ADVANCES IN
Immunology
V O L U M E 30
CONTRIBUTORS TO THIS VOLUME
M. J. BUCHMEIER HENRYN. CLAMAN PAULJ. CONLON F. ...
10 downloads
866 Views
20MB 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
V O L U M E 30
CONTRIBUTORS TO THIS VOLUME
M. J. BUCHMEIER HENRYN. CLAMAN PAULJ. CONLON F. J. DUTKO FRANCIS LOOR STEPHEND. MILLER JOHN W. MOORHEAD M. B. A. OLDSTONE WILLIAM0. WEIGLE R. M. WELSH
ADVANCES IN
Immunology E D I T E D BY
FRANK J. DIXON
HENRY
Scripps Clinic and Research Foundation La Jalla, California
G. KUNKEL
The Rockefeller University N e w Yo&, N e w Yo&
V O L U M E 30
1980
ACADEMIC PRESS A Subsidiary of Harcoud Bmce Jovanovich, Publishers
New York
London
Toronto
Sydney
San Francisco
COPYRIGHT @ 1980, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY F O R M 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.
111 Fifth Avenue, New York, New York 10003
United Kingdom Edition published by ACADEMIC PRESS, INC. ( L O N D O N ) LTD.
24128 Oval Road, London N W I 7DX
LIBRARY OF
CONGRESS CATALOG CARD
NUMBER: 61 - 17057
ISBN 0-12-022430-5 PRINTED IN THE UNITED STATES OF AMERICA
80 81 82 83
98 7 65 4 3 2 1
CONTENTS LIST OF CONTFUBUTORS ..................................................... PREFACE....................................................................
vii ix
Plasma Membrane and Cell Cortex interactions in Lymphocyte Functions
FRANCIS LOOR I. Introduction ....................... 11. Structural Relationships of the Plasm ... .............................. Cell Cortex.. ..... 111. Uropode Formation and Capping ...................................... IV. Microvilli Formation and Shedding.. ................................... V. Cell Cortex Control of Recognition Phenomena That Take Place at the Plasma Membrane (PM) Level ................................... VI. Cell Cortex and Plasma Membrane Functions in the (Mitogenic) Activation of Lymphocytes . ....................... dies ....................... VII. Activation of Lymphocytes b VIII. To Cluster but Not to Cap-Is That What Triggers? ..................... .................... IX. About Activation of Nonlymphoid Cells X. Concluding Remarks . . . . . . . . . . . . . . . . . . .................... References . . . . . . . . . . . . . . . .......................................
1
2 5 8 12 31 72 78 87 100 102
Control of Experimental Contact Sensitivity
HENRYN. CLAMAN, STEPHEND. MILLER,PAULJ. CONLON, AND JOHN
w. MOORHEAD
I. Introduction .......................................................... 11. Contact Sensitivity: Induction and Elicitation ........................... 111. Control of Contact Sensitivity .......................................... IV. Tolerance to Contact Sensitivity.. ...................................... References ..........................................................
12 1 122 130 133 153
Analysis of Autoimmunity through Experimental Models of Thyroiditis and Allergic Encephalomyelitis
WILLIAM0. WEICLE
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Mechanism of Self-Tolerance .......................................... 111. Types of Acquired Immunologic Tolerance ............................. IV. Relationship between Experimentally Induced Tolerance and Self-Tolerance: Implications in Autoimmunity .......................... V
159 161 162 178
vi
CONTENTS
V. Experimental Autoimmunity ........................................... VI . Concluding Remarks .................................................. References ...........................................................
184 251 253
The Virology and lmrnunobiology of Lymphocytic Choriorneningitis Virus Infection
M . J . BUCHMEIER. R . M . WELSH.F. J . DUTKO. AND M . B . A. OLDSTONE I . Prologue ............................................................. I1 . Virus and Host CeIl Interactions ....................................... 111. LCMV-Induced Acute Immune Response Disease ...................... IV LCMV-Induced Chronic Immune Response Disease ..................... V. Epilogue ............................................................. References ...........................................................
275 282 302 313 325 326
.....................................................................
333 337
.
INDEX
CONTENTS OF PREVIOUS VOLUMES..........................................
LIST OF CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors’ contributions begin.
M. J. BUCHMEIER,Department of Zmmunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037 (275) HENRYN. CLAMAN, Departments of Medicine and of Microbiology and Immunology, University of Colorado Medical School, Denver, Colorado 80262 (121) PAULJ. CONLON, Departments of Medicine and of Microbiology and Immunology, University of Colorado Medical School, Denver, Colorado 80262 (121)
F. J. DUTKO,Department of Zmmunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037 (275) FRANCIS LOOR,*Basel Institute for Immunology, Basel, Switzerland, Sandoz Preclinical Research Department, Basel, Switzerland, Zmmunology Unit, Faculty of Medicine, State University, Mons, Belgium, and Zmmunology Department, Faculty of Sciences, Louis Pasteur University, Strasbourg, France (1) STEPHEND. MILLER,Departments of Medicine and of Microbiology and Immunology, University of Colorado Medical School, Denver, Colorado 80262 (121) W. MOORHEAD,Departments of Medicine and of Microbiology and Immunology, University of Colorado Medical School, Denver, Colorado 80262 (121)
JOHN
M. B. A. OLDSTONE,Department of Zmmunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037 (275) WILLIAM 0. WEIGLE, Department of Zmmunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037 (159)
R.M . WELSH,^ Department of Zmmunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037 (275)
* Present address: Departement d’Immunologie, Institut de Chimie Biologique, Universite Louis Pasteur, 11 rue Humann, F 67085 Strasbourg Cedex, France. f Present address: Department of Pathology, University of Massachusetts Medical School, Worcester, Massachusetts 01605. vii
This Page Intentionally Left Blank
PREFACE
The subjects presented in this volume emphasize the breadth of immunology in the areas of science to which it contributes and from which it draws information and technology. In this interdisciplinary exchange, it is becoming apparent that the cellular and humoral events involved in immunologic reactions are similar in many ways to their counterparts in other systems. The lymphocyte has become one of the most successfully exploited subjects of cell biology and that exploitation has ultimately shed light on immunologic processes. Consideration of the part played by interactions between the plasma membrane and cell cortex in lymphocyte function brings the cellular events of immune responses within the framework of cellular activities in general. Only recently have the elaborate, multiple control mechanisms governing immunologic responses begun to be appreciated, and these also seem to fit within the general pattern of controls operating elsewhere. Analysis of controls operating in the contact sensitivity response indicates that they include the same elements, with perhaps some variation in emphasis, that operate in other immune responses. The discussion of self-tolerance and possible reasons for its breakdown with consequent development of autoimmunity also deals with the same regulatory processes that monitor usual immune responses. Even the manifestations of autoimmune disease are produced by a perversion of ordinarily protective mechanisms operating either singly or in combination. The description of host-viral relationships in LCM infections indicates the complexity of mechanisms related both to viral propagation and to host defense which interact to determine the nature of the largely immunologic disease that results. Surfaces of lymphocytes and macrophages in large part determine and control the multiplicity of cellular interactions that characterize immune function. The organization and physiology of cell surfaces are in turn determined not only by intrinsic plasma membrane markers and receptors but also by the state of the underlying cell cortex which contains the cytoskeleton and cytomusculature. This critical interrelationship between cell cortex and plasma membrane and its role in determining immunologic function are reviewed in the first chapter of this volume by Dr. Francis Loor. Two general kinds of plasma membrane interactions are considered in detail. The first is an active process of motion and/or recognition initiated in the cortex and usually involving formation of uropods and microvilli where receptors are ix
