected V gene is replaced by a second choice. Particular antibody specificities appear in a predictable sequence during ontogeny, and recent studies indicate that this is mediated by a nonrandom selection among the seven families of variable region genes (338, 499). Some aspects of the orientation and chromosomal location of these gene families are controversial, and the order of expression may not be identical with all strains of laboratory mice (P. Tucker and C. J. Paige, personal communications, 495b, 495c). However, families situated nearest the constant region gene segments tend to be used first (7, 388, 499). In contrast, B cells in mature animals appear to utilize all of the gene families approximately the same. The model proposed above for V, replacement would only permit progression in an upstream (5’)direction, and it was suggested that this may contribute to the order of gene utilization (200). A common rearrangement mechanism may be involved in configuring light and heavy Ig chains as well as antigen receptors on T cells (500). The signals that initiate this process are not known, but it has been hypothesized that access of a recombinase enzyme to Ig gene segments controls the order and rate of rearrangement events (7). This is based in part on findings that Ig genes are transcriptionally active prior to assembly (184, 359, 467). Short transcripts of one of the variable region gene families have been detectable in fetal liver, and actual synthesis and release of “truncated” p chains has been demonstrated in adult animals (258, 498). The latter are made without variable regions by bone marrow cells and even by plasma cells which synthesize complete Ig molecules. Transcripts or products of incompletely assembled Ig genes may or may not have any functional significance, but it is clear that chromatin is “open” at a time when the rearrangement steps are proceeding. Most of the available information on Ig gene rearrangements has come from studies of tumor cell lines. Coffman and Weissman (45) addressed this question with sorted bone marrow cells which lacked surface Ig, but which were positive for Ly-5(220). All of these cells had rearranged both alleles at least to the D-J stage. Many small cells, but no large cells, had begun rearrangement of light chain genes. In a similar study with Dr. S. Akira, we found that Ly-5(220) cells in fetal liver were already undergoing D-J rearrangement by 16 days of gestation (unpublished observations). However, these and phenotypically similar cells isolated from adult marrow still retained some germ line IgH genes. Essentially all of our cells had a lymphoid morphology, and it may be that the different monoclonal antibody and selection techniques that we used permitted isolation of some earlier cells in the B lineage series which were not included in the Coffman and Weissman analysis. Early Ly-5(220) cells, which possibly express TdT (see Section +
+
B LYMPHOCYTE FORMATION
205
111), could be comparable to B lineage leukemias which have only one rearranged heavy chain allele (209). It is also possible that D to J joining has not occurred on either chromosome in some of these cells, and we need better ways to isolate them from normal tissues. In contrast to the situation with light chain genes, both Ig, alleles are rearranged in B cells in peripheral lymphoid tissues (161, 303). It would be advantageous to have additional information about the Ig gene configuration of normal bone marrow cells resolved on the basis of various markers. Long-term bone marrow cultures can be manipulated such that precursors with Ig genes in germ line configuration undergo most or all of these rearrangement steps (Section VI), and it will be interesting to learn if this can be controlled with particular stimuli. Other issues include the necessity for rearrangement of both heavy chain alleles before light chain genes and the frequency of abortive clones among unselected populations. Bias in the use of 1’ region gene families might also be seen in cells with an “early” rather than a “late” pre-B phenotype. Expression of a fully assembled Ig gene is thought to be regulated by cisand trans-acting nuclear factors and control sequences within the gene (268, 392, 422, 476). The latter include at least the promoter and enhancer regions and possibly an additional “facilitator” segment (117, 128, 130). Initiation of transcription may depend on factors already present in the cell, but whose activity is blocked by labile repressor molecules (469), e.g., cyclohexamideinduced transcription of a human heavy chain which had been transfected into mouse fibroblasts (164). Brief interruption of protein synthesis also induced light chain gene expression in a pre-B cell line (393, 469). Nuclear factors from heavy chain-synthesizing pre-B and B cells have also been used to signal transcription after microinjection (250). High-resolution DNA “footprinting” techniques, retardation of mobility of DNA fragments, and functional assays are being used to obtain detailed information about regulatory proteins and regions of the Ig genes with which they interact (10, 96). V. Population Dynamics
Depending on the species and age of an animal, bone marrow contains small numbers of T lymphocytes, some long-lived B lymphocytes of peripheral origin (presumably memory cells), and some cells involved in active immune responses (204, 205, 296, 369). However, the vast majority of lymphocytes are made locally and within a %day time span (315). Much of our understanding of the proliferative events which generate B lymphocytes within this #centrallymphoid tissue has derived from radioautographic studies of tritiated thymidine incorporation into B lineage precursors. A number of generalizations can be drawn from those reports and some will be briefly
206
PAUL W . KINCADE
summarized here. However, it is important to realize that the arguments made are essentially statistical and, while one can say that the data are compatible with particular hypotheses, it is not possible to conclude that they are proved. For example, it has been frequently stated that lymphocytes leave the pool of small marrow cells randomly rather than as a function of postmitotic age (221, 317, 369). This would make lymphocytes different from other blood cells, which normally exit the marrow at maturity (348). However, it is difficult to distinguish between those cells physically leaving via the circulation and those dying locally. As marrow cells are resolved into smaller and smaller subpopulations (Section 111), it becomes ever more difficult to determine the rates at which each compartment self-renews, is replenished with less differentiated cells, and relinquishes cells to the next compartment in sequence. The analyses become almost hopelessly complex if the issue of separate lineages of B cells is raised (136). For purposes of this brief discussion, it will be assumed that there is but one linear differentiation pathway in marrow, with a series of obligatory steps through which all B cells are made. A delay of -12 hours between 3H-labeled TdR infusion and the appearance of labeled small (37T
70-75 kDa (reduced) 60-64 kDa (nonreduced) Ca2+, temperature >30"C
loo A
160 A 0.4-6 nSb Voltage-resistant, open state favored, slow channel kinetics Permeable to monovalent and divalent ions, Lucifer Yellow, sucrose, and glucosamine (+) Antiserum to reduced C9 (-) Antiserum to nonreduced C9 Hemolytic; cytolytic to a variety of tumor cell lines; cytolysis requires Ca2+
0.2-4 nS Voltage-resistant, open state favored, slow channel kinetics
Functional size
Permeable to monovalent and divalent ions, Lucifer Yellow, and sucrose
Antigenic cross-reactivity
(+) Antiserum to reduced perforin (-) Antiserum to nonreduced perforin Hemolytic activity requires activation of C5b-C8; cytolytic to a variety of tumor cell lines
Cytotoxicity
Lymphocyte PFP
Lymphocyte PFP/perforin refers to material purified from mouse cytotoxic T cell lines and NK-like lymphocytes, whereas C9 refers to human material. This range of channel sizes includes data obtained on channels formed by lymphocyte PFP/perforin added directly to the aqueous phase of planar bilayers at room temperature and also by polyperforin polymerized in lipid vesicles at 37°C prior to incorporation into planar bilayers. The range of channel sizes for C9 shown here pertains only to poly C9 complexes previously formed and transferred to planar bilayers. From Young et al. (19861). a
b
LYMPHOCYTE AND COMPLEMENT-MEDIATED CYTOTOXICITY
309
reduced and alkylated (Table I). These results indicate that the cross-reactivity observed between the two species is restricted to cysteine-rich domains which are normally masked by disulfide bridges and become antigenically exposed only upon chemical reduction. Recent studies by Tschopp and his colleagues (1986b) showed that antibodies raised against a synthetic peptide prepared to mimic the region of homology (in the cysteine-rich area, beresidues 101-111: Asp-Asn-Asp-Cys-Gly-Asp-Phe-Ser-Asp-Glu-Asp) tween C9 and LDL receptor also react against mouse PFP. These authors also showed that the synthetic peptides inhibit the hemolytic activity of granule PFP/perforin. The immunological cross-reactivity is not limited to C9 and PFP. Recent studies have shown that the lymphocyte PFP is also immunologically related to C5b-6, C7, and C8 (Young et al., 19863’). The antigenic epitope(s) shared by these proteins is also restricted to the cysteine-rich domain(s). In parallel studies, antibodies raised against the synthetic peptide mentioned above also react against the other components of the MAC and the polymerized complexes of C5b-9 (Tschopp and Mollnes, 1986; Tschopp et al., 1986b). The cysteine-rich domain that appears to be conserved in these molecules may be exposed following major structural rearrangement of these proteins, which is thought to accompany membrane insertion and polymerization. This region of homology may play some function related to their attachment and/or in their subsequent pore formation in the membrane. Recently, rabbit polyclonal antisera directed against human C9 have been used in the affinity purification of a PFP localized in human large granular lymphocytes (Zalman et al., 1986a) and in human peripheral blood NK cells separated by using a panel of monoclonal antibodies directed against N K and non-NK surface antigens (Liu et a l . , 1986). Zalman and her colleagues (198613) showed, furthermore, that monoclonal antibodies directed against C9 which also react against the PFP of LGLs are capable of blocking killing of K562 cells by LGLs. Polymerization of the isolated polypeptide resulted in the formation of two different circular structures, with internal diameters of 6 and 12.5 nm. Polymerized PFP in liposomes gave rise to channels of two functional diameters of 5-9 and 10.2 nm, as determined by liposome marker-retention assays. Polyclonal antibodies raised against residues 101-111 of C9, mentioned above, also block killing by LGLs (Zalman et al., 1986a). In our studies with purified NK cell populations, a large scatter of channel sizes was observed by morphological analysis, but rings with 16-nm internal diameter were the most commonly observed lesions (Liu et al., 1986). In our hands, the functional channel sizes formed by the NK cell polypeptide in planar lipid bilayers are also quite heterogeneous, showing a large scatter which may correspond to the different aggregation or polymerization states. Functionally, the lymphocyte PFP and poly C9 form ion nonselective
310
JOHN DING-E YOUNG AND ZANVIL A. COHN
channels which remain permanently open (Table I). The pores formed by these two complexes resist to closing induced by changes in the membrane potential. This behavior indicates that stable and voltage-resistant channels are formed by these protein complexes, which are attributes that would favor an active role of these channels in cytolysis. Functional sieving experiments have suggested that the pores formed by these two entities are heterogeneous in size, with sizes attaining 6-8 nm in functional diameter. Both these complexes are thought to form pores by a barrel stave model, with the pores consisting of discrete membrane nuclei that enlarge in size through the uptake of monomers. The recent demonstration that lymphocyte granules contain potent serine esterase activities which may play some role related to cytotoxicity (Pasternak and Eisen, 1985; Pasternak et al., 1986; Masson et al., 1986; Young et al., 1986a)has further suggested the connection between lymphocyte-mediated killing and cytolysis mediated by C. This analogy has recently been discussed in more detail by Reid (1986). It is now well known that the assembly of C lesions also depends on the activity of the complex serine proteases C4b2a3b and C3bBb3b that are involved in the generation of C5b. Moreover, C6 has also recently been described as a serine protease, and its enzymatic activity appears to be linked with the formation of C5b-9 complexes (Kolb et al., 1982). Like the C serine proteases, the granule serine proteases may also play an intermediate processing function, perhaps in the conversion of other lytic granule proteins into their active forms. The collective results obtained in several laboratories further support and extend the notion of “complement supergenes” initially proposed for C6 and C7 (Lachmann and Hobart, 1978; Podack et al., 1979b). On the basis of structural and functional similarities, a genetic relationship has also been proposed for several other complement proteins, including Clr and Cls, C3, C4, and C5, and C2 and factor B (reviewed by Campbell et al., 1986). The observed immunological similarities between lymphocyte PFP and several C components of the MAC suggest the possibility that the lymphocyte PFP may be an additional member of the complement supergene family. It is possible that all these effector molecules may have emerged from the same ancestral protein during evolution, but diverged and became specialized later to carry out either humoral or cellular immune responses. Studies on C-mediated cytolysis have been highly controversial over the years until the recently unified views on the molecular nature of C lesions. It is also expected that future studies on lymphocyte-mediated killing will also generate controversy and disaccord. Thus, Berke and his colleagues (as cited in a review article by Marx, 1986) have claimed that lymphocyte-damaged membranes do not exhibit the ring-like lesions described by other workers. This observation was taken as a piece of evidence against a role for pore
LYMPHOCYTE A N D
COMPLEMENT-MEDIATED CYTOTOXICITY
311
formation in lymphocyte-mediated killing. It might be noteworthy to point out, however, the differences between functional channel formation and ultrastructural observations of structural tubular lesions. The lymphocyte PFP is thought to damage membranes by forming functional channels on target membranes. Under certain favorable conditions (protein density, time of incubation, temperature), the protein may aggregate to form circular lesions of high PFP multiplicity. However, in our opinion, formation of the macromolecular tubular lesions is not a requisite for functional channel formation, and therefore, the morphological criteria may not be taken as synonymous of membrane damage by channel formation. Functional channels are, in faci, thought to be formed prior to complete circular polymerization. Similar notions have been put forth in the C field.