X
PREFACE
concentrated. The second is passive recognition in which resting lymphocytes are exposed to ligands that bind to and cluster receptors thereby perturbing the cells’ plasma membranes and initiating the biochemical events of activation. In the second chapter Drs. Claman, Miller, Conlon, and Moorhead discuss controls operative in the expression of experimental contact sensitivity, drawing on their own considerable work in this field. The events leading to contact sensitivity including the essential reactivity of the antigen with self molecules, the involvement of Langerhans cells of the skin, and the eventual development of effector T lymphocytes are clearly presented. This chapter then focuses on several mechanisms influencing the extent of a sensitivity response. Controlling factors include route and amount of antigen presentation, which to a great degree determine whether sensitization or tolerance will develop, production of soluble suppressor substances that may accompany tolerogenic regimes of antigen administration, formation of inhibitory antiidiotypic antibodies, and development of T suppressor cells. These multiple regulatory mechanisms operating in contact sensitivity are quite similar to those involved in other cellular immune responses as well as in humoral responses and are in keeping with the general elaborate regulation controlling immunity. The relationship between the normal state of self-tolerance and its abnormal corollary, autoimmunity, is presented in the third chapter by Dr. Weigle, a long time authority on this subject. The several possible bases for loss of self-tolerance, including failure of immune regulation and abnormal presentation of potential self-antigens, are considered in terms of current understanding of cellular and humoral immune processes. A particularly thorough evaluation of suppressor cells and their possible role in self-tolerance and autoimmunity is given. Finally, the pathogenic mechanisms that may be involved in autoimmune disease are examined in the light of two familiar experimental models, allergic encephalomyelitis and autoimmune thyroiditis. The pathogenic potential of cellular autoimmune responses in encephalomyelitis, on the one hand, and of autoantibody responses in thyroiditis, on the other, indicate the very different pathogeneses which may operate in autoimmunity. Study of LCM viral infections has led to a surprising number of conceptual immunologic advances. It was initially postulated that infection occurring prior to maturation of the immunologic system might culminate in tolerance to the infecting virus via clonal elimination of responsive lymphocytes-a concept later extended to account for normal self tolerance. Eventually, research showed that complete im-
PREFACE
xi
munologic tolerance could not be induced by viral infections initiated at any time and that the diseases accompanying LCM infection were caused by host-T cell responses in acute infection and antibody-LCM virus immune complexes in chronic infections. In the fourth chapter Drs. Buchmeier, Welsh, Dutko, and Oldstone, who have contributed greatly to this subject, present in molecular terms the explanation for immunologic events accompanying LCM infection, including the dual recognition of viral and histocompatibility antigens essential for T cell action and the modulation of viral expression resulting from the immune reponse. Finally, they discuss the important but less well appreciated aspects of nonimmunologic regulation of viral infection particularly the role of defective interfering virus. We wish to thank the authors for their efforts and care in the preparation of their chapters and the staff of Academic Press for their usual expeditious preparation of the volume.
FRANK J. DIXON HENRYG. KUNKEL
This Page Intentionally Left Blank
ADVANCES I N IMMUNOLOGY. VOL. 30
Plasma Membrane a n d Cell Cortex Interactions in Lymphocyte Functions FRANCIS LOOR’ Bosel Institute for Immunology, Basel, Switzedond
I. 11. 111. IV. V.
VI.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structural Relationships ofthe Plasma Membrane (PM) with the Cell Cortex Uropode Formation and Capping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microvilli Formation and Shedding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cell Cortex Control of Recognition Phenomena That Take Place at the Plasma Membrane Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . A. Spontaneous and Lectin-Induced Lymphocyte Aggregation . . . . . . . . . . B. Rosetting .... C. Migration of Lymphocytes from Blood to Lymphoid Tissue . . . . . . . . . . D. Macrophage-Lymphocyte Interactions . . . . . . . . . . . . . . . . . . . . . . .... . . .. .............. E. Killer-Target Interactions . . . . . . . . . . . . . .. . Cell Cortex and Plasma Membrane Functions i enic) Activation of Lymphocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Effects of Classical Drugs That Affect the Cytomusculature . . . . . . . . . . . B. Effects of Classical Drugs That Affect the Cytoskeleton . . . . . . . . . . . . . . C. Effects of Other “Factors” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Activation of Lymphocytes by Anti-Ig Antibodies . . . . . . . . . . . . . . . . . . . . . .. To Cluster but Not to Cap-Is That What Triggers? . . . . . . . . . . . . . . . About Activation of Nonlymphoid Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Fibroblasts.. . ........... .................
.
.
..
VII. VIII. IX.
.
. . .. .. . . . . ..
...................
1 2 5
8 12 12 15 16 19 23
31 33 35 40 72 78 87 87 87
C. Basophils and Mast Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 100 . . . . . . . . . . . . . . ......................... X. Concluding Remarks References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 .I..
I. Introduction
Structural and dynamic aspects of the lymphocyte surface physiology have been the subject of a few recent extensive reviews (125,363, 438,439, 570). During the last few years a number of observations have shed new light on our comprehension of the phenomena. Various structures in the lymphocyte cortex influence the organization of the lymphocyte surface, and this is especially evident for the expression of microvilli (MV), the formation of the cap and the uropode, and gross cell locomotion. The influence of the lymphocyte cortex can be de1 Present address: Dbpartement d‘hmunologie, Institut de Chimie Biologique, Universiti. Louis Pasteur, 11 rue Humann, F 67085 Strasbourg Cedex, France.