VI. Other Cytolytic Pore-Forming Proteins
A. OTHERPORE-FORMING PROTEINS I N CELL-MEDIATED KILLING 1 . Eosinophil Cationic Protein Eosinophils play an active role in allergic reactions and in the antibodydependent killing of a number of helminthic parasites (reviewed by Dessein and David, 1982; Venge et al., 1980; Spry, 1985; Gleich and Loegering, 1984). Eosinophil granule proteins have long been implicated in this type of cxtotoxicity. A number of cationic proteins have been isolated from eosinophi1 granules of several species and have been partially characterized. A major basic protein (MBP) of 9-11 kDa has been isolated from eosinophils of several species by Gleich and his colleagues (Gleich et al., 1973, 1974, 1976). A major eosinophil cationic protein (ECP) with a molecular mass of 21 kDa has also been isolated from human eosinophil granules (Olsson and Venge, 1974; Olsson et al., 1977). Other basic proteins that have been isolated from eosinophil granules include EP-X (Peterson and Venge, 1983), eosinophil peroxidase (Carlson et al., 1985), and the eosinophil-derived neurotoxin (EDN) (Durack et al., 1979, 1981). The strategy used by these investigators to isolate eosinophil proteins that are highly basic in nature has consisted of solubilization of granule contents in low pH and purification of proteins using cation-exchange chromatography. Since all these proteins represent abundant proteins within eosinophil granules (Ackerman et aZ., 1983), their role in parasite and microbial killing has been investigated in several laboratories. MBP has been shown to damage parasites (Buttenvorth et al., 1979; Wassom and Gleich, 1979) and mammalian cells (Gleich et al., 1979) at concentrations exceeding lop5 M . The human ECP, however, has been shown to damage schistosomula larvae
312
JOHN DING-E YOUNG A N D ZANVIL A. COHN
of the intravenous parasite Schistosoma mansoni at concentrations as low as 10-7 M (McLaren et al., 1981). ECP also produces the classic paralytic syndrome known as the Gordon phenomenon after intrathecal injection into guinea pigs (Fredens et al., 1982). Furthermore, E C P has been detected by radioimmunoassays in the supernatants of human degranulated granulocytes stimulated via the Fc-linked mechanism (Venge et al., 1980), and a secretory form of ECP has recently been identified in the granules of eosinophils (Tai et al., 1984). Recent studies with purified ECP suggest that this protein may form functional channels in lipid bilayers (Young et al., 19863’). ECP depolarizes the membrane potential of cultured nucleated cells and induces ion flow through model lipid bilayers. Purified ECP forms channels which show characteristics similar to those produced by C9 and lymphocyte PFP. ECP channels are resistant to closing by high transmembrane voltages and appear to be stable transmembrane entities, remaining permanently open once inserted into the bilayer. Ion-selectivity experiments show that ECP channels are relatively nonselective to all the monovalent ions tested, being slightly more permeable to anions. The channel-forming activity is only observed when the acid form of ECP is diluted in the presence of target membranes in a neutral environment. Extensive dialysis of E C P against buffers of neutral pH drastically lowers its channel-forming activity, suggesting that this protein may form inactive aggregates that are no longer membrane active. ECP has recently been observed to form ring-like lesions on the surface of liposomes (Young, Peterson, and Venge, unpublished observations). The lesions formed by ECP are heterogeneous in size, varying between 2 and 5 nm. It is not known whether purified MBP which is also cytolytic is capable of assembling membrane channels like ECP. The functional similarities between ECP and lymphocyte PFP and C9-mediated lesions suggest that E C P may also damage cells by a colloid osmotic mechanism.
2 . PFP of Amoebas Entamoeba histolytica is the enteric human parasite responsible for dysenteric amebiasis (reviewed by Ravdin and Guerrant, 1982). This infection is characterized -by an invasive enteric illness that may spread to multiple organs. In culture, E . histolytica is cytolytic to a variety of cell types, including leukocyte. Several laboratories have shown that the cell killing mediated by the amoeba is surface contact dependent (reviewed by Ravdin and Guerrant, 1982; Gitler et al., 1984; Young and Cohn, 1985). Following contact, the amoeba may rapidly ingest the target cell. Eaton et al. (1969) have suggested that a surface triggering mechanism may be involved in killing in which lysosomal contents are released at the site of surface contact between the cells. This suggestion has been substantiated by more recent cin-
LYMPHOCYTE A N D COM PLE M ENT-M E DIATE D CYTOTOXICITY
313
emicroscopic and kinetic studies (Ravdin et al., 1980; Ravdin and Guerrant, 1981) indicating that the cytolysis mediated by the amoeba may precede the phagocytic event, therefore raising the possibility of an extracellular cytolytic event triggered upon surface contact. In many respects, the mode of killing mediated by amoebas resembles that produced by immune cells described earlier in this review. A PFP of amoebas has recently been isolated (Lynch et al., 1982; Young et al., 1982; Gitler et al., 1984; Young and Cohn, 1985). The isolated polypeptide assumes a molecular mass of 15 kDa under reducing conditions. Under nonreducing conditions in the presence of the nonionic detergent p-D-OCtylglucoside, it assumes an apparent molecular mass of 28-30 kDa as determined by molecular sieving chromatography (Young and Cohn, 1985). The amoeba PFP forms voltage-dependent ion channels in liposomes and in planar lipid bilayers, and several of its biophysical properties have been studied in some detail. One of the more remarkable properties of this PFP is its tendency to aggregate in the lipid bilayer to assume channels of multiple sizes that function synchronously as individual units. This behavior is reminiscent of the barrel stave model. Following stimulation of their surface with the calcium ionophor A23187, LPS, or concanavalin A, amoebas release PFP rapidly into the the extracellular medium (Young et al., 1982). The amoeba PFP lyses a variety of tumor cell lines as assayed by conventional 51Cr release assays (unpublished observations).
B. OTHERTOXINSAS POREFORMERS A number of other toxins are present in nature, and their modes of action have been extensively studied in several laboratories (for reviews, see Rogolsky, 1979; Alouf, 1980; Latorre and Alvarez, 1981; Bernheimer and Rudy, 1986; Bhakdi and Tranum-Jensen, 1986b). Here, we will only discuss toxins that mediate target membrane damage by means of pore formation and, in particular, those aspects related to C and lymphocyte pore formers. For the sake of brevity, the toxins will be subdivided into several broad arbitrary categories as follows. 1 . Small Peptides Included in this category are small peptides that produce cytolysis by forming either aqueous channels which span the membrane bilayer or other undefined structures which increase membrane permeability. a . Melittin. Melittin is the prototype of this group and is the most widely studied model membrane lytic peptide. Melittin is the main component of the venom of the honeybee Apis mellt$era and may comprise up to 50% of the bee venom by weight (Haberman, 1972). Melittin is a small basic peptide of 26 amino acids (Fig. 3). The amphipathic nature of this peptide is
314
JOHN DING-E YOUNG A N D ZANVIL A. COHN
clear from its primary sequence. It contains a hexapeptide at the carboxylterminus carrying a cluster of positive charges, followed by a long stretch of hydrophobic amino acids (Fig. 3). Such distribution of amino acids and charges has been observed with a number of integral membrane proteins that span the bilayer. The peptides of two other species of honeybee have recently been sequenced, and they all show conservative changes in their structure, maintaining the above-mentioned segregation of charges (Kreil, 1973; Fig. 3). When the hexapeptide segment is removed from melittin, the remaining 20 amino acid segment, while capable of binding to erythrocytes, does not lyse them (Schroeder et al., 1971). It has recently been suggested that these peptides tend to form amphiphilic a helical structures, with the amino acid side chains segregated on either a hydrophobic or a hydrophilic side (Kaiser and Kezdy, 1983, 1984). Kaiser and Kezdy have proposed that the ability to form such amphiphilic secondary structures is vital to the biological function of a number of peptides, including hormones, apolipoprotein A-I, and melittin. Thus, it is thought that the amphiphilic a helix may be essential for the lytic activity of melittin,
k b (z?)
S
s
d
dLp j
.i
Determinants expressed
Regulatory genes ,
1' 1'
(Z?)
1' 1'
2' 2' 2'
3' 3' 3' 3' 3'
None
4'
4' 4' (4'?)
1 1
4 4
2 2 2
3
(4')
H-2Lhas not been tested for determinants 2 or 4 but appears to share determinant 1 with H-2', and H-2s. H-2J has not been tested for determinant 4. Determinant 4 was detected only when using (NZB X B6)FL mice as hosts, which reject H-2k and H-2k/H-2", but riot H-2k X H-2s, H-2q, H-2r, or H-2k X H-2d marrow cells. C3H lymphoid cells may express a similar determinant (Eastcott et al., 1981).
402
MICHAEL BENNETT
antigen on the BMC and will accept parental A or B BMC grafts; (4) F, hybrids between strains A and C with nonidentical Hh-1 antigens will not express the Hh-1 antigens of A or C (exception is determinant 2 i n j x d F,s) and will usually reject parental A or C grafts; (5) rejection of marrow grafts from strains A and B with identical Hh-1 antigens will be inhibited by the infusion of strain A or B Hh-1 antigen + tumor cells (in uiuo cold target cell competition). This model appears to conflict somewhat with the idea that H h - W B , Hh-20 2 , and Hh-NZB loci map 4, 16, or 32 centimorgans to the right of H - 2 (Cudkowicz and Nakamura, 1983). Perhaps genes in those positions affect hybrid resistance, and could even be the structural genes discussed above. Those mapping studies were used to locate the genes regulating hybrid resistance to BMC grafts, whereas the model presented above functions for both allogeneic and hybrid resistance. On balance, Nakamura et al. (1986) did detect Hh-lb antigens on WB x B6 F, cells. If the Hh-1 genes are regulatory genes, what gene products might they “control” or “modify” on the cell surface of stem cells or other progenitor cells? The nearest (genetically speaking) gene products could be class I or class I1 antigens themselves. Singh and and David (1983) suggested that enzymes encoded by genes in the S / D region of H - 2 might alter Ia antigens so as to affect immune functions. Class I, but not class 11, antigens are expressed very strongly on stem cells such that anti-H-2kb antibodies have been used to purify mouse marrow stem cells (Visser and Eliason, 1983; Visser et al., 1984). Therefore, Hh-1 genes could modify class I antigens, perhaps enzymatically (Warner, 1978), so as to prevent their recognition by host effector cells, probably NK cells. O’Neill and Blanden (1979) measured the ability of spleen cells of parental BlO.A(BA) and BlO.A(SR) or F, hybrid mice to absorb anti-H-2 antibodies and the ability of their macrophages to serve as targets for cytolytic T lymphocytes. There was a disproportional loss of expression of H-2Kb and H-2Dd antigens on the F, hybrid cells. Such data are in keeping with the possibility that Hh-1 can d e c t expression of class I antigens. However, preformed anti-class I antibodies and presensitized T cells can lead to rejection of BMC, indicating that class I antigen recognition would have to be different by N K cells. Milisauskas et al. (1986) recently observed that loss of Db antigens from RBL-S tumor cells was not associated with loss of Hh-lb antigens. Other candidates include drug receptors; Byron (1972, 1973) detected P-adrenergic and cholinergic receptors on CFU-S, which regulate the cell cycle status of these cells. Insulin receptors have been associated with class I H-2 and HLA antigens (Edidin, 1986; Due et al., 1986). Therefore, it is conceivable that Hh-1 genes function to affect hormone or drug receptors which may or may not be associated with H-2 antigens. An appealing possibility is that viral gene products constitute the
HYBRID RESISTANCE
403
antigen recognized and that the Hh-1 genes determine the expression of the viral products on stem cells. Pampeno and Meruelo (1986) detected retroviral sequences in the T l a region of H-2; it is conceivable that similar gene products are present throughout the MHC region of mice. These could be acted upon by the Hh-1 regulatory genes. The H - 2 D region can determine the rate of synthesis of Friend leukemia virus by cultured leukemic cells (Bubbers et al., 1977; Freedman et al., 1978). Hyman and Cunningham (1986) described a trans-acting gene which regulates cell surface expression of Thy-1 by controlling transcription. Another possible target for the putative Hh-1 regulatory genes could be the large gl ycoproteins that are thought to constitute the target structures on NK-sensitive cells (Roder et al., 1979; Obexer et al., 1983; Henkart et al., 1986). The ability of laminin to block NK cell function, including binding of NK cells to their targets, suggested that laminin receptors could be target structures on cells recognized by N K cells (Hiserodt et al., 1985). If these surface molecules are modified by Hh-1 gene products, they could be involved in hybrid resistance. The extreme sensitivity of H-2-negative YAC-1, RBL-5, and EL-4 tumor cells to N K cell-mediated rejection (Ljunggren and Karre, 1985; Pointek et al., 1985; Karre et al., 1986) could be explained by a loss of function of the down-regulating trans genes at Hh-1. Perhaps such cells express all H-2controlled Hh-1 determinants (and perhaps other Hh antigens) and are thus much better targets for NK cell-mediated lysis. This could explain why a B6 (H-2b/Hh-lb) mouse might recognize H-2-negative EL-4 cells almost as well as an F, hybrid; those cells would express determinants 2 and 3 as well as determinant 1. In the human system, loss of HLA was associated with an increased sensitivity to N K cell lysis (Hard-Bellan et al., 1986). Bach (1978) maintained that hybrid resistance was governed by density of class I antigens expressed on homozygous versus heterozygous hemopoietic cells.