1 Copyright @ 1980 by Academic Ress, Inc. All rights of reproduction in any form reserved. ISBN 0-12422430-5
2
FRANCIS LOOR
duced from the effects of drugs known to interfere with its activity, from observations with the electron and fluorescence microscopes, and more recently from the biochemical analysis of lymphocyte surface components. Here I will attempt to summarize the most critical progress made in the understanding of the organization of the plasma membrane and of the cell cortex. This will be developed in extenso elsewhere, together with the building of a hypothetical model explaining how interactions of plasma membrane components and cell cortex elements may result in the gross morphological changes frequently associated with cell activation (367). My purpose here is to review the experimental evidence showing that both the plasma membrane organization and the activity of the cell cortex elements are implicated in the immunological function of the lymphoid cells. II. Structural Relationships of the Plasma M e m b r a n e (PM) with the Cell Cortex
All eukaryotic cells seem to contain filaments and tubules, i.e., a cytomusculature (Cm) and a cytoskeleton (Cs). Thin microfilaments (MF, 5-8 nni diameter, made of actin) and microtubules (MT, 24 nm outside diameter, made of tubulin) are regularly detected, but thick filaments (13-25 nm diameter, made of myosin) usually are not, and intermediate filaments (10 nm diameter) are as yet not enough studied to decide their ubiquitous distribution (196). Even if all Cs and Cm elements were present in all types of eukaryotic cells, there could be important differences in both their content and their organization. As an example, actin represents u p to 20-30% of the protein content of mobile cells, but only 1-2% in some tissue cells (327). Furthermore, even though polymorphism of the two major proteins, actin and tubulin, is limited, there can be important heterogeneity, even within a single cell, of the content, distribution and organization of actin-made M F and tubulin-made MT. For instance, in fibroblasts, two distinct sets of M F seem to exist, which have different properties and assume different functions; similarly, not all the MT of a cell show the same sensitivity to a drug-induced depolymerization (363, 364). The cell has indeed multiple tools to control the state of assembly and the activity of such structures. Among these are, in the case of MF, the myosins, which are more polymorphic (663), and the accessory proteins such as the tropomyosin, the troponins, the filamin, the profilin (72), and in the case of MT, the not yet well defined microtubule associated-proteins (MAPS) found along MT (595, 596). If, in the cell, M T and M F show some physical association, such an interaction
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
3
would probably be regulated by the accessory proteins. At least, in vitro, there is evidence for it (222). It is, however, still difficult to integrate the presently known data (for review, see 367) into a general model of cell motility, and to understand fully those activities of lymphocytes that depend on various kinds of membrane movements. Particularly, the information available for lymphocytes and, more generally, leukocytes is still limited. In mobile eukaryotic cells, the principal constituent of MF, actin may represent 10-30% of the total protein of the cell (327).Actin is a major component of the normal lymphocyte (27); it represents some 6%of the total protein of a guinea pig B cell leukemia line (160).Actin is usually a major component of purified PM of lymphocytes (26),and the amount of PM-bound actin may represent some 10% of the total cellular actin (168). Myosin can represent up to some 5% of the total protein of nonmuscle cells (468). In polymorphonuclear leukocytes, a value of 2.4%of total protein has been reported (623). To my knowledge, there are no data for lymphocytes in general, but in a particular type of lymphocyte (acute lymphoblastic leukemia) it represents only some 0.63%of the total cell protein (468), and in another type (B cell leukemia cell line) it represents from 0.68% to 0.83% of total cell protein (160). Other proteins, such as a-actinin, troponiyosin, and troponins (having structural and regulatory functions in muscular fiber contraction), have, not been quantitatively determined in nonmuscle cells, but some of them have been detected in lymphocytes. As far as M T are concerned, their major constituting proteins, the tubulins, represent 0.5%to a maximum of 1.0%of the total cell protein (594). The possible association of tubulin with the PM is still controversial (226). The high molecular weight proteins associated with MT have been detected in fibroblasts (596). Recently developed methods of immunofluorescence (IFM) have allowed the detection of some of those Cs/Cm proteins in lymphocytes, although most extensive studies have so far been performed with fibroblasts, epithelial cells, macrophages, polymorphonuclear neutrophils, etc. (e.g., 203). The methods make use of various specific antibodies (anti-actin, anti-tubulin, and so on) that may show different specificities from batch to batch, thus possibly explaining some discrepancies. Two disadvantages of these methods are their relative lack of sensitivity when directly labeled antibodies are used (direct IFM), and the difficulty of trying to look simultaneously at two different CslCm and membrane components when indirect IFM is used. Moreover, although there is more actin than myosin in the cells, the specific antibodies usually give rather poor detection of actin (see 150, 154),
4
FRANCIS LOOR
compared to an effective detection of myosin, which may be due to a difference of immunogenicity, actin being more evolutionarily conserved than myosin. An elegant nonantibody method can be used to detect actin: it consists of detecting actin through the binding of biotinylated heavy meromyosin, which is itself revealed by binding of fluoresceinated avidin (which binds avidly to biotin) (254).This allows great sensitivity of actin detection, possibly simultaneously with detection of another Cs/Cm or PM component by sensitive indirect IFM. The Cs/Cm elements can be detected only after fixation and lipid extraction to make the membrane permeable for the reagents, but no binding is detected on viable cells, showing that no antigenic determinants specific to Cs/Cm components are exposed on the cell surface. There are some controversial cases, however, that were reported for myosin exposed on the outside of the cell membrane (discussed in 327). In at least one case, the cell surface component that was detected b y some of the antibodies in the anti-myosin serum appeared, however, to be unrelated to myosin (142). In resting, rounded lymphocytes, all Cs/Cm components that have been detected so far (actin, u-actinin, myosin, and tubulin) were distributed homogeneously as a faint rim at the periphery of the cell (142, 693). Thus, in contrast to the spectacular IFM detection of Cs/Cm components in the flat interphase fibroblast, one does not find in resting lymphocytes splendid microtubular networks or microfilamentous stress fibers. However, when the interphase fibroblast is rounded, most of the IFM staining that detects M F and MT is then diffuse, with some concentrated at the periphery, near the surface in the cortical cytoplasm of the cell (184). Resting lymphocytes are thus similar to rounded interphase fibroblasts in that respect. Alterations of such homogeneous distribution of Cs/Cm proteins in the lymphocyte cortical cytoplasm can be observed under two main circumstances: ( a ) when the lymphocyte develops long microvilli (MV), either spontaneously or induced, actin is found within the MV (151-153, 693-695); and (b) when the lymphocyte shows a uropode, either spontaneously or induced, actin is found to be more concentrated within the uropode, together with myosin and tubulin (572,634) (see Sections I11 and IV). Before IFM methods became popular for the detection of Cs/Cm proteins, electron microscope methods had already shown the way. Thin M F filaments 4-6 nm in diameter had regularly been observed under the PM of lymphocytes, forming a kind of narrow, dense network (124). Such thin MF are likely to be made of actin. Thick bundles of MF were also found within MV and pseudopode-like structures of