B. A MISMATCHINGOF H-2 ANTIGEN BETWEEN DONORAND HOST PREVENTS OPTIMALGROWTHOF TRANSPLANTED CELLS When dealing with transplants of hemopoietic cells, even into irradiated recipients, it is hard to ignore the potential for host-versus-graft reactions. Preformed natural antibodies, N K cells, and presensitized T cells can function under those conditions. Nonetheless, it is equally difficult to disregard the data of Hellstrom and Lengerova and their colleagues, which suggest that mismatching of H-2 antigens can lead to poor growth of transplanted bone marrow or leukemia/lymphoma cells. Snell(1976a, b, 1979) envisioned that a mismatch of H-2 antigens between donor stem cell and host effector cell could trigger a destructive reaction by the effector cell. Carlson et al. (1984b) suggested that “cell positioning” of doncr stem cells with respect to
404
MICHAEL BENNETT
host organ stroma or microenvironment could be the basis of poor growth. Karre et al. (1986a,b) conceived the idea that loss of “self” H-2 antigens on tumor cells or a lack of self H-2 antigens on stem cells would “unmask’ cell surface target structures that could be recognized and attacked by NK cells. One way to test Karre’s hypothesis would be to transplant (A x B)F, marrow cells into irradiated C hosts, where A, B, and C are different H-2/Hh-1 types. Several years ago in Boston, Michael Williams and I observed that B10.BR (H-2k) hosts rejected B10 (H-2b) or B10.D2 (H-2d) but accepted (B10 x BlO.D2)F, marrow grafts, using a 4-day 1251UdRassay. This results conflicts with the “non-self’ hypothesis but Lengerova et al. (1973a) have observed rejection of small numbers of H-2 heterozygous marrow cells, using the spleen colony assay after day 8. While writing this review I became more aware of the possibility that T cells might be able to function in marrow allograft rejection, even during the first week after cell transfer. Therefore H-2k/H-213(C3H x B6)F, marrow cells were infused into irradiated H-2d BALB/c or SCID mice, and growth of marrow was assessed by measuring splenic 1251UdRuptake on days 4 or 8. In BALB/c mice, (C3H x B6)F, mice grew without impairment on day 4 but were rejected by day 8. In contrast, the F, cells grew well in SCID mice on both days 4 and 8. These preliminary experiments performed by Bill Murphy do not support the idea that lack of “self’ H-2 antigens is the trigger for rejection of marrow stem cells. Curtis and Rooney (1979) observed that epithelial cells from adult kidneys of mice exhibit contact inhibition when grown in vitro. By mixing cells or different H-2 types, they noted an enhanced degree of contact inhibition, especially if the two cell types differed for class I antigens, i.e., H-2D or H-2K antigens. If a similar phenomenon occurs in spleens or other hemopoietic organs of mice after transfer of stem cells, a deficient cell interaction could lead to “CFU-S repression” (McCulloch and Till, 1963). C. A MODELTO EXPLAINH o w HEMOPOIETIC CELL GRAFTSARE REJECTED The existence of natural antibodies or the induction of antibodies by deliberate or accidental immunization can lead to rejection by targeting stem cells for lysis by NK cells or other cells capable of antibody-dependent cellular cytotoxicity (ADCC). The ability to sensitize T cells in irradiated mice can lead to direct lysis of stem cells by CTL. The mechanism of rejection by these two effective immune systems is straightforward and requires no new model to explain how marrow grafts might be rejected. One should not underestimate the importance of T cell-mediated responses to BMC allografts or the existence of preformed antibodies. The mechanism of rejection of BMC grafts by NK cells without the aid of antibodies has not been determined. Let us propose the idea that NK cells
HYBRID RESISTANCE
405
have specific receptors for Hh-1 and other Hh antigens that are expressed on donor stem cells. There is a limited degree of polymorphism so far detected for Hh-1 antigens. If the model described above is correct, no more than 4 different Hh-1 determinants exist. However, there are Hh-3 and Hh-DBA “minor” antigens and potentially different Hh antigens on normal lymphoid cells. Therefore, N K cells may have clonally distributed anti-Hh-1 receptors. The other alternative is that each NK cell expresses receptors for all Hh antigens of the species. Since boina fide N K cells do not express T cell receptors, N K cells must utilize their own type of receptor for Hh antigens. What triggers the destruction of incoirnpatible stem cells by host NK cells? The findings that antibodies to interferon-a/p inhibit hybrid resistance and that macrophages are necessary for the secretion of interferon (Afifi et al., 1985) indicate that NK cells by themselves are unable to mediate graft rejection. Macrophages or products of macrophages are apparently required to stimulate NK cells in some way that endows them with the ability to reject stem cells. Mice from poor responder strains for a given Hh-1 antigen fail to reject normally, but will reject if their NK cells are stimulated by interferon or interferon inducers. Therefore, an early recognition event must lead, in the good responder mice, to the elaboration of at least interferon by macrophages. The finding that athymic nude mice and SCID mice can reject BMC allografts well reduces the cell types necessary for the response to N K cells and macrophages. ‘The process of rejection does not appear to be a simple, rapid destruction of stem cells by NK cells similar to the lysis of YAC-1 cells. Using the antiasialo GM1 serum which inactivates NK cells immediately in uiuo, Sentman et al. (1987) observed that 24 hours were required for rejection of parental marrow cells, even in mice stimulated with polyinosinic:polycytidylic acid. Hybrid resistance was inhibited even when the antiserum was injected 24 hours after marrow cell infusion. Therefore, the series of events that may occur during hybrid resistance to H-2”IHh-lb BMC by B6 x I>BA/2 F, hybrid mice could be the following:
1. NK cells with anti-Hh-lh receptors recognize BMC expressing Hh-lh and are stimulated in some way, but cannot yet mediate rejection. 2. NK cells so stimulated by recognizing Hh-ll’ antigens interact with syngeneic macrophages, stimulating them to secrete interferon-a/@. This event is obviously not understood, but may well be the critical one that makes a mouse a good responder or a poor responder to a given Hh-1 -incompatible inarrow graft. 3. Macrophages can be stimulated in other ways to secrete interferon, e.g., by the administration of agents that induce interferon secretion by macrophages.
406
MICHAEL BENNETT
TABLE IV CELLSURFACE ANTIGENS EXPRESSED BY NK CELLSA N D OTHERPOTENTIALLY RELATED CELL TYPES
NK cells Murine Ly-1 Ly-2
L3T4 Mac-I (CR3) Asido GM1 Thy-1 NK-1.1 NK-2.1 Qa-2 Qa-5 Human CD2 CD3 CD4 CD7 CD8 C D l l (CR3) CD16 Leu 7 Leu 19 Leu M3
-
+ + +I+ + + + +
-
T cells
Monocytes
Granulocytes
+ + + + + + -a -a
+ + + + + + + + -b + +
a Has been identified on some Ly-2+ cell lines with or without non-MHC-restricted cytotoxicity (Brooks et al., 1982a,b). b A rare subset of T cells has been reported to express low-density CD16 antigen (Lanier et a!., 1985).
4. The interferon augments the lytic function by NK cells with specific anti-Hh-lb receptors. Interferon recruits pre-NK cells and enhances the kinetics of lysis (Silva et al., 1980). 5 . The Hh-lb-specific NK cells lyse stem cells which express Hh-lb antigens.
Even though interferon can definitely stimulate NK cells to reject BMC allografts, there are other agents which could have similar activities. Thymosin a-1 synergizes with interferon to stimulate NK cell maturation (Favalli et al., 1985) and function (Henney et al., 1981; Umeda et al., 1983). Interleukin 2 augments NK cell function (Kalland, 198613; Koo and Manyack, 1986), and tumor necrosis serum (Chun et al., 1979) has a distinctive stimulating effect on NK cells. Perhaps tumor necrosis factor can also activate
407
HYBRID RESISTANCE
TABLE V STRAINDISTRIBUTION OF N K ALLOANTIGENS~ Strains
NK-l.lb
NK-2.1c
C58/J C57BL/6J B1O.A CE/J C57Brl J SJL Ma/My NZB/J SM/J A/J BALB/c C3H/J Ba1b.H-26 CBA/J RF AKR AU LP/J 129/J DBA/2 DBA/1
+ +d + + + + + + +
+ + -
-
+ + + + + + + + + -e + +e
As determined by complement elimination. (C3H X BALB/c)FI anti-CE; C3H anti-CE; CBA anti-CE; (C3H X DBA/2)FI anti-CE CE anti-DBA; (NZB X CE)F1 anti-CBA; (CE X CBA)F1 X CE anti-CBA; NZB anti-BALB/c. Also on B6.H-2k; B6.Lyt-1.1, 2.1, 3.1; B6.PC+; B6.Gix+; B6.Tla; B6.Kl. By FACS analysis, 129/J lacks detectable NK-2.1 (NZB anti-BALB/c) and DBA/1 is positive for both NK-1.1 and NK-2.1 (Pollack and Emmons, 1982. f n t . Not tested. a
N K cells. Wakowiak et al. (1986) described a Thy-l+ Lyt-l+ cell destroyed by 5-flourouracil that appears to participate in hybrid resistance to BMC grafts. The role of this latter cell type and other cytokines/lymphokines in the rejection process have yet to be tested in any detail. The lytic molecules of N K cell:rmphocyteantigen and, 3, 10, 13 T cell subsets and, 41, 50 Cell-mediated immunity, hybrid resistance and, 370 Cell-mediated killing, 319, 320 granule prciteins and cell lines, 286, 287 cytoplasnnic granules, 287-291 lymphotoxin, 298 membrane attack complex, 310 proteoglycans, 298, 299 serine esterases, 297, 298 TNF-related polypeptides, 298 pore-forming proteins and amoebas., 312, 313 biochemical properties, 292-295
451
eosinophil cationic protein, 311, 312 membrane binding, 295, 296 purification, 291, 292 Cell-mediated lympholysis hybrid resistance and, 355, 364, 394-396 T cell subsets and, 43, 45, 51, 57, 108 Cell-mediated lysis, hybrid resistance and, 407 Cell surface molecules T cell activation and, 1, 2 accessory molecules, 14, 15 IL-1 receptor, 13, 14 T cell antigen receptor, 2-8 T1, 13 T11, 8-10 Thy-I, 10, 11 Tp44, 11-13 T cell subsets and, 40, 43-48 accessory molecules, 49, 50 T cell receptor, 40-43 Cellulose acetate, hybrid resistance and, 372 C-fos, T cell activation and, 27, 28 Channel formation cytolytic protein and, 318, 319 membrane attack complex of complement, 301-303, 306, 307, 311 Chemotherapy, hybrid resistance and, 350 Chicken ovalbumin, T cell subsets and, 53 Chloroquine, 56, 57, 281 Cholesterol, cytotoxicity and, 279, 300, 302, 317, 318 Chondroitin sulfate A, cell-mediated killing and, 298, 299 Chromatin B lymphocyte formation and, 204, 221 cytolysis and, 283 Chronic myeloid leukemia, B lymphocyte formation and, 188 Cleavage cytolytic protein and, 318 membrane attack complex of complement and, 299, 306 Clones B lymphocyte formation and, 182, 184, 185, 238 B cell precursors, 191, 196, 200, 201 bone marrow cultures, 213, 216, 217, 219, 220 C3H/HeJ mice, 231 Ig genes, 203, 205
452
INDEX
inducible cell line, 220-222 lymphohemopoietic tissue organization. 187 NZB mice, 226, 229 population dynamics, 207 SCID mice, 225, 226 WIW anemic mice, 224 cytolysis and, 274, 276, 280, 281, 285 cytotoxic T lymphocytes and, 167 amino acid changes, 162 carbohydrate moieties, 154 exon shuffling, 138, 142, 144-146, 149 HLA class I antigens, 149-151 Pz-microglobulin, 156 monoclonal antibodies and, 162, 166, 167 cytotoxicity and, 319, 320 granule proteins, 286, 287, 291, 297 nature of mediation, 271, 273 pore formers, 317, 319 hybrid resistance and antibodies, 377 antigen expression, 405 in uitro assays, 394 marrow engraftment, 387 NK cells, 375 T cells, 383 T cell activation and, 1 gene regulation, 27 IL-1 receptor, 13 receptor-mediated signal transduction, 20-22, 24 synergy, 17 T cell antigen receptor, 2, 3, 6, 7 T11, 9 Thy-1, 10 T cell subsets and cell surface molecules, 40, 49. 50 H-2 alloantigen recognition, 79, 87, 94 H-%restricted antigen recognition, 52, 53, 57, 59-61, 64,66, 71, 72, 74, 75 C-myc, T cell activation and, 27, 28 Cognate help, T cell subsets and, 77 Coimmunoprecipitation, T cell activation and, 4, 5 Colchicine, cytolysis and, 282 Cold target cell competitors, hybrid resistance and, 396, 402
Colicins, cytolysis and, 318, 319 Colloid osmotic killing membrane attack complex of complement, 300 membrane damage, 280 pore formers, 315, 318 pore-forming protein, 295, 312 Colony-forming cells B lymphocytes and, 182, 191, 200, 230, 237 hybrid resistance and bone marrow cells, 336, 344 leukemia/lymphoma cells, 359 NK cells, 376 T cells, 381, 382 Colony-forming unit, hybrid resistance and, 336, 396 Colony-stimulating factor B lymphocyte formation and, 219, 228230, 234, 236, 237 hybrid resistance and, 382 Complement-mediated cytotoxicity, see Cytotoxicity, lymphocyte and complement-mediated Complementarity-determining regions, T cell subsets and, 52 Complete Freund’s adjuvant, T cell subsets and, 70, 71 Concanavalin A B lymphocyte formation and, 198, 229 hybrid resistance and, 375, 409 pore-forming protein and, 313 T cell activation and, 8, 28 T cell subsets and, 63-65 Concanavalin A-activated spleen cells, T cell subsets and, 64,65, 70 Conjugate formation cell mediation and, 272 cytolysis and, 274-276, 278, 282, 283 Corticosteroids, B lymphocyte formation and, 197 Cortisone, hybrid resistance and, 338, 359, 360 Corynebacterium parvum B lymphocyte formation and, 208 hybrid resistance and bone marrows, 340, 347 macrophages, 370 marrow microenvironment, 391, 392 syngeneic stem cell functions, 389
INDEX
Cross-linkage cell mediation and, 272 T cell activation and, 4, 5, 9, 11, 12, 25 T cell subsets and, 62, 111 Cross-priming, T cell subsets and, 74 Cross-reactivity lymphocytes and, 144, 147, cytotoxic ’I 162, 170 hybrid resistance and, 357 membrane attack complex of complement and, 307, 309 T cell subsets and, 79, 88,112 Cyclic neutropenia, B lymphocyte formation and, 222, 228-230, 234, 237 Cyclophosphamide, hybrid resistance and bone mamow cells, 347, 348 effector mechanisms, 370, 388, 390, 393 leukemia/lymphoma cells, 366, 368 lymphoid cells, 354 Cyclosporin, 60, 350 Cyclosporin .4, T cell activation and, 29, 30 Cysteine cytotoxic I’lymphocytes and, 148, 163 membrane attack complex of complement and, 307, 309 Cytochalasin B, cytolysis and, 282 Cytochrome c, T cell subsets and, 53 Cytolysin, cell-mediated killing and, 291 Cytolysis af27 polarity, 273-275 cell mediation, 272 colloid osmotic killing, 280 cytoplasmic granules, 287, 288, 291 granule exocytosis, 281, 282 hydrolytic enzymes, 273-275 intracellular damage, 282-284 leukoregulin, 286 lymphotoxin-like molecules, 284-286 membrane attack complex of complement, 299, 300, 310 membrane damage, 277-281 polypeptide toxins, 316-319 pore-forming proteins amoebas, 312, 313 cytolytic proteins, 316-319 eosinopliil cationic protein, 311, 312 polypeptide toxins, 316-319 small peptides, 313-316 reactive oxygen metabolism intermediates, !286
453
Cytolytic T cell clones, 9, 12, 14 Cytolytic T cells, 365, 367, 368 Cytolytic T lymphocytes, hybrid resistance and, 364, 380, 395, 396, 402, 404 Cytoplasmic granules, cytotoxicity and, 281, 287-291 Cytosol, T cell activation and, 17, 22 Cytotoxic T cells, hybrid resistance and, 385 Cytotoxic T lymphocyte cell lines cell-mediated killing and, 287-289, 291, 297 cytolysis and, 273 Cytotoxic T lymphocyte differentiation factor, T cell subsets and, 64, 76 Cytotoxic T lymphocytes, 269, 270 granule exocytosis, 281 granule proteins, 286, 287 cytoplasmic granules, 287 pore-forming protein, 291, 296, 297 serine esterase, 297, 298 hydrolytic enzymes, 275-277 intracellular damage, 283 lymphotoxin-like molecules, 284, 285 mediation, 270-273 polarity, 273-275 recognition by MHC molecules, 135-138 carbohydrate moieties, 152-154 exon shuffling, 138-149 H-2 mutant strains, 158-163 HLA class I antigens, 149-152 HLA subtypes, 163-165 Pz-Microglobulin, 154-156 monoclonal antibodies, 156-158 somatic cell class I variants, 165-167 T cell subsets and, 39, 51 activated T cells and hybridomas, 54, 56, 57 alloreactivity, 78, 81, 82 effector phase, 75, 92, 94, 95 H-2 molecules in thymus, 97, 99-103, 108 resting T cell subsets, 83-88 T accessory molecules, 60 unprimed and resting T cells, 65, 71-75 Cytotoxicity hybrid resistance and, 359, 361, 362, 366, 372, 393, 394 lymphocyte and complement-mediated, 269, 270, 319, 320 at27 polarity, 273-275
454
INDEX
colloid osmotic killing, 280 CTL cells, 270-273 granule exocytosis, 281, 282 hydrolytic enzymes, 275-277 intracellular damage, 282-284 leukoregulin, 286 lymphotoxin-like molecules, 284-286 membrane attack complex of complement, 299-307 membrane damage, 277-281 NK cells, 270-273 pore-forming proteins, 311-319 reactive oxygen metabolism intermediates, 286 Cytotoxins, 269, 270
D Degranulation, cytotoxicity and, 282, 296, 297, 312 Delayed hypersensitivity, hybrid resistance and, 383 Delayed-type hypersensitivity, T cell subsets and, 76, 77, 94 Deletion B lymphocyte formation and, 203 cytotoxic T lymphocytes and, 139, 167 Dendritic cells, T cell subsets and H-2 alloantigen recognition, 89-92 H-2 molecules in thymus, 103-106, 108110 H-2-restricted antigen recognition, 74, 75, 77 Deoxyguanosine (dGuo), T cell subsets and, 104, 106, 108 Deoxynucleotidyltransferase, B lymphocyte formation and, 194, 203, 204, 220, 231 Depolarization cell-mediated killing and, 273 pore-forming protein and, 293, 294 Desotope, T cell subsets and, 53, 58 Dexamethasone, hybrid resistance and, 387 Dexter’s culture system, B lymphocyte formation and . bone marrow, 208-210, 213, 215-217, 219, 220 W/W anemic mice, 224 Diacylglycerol, T cell activation and, 15-19, 30 Diamphotoxin, cytolysis and, 318
Differentiation B lymphocyte formation and, 181, 235237, 239 B cell precursors, 188, 189, 191-192, 194, 195, 197, 199-201 bone marrow cultures, 210, 212, 214218, 220 genetically determined defects, 223226, 230, 231 inducible cell line, 220, 221, 223 lymphohemopoietic tissues, 185, 187 population dynamics, 206, 207 soluble mediators, 232, 235 cytolysis and, 270 hybrid resistance and antigen expression, 400 bone marrow cells, 336 leukemia/lymphoma cells, 369 marrow microenvironment, 390, 392 NK cells, 372, 373 syngeneic stem cells, 388 T cells, 381 T cell activation and, 22 T cell subsets and H-2 molecules in thymus, 98-102, 104, 105, 107, 108 H-2-restricted antigen recognition, 52, 64,66, 76, 77 DNA B lymphocyte formation and, 183, 184, 203, 217, 235 cell-mediated killing and, 298 cytolysis and, 277, 283-285 hybrid resistance and, 343, 377, 378 T cell activation and, 29 T cell subsets and, 41, 44, 46-48, 78 Dog leukocyte antigen, hybrid resistance and, 349, 350
E Effector cells cell-mediated killing and, 296 cytolysis and, 274, 275 granule exocytosis, 281 hydrolytic enzymes, 276 intracellular damage, 283, 284 membrane damage, 277 cytotoxic T lymphocytes and, 148, 150, 170
INDEX
cytotoxicit) and, 269, 270, 272 hybrid resistance and, 334, 370 antibodies, 376, 378, 379 antigen expression, 402, 403 bone marrow cells, 346 in uitro :assays, 393-396 leukemia/lymphoma cells, 359, 366, 368 lymphoid cells, 353 macrophages, 371 marrow imicroenvironment, 390 NK cells, 372, 373 syngeneic stem cells, 390 T cells, 382-384 Effector mechanisms of hybrid resistance, 369, 370 antibodies, 376-380 in oitro assays, 393-396 macrophages, 370-372 marrow engraftment, 384-388 marrow microenvironment, 390-393 NK cells, 2872-376 syngeneic stem cell functons, 388-390 T cells, 380-384 Effector phase, T cell subsets and, 75-77, 92-95 Electron microscopy B lymphocyte formation and, 212 cell-mediated killing and, 292, 294, 295 cytolysis arid, 282 membrane attack complex of complement and, 300, 306 T cell subsrets and, 106 Electron-spin resonance, membrane attack complex of complement and, 303 Embryo, B lymphocyte formation and, 181, 193, 199, 201, 224, 227 Endocytosis, cell-mediated killing and, 298 Endothelid cells B lymphocyte formation and, 186, 222225, 235, 236 hybrid resistance and, 400 T cell subsets and, 46, 90 Endotoxin, B lymphocyte formation and, 224, 230 Eosinophil cationic protein, 311, 312 Eosinophil-derived neurotoxin, 311 Eosinophils, Icytotoxicity and, 269, 270, 291, 312 Epithelial cells hybrid resistance and, 335
455
T cell activation and, 1, 10 T cell subsets and cell surface molecules, 46 restricted T cells, 103-106 T cell development, 96-98 T cell specificity, 113 tolerance induction, 107-100 Epitopes B lymphocyte formation and, 191, 192 cytotoxic T lymphocytes and, 138, 169, 170 amino acid changes, 159, 162, 163 carbohydrate moieties, 152, 154 exon shuffling, 139-142, 144, 145, 147 HLA class I antigens, 150, 151 Pz-Microglobulin, 155 monoclonal antibodies, 156-158, 165167 T cell activation and, 1, 9, 11, 20 T cell subsets and cell surface molecules, 44, 47 H-2 alloantigen recognition, 79, 80, 83 H-%restricted antigen recognition, 53, 54, 56, 62 T cell specificity, 111, 112 Epstein-Barr virus, cytotoxic T lymphocytes and, 144, 165, 167 Erythroblasts, hybrid resistance and, 376 Erythrocytes B lymphocyte formation and, 189, 208 cell-mediated killing and, 289, 290, 293, 295 cytolysis and, 277, 278, 280 hybrid resistance and, 337, 376, 386, 388, 400 membrane attack complex of complement and, 299-301, 303, 304, 306 pore formers and, 314, 316, 317, 319 Erythroid cells, B lymphocyte formation and, 186, 213 Erythroleukemia, hybrid resistance and, 336, 372, 378 Erythropoiesis, hybrid resistance and, 336, 337, 339, 348 Erythropoietin, hybrid resistance and, 348 Escherichia coli, 302, 318 Esterase B lymphocyte formation and, 212, 213 cell-mediated killing and, 287 Estradiol, hybrid resistance and, 374, 389392
456
INDEX
Exogenous help, T cell subsets and, 84, 85, 87, 112 Exon shuffling, CTL and, 144-148, 169 altered cytoplasmic regions, 148, 149 class I gene transfection, 138, 139 CTL recognition, 141-144 monoclonal antibodies, 156, 165 serology, 139-141
marrow engraftment, 387 NK cells, 376 Friend spleen focus-forming virus, hybrid resistance and, 389 Friend virus, 168, 378, 389 Fungi, cytotoxicity and, 269, 270
F
G protein, CTL and, 154 Gene dosage, hybrid resistance and, 357 Glucose, membrane attack complex of complement and, 300 Glutaraldehyde, T cell subsets and, 53 Glycans, CTL and, 152, 153 Glycogen, cell-mediated killing and, 287 Glycolipid, membrane attack complex of complement and, 301 Glycoprotein B lymphocyte formation and, 189-193, 227 CTL and, 152 human T lymphocyte activation and, 2, 4, 8 hybrid resistance and, 378, 403 T cell subsets and, 43, 44, 46, 49 Glycosylation B lymphocytes and, 190, 192, 215 CTLand, 135, 138, 152, 153 cytotoxicity and, 272 T cell subsets and, 41, 50 Golgi apparatus cytolysis and, 275, 281, 282 T cell subsets and, 57 Graft, hybrid resistance and antigen expression, 397, 399, 401, 402, 404-407, 410 bone marrow cells, 336-351 effector mechanisms antibodies, 376, 377 in oitro assays, 393, 395 macrophages, 371 marrow, 384-388, 390 NK cells, 373-375 syngeneic stem cell functions, 388, 390 leukemia/lymphoma cells, 362, 364-367 lymphoid cells, 352-355 Graft rejection, hybrid resistance and antigen expression, 404-407 bone marrow cells, 344, 350