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
5
lymphocytes (692). Early studies on capping had also indicated the particularly high concentration of M F in the tail of lymphocytes capped with anti-Ig antibodies (510). Thick filaments made of myosin have not been reported to occur in lymphocytes, but a few filaments 15 nm thick and 300 nm long were observed in glycerinated horse neutrophil leukocytes (587), and similar filaments form in vitro from the myosin extracted from a B cell leukemia line (160). Lymphocytes at 37°C have numerous MT assembled, most or perhaps all of which would be connected to the centrioles (125). Apart from these, the distribution of the few MT in the cytoplasm of lymphocytes appears to be random, rather than specifically associated with the membrane (125, 692). Although some MT have been observed running under the PM (124, 692), they remain separated from it by the cortical layer of M F (124).With the exception of MT, which are found in the uropode, MT do not seem to be associated with the PM (124). Rare MT were observed to extend perpendicularly to the PM (125,692).The total number of M T that terminate after the PM in a cell was estimated to be probably one hundred at most (125). Their presence within “microspikes” (629) or within each MV present on the uropode of the motile lymphocytes (399) had been reported in the past, but the nature of these structures (MT or MF) remains controversial. Treatments or drugs known to affect MT integrity do not interfere with the expression of lymphocyte surface structures such as MV (368, 444, 445). If anything, drugs like colchicine and vinblastine actually increase the development of a uropode and the extent of capping. Conversely, the various cytochalasins (A-E), which inhibit M F function in different ways, can abolish the emergence of MV on lymphocytes (as well as on other cells) (127, 152,368,444,445)and they can inhibit the capping in varying degrees (123,124,361,362,368,371,639),cytochalasin A in my experience showing most inhibition, and cytochalasin C showing the least (F. Loor and L. Angman, 1980, E x p . Cell Res., in press). 111. Uropode Formation and Capping
The general characteristics of the capping phenomenon, especially its requirement for energy, its inhibition by low temperature, its inhibition and its reversal by cytochalasin, and the accumulation of M F under the cap, have led to the suggestion that capping is probably a contractile phenomenon: it would be mediated by contractile M F bringing patches of aggregated membrane components to an area of the cell where they would be endocytosed and digested, or be shed (54, 55, 125, 363, 365, 371, 438, 570). The role of MT appears to be
6
FRANCIS LOOR
more complex. Apparently, the motility of M F would not be synergized, but rather antagonized, by MT, whose role is to establish cell shape and to maintain its compartmentalization. Evidence for the inhibitory role of M T comes essentially from extensive experience with concanavalin A (Con A) capping. In some circumstances (reviewed in 125, 363, 438,439, 570), Con A does not cap; on the contrary, it even inhibits capping of other cell surface macromolecules by their own antibodies. Such inhibition is abolished b y treatments that depolymerize MT (141, 362, 368, 374, 646, 691, 692). Along the same lines, mIg or Con A sites cap slowly on CBA mouse lymphocytes and rapidly on A mouse lymphocytes, but after colchicine treatment, cells from both mouse strains show fast capping (173). In the Beige mouse strain, fast capping of Con A on polymorphonuclear leukocytes occurs in absence of drugs that depolymerize MT (457).That this were due to a genetic defect of MT polymerization (457) has been questioned (174),but it remains possible that some MT only present anomalies of polymerization or that the association of MT with the PM is altered in Beige mouse leukocytes. A synergy for capping inhibition was obtained by simultaneous use of colchicine or vinblastine with cytochalasin (123, 124, 646). In view of more recent data reviewed below, this phenomenon probably results from M T disruption that provokes a complete reorganization of the Cm leaving in the cell cortex surrounding the nucleus much less MF, which would be readily inhibited b y cytochalasin. A clue for understanding whether M T antagonize or synergize with M F for capping could be found u priori in their overall structural location in the cell. As indicated above, IFM detection of MT and M F does not help that purpose very much. In the flat interphase fibroblasts, such methods indicate that MT and M F have a different distribution; i.e., most of them are not localized in the same parts of the cells and, u fortiori, not in the same parallel fiber system or other kind of network (184, 255). However, their organization in rounded fibroblasts is not known and may eventually be revealed to show interactions. Recently a number of researchers have looked for the possible existence of transmembrane linkages between Cs/Cm elements and PM components exposed on the outside. These links may occur either directly between a transmembrane glycoprotein [such as HLA (656)] and the Cs/Cm elements, or indirectly via other integral membrane proteins if the PM component exposed at the outer cell surface does not cross the membrane [such as mIg (656)], or indirectly via Cs/Cmassociated elements, such as the a-actinin molecules associated with M F or perhaps the MAPS (MT-associated proteins), but this is purely
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
7
speculative. The first attempts of this kind did not show any concomitant redistribution of a-actinin, of cytoplasmic myosin, and of actin in lymphocytes when capping of H-2, Con A binding sites, or mIg was induced on the cell surface (142). However, there is now a substantial amount of evidence to the contrary. mIg capping on mouse B cells cocaps tubulin and actin, which are now found under the mIg cap in the cytoplasm of the uropode (185), a similar result being obtained when myosin is detected (572).Cocapping of Con A and actin has also been observed for fibroblasts (634),for rabbit neutrophils (459),and for human leukocytes (6). Immediately the question arises whether cocapping actually demonstrates a linkage of the surface components with the Cs/Cm structures that are coredistributed in the uropode, or whether it has some trivial explanation, e.g., that capping stimulates uropode formation. Whichever is the case, most of the lymphocyte cytoplasm flows from its thin rim distribution around the nucleus to the uropode, bringing with it all cell organelles, including Cs/Cm proteins. The fact is that a single capping process seems to bring all the actin, myosin, and tubulin that is detectable by IFM into the cap/uropode area, with practically no Cs/Cm proteins detectably left in the rest of the cell. Hence, one can induce a lymphocyte to make two or three successive cappings of PM components at short time intervals. This may be d u e to recruitment, to rapid recycling of the myosin. Moreover, the number of cells showing cocapping of myosin and mIg consistently represents only 70-80% of all cells showing mIg capping (572).Thus, there can be capping of PM components without detectable coredistribution of Cs/Cm components (142,572).Inversely, there are always a proportion of cells spontaneously forming a uropode where W C m proteins detectable by IFM accumulate (6-12% of T cells) (572). Capping of Con A on leukocytes (depleted from M T by drugs) occurs at one cell pole on a protuberance that consists of highly plicated membrane sustained by a network of densely packed MF, each plication containing a bundle of M F (6).The formation of such a structure (which I would call a uropode studded with microvilli) is not induced by the Con A, but simply follows the disassembly of MT per se. The migration of M F does not bring detectably more membrane Con A binding sites on the uropode, and cells having a preformed uropode can still redistribute their Con A sites initially dispersed on the whole cell (6). This shows that the bulk of M F or M T found in the uropode is not involved, or at least is not needed for dragging liganded membrane components to the cap. This provides a possible explanation for the synergy between colchicine and cytochalasin for capping inhibition:
8
FRANCIS LOOR
in presence of colchicine, only few M F are left in the cortical cytoplasm of the cell, and the effects of cytochalasin can be more dramatic. On T cells, not only the capping of two different PM components (H-2 and T 25) coincides with the redistribution of myosin and actin as subcaps, but there is also a coincidence at the earlier stage of patching: the intracellular PM-associated actin or myosin becomes clustered into submembranous patches located directly under the external PM component patches (54,55).Similar results are obtained for the patching and capping of Con A on HeLa cells; these cells are larger than lymphocytes and allow a better distinction by IFM of PM bound and of cytoplasmic actin. When actin is detected by use of the sensitive fluoresceinated avidin-biotinylated myosin method, it then appears that only a fraction of the actin is redistributed as subcaps under the Con A caps and a substantial amount of the intracellular actin remains in the rest of the cell (54). Such association of subpatches of actin and myosin with patches of external PM components is obtained for a variety of different PM components in various metabolism-inhibited cells (16, 17, 54-56). It shows that external clustering of PM components by multivalent ligands leads to an energy-independent association with mechanochemical proteins prior to redistribution of the latter in the uropode, and it suggests that these Cs/Cm proteins actually induce movement of the patches into the cap area. Not all the actin is brought to the cap, and it would have been surprising indeed if all of a protein that represents 10-20% of total cell protein, and 10%of which is PM associated, were required for a process such as capping. IV. Microvilli Formation and Shedding
Besides capping, microvilli (MV) formation and dynamics is one of the most remarkable activities of the cell cortex. The expression of these labile structures by lymphocytes is extremely variable. The causes of such variability are not all established; most of them seem to depend on environmental conditions (reviewed by 363, 531; see also 696), but others depend on the level of differentiation of the cell and the given phase of the cell cycle (148, 151,322).In uitro, it is possible to modulate MV expression by various means. Expression of M V increases in the presence of a metabolic inhibitor such as 10 mM NaN3 (127, 128,368,411) and generally when the level of ATP of the cell is decreased by use of inhibitors of respiration and glycolysis (127, 128). The presence of cytochalasin B in the medium can completely abolish the expression of MV on lymphocytes (368, 445) as well as on other cells (152, 444, 445). When cells expressing numerous long MV are treated with the drug, first their M V collapse, but later a number of
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
9
membranous “blebs” appear on the cell surface (368).On the contrary, colchicine or vinblastine (even at M ) d o not detectably modify preformed MV (368, 445) or NaN,-induced increase of MV expression on cells that have been processed at physiological temperatures (368). Processing lymphocytes at 0°C (in an ice bath) results in decreased MV expression as compared to processing at 37°C (353, 368). Chilled lymphocytes are able to reexpress MV in culture, but drugs affecting MT interfere with any recovery from chilling (368). However, this does not necessarily suggest an involvement of M T in the process of MV reformation; it may be due to the drastic changes caused by the drugs in the normal distribution of M F in the cell cortex, i.e., the redistribution of most of them in the uropode (see Section 111);it may also be due to membrane properties of the drugs: both colchicine and vinblastine are molecules that are “cup fonners” for erythrocytes (576); another such molecule, chlorpromazine, can also reduce MV formation, although that local anesthetic has no effect on MT assembly; the inhibition of MV formation by the cup-former molecule can be counterbalanced by the addition of TNP, a crenator molecule (although this one does not, by itself, significantly increase MV expression) (592). Thus, agents that could modulate curvature can modulate MV expression, but the mechanism of action is far from being clear. Do such agents modulate M F activity (e.g., via their effect on Ca2+availability), or do they modulate membrane structure in such a way that M F cannot interact with it to form MV? Do they act at all on the M F themselves? Another local anesthetic, lidocaine, which is also a cup former, cationic amphipathic molecule (576), was recently shown to bind not to MF, but to MT, and even competitively with colchicine (375). The most likely explanation remains, however, that MV expression is essentially controlled by M F activities. Immunofluorescence studies on lymphocytes from normal peripheral blood and from lymphoid lines established in culture have further substantiated that the MV substructure is made of actin MF. When using anti-actin antibodies no binding is detected on live cells, but when their membrane is made permeable b y fixation and lipid extraction, then actin is detected inside MV (151, 153). Such actin is organized as bundles of M F that are revealed in each MV by thin-section electron microscopy (692). As far as actin MF-associated proteins are concerned, very little is known as to lymphocytes, but it is probably similar to what is found in other cells. Thus, the microvilli that constitute the brush border of various epithelial cells (intestine, kidney) contain little or no tropomyosin or filamin (61),or myosin (53, 61), although these proteins are detectable in the whole cell (61). Brush border MV contain, besides actin as a major
10
FRANCIS LOOR
component, an approximately ten times lesser amount of a protein that has been tentatively identified as a-actinin, but may not be so (52,61, 62). This would fit the model (363) showing only actin and a-actinin in the MV tips, which was based on electron microscope observations (416,417).Among the other proteins, myosin molecules are detected; they are not in the MV itself, but they seem to be present at their bases (418), and, when isolated, they can form in nitro bipolar filaments of -300 nm by 11 nm. These filaments are similar to those observed by electron microscopy at the MV bases (417). Microfilaments 15 nm thick and of similar length (-300 nm) have been reported to occur in neutrophil leukocytes (587), and they were also obtained in vitro from the polymerization of the myosin isolated from those cells (587), as well as from lymphocytes (160).It is therefore permissible to speculate that short bipolar myosin M F might be involved in the expression and mobility of the bundles of actin MF. The lymphocytes may express no MV, but on the contrary exhibit a smooth surface, for reasons that are still largely undefined (reviewed in 363,531). It seems that factors such as the osmolarity of the medium and the type of ligand used to label the PM