Fc receptors B lymphocyte formation and, 194 cytotoxicity and, 269, 271, 279, 296 hybrid resistance and, 371, 376-378, 394 T cell activation and, 9 Fibroblasts B lymphocyte formation and, 235, 236, 238 genetically determined defects, 230 Ig genes, 205 lymphohemopoietic tissue organization, 186, 188 CTL and, 145, 150-153 hybrid resistance and antigen expression, 400 in vitro assays, 393, 394 leukemia/lymphoma cells, 359, 361 NK cells, 372 T cell activation and, 10 T cell subsets and, 46, 89-91, 94 Fibronectin, B lymphocyte formation and, 215 Fibrosarcoma CTL and, 45 hybrid resistance and, 359, 361, 363 Fluorescein isothiocyanate, CTL and, 166, 167 Fluorescence B lymphocyte formation and, 196 cell-mediated killing and, 298 Fluorescene-activated cell sorter, T cell subsets and, 84 Friend erythroleukemia, hybrid resistance and, 389 Friend leukemia virus, hybrid resistance and antigen expression, 400, 403 hone marrow cells, 345 lymphoid cells, 354
G
INDEX
effector mechanisms, 370, 371 antibodies, 376, 378 marrow engraftment, 385, 388 NK cells, 375 syngeneic stem cells, 390 Graft-versus-host disease (GVHD), T cell subsets and cell surface molecules, 43 H-2 alloantigen recognition, 83, 84, 86, 88, 92, 95 H-&-restricted antigen recognition, 71, 72 Graft-versus-lhost reaction, hybrid resistance and antigen expression, 403 bone marrow cells, 335, 336, 340, 350 in vitro assays, 394 leukemia/lymphoma cells, 360, 367 lymphoid cells, 351-356 marrow engraftment, 385-387 T cells, 380 Gramicidin A , cytolysis and, 316 G r a d e exocytosis, cytotoxicity and, 273, 281, 282 Granule proteins, cell-mediated killing and cell lines, 286, 287 cytoplasmic granules, 287-291 lyrnphotoxin, 298 membrane attack complex of complement, 310 pore-forming protein, 291-297 proteoglycans, 298, 299 serine esterases, 297, 298 TNF-related polypeptides, 298 Granulocytes B lymphocytes and, 238 B cell piocursors, 191, 195 bone marrow cultures, 209, 210, 219 geneticallly determined defects, 229, 230 cell-mediated killing and, 312 hybrid resistance and, 336, 370, 380 Granulopoiesis, hybrid resistance and, 336
H H-2 moleculm alloantigen recognition and, 78 alloreact ivity, 78-83 antigen-presenting cells, 88-92 effector phase, 92-95 resting T cell subsets, 83-88
457
antigen recognition and, 51, 52 effector phase, 75-77 T accessory molecule function, 59-62 triggering of activated T cells and hybridomas, 52-59 triggering of unprimed and resting T cells, 62-75 B lymphocyte formation and, 213 CTL and, 136, 170 carbohydrate moieties, 152-154 exon shuffling, 138-148 HLA class I antigens, 149, 152 Pz-Microglobulin, 155, 156 monoclonal antibodies, 156, 157, 166, 167 mutant strains, 158-163, 165 cytotoxicity and, 269, 279-281 hybrid resistance and antibodies, 376-378 antigen expression, 397-404 bone marrow cells, 336-348 in oitro assays, 395, 396 leukemia/lymphoma cells, 358-369 lymphoid cells, 352-354 marrow engraftment, 386-388 NK cells, 375, 376 syngeneic stem cells, 390 T cells, 381-383 T cell subsets and, 40, 43-48 thymus and, 95, 96 development, 96-99 restricted T cells, 99-107 tolerance induction, 107-110 Haplotype CTL and, 159 hybrid resistance and antigen expression, 397-399, 409 bone marrow cells, 341, 343 leukemia/lymphoma cells, 364 lymphoid cells, 353 T cell subsets and, 47, 48, 78, 79, 102 Helper T cells CTL and, 136, 137 cytolysis and, 276 Hemagglutinin, T cell subsets and, 57 Hemolysis, cell-mediated cytolytic proteins and, 317, 318 cytoplasmic granules, 288-290 membrane attack complex of complement and, 299, 305, 306, 309
458
INDEX
membrane damage and, 279 pore-forming proteins and, 291, 293, 295, 296 proteoglycans and, 299 Hemopoiesis, B lymphocyte formation and, 181, 182, 235-238 B cell precursors, 188, 191, 196, 197, 200, 201 bone marrow cultures, 208-210, 212, 214, 215, 217, 219, 220 genetically determined defects, 224, 225, 228, 229, 231, 232 inducible cell line, 220, 221, 223 lymphohemopoietic tissues, 185-188 soluble mediators, 233 Hemopoietic cells, hybrid resistance and antigen expression, 400, 403, 405-407, 410 bone marrow cells, 335-351 effector mechanisms, 369, 370 in witso assays, 396 macropbages, 371, 372 marrow engraftment, 385-388 marrow microenvironment, 391, 392 NK cells, 373, 374, 376 syngeneic stem cell functions, 388-390 T cells, 380 leukemia/lymphoma cells, 361, 363 Hemopoietin 1, B lymphocyte formation and, 234 Heparin cell-mediated killing and, 295, 299 hybrid resistance and, 371 Hepatocytes, T cell activation and, 16 Herpes simplex virus, 153, 375, 376 Heterogeneity B lymphocyte formation and, 196, 216, 217 CTL and, 138, 139, 152 cytoxicity and cell-mediated killing, 312 mediation, 271 membrane attack complex of complement, 305, 306, 309, 310 hybrid resistance and, 372 T cell activation and, 27 T cell subsets and, 43, 79, 98 Histocompatibility antigens (HA) H-2 alloantigen recognition and, 82, 87 H-2 molecules in thymus and, 99, 100
H-%restricted antigen recognition and, 51, 57, 71-75 hybrid resistance and, 333, 334, 350 Histiocytes, hybrid resistance and, 335 Histotope, T cell subsets and, 53, 58 Homeostasis, B lymphocyte formation and, 185, 208, 236 Homogeneity B lymphocyte formation and, 221, 228 cytotoxicity and, 270, 284, 287, 300 hybrid resistance and, 338, 339 T lymphocyte activation and, 13 Homology CTL and, 136 exon shuffling, 140-142, 144-146 &-Microglobulin, 154, 155 monoclonal antibodies, 156, 158 cytotoxicity and, 285, 307, 309, 314, 316 human T lymphocyte activation and, 3 accessory molecules, 14 gene regulation, 29 synergy, 17 T cell antigen receptor, 4 thy-ll,lO hybrid resistance and, 335 T cell subsets and, 41, 46, 49, 50 Hormones B lymphocyte formation and, 196, 229 hybrid resistance and, 402 pore formation and, 314 Horse red blood cells, T cell subsets and, 68 Horse serum, B lymphocyte formation and, 209, 219 Human leukocyte antigen (HLA) CTL and, 168, 170 amino acid, 163-165 carbohydrate moieties, 153 exon shuffling, 138, 139, 141, 142, 144 Pz-Microglobulin, 155, 156 monoclonal antibodies, 158, 165, 167 transfected cells, 149-152 cytotoxicity and, 270 hybrid resistance and, 350, 369, 400403 Hybrid histocompatibility antigens antigen expression, 397-399, 401, 405407, 409, 410 trans gene model, 401-403 bone marrow cells, 340-343, 345, 347
459
INDEX
genetics of expression, 397 in oitro assays, 394-396
leukemia/lymphoma cells, 360-367 lymphoid cells, 352, 353 marrow engraftment, 385, 386 Hybrid hyperreactivity, 360, 379 Hybrid resistance, 333, 334 antigen expression genetics, 397-401 hemopoietic cell grafts, 404-407 Hh-1, 401-403 marrow allograft reactivity, 408-410 transplated cells, 403, 404 effector mechanisms, 369, 370 antibodies, 376-380 in oitro assays, 393-396 macrophages, 370-372 marrow engraftment, 384-388 marrow microenvironment, 390-393 N K cells, 372-376 syngentic stem cell functions, 388-390 T cells, 380-384 hemopoietic cells hone m.arrow cells, 335-351 lymphord cells, 351-358 leukemia/ lymphoma cells, 358-369 Hybridization cytotoxicity and, 276, 288 T cell subsets and, 41, 100, 101 Hybridomas B lymphocyte formation and, 223, 225, 237 CTL and, 150, 168, 169 T cell activation and, 1, 23, 25 T cell subsets and H-2 alkiantigen recognition, 79, 82, 91 H-2-restricted antigen recognition, 5260, 62, 70 T cell receptor, 41, 42 T cell triggering, 111 Hybrids B lymphocyte formation and, 199, 226 CTL and, 169 exon shuffling, 139, 142, 144, 145, 147 HLA class I antigens, 150, 151 &-Microglobulin, 156 monoclonal antibodies, 158 Hydrocortisone, hybrid resistance and, 387, 390 Hydrolysis, T cell activation and, 15, 16, 23, 25, 30
Hydrolytic enzymes, cytolysis and, 275-277, 281 Hydrophobic domains human T lymphocyte activation and, 5, 13, 25 T cell subsets and, 45, 54 Hydrophobicity B lymphocyte formation and, 197, 198 cytotoxicity and cell-mediated killing, 295 cytolysis, 279, 280 membrane attack complex of complement, 299-302 pore formers, 314-317 Hypersensitivity, B lymphocyte formation and, 183
1 Immune response genes hybrid resistance and, 361 T cell subsets and, 44, 51, 54, 55 Immunodeficiency, B lymphocyte formation and, 199, 200, 207, 216 genetically determined defects, 224-226, 231 Immunodepression, hybrid resistance and, 385 Immunofluorescence B lymphocyte formation and, 188, 221 cytotoxicity and, 275, 298 Immunoglobulin B lymphocyte formation and, 182, 183, 237, 238 B cell precursors, 188, 189, 191, 193195, 199-202 bone marrow cultures, 209, 217, 218 CBA/N mice, 226 inducible cell line, 221, 223 lymphohemopoietic tissue organization, 187 population dynamics, 206, 207 SCID mice, 225 cell-mediated killing and, 291 CTL and, 168 hybrid resistance and, 357, 365, 366, 378, 379 Immunoglobulin G B lymphocyte formation and, 223, 237 cytotoxicity and, 271
460
INDEX
hybrid resistance and, 336, 355, 357 effector mechanisms, 376-378, 394 Immunoglobulin genes human T lymphocyte antigen and, 3, 10 T cell subsets and, 40, 41, 46, 52 Immunoglobulin H, B lymphocyte formation and, 198 Immunoglobulin M B lymphocyte formation and, 183 B cell presursors, 194, 200 inducible cell line, 221, 222 population dynamics, 206 soluble mediators, 234 hybrid resistance and, 336, 377, 378 Immunoglobulin receptors, T cell subsets and, 77 Immunoprecipitation B lymphocyte formation and, 221 T cell activation and, 19 Immunosuppression, hybrid resistance and, 340, 353, 354, 386-388 Inflammatory response B lymphocyte formation and, 208, 237 lymphoid cells and, 354 Influenza virus CTL and, 136, 168 amino acid, 165 exon shuffling, 144, 146 HLA class I antigens, 151, 152 monoclonal antibodies, 151, 166, 167 hybrid resistance and, 375 Inhibition, see also Allogeneic inhibition B lymphocyte formation and, 236 B cell precursors, 91, 92 bone marrow cultures, 216 inducible cell line, 222, 223 CTL and, 168, 169 amino acid, 162 carbohydrate moieties, 152-154 exon shuffling, 139 HLA class I antigens, 149-151 monoclonal antibodies, 157, 158 cytotoxicity and cell-mediated killing, 295, 296 cytolysis, 276, 281, 282 cytolytic proteins, 317 membrane attack complex of complement, 309 hybrid resistance and antibodies, 378
antigen expression, 397, 402, 404, 409 bone marrow cells, 337-340, 347, 348 effector mechanisms, 370 in vitro assays, 394-396 leukemia/lymphoma cells, 358, 364, 367 lymphoid cells, 351, 352, 354, 356 macrophages, 371, 372 marrow engraftment, 387 marrow microenvironment, 390-392 N K cells, 374-376 syngeneic stem cells, 388, 390 T cells, 381, 382 T cell activation and, 9, 22, 26 T cell subsets and H-2 alloantigen recognition, 79, 82, 85, 86, 90,91 H-2 molecules in thymus, 109, 110 H-%restricted antigen recognition, 60, 61, 74, 76 Inositol triphosphate, T cell activation and, 15, 16, 21, 24, 26, 30 Insulin B lymphocyte formation and, 197 hybrid resistance and, 402 Interdigitating cells, hybrid resistance and, 356, 357 Interferon B lymphocyte formation and, 228, 234, 265 hybrid resistance and, 368, 377 antigen expression, 405, 406 bone marrow cells, 345 macrophages, 371, 372 marrow microenvironment, 392 syngeneic stem cells, 388-390 Interferon-? B lymphocyte formation and, 197, 222, 230 cytotoxicity and, 272, 284 hyman T lymphocyte antigen and, 6, 11, 27-29 hybrid resistance and, 371 T cell subsets and H-2 alloantigen recognition, 89, 90, 94, 95 H-2 antigen recognition, 64, 66, 69, 76, 77 H-2 molecules in thymus, 106 Interleukin- 1 B lymphocyte formation and, 192, 208, 222, 236, 237
461
INDEX