can modulate MV expression. Depending on the circumstances, one can find the ligand bound to MV on a villous surface, or homogeneously distributed on a smooth cell surface or PM invaginations, or included in a number of pinocytosis vesicles. Formation of MV and formation of pinocytosis vesicles are mutually exclusive, but in both cases the ligand is not restricted to these structures, but rather is present also on the smooth cell surface, although in lesser amounts (363, 368). An asymmetrical distribution of MV at one pole of the motile lymphocyte has been frequently reported (127, 128, 151, 303, 353, 368, 399, 523, 617): “The greater portion of the cell is devoid of MV whereas the tenninal portion of the uropod is studded with MV” (399). Moreover, a flow of MV toward one pole of the lymphocyte was occasionally reported to occur in conditions where no specific attempt to destroy M T had been made (127,303,368,523,617).Such a migration is reversibly inhibited by the presence of NaN3 and by cytochalasin B, but not by colchicine or vinblastine (368). Actually, the drugs that destabilize MT favor the formation of the uropode and the segregation of the MV on that part of the cell (5, 127). Interestingly, when pinocytosis vesicles, rather than MV, are formed, one can also observe a similar tendency to migrate to the cell pole, a migration that can be reversibly inhibited by azide or cytochalasin (368). The similarity between the process of segregation of MV or pinocytosis vesicles to one cell pole and the segregation of spots to the
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
11
cap suggests a common mechanism, even to the extent that all these types of redistribution may be the same phenomenon (368).Th’is contention is further supported by the fact that mIg are more concentrated on the MV than on the rest of the cell body (128,356,368,523) and up to sixfold in azide-treated cells (128). Such an observation, together with the ease with which one can confuse MV or pinocytosis vesicles in IFM with real spots (368),may explain the cases of “selective spontaneous capping” of mIg that occur on those few B cells that spontaneously form a uropode (571),whose number is increased in lipopolysaccharide (LPS)-treated cells in culture (571) and even more when M T are depolymerized by drugs (126). In contrast to mIg, other membrane components analyzed so far have been found to be distributed equally well on the membranes of both the MV and the cell body. Those other components d o not represent a restricted molecular species: they are heterogeneous speciesspecific antigens detected by anti-lymphocyte serum (152, 153, 445), Con A binding sites (127,368),and phytohemaggutinin (PHA) binding sites (127),and differences in the density of some of them would not b e easily detectable. Actually, in contrast to their segregation to MV, a similar presence of mIg on both smooth and villous parts of the membrane has also been reported by some authors (357, 411, 523). I have already discussed possible reasons for such discordant results which concern the natural distribution of some membrane antigens: it can be affected by a number of environmental factors, and by the activity of the cell cortex and the process of cell membrane growth (363, 364). I would try to explain the segregation of mIg on MV membrane as follows. Increased MV expression (due to increased actin polymerization into M F bundles) would be associated with an (increased?) phase separation of the membrane lipids, leading to differential fluidities on MV and on the cell body. B cell mIg would be a type of integral membrane glycoprotein whose solubility would be very much dependent on membrane lipid composition and fluidity, and they would tend to localize on MV, when these are expressed. This interpretation fits with the observation that mIg on MV are still mobile, since they can still be clustered by divalent Ab in the plane of the membrane (128, 680). Incidentally, at the electron microscope level there is no possible confusion of such “spots” of PM components, with the larger “patches” that can form when cross-linking ligands bridge and agglutinate several adjacent MV (128, 129, 368). The segregation of mIg on MV is directly relevant to their process of shedding from the B cell membrane. It would explain why shedding of mIg is faster than shedding of other B cell membrane components
12
FRANCIS LOOR
(reviewed in 363,364). Shedding may occur by pinching off MV (654), and there is substantial evidence that this is a real, physiological process for multiple cell types (reviewed in 363, 364). Shedding of B cell mIg complexed with anti-Ig was observed on capped lymphocytes, and other cases of shedding of other lymphocyte membrane components have been reported (363, 364). Thus, if the material shed is constituted of bits of MV, one can expect to find mIg associated with the actin of M F within the material, as was recently shown (168). In other cell types with a different membrane composition, there will be other membrane antigens that are more soluble in the MV membrane than in the cell body membrane. Whenever such segregation exists, the shed material will be found to be enriched in a particular membrane component, e.g., on the very villous P815 cells. (323). Conversely, membrane components that would usually segregate to the cell body membrane would usually be depleted or absent from the shed material. A more detailed analysis of structural aspects of the lymphocyte surface organization will appear elsewhere (367). V. Cell Cortex Control of Recognition Phenomena That Take Place at the Plasma Membrane (PM) level
There is substantial evidence from a variety of systems that cells identify each other by means of MV. Interactions between “inductor” cells and “inducible” cells in differentiation phenomena, in embryology, are commonly seen as mediated by virtue of such microprojections of the cell surface. Lymphocytes too use MV to establish contact with other lymphocytes and with other types of cells. This is shown by morphological observations with optical and electron microscopes, and it can be also inferred from the effects of drugs that interfere with the Cs/Cm elements, on such recognition. Thus, lymphocyte MV, either distributed on the whole cell periphery or localized on the uropode, are involved in a variety of intracellular interactions, some of which have an evident function in immunological phenomena. A. SPONTANEOUS AND LECTIN-INDUCED LYMPHOCYTE AGGREGATION
The amount of available data concerning lymphocytes is rather limited, but the mechanism of cell agglutination has been the object of extensive studies on various types of fibroblasts (438, 439). There is still much controversy about the suggested involvement of MV in the agglutination ofcells by lectins (361,496),for cells showing numerous MV were not always found to be more easily agglutinable
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
13