genetically determined defects, 228, 230 soluble mediators, 234, 235 cell-mediated killing and, 288 T cell activation and cell surface molecules, 2, 8, 13, 14 receptor-mediated signal transduction, 20 22, 23, 26 T cell subsets and, 63-66, 84, 85, 107, 111 In terleukin-2 B lymphocyte formation and, 193, 228, 237 cell-mediated killing and, 287, 288 human T lymphocyte activation and, 1, 6, 11-14 gene regulation, 27-300 receptor-mediated signal transduction, 22 25 synergy, 15, 19 T11, 9, 10 hybrid resistance and, 351, 380, 382, 396 T cell subsets and H-2 alloantigen recognition, 87, 93-95 H-2 molecules in thymus, 98, 99, 107 H-2-restricted antigen recognition, 52, 53, 56, 62-66, 72-74, 76, 77 T cell triggering, 111, 112 Interleukin-3 B lymphocyte formation and, 228-230, 232-234, 236, 237 hybrid resistance and, 389 Inulin, cytotoxicity and, 304, 306 Iodo-2’-deouynridine (IUdR), hybrid resistance and antigen eupression, 397, 404 bone marrow cells, 340, 342, 345, 347 effector mechanisms, 370 leukernia/’lymphorna cells, 364-368 lymphoid cells, 352, 354 T cells, 380 Ionomycin, T cell subsets and, 65, 98 Irradiation, hybrid resistance and antigen eupression, 397, 403, 404, 408410 bone marrow cells, 339, 344, 346-348, 350 effector mechanisms, 369, 370 antibodies, 376, 377, 379 in uitro assays, 394, 396
marrow, 384-386, 388, 390 NK cells, 373-375 syngeneic stem cells, 389, 390 T cells, 380-384 leukemia/lymphoma cells, 358, 359, 361, 363-367 lymphoid cells, 352, 354-357
K Keratinocytes, B lymphocyte formation and, 234 Keyhole limpet hemocyanin, T cell subsets and, 70, 71 Kidney CTL and, 149 hybrid resistance and, 338, 348, 356, 404
1 L cells B lymphocyte formation and, 198 CTL and amino acid, 163 exon shuffling, 138-140, 144, 147, 148 HLA class I antigens, 149-152 Laminin, hybrid resistance and, 403 Langerhans cells, T cell subsets and, 69, 92, 94 Large granular lymphocytes cytotoxicity and, 271, 287, 289, 291, 309 hybrid resistance and, 376, 380 Lectin B lymphocyte formation and, 201, 210 CTL and, 150 human T lymphocyte activation and, 7-9, 12, 28 T cell subsets and, 60, 76 Lesions, cytotoxicity and, 288-291, 305307, 318 Leucine, CTL and, 166 Leukemia B lymphocyte formation and, 185, 191, 193, 205, 220, 233 hybrid resistance and, 358-369 antibodies, 378, 380 antigen expression, 401, 403 bone marrow cells, 346, 348 Leukocytes B lymphocyte formation and, 192, 194, 198, 229
462
INDEX
cytotoxicity and, 270, 312 hybrid resistance and, 349, 350 T cell subsets and, 108 Leukopenia, hybrid resistance and, 388 Leukoregulin, cytolysis and, 286 Leukotrienes, hybrid resistance and, 360 LFA CTL and, 136, 139, 150-152, 154 cytotoxicity and, 271, 272 LFA-1, B lymphocyte formation and, 194 Ligand CTL and, 168, 169 carbohydrate moieties, 154 exon shuffling, 139, 142, 147 HLA class I antigens, 149-152 monoclonal antibodies, 158 cytotoxicity and, 307 T cell activation and, 30 cell surface molecules, 2, 7, 8, 10-13 gene regulation, 28, 30 receptor-mediated signal transduction, 20, 22 synergy, 17 T cell subsets and, 40, 50-53 Ligand-receptor interactions, T cell activation and, 1, 26 Lineage fidelity, B lymphocyte formation and, 195 Lipid B lymphocyte formation and, 197, 213 cell-mediated killing and, 290, 291, 294296, 312, 313 CTL and, 169 cytolysis and, 275, 278-280 membrane attack complex of complement and, 299-303, 305, 306 pore formers, 315, 317, 318 Lipopolysaccharide B lymphocyte formation and, 196, 200, 221, 222, 231 pore-forming protein and, 313 Liposomes CTL and, 154 cytotoxicity and cytolytic proteins, 319 membrane attack complex of complement, 300, 302, 305, 309 pore-forming proteins, 312, 313 Liver B lymphocyte formation and, 181, 182, 224, 225
B cell precursors, 191, 193 bone marrow cultures, 216 Ig genes, 204 lymphohemopoietic tissue organization, 187 population dynamics, 207 CTL and, 149 hybrid resistance and, 335, 351, 356, 363, 366, 373, 386 T cell subsets and, 67, 69 Liver cell grafts, hybrid resistance and, 333 Localization, B lymphocyte formation and, 195, 196 Low density lipoprotein B lymphocyte formation and, 212, 213 cytotoxicity and, 295, -307, 309 Lung, hybrid resistance and, 356, 366, 373 Lymph nodes B lymphocyte formation and, 191, 223 hybrid resistance and, 359, 384, 393 bone marrow cells, 337, 338 lymphoid cells, 352, 353, 355, 356 NK cells, 372 T cells, 381 T cell subsets and H-2 alloantigen recognition, 78, 92 H-2 molecules in thymus, 99 H-%restricted antigen recognition, 6771 Lymphoblasts, CTL and, 139 Lymphochoriomeningitis virus, hybrid resistance and, 355, 390 Lymphocytes complement-mediated cytotoxicity and, see Cytoxicity, lymphocyte and complement-mediated hybrid resistance and antigen expression, 400 bone marrow cells, 348, 350 leukemia/lymphoma cells, 361, 364 lymphoid cells, 353, 355-357 marrow, 387, 390 T cells, 380 T cell subsets and, 70, 81, 110 Lymphohematopoietic cells, T cell subsets and, 107-109 Lymphohemopoietic tissue organization, B lymphocyte formation and, 185-188 Lymphoid cells B lymphocyte formation and, 196, 213, 220, 225, 226
463
INDEX
hybrid resistance and, 333 antibodies, 376, 380 antigen expression, 401, 405 hone marrow cells, 335, 338, 339 in uitro assays, 393, 394 leukemia/lymphoma cells, 359, 369 marrow engraftment, 386 NK cells, 374, 375 normal hemopoietic cells, 351-358 T cells, 380, 383 T cell subsets and, 44, 78, 90, 96, 101, 109 Lymphokines B lymphocyte formation and, 199, 222, 231 cytolysis and, 284, 286 human T lymphocyte antigen and, 1, 6, 20 gene regulation, 27, 29 synergy, 15 T11, 9, 10 hybrid resistance and, 407 T cell subsets and H-2 alloantigen recognition, 94 H-2 molecules in thymus, 107 H-2-restricted antigen recognition, 52, 62--66, 76, 77 T cell trigger, 112 Lympholysis, hybrid resistance and, 388 Lymphoma cells B lymphocyte formation and, 202 cell-mediated killing and, 293 hybrid resistance and, 334, 358-369 antigen expression, 401, 403 NK cellk, 373, 374 Ly mphopoiesis B lymphocyte formation and, 213, 229231, 237, 238 hybrid resistance and, 339 Lymphoproliferation, B lymphocyte formation and, 193, 208 Lymphosarcoma cells, hybrid resistance and, 358 Lymphotoxin B lymphocyte formation and, 236 cytotoxicity and, 284-286, 298 hybrid re:jistance and, 400, 407 Lymphotoxin-like molecules, cytolysis and, 284-286 Lysine, cell-mediated killing and, 297
Lysosomes CTL and, 168 cytotoxicity and, 276, 281, 288
M Macrophages B lymphocyte formation and, 186, 208, 234-236, 238 B cell precursors, 191, 192, 194, 195 bone marrow cultures, 212, 213, 218 genetically determined defects, 230232 cytotoxicity and, 269, 285, 293 hybrid resistance and, 340, 356, 366 antigen expression, 402, 405 effector mechanisms, 370-372, 374, 387, 392, 393 T cell subsets and cell surface molecules, 46 H-2 alloantigen recognition, 89, 90,94 H-2 molecules in thymus, 103-106, 108-110 H-2-restricted antigen recognition, 51, 69-71, 74, 75, 77 Magnesium, cytotoxicity and, 271, 294 Major basic protein, cell-mediated killing and, 311, 312 Major histocompatibility complex (MHC) B lymphocyte formation and, 194, 197 cytotoxicity and, 270 human T lymchocyte activation and, 2, 3, 5, 14 hybrid resistance and, 348, 357, 361, 369, 378, 403 T cell subsets and, 39, 71, 78, 81, 101, 108 cell surface molecules, 43, 44, 50 class I molecules, cytotoxic T lymphocytes and, see Cytotoxic T lymphocytes Marrow engraftment, hybrid resistance and hemopoietic cells, 386-388 specific unresponsiveness, 384-386 Marrow microenvironment, hybrid resistance and, 390-393 Mast cell tumors, hybrid resistance and, 361 Melittin, pore formation and, 313-316 Membrane attack complex of complement amphiphilic nature of C5b-9, 299-304 analogues, 307-311 subunit composition, 304-307
464
INDEX
Membrane-bound IL-1, T cell subsets and, 90 Membrane damage cytolysis and, 277-281 pore formation and cytolytic proteins, 316-319 small peptides, 313-316 Metaphase, B lymphocyte formation and, 206 Methotrexate serum, hybrid resistance and, 350 Micelle, cytotoxicity and, 303 &-Microglobulin B lymphocyte formation and, 198 CTL and, 135, 138, 165, 170 exon shuffling, 138, 139, 147 HLA class I antigens, 149, 150 T cell recognition, 154-156 T cell subsets and, 44, 75 Microtubule organizing center (MTOC), cytolysis and, 275, 282 Mitochondria, cell-mediated killing and, 288 Mitogen B lymphocyte formation and, 192, 200, 228, 231 cytotoxicity and, 272, 284, 285 human T lymphocyte activation and, 7, 8, 11, 12, 26, 27 hybrid resistance and, 350, 393, 395 T cell subsets and, 60, 63-66, 111 Mitosis, B lymphocyte formation and, 184 Mixed-lymphocyte reaction, T cell subsets and cell surface molecules, 43 H-2 alloantigen recognition, 81, 83-87 APC, 88-92 H-2 molecules in thymus, 107, 108 Moloney leukemia virus, 364, 400 Moloney sarcoma-leukemia virus, 153 Moloney sarcoma virus, 378 Monensin, 281 Monoclonal antibodies B lymphocyte formation and, 181, 182, 235, 238 B cell precursors, 188, 189 bone marrow cultures, 213 cell size changes, 202 functional assays, 200, 201 Ig genes, 204 Ly-5 family of glycoproteins, 189, 192
lymphomohemopoietic tissues, 187 markers, 193, 194 NZB mice, 227 PI-linked lymphocyte antigens, 198 SCID mice, 225 tumor cell lines, 195 CTL and, 169, 170 amino acid, 163 blocking, 156-158 exon shuffling, 140, 141, 145, 148 HLA class I antigens, 150 Pz-Microglobulin, 155, 156 somatic cell class I variants, 165-167 cytotoxicity and, 285, 291, 307, 309 hybrid resistance and, 336, 374, 377, 378, 386 T cell activation and, 30 accessory molecules, 14 cell surface molecules, 1, 2 gene regulation, 28 IL-1 receptor, 14 receptor-mediated signal transduction, 19-22, 24-26 T cell antigen receptor, 2, 4-8 T1, 13 T11, 9, 10 Thy-1, 11 Tp44,ll T cell subsets and H-2 alloantigen recognition, 79, 84, 89, 93 H-2-restricted antigen recognition, 53 T cell receptor, 40, 43 Monocytopenia, hybrid resistance and, 392 Morphology B lymphocyte formation and, 182, 186, 204, 213, 219, 238 cytotoxicity and, 271 cytolysis, 274, 276, 281-283 cytolytic proteins, 317 granule proteins, 287, 291, 294, 295 membrane attack complex of complement, 300, 304, 305, 309, 311 mRNA CTL and, 148 T cell activation and, 28 T cell subsets and cell surface molecules, 41-43, 45, 47, 50 H-2 molecules in thymus, 97, 99
465
INDEX
Murine cytomegalovirus, hybrid resistance and, 3.59 Murine hepatitis virus, hybrid resistance and, 375, 376 Murine system, T lymphocyte activation and, 1 cell surface molecules, 10, 11, 13, 14 receptor- mediated signal transduction, 23, 25 Murine T cells, 64,79 Mutation B lymphocyte formation and, 185 B cell precursors, 195, 198 genetically determined defects, 224, 22!6, 228, 231, 232 CTL and, 138, 169, 170 amino acid, 158-163, 165 carbohydrate moieties, 153, 154 exon shuffling, 141, 142, 145, 148 &-Microglobulin, 155 monoclonal antibodies, 156, 157, 165, 166 hybrid resistance and, 343, 368, 374, 397 T cell activation and, 5, 8, 12, 20 T cell subsets and, 42, 46 H-2 alloantigen recognition, 81, 82, 85, 86, 88, 93 Myeloid cells B lymph,ocyte formation and, 237 B cell precursors, 189-191 bone marrow cultures, 209, 210, 212, 2113, 219, 220 cyclic neutropenia, 229, 230 inducible cell line, 220 lymphohemopoietic tissue organization, 87 NZB mice, 229 SCID mice, 225 soluble mediators, 232, 233 W/W anemic mice, 224 cytolysis and, 285 hybrid resistance and, 373, 376 Myeloma cells B lymphocyte formation and, 220 hybrid resistance and, 378 Myelopoiesis, B lymphocyte formation and, 186, 187, 213, 219 Myosin cytolysis and, 275 hybrid resistance and, 378