than cells with few MV (363, 364,438). With lymphocytes, it appears that, provided the dose of lectin used to agglutinate them is low enough, lectin-induced agglutination can be efficiently inhibited, or even blocked, when cytochalasin B is present at the initiation of the process (45,46, 195,262,361). Actually, a fraction of the lymphocytes is usually found in the form of small clusters in absence of any added lectin, and such a spontaneous aggregation of the cells is also inhibited by the drug (361). Such clustered cells express more MV than isolated cells, as revealed by scanning electron microscopy (SEM) (29). Both spontaneous aggregation and lectin-induced agglutination are also inhibited by NaN, (361).Although cytochalasin B was shown to destroy MV, NaN, was found to increase MV expression (128, 368). It was suggested (363,364) that MV provide a useful or necessary, but insufficient, step for successful aggregation: the ATP depletion caused by NaN, may result not only in increased M V expression, but also in a blockage of their retraction so that they do not bring the contacted cells close to each other as a cluster. The involvement of MV in the initiation of cell agglutination by low, mitogenic doses of lectins is further supported by the observation of an increased MV expression on cells treated by such low doses (49, 103, 151, 368). Hence, larger doses of lectin that do not stimulate M V formation (368) also agglutinate the cells, but this process is not detectably sensitive to NaN, or cytochalasin B (361).This shows that it proceeds on a purely passive basis, i.e., a physical bridging of different cells by lectin molecules, rather than on a biologically active process (361). I believe that much of the controversy on the involvement of MV in the agglutination of fibroblasts and other cells comes from the fact that the doses of lectins that are usually engaged to agglutinate the cells are so high (100-200 pglml) that they can also agglutinate fixed cells or glycoprotein-coated beads. When cells are coated with such large amounts of lectins, any active aggregation process (cell-to-cell recognition?) would be blurred by a fantastic passive agglutination process (simple physical cross-linking by polymeric ligands), however villous the cells can be. When the cells are made more villous by various tricks (depletion of ATP, increase in CAMPcontent), still it would be useful to check that such induced MV are still mobile and physiologically functional, not just some kind of spikes in a frozen, rigid state. Even when low doses of lectins have been used, the cell-cell associations in the agglutinates may show large flat areas of opposed membrane (436), rather than the “tangles of interdigitated MV at points of cell-cell contact” (643, 678). A clue for such a discrepancy may be
14
FRANCIS LOOR
found in that agglutination is a dynamic phenomenon: The initiation of the intercellular contracts is mediated by MV (38, 505). The MV may connect cells which are far apart (distant of several diameters) and become shorter as the cells come closer to each other (38). The final aggregate of packed cells may show no MV left. If MV have such a “mechanical” role of bringing distant cells together, the reasons why both metabolic inhibitors and agents that affect Cm integrity can block aggregation are easy to imagine, and they are different. The cytoagglutination of Novikoff tumor cells was studied with variable doses of lectins; the inhibitory effects of metabolic inhibitors of ATP formation and of M F active drugs were also evident at low lectin doses (198). Finally, and though this is not clearly related to MF, it has been found that the aggregation of lymphocytes by Con A was strongly M ) (162),a local anesthetic inhibited by chlorpromazine (at -5 x known to abolish MV on lymphocytes (592). There is one more point to be stressed: the agglutination (induced by lectins) can be completely reversed by its ligand (or at least by a sugar having a high affinity for that lectin) only if it is added shortly after the initial formation of multicellular aggregates; thereafter, there is irreversibility of the lectin-induced agglutination by the sugar, even in conditions where most of the bound lectin can be eluted in that way from the surface of the agglutinated cells (70, 437,545).This suggests either that the lectin molecules involved in cell-cell bridging cannot be eluted because of a protective environment or of a particularly well fitting determinant conferring very low dissociation rate to the lectin, o r - q u i t e a different suggestion-that no lectin molecule is involved in the bridging, i.e., that other cell surface components are the “adhesive” molecules that keep the cells attached to each other: lectin binding to the cell surface would only “induce” the expression of such sticky molecules or their segregation on the cell surface, e.g., to MV tips. In support of this idea is the observation that, within minutes after binding, the surface-bound lectin is rapidly withdrawn from the MV (67, 615, 643, 670). If a-actinin is concentrated in MV tips, this could indicate that adhesion sites are indeed located on the overlying PM, by analogy to the higher concentrations of a-actinin under the PM of the fibroblast at the points of adhesion with the substrate. The existence of “adhesive sites” is just one among various possible forces that would keep the cells together, independently from those contributed by the lectins that have initiated the cell-to-cell bridging. Among these, one could suggest the involvement of some “membrane lectins” (312, 313), of surface charge, or of hydrophobic interactions between components on adjacent cells (reviewed in 438,657) and also
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
15
gross cell surface deformability with formation of areas of close opposition between adjacent cells as recently observed (441). Such gross cell surface deformability will be influenced very much by a number of treatments, such as metabolism inhibition, destruction of the cytomusculature or of the cytoskeleton, alteration of the cholesterol/ phospholipid ratio that controls membrane viscosity, etc. In a study performed with erythrocytes enriched or depleted in cholesterol to increase or decrease membrane viscosity, the agglutinability by lectins was decreased or increased, respectively (7).As erythrocytes have a membrane made much more rigid than the lymphocyte membrane b y a cortical network of spectrin (608),they do not develop microvilli, and it seems that in this case a difference of agglutinability may indeed depend on a different ability of cells to adapt to the shape of one another to allow appropriate alignment of ligands giving efficient cell-cell cross-bridging.
B. ROSETTINC The constitution of T cell rosettes [E rosettes, nonimmune specific, between human T cells and sheep erythrocytes (SRBC)] activates the emergence of MV (39, 291, 351, 352, 494, 495). The MV increase in number and size in the area ofcontact between lymphocytes and erythrocytes only on rosetting cells, not on the others (291, 351). At variance with rosettes formed through complement-component and complenient-receptor, which show broad contact zones (39), nonimmune T cell rosetting is mediated mainly through point contacts (39, 495), the T cell membrane and the SRBC membrane remaining separated by a gap of 50 nm or more (39). T cell rosetting works better at low temperatures (39, 311): twothirds of the T cells form rosettes at +4"C or +24"C, but only onefourth can rosette at 0" or at 37°C (311). Colchicine and vinblastine M ) d o not affect rosetting efficiencies (101, (even at high doses, 311), but cytochalasin B (101, 311), cyanide (60), and azide (39) cause its reversible inhibition. Both need to be present during the incubation of the T cells with SRBC. Thus, factors that typically affect MV function (MV-mediated adherence) inhibit the nonimmune T cell rosetting. I suggest that the decreased rosetting capacity observed at 37°C is not due to decreased binding, but rather to increased loss of bound erythrocytes, by some shedding process. Along these lines are the recent observations that Zn2+ions in the M range markedly enhance the E rosetting capacity of human T lymphocytes, with a retardation of the temperature-dependent spontaneous decay and of the capping (400). The effects of Znz+are analyzed in Section V1,C.
16
FRANCIS LOOR
Finally, rosette formation is inhibited when the lymphocytes are treated with an agent that reacts with amino groups (fluorescamine), but it is increased when they are treated by a sulfhydryl binding reagent (p-chloromercuriphenylsulfonate) (639). Although neither of these agents penetrates through the membrane, both may affect cell cortex activities, but it is presently still unclear what are their mechanisms of inhibition or of enhancement.