N Natural cytotoxic cells, hybrid resistance and, 373, 374 Natural killer cells B lymphocyte formation and, 192, 199, 225, 231, 236, 237 cytotoxicity and, 269-273 cytolysis, 275-277, 280-286 granule proteins, 286, 287, 289, 291, 297, 298 membrane attack complex of complement and, 309 hybrid resistance and, 334 antibodies, 376-380 antigen expression, 402-410 bone marrow cells, 336, 338, 340, 351 effector mechanisms, 372-376 leukemia/lymphoma cells, 361, 363-369 lymphoid cells, 354, 357 macrophages, 371, 372 marrow engraftment, 386-388 marrow microenvironment, 390-393 syngeneic stem cell functions, 388-390 T cells, 382 T cell activation and, 9, 10, 24 T cell subsets and, 49 Natural killer cytotoxic factor, 285, 296 Natural resistance, hybrid resistance and, 334 Neoplasia, hybrid resistance and, 387 Neuraminidase hybrid resistance and, 400 T cell subsets and, 64,91 Neutropenia, hybrid resistance and, 376 Neutrophil B lymphocyte formation and, 229, 230 cytotoxicity and, 269 New antigenic determinants, T cell subsets and, 51 Newcastle disease virus, 166, 167, 389 Nucleotide B lymphocyte formation and, 191, 194, 197, 203 CTL and, 159 T cell activation and, 18, 26 T cell subsets and, 42, 43, 46, 47
0 Oligomerization, cytotoxicity and, 306, 315, 317, 318
466
INDEX
Oligosaccharides, CTL and, 135, 152, 153 Osteopetrosis B lymphocyte formation and, 232 hybrid resistance and, 374, 390, 392, 393 Ovalbumin CTL and, 167, 168 T cell subsets and, 55, 70
P Parasites, cytotoxicity and, 269 Peanut agglutinin, T cell subsets and, 97 Peptide B lymphocyte formation and, 215 CTL and, 136, 137, 167-169 amino acid, 162 exon shuffling, 148 HLA antigens, 152 cytotoxicity and, 279, 280, 309, 313 pore formation and, 313-319 T cell subsets and, 41, 47, 110 H-%restricted antigen recognition, 5255, 57, 58, 74, 75, 77 T lymphocyte activation and, 3 Perforin, cytotoxicity and, 280, 281, 291, 305, 307-309 Phagocytosis B lymphocyte formation and, 212, 213 hybrid resistance and, 340, 357, 370 pore-forming proteins and, 313 Phenotype B lymphocyte formation and B cell precursors, 192, 201 Ig genes, 204, 205 lymphohemopoietic tissues, 187 NZB mice, 227, 229 W/W anemic mice, 224 cytotoxicity and, 273, 287 hybrid resistance and, 338, 374, 382, 400, 401 T cell subsets and cell surface molecules, 49, 50 H-2 alloantigen recognition, 84, 87, 88 H-2 molecules in thymus, 96 H-%restricted antigen recognition, 60, 69, 73 T lymphocyte activation and, 1, 7, 28 Phenylalanine, CTL and, 162, 166 Phorbol esters B lymphocyte formation and, 222 cell-mediated killing and, 14-19, 24, 26
Phorbol myristic acetate, T cell activation and, 8, 9, 11, 12, 14, 19 gene regulation, 28, 30 receptor-mediated signal transduction, 20, 22, 23, 25 Phosphates CTL and, 153 T cell activation and, 16, 24 Phosphatidylcholine, cytotoxicity and, 275, 276 Phosphatidylinositol, B lymphocyte formation and, 197, 198, 202, 222, 239 bone marrow cultures, 215, 216 Phosphatidylinositol hiphosphate, T cell activation and, 15, 16, 23, 25 Phosphatidylinositol-specific phospholipase, B lymphocyte formation and, 198, 214, 215 Phospholipase B lymphocyte formation and, 198, 216, 239 cytotoxicity and, 300, 314 Phospholipase A, cytolysis and, 275, 276 Phospholipase Az, pore formers and, 314316 Phospholipid cytotoxicity and, 276, 300, 302, 303 pore formation, 314, 315, 318 T cell activation and, 15-19 Phosphorylation CTL and, 135, 148 cytotoxicity and, 273 T lymphocyte activation and, 4, 16, 18, 19, 23, 24 Phytohemagglutinin (PHA) B lymphocyte formation and, 198 CTL and, 50 hybrid resistance and, 359, 361 T cell activation and, 8, 14, 22, 26-28 T cell subsets and, 60, 63 Pinocytosis, B lymphocte formation and, 212, 213 Plaque-forming cells, hybrid resistance and, 352 Plasmacytonias, hybrid resistance and, 367, 368 Polarity cytolytic mechanisms and, 273-275, 282 T cell subsets and, 41 Polyinosinic: polycytidylic acid, hybrid resistance and, 371-372, 388-390
INDEX
Polymerization cell-mediated killing and, 290, 292-295, 298 membrane attack complex of complement and, 304, 305, 309, 311 Polymorphism B lymphocyte formation and, 193, 197, 223, 224, 229 CTL and, 135-137, 169 amino acid, 158-163, 165 exon shuffling, 141, 142 HLA class I antigens, 151 monoclonal antibodies, 158, 165 hybrid resistance and, 367, 400, 401, 405 T cell subsets and cell surface molecules, 41, 44-47 H-2 alloantigen recognition, 182, 183 H-2-restricted antigen recognition, 58, 59, 62 Polypeptides B lymphocyte formation and, 190 cell-mediated killing and, 291, 292, 298, 313 CTLand, 168 cytolysis and, 276, 279, 285 cytotoxicity and, 309, 319 pore formers and, 316, 317 T cell activation and, 10, 17, 23 T cell subsets and, 41, 43, 50, 51, 75 Polyperforin 1, cytotoxicity and, 280, 289 Polyperforin 2, cytotoxicity and, 280 Polyribosomf:s, B lymphocyte formation and, 1828 Poly-2-vinylpyridine-N-oxide (PVNO), hybrid resistance and, 371 Population dynamics, B lymphocyte formation and, 205-208 Pore formation, cytotoxicity and, 313-320 Pore-forming protein (PFP), cytotoxicity and, 269, 270, 319 cell-mediated killing, 291-299 cytolytic, 311-319 membrane attack complex of complement, 299, 305, 307-309, 311 Potassium cytotoxicity and, 294, 318 Priming hybrid resistance and, 370 T cell subsets and, 72, 74, 100 Procarbazine serum, hybrid resistance and, 350
467
Proliferation B lymphocyte formation and, 183, 184, 186, 187 B cell precursors, 192, 200, 202 bone marrow cultures, 210, 212 genetically determined defects, 225, 229, 231, 232, 234 inducible cell line, 220 population dynamics, 205-207 cytotoxicity and, 269 hybrid resistance and, 384, 390, 391, 397 bone marrow cells, 335, 336, 338, 340, 349 lymphoid cells, 352, 353 T cell subsets and H-2 alloantigen recognition, 85, 86, 88, 92 H-2-restricted recognition of antigen, 51, 52, 60, 62-66, 70, 72 T lymphocyte activation and, 1, 6, 7, 9, 10, 12-15, 26 Promonocytes, hybrid resistance and, 371 Promyelocytic leukemia cells, hybrid resistance and, 348 Pros taglandin B lymphocyte formation and, 208, 236 hybrid resistance and, 360, 387 Proteases, cytotoxicity and, 282, 298 Protein B lymphocyte formation and, 205, 216, 239 B cell precursors, 191, 197, 198 inducible cell lines, 221, 222 CTL and, 135, 148, 153, 154, 159, 165, 168 cytotoxicity and, 269, 270, 320 cell-mediated killing, 289-295, 297, 311, 312 cytolysis, 277-281, 316-319 mediation, 270-273 membrane attack of complement and, 299, 301, 303, 307, 309-311 small peptides, 314 human T lymphocyte activation and, 1, 4, 8, 18, 23 gene regulation, 29, 30 IL-I receptor, 13, 14 hybrid resistance and, 393 T cell subsets and, 90,97, 99 cell surface molecules, 45-47, 50
468
INDEX
H-2-restriction antigen recognition, 52, 54, 56-58, 62, 70, 74, 75 T cell receptor, 41-43 Protein kinase C B lymphocyte formation and, 222 cell-mediated killing and, 296, 297 T cell activation and, 8, 31 Ca2+, 15-19 gene regulation, 28-30 receptor-mediated signal transduction, 19, 20, 23, 25, 26 T cell subsets and, 65 Proteoglycans, cell-mediated killing and, 298 Proteoliposomes, cytotoxicity and, 274, 278, 303 Proteolysis cytotoxicity and, 275, 276, 301, 305 T cell activation and, 18 T cell subsets and, 53
R Radioactivity, hybrid resistance and, 349 Radiautoautography, B lymphocyte formation and, 205 Radioresistance, hybrid resistance and, 365, 376, 380, 383, 390, 393, 409 Radiosensitivity, hybrid resistance and, 350, 368 Reactive oxygen metabolism intermediates, cytolysis and, 286 Receptor-mediated signal transduction, T cell activation and, 19-26 Red blood cells, hybrid resistance and, 337, 342, 349, 351, 352, 359 Replication, B lymphocyte formation and, 183-185, 236 B cell precursors, 200 bone marrow cultures, 210, 213, 216, 217 genetically determined defects, 228, 231 population dynamics, 206, 207 soluble mediators, 232-234 Retrovirus B lymphocyte formation and, 195, 233 hybrid resistance and, 378, 400, 403 RNA B lymphocyte formation and, 191, 198, 22 1 hybrid resistance and, 377, 80 T cell activation and, 27
s Sarcomas, hybrid resistance and, 358, 359, 363 Salt, cytotoxicity and, 290, 301 Saccharomyces cereuisiae, cytolysis and, 319 Second messengers B lymphocyte formation and, 197 T cell antigen receptor and, 2, 26, 27 Serine, CTL and, 135, 148 Serine esterases, cytotoxicity and, 267, 288, 297, 298, 310 Serine proteinases, cytotoxicity and, 277, 297, 310 Serology, CTL and, 139-141 Severe combined immunodeficiency disease (SCID) B lymphocyte formation and, 224-226, 237 hybrid resistance and, 351, 365, 366, 379381, 384 antigen expression, 404, 405, 408, 409 Sheep red blood cells cytotoxicity and, 293 hybrid resistance and, 352 T cell subsets and, 67-71, 77, 90, 92 Silica, hybrid resistance and, 350, 368, 371373, 395 Skin allografts, T cell subsets and, 86, 88, 93 Skin graft B lymphocyte formation and, 225 CTL and, 158 hybrid resistance and, 369, 376, 383 Skin graft rejection CTL and, 159, 165 T cell subsets and, 71, 88, 93, 94 Somatic hypermutation, T cell subsets and, 42, 88, 106 Specific unresponsiveness, hybrid resistance and, 384-386 Spleen B lymphocyte formation and, 181, 182, 184, 238 B cell precursors, 198, 200-202 genetically determined defects, 229 inducible cell line, 223 population dynamics, 207, 208 CTL and, 138 hybrid resistance and, 334 antigen expression, 397, 402, 404, 408
469
INDEX
bone marrow cells, 335, 336, 338-340, 342, 345, 347, 348 effector mechanisms, 370 in vitro assays, 394-396 leukemia/lymphoma cells, 358, 361-368 lymphojid cells, 351-356 marrow, 384-388, 391, 393 NK cells, 372, 373, 375 syngeneic stem cells, 388-390 T cells, 380-384 T cell subsets and H-2 akiantigen recognition, 78, 86, 89, 91, 92 H-2 molecules in thymus, 95, 102 H-2-restricted antigen recognition, 67, 69, 72 Spleen colony formation, hybrid resistance and antigen expression, 402, 404 bone marrow cells, 337, 347 syngeneic stem cells, 388-390 T cells, 362 Staphylococcal LX toxin, cytolysis and, 317 Staphylococcal &toxin, cytolysis and, 315, 316 Stem cells B lymphocyte formation and, 181, 182, 184, .187, 206, 237 B cell precursors, 188, 191, 195, 199, 200 bone marrow cultures, 209, 217, 219 genetically determined defects, 220, 224, 225 soluble mediators, 232-234 hybrid resistance and, 352, 362, 367 antigen expression, 402-407 bone m.arrow cells, 336-340, 344, 346 effector mechanisms, 370-373, 377380, 383, 386, 388-391 Steroids B lymphocyte formation and, 209, 219, 236, 238 hybrid resistance and, 348, 387 Streptolysin 0, 317, 318 Stroma, hybrid resistance and, 336, 404 Stromal cell!;, B lymphocyte formation and, 184, 185, 237, 238 B cell precursors, 198, 200, 203 bone marrow cultures, 209, 210, 212-216, 219, 220
genetically determined defects, 227, 229, 231, 232, inducible cell lines, 223 lymphohemopoietic tissue organization, 186-188 soluble mediators, 234 Subclones B lymphocyte formation and, 210 CTL and, 139 Sucrose, cytotoxicity and, 290, 304 Supergenes, cytotoxicity and, 310 Suppression, hybrid resistance and, 380, 381, 384, 387, 389, 395 Supressor cells, hybrid resistance and, 340, 368, 410 marrow, 385-387, 391, 392 Suppressor T cells, T cell subsets and, 102, 107, 110 Synergy B lymphocyte formation and, 222, 234 cytotoxicity and, 284, 296 hybrid resistance and, 406 T cell activation and, 20, 23, 28, 30 Ca2+ ionophores, 15-19 cell surface molecules, 2, 12, 14 Syngeneic preference, hybrid resistance and, 334, 358 Syngeneic stem cell functions, hybrid resistance and, 388-390
T T cell B lymphocyte formation and, 236 B cell precursors, 189, 191, 192, 198, 199 genetically determined defects, 226, 227, 230, 231 Ig genes, 204 soluble mediators, 233-235 CTL and, 165 cytotoxicity and, 271-273, 277, 297 hybrid resistance and, 337, 364366, 368 antigen expression, 402-405, 408-410 effector mechanisms, 372-374, 376388, 392, 395 T cell activation, 1, 30, 31 cell surface molecules, 1-15 gene regulation, 26-30 intracellular signals, 26
470
INDEX
receptor-mediated signal transduction, 19-26 synergy, 15-19 T cell antigen receptor, 2-8 T cell receptor B lymphocyte formation and, 194, 199, 225, 226, 237 CTL and, 136, 137, 139, 152, 167, 168, 170 cytotoxicity and, 271, 272, 274 T cell subsets and, 40-43, 46, 111 H-2 alloantigen recognition, 79-81, 88 H-2 molecules in thymus, 97, 99, 106, 107 H-2-restricted antigen recognition, 5154, 56, 60-65 T cell subsets in mouse, 39, 110-13 cell surface molecules, 40-50 H-2 alloantigen recognition, 78 alloreactivity, 78-83 antigen-presenting cells, 88-92 effector phase, 92-95 resting T cells subsets, 83-88 H-2 molecules in thymus, 95, 96 development, 96-99 restricted T cells, 99-107 tolerance induction, 107-110 H-2-restricted antigen recognition, 51, 52 effector phase, 75-77 T accessory molecule function, 59-62 triggering of activated T cells and hybridomas, 52-59 triggering of unprimed and resting T cells, 62-75 T helper cells cytotoxicity and, 274, 297 T cell subsets and, 40 H-2 molecules in thymus, 101-103, 105 H-%restricted antigen recognition, 67, 70-72, 75, 77 T killer cells cytolysis and, 280 hybrid resistance and, 382, 383, 400 T cell subsets and, 40, 54 T lymphocytes B lymphocyte formation and, 205, 212, 225 hybrid resistance and, 353, 364 T cell activation and, 17, 19, 21, 22 T cell subsets and, 44
T suppressor cells, hybrid resistance and, 379, 386, 410 Target cells CTL and, 137, 158, 162, 168 carbohydrate moieties, 153, 154 exon shuffling, 142, 144, 149 HLA class I antigens, 149-152 cytotoxicity and, 269-272, 300, 319, 320 cell-mediated killing, 294, 296-298, 312 cytolysis, 273-284 hybrid resistance and, 363, 369, 372, 373, 383 antibodies, 376-378 antigen expression, 400, 404 in vitro assays, 393-396 marrow, 387, 391, 392 T cell subsets and, 76, 94, 95 T lymphocyte activation and, 1, 14, 30 Thiol-activated lysins, 317, 318 Thoracic duct lymphocytes, hybrid resistance and, 356-358 Thymocytes B lymphocyte formation and, 198, 199 cytolysis and, 276, 279 hybrid resistance and, 337, 376, 409 T cell subsets and, 96-98, 105-108, 112 T lymphocyte activation and, 4, 8, 10, 24, 25 Thymus B lymphocyte formation and, 191, 223, 236-238 bone marrow cultures, 212, 214 genetically determined defects, 225227, 231 population dynamics, 207, 208 hybrid resistance and, 337, 352 effector mechanisms, 372-374, 384, 390-392 T cell subsets and, 43, 44, 112 H-2 molecules, 95--10 H-2-restricted antigen recognition, 51, 59, 69 Tolerance induction hybrid resistance and, 356, 361, 409 T cell subsets and, 107-110 Total lymphoid irradiation, hybrid resistance and, 386, 387 Transcription B lymphocyte formation and, 192, 198, 204, 205, 221, 222, 226
471
INDEX
cytotoxiciiy and, 271, 276, 277, 297 hybrid resistance and, 380, 403 T cell subsets and, 41, 43, 48, 97 T lympho'cyte activation and, 3, 27-31 Transfection B lymphocyte formation and, 198, 205 CTL and, 138, 139, 144, 149-152, 163 cytotoxicity and, 272 T cell subsets and, 58, 59, 62, 80, 83 T lymphocyte activation and, 3, 5, 8, 11, 25 Transferrin, B lymphocyte formation and, 193, 212 Translocation, T cell activation and, 17-19, 22, 23 Transmembrane channel, cell-mediated kill ing and, 298 Transmembrane domain, CTL and, 135, 146-148 Trypsin, 2101, 227, 297, 298 Tubulin, 275, 378 Tumor B lymphocyte formation and, 186, 204, 230, 233, 237 B cell precursors, 188, 193, 194, 196 bone m.irrow cultures, 214, 217 inducible cell line, 220, 221 lineage fidelity, 195, 196 CTL and, 135, 138, 139, 154 cytotoxicity and, 270, 284-286, 293, 313 hybrid resistance and, 338 antigen expression, 397, 402-404 effector mechanisms, 375, 378, 380, 382, 387, 393, 394, 396 leukemia/lymphoma cells, 358-369 T cell activation and, 6 T cell sub4jets and, 74, 91
Tumor necrosis factor, 235-237, 285, 298, 374, 400, 401 Tumor necrosis serum, 406 Tyrosine, CTL and, 135, 148, 162
V Venous sinuses, B lymphocyte formation and, 187, 214, 236 Vesicular stomatitis virus, 142, 148, 153, 154, 163, 168 Virus B lymphocyte formation and, 209 CTL and, 135, 142, 153, 154, 165 cytotoxicity and, 269, 270 hybrid resistance and, 375, 378, 390, 400, 402, 403 T cell subsets and, 54, 56, 57, 74-76, 101
W White blood cells, hybrid resistance and, 349, 356 Whitlock-Witte cultures, B lymphocyte formation and, 209, 216-220, 231
Y Yeast, cytotoxicity and, 269, 270, 319
Z Zinc, cell-mediated killing and, 293, 294
This Page Intentionally Left Blank
CONTENTS OF RECENT VOLUMES
Volume 3’1
Protein A of Staphylococcus aureus and Related immunoglobulin Receptors Produced by Streptococci and Pneumonacocci JOHN J. LANGONE
The Regulatory Role of Macrophoges in Antigenic Stirnulotion Part Two: Symbiotic Relationship between Lymphocytes and Macrophoges E M I L R. U N A N U E
Regulation of Immunity to the Azobenzenearsonote Hapten MARKI. GREENE, MITCHELL J . NELLES, MAN-SUNS Y , A N D ALFRED NISONOFF
T-cell Growth Factor and the Culture of Cloned Functional T Cells KENDALLA. SMITH A N D F R A N C I S w. RUSCETTI
Immunologic Regulation of Lymphoid Tumor Cells: Model Systems for Lymphocyte Function ABUL K. ABBAS
Formation of B Lymphocytes in Fetal and Adult Life P A U L W. KINCADE Structural Aspects and Heterogeneity of ~mmunoglobdinFc Receptors JAY c. UNKELESS, HOWARD FLEIT, A N D IRAS . MELLMAN
INDEX
The Autologous Mixed-Lymphocyte Reaction MARC E. WEKSLER, CIIAHLES E. M o o i w , Jn., A N D ROBERT W. KOZAK
Volume 33 The CBA/N Mouse Strain: An Experimental Model Illustrating the Influence of the X-Chromosome on Immunity IRWINSCHER
INDEX
Volume 32 Polyclonal B.Cell Activators in the Study of the Regulation of immunoglobulin Synthesis in the Human System THOMAS A. WALDMANNA N D SAMUEI. BHODEH
The Biology of Monoclonal Lymphokines Secreted by T Cell Lines and Hybridomas AMNONALTMANA N D DAVIDH . KATZ
Typing for Human Alloantigens with the Prime Lymphocyte Typing Technique N. MORLINC, B. K. JAKOBSEN,P. PLATZ, L,. P. RYDER, A. SVEJGAARD, A N D M . THOMSEN
Autoantibodies to Nuclear Antigens (ANA): Their lmmunobiology and Medicine ENGM. TAN The Biochemistry and Pathophysiology of the Contact System of Plasma CHARLES G . COCHRANE A N D JOHN H. GRIFFIN
473
474
C O N T E N T S OF R E C E N T VOLUMES
Binding of Bacteria to Lymphocyte Subpapulations MARIUSTEODORESCU ANI) EUGENEP. MAYER
INDEX
Immunoglobulin RNA Rearrangements in B Lymphocyte Differentiation JOHN ROGERSA N D RANDOLPH WALL Structure and Function of Fc Receptors for IgE on Lymphocytes, Monocytes, and Macrophages HANSL. SPIEGELBEHC
Volume 34 T Cell Alloantigens Encoded by the IgT-C Region of Chromosome 12 in the Mouse F. L. OWEN Heterogeneity of H-2D Region Assaciated Genes and Gene Products KEIKO OZATO, A N D TED H. HANSEN, DAVIDH. SAC115 Human Ir Genes: Structure and Function THOMAS A. GONWA,B. MATIJA PETEHLIN,A N D J O H N D. STOBO Interferons with Special Emphasis on the Immune System A N D STEFANIE ROBEHTM. FHIEDMAN N. VOCEL Acute Phase Proteins with Special Reference to C-Reactive Protein and Related Proteins (Pentaxins) and Serum Amyloid A Protein M . B. PEWS A N D MARILYN L. BALTZ Lectin Receptors as Lymphocyte Surface Markers NATHAN SHARON
INDEX
Volume 35
The Murine Antitumor Immune Response and Its Therapeutic Manipulation ROBERTJ. NORTH Immunologic Regulation of Fetal-Mater-
rial ~~l~~~~ DAVIDR. JACOBY, LARSB. OLDIN(;, MICIIAELB. A. OLDSTONE
AND
The Influence of Histamine on Immune and Inflammatory Responses J. BEER,STEVENM. DENNIS, MATLOFF,A N D Ross E. ROCKLIN
INDEX
Volume
36
Antibodies of Predetermined Specificity in Biology and Medicine RICIIAHIJALANLEHNEH A Molecular Analysis of the Cytolytic Lymphocyte Response STEVENJ. BUHAKOFP, OPHA WEINBEHGEH, ALAN M . KHENSKY,A N D CAHOLS. REISS
The Human Thymic Microenvironment BARTON F. HAYNES The Generation of Diversity in Phosphorylcholine-Binding Antibodies Aging, ldiotype Repertoire Shifts, and ROGER M . PEHLMUTTEH, STEPHENT. Compartmentalization of the Mucosal-AsCHEWS,RICIIARDDOUGLAS, GREG SORENSEN, NELSONJ O H N S O N , NADINE sociated Lymphoid System ANDREW W. WADEA N D MYRONR J. GEARHART, AND NIVERA,PATRICIA LEROYHOOD SZEWCZUK
CONTENTS OF R E C E N T V O L U M E S
A Major Role of the Macrophage in Quantitative Genetic Regulation of Immunoresponsivenessand Antiinfectious Immunity GUIDOBIOZZI,DENISEMOUTON, CLAUDESTIFFEL,A N D YOLANDE BOUTIIILLIEH
INDEX
Volume 37 Structure, Function, and Genetics of Human Class II Molecules ROBERTC. GILESA N D J. DONALD CAPHA
475
Volume 38 The Antigen-Specific, Major Histocompatibility Complex-Restricted Receptor on T Cells PHILIPPAMARRACK A N D JOHN
KAPPLER Immune Response ( I r ) Genes of the Murine Major Histocompatibility Complex RONALDH. SCHWARTZ The Molecular Genetics of Components of Complement A. D. CAMPBELL, M. C. CARROLL, A N D R. R. PORTER Molecular Genetics of Human B Cell Neoplasia CARLOM. CROCEA N D PETERC. NOWELL
The Complexity of Virus-Cell Interactions in Abelson Virus Infection of Lymphoid and Other tiematopoietic Cells CIIEHYL A. WIIITLOCK AND OWEN N . WllTE
Human Lymphocyte Hybridomas and Monoclonal Antibodies DENNISA. CARSONA N D BRUCED
Epstein-Barr Virus Infection and Immunoregulation in Man GIOVANNA TOSATOA N D R. MICHAEL BLAESE
Maternally Transmitted Antigen JOHNR. RODGERS, ROGER SMITH111, MARILYN M. HUSTON,A N D ROBERTR. RICH
The Clossiccil Complement Pathway: ACtivation ond Regulation of the First Complement Component NEIL R. COOPER Membrane Complement Receptors Specific for Bound Fragments of C3 GORDOND. Ross A N D M. EIIWARD MEDOF Murine Models of Systemic Lupus Erythemator,us ARGYHIOS N . THEOFILOPOULOS A N D FRANK .I. DIXON
INDEX
FREIMARK
Phagocytosis of Particulate Activators of the Alternative Complement Pathway: Effects of Fibronectin JOYCEK. CZOP
Volume 39 Immunological Regulation of Hematopoietic/Lymphoid Stem Cell Differentiation by lnterleukin 3 JAMESN. IIILEA N D YACOB WEINSTEIN Antigen Presentation by B Cells and Its Significance in T-B Interactions ROBERT W. C I I E S N U T A N D HOWARD M . GREY
476
C O N T E N T S OF R E C E N T V O L U M E S
Ligand-Receptor Dynamics and Signal Amplification in the Neutrophil LARRY A. SKLAH Arachidonic Acid Metabolism by the 5Lipoxygenase Pathway, and the Effects of Alternative Dietary Fatty Acids TAKH. LEE A N D K. FRANK AUSTEN The Eosinophilic Leukocyte: Structure and Function GEALV J. GLEICll A N C i l E H Y L R. ADOLPIISON ldiotypic Interactions in the Treatment of Human Diseases RAIF S. GEHA Neuroimmunology DONALD G. PAYAN, JOSEPII P. MCGILLIS, A N D EDWAHV J. GOETZL
Index
Volume 40 Regulation of Human B Lymphocyte Activation, Proliferation, and Differentiation DIANEF. JELINEKA N D PETERE.
LIPSKY
Biological Activities Residing in the Fc Region of Immunoglobulin EDWAHDL. MoRAN A N V W I L L I A M 0. WEICLE Immunoglobulin-Specific Suppressor T Cells RICIIAHDG . LYNCII Immunoglobulin A (IgA): Molecular and Cellular Interactions Involved in IgA Biosynthesis and Immune Response JIHI MESTECKY A N D JEHHYR. MCGIIEE The Arrangement of Immunoglobulin and T Cell Receptor Genes in Human Lymphoproliferative Disorders TIIOMAS A. WALVMANN Human Tumor Antigens RALPH A. REISFELD A N D DAVIVA CIIEHESII Human Marrow Transplantation: An Immunological Perspective PAULJ. MARTIN,JOIIN A. HANSEN, RAINEHSTOHB,A N D E. DONNALL TIIOMAS
Index