c. MIGRATIONOF LYMPHOCYTES FROM BLOODTO LYMPHOID TISSUE B and T lymphocytes home in different compartments of the lymphoid organs, and to do so they have to leave the peripheral blood by passing through the walls of the blood vessels and migrate within the lymphoid tissue. Thus, cell locomotion is expected to be an absolute requirement for such a chemotaxis. Further, since B cells and T cells go to their respective homing area, they must be “told” by the local environment which way to go, where to leave the blood. This implies recognition by the lymphocytes of signals from the environment; for example, some membrane components on the cells that make the blood/tissue barrier or some gradient of soluble chemotactic agent, and it further implies the existence on the lymphocyte membrane of appropriate receptors to recognize such signals. Thus, many experiments were performed trying to alter the lymphocyte surface by various enzymic treatments or by blind folding membrane components with a variety of ligands and looking for the pattern of such treated cells upon in vivo reinjection (e.g., 179-181). One of the first such attempts indicated that trypsin-sensitive structures on the lymphocytes were important for their appropriate homing in the body (682). These studies, however, told little about the mechanisms of such a migration. In the lymph node this happens at the level of the postcapillary venules (PCV). Studies of lymphoid tissue in situ by transmission (TEM) and scanning (SEM) electron microscopy reveal interesting features of such a migration. While circulating in the blood, the lymphocytes (both B and T) are covered with microvilli (MV) (647).In the PCV they adhere to the high endothelial cells (HEC) (which form the PCV wall) by means of their MV, usually redistributed on a “uropode” (647). The MV form contact points that resist hydrodynamic and osmotic shearing forces, but require divalent cations, intact surface glycoproteins on the lymphocytes, and an intact endothelial glycocalyx (11). The lymphocytes lose their MV and assume a motile configuration while they traverse the PCV wall between the HEC (11, 647). They are smooth when arriving in the nodal cortex. Lymphocyte adherence
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
17
through MV, focusing toward the PCV wall, suggests that MV may play a role in the recognition of HEC. Since B and T cells leave the PCV at different sites to migrate directly in their respective homing area, there could be differently recognizable B or T specific structures on the HEC (perhaps in their glycocalyx). The recognition of gradients of specific chemotactic factors is another possibility, since a macromolecular marker (peroxidase) injected within the lymphatic stroma becomes established as a decreasing gradient from stroma to blood, especially in the area around the lymphocytes on their way through the PCV wall (11). In an in vitro model (575), the chain of lymph nodes from the rat mesentery is isolated and the lymphocytes are perfused via cannulas in the superior mesenteric vessels, allowing the study of their migration through the lymph nodes independently of accessory phenomena that can interfere with the process in vivo; for example, the cells would normally migrate at the level of the PCV, were they not taken up by the liver before they can reach the lymph nodes. In such a system, the perfused lymphocytes localize in and around the PCV, in the paracortex of the nodes, as they do in vivo soon after an intravenous injection, and other parameters of such homing indicate that this in vitro model is a good one (575). In such a system, there is no adhesion of lymphocyte membrane vesicles to the HEC of PCV, which, among many other possible interpretations, may indicate that adhesion requires either live lymphocytes or that the expression of the recognition sites for specific adhesion depends on a cytoplasmic component control (575).Lymphocytes exposed to trypsin show a deficient capacity to migrate, which suggests that some membrane (glyco)protein(s)may be crucial for adhesion to HEC (170). This does not say whether the trypsin-sensitive structure is the recognition site itself, since the enzyme may as well affect a nonspecific surface component required for adhesion. Neuraminidase treatment has no effect on lymphocyte migration, suggesting that gross, overall surface charge is not crucial for (this type of) adherence and that the neuraminidase-sensitive sialylated components of the lymphocyte membrane are not essential either (170). The altered migration of neuraminidase-treated lymphocytes to lymph nodes, in uivo, which had been reported in the past (683) can now be attributed to their sequestration in the liver, preventing them from reaching the nodes (170).Thus, caution must be exercised in the interpretation of in vivo results. Another in vitro model (618,684)makes use of glutaraldehyde-fixed lymph node sections (thus without active participation of HEC) and of thoracic duct lymphocytes (i.e., recirculating, mostly mature T lympho-
18
FRANCIS LOOR
cytes). Lymphocytes can specifically adhere at 7°C to the HEC. This requires Ca2+ but not Mg2+ and depends on lymphocyte structures (glycoproteins?) sensitive to trypsin but resistant to neuraminidase. It requires metabolically active lymphocytes (reversible inhibition by 1-10 mM azide, irreversible inhibition by 3-30 mM iodoacetate), A high colchicine concentration for even 3 hours at 37°C has no effect, but 5 minutes with cytochalasin B is enough to inhibit the adherence completely and irreversibly (an effect not due to glucose, since it is also obtained in glucose-free medium). A calcium ionophore, however, gives only partial inhibition. Thus, in this system too, adherence seems to be mediated through MV, being inhibited by factors known to affect their function. A series of assays have recently been performed in vivo using various inhibitors of ATP metabolism, agents modulating the CAMP level in the cell, MF active agents, and MT active agents. It was hoped that some of the effects of such agents on cell cortex activities, such as MV expression, would still be present, in spite of their reversibility (e.g., ATP depletion by NaN3 increases MV expression, but this can be reversed by washing; cytochalasin B makes MV collapse, but after its removal the MV can be reexpressed). Thus, lymphocyte emigration from blood to lymph node was impaired when the cells were treated by agents that increase the cellular CAMP level, but it was enhanced when the cell treatment was depleting their CAMP level (178).Migration was also impaired after azide treatment of the cells, as they remained longer in the blood than control cells (178,614). Cytochalasin B treatment of lymphocytes would also impair their migration through the PCV of lymph nodes, and it would do so because of an abnormal interaction of the treated cells with the HEC of the PCV (178). Not being a specialist on cell traffic, I do not see how such a precise conclusion can be brought from the general pattern of cell homing in the other organs of the tested animal. Cytochalasin A treatment of lymphocytes also impairs their blood to node migration (12). In this case, autoradiography, SEM, and TEM would all indicate that the cytochalasin A-treated lymphocytes had actually homed to the PCV but did not migrate into the nodal parenchyma. This is interpreted as showing that the damages caused by cytochalasin A affect lymphocyte mobility but not recognition and adhesion to the HEC of PCV (12). This differs from our conclusions as to in vitro experiments described above, as does also the other in uiuo observation that colchicine-treated lymphocytes failed to home in the nodes (12). Both the latter studies are still preliminary, however, and will have to await confirmation. The adherence of lymphocytes to the HEV of PCV in lymph nodes is
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
19
cell specific (both with respect to the cell that can migrate and the place to which it does so). It also shows an interesting species specificity: by using the aforementioned in vivo and in vitro tests, the homing capacity in mouse lymph nodes was evaluated for lymphocytes from various vertebrate species. It was found to decline exponentially as the evolutionary separation of the lymphocyte donor and the host (mouse) increased (71). This suggests that the changes in the structures involved in the lymphocyte-HEC recognition have occurred at a constant rate during evolution, as could be the case if the recognition structures were of protein nature. It seems that a new type of approach to the nature of the recognition structures may be to test the migration, in mouse lymph nodes, of human lymphocytes into the membrane of which various mouse lymphocyte membrane components (from whole PM vesicles to defined glycoproteins or glycolipids) would have been introduced.
D. MACROPHAGE-LYMPHOCYTE INTERACTIONS Macrophages play complex regulatory roles in the development of an immune response and their interaction with lymphocytes have been the topic of extensive studies. Some types of macrophagelymphocyte interactions may be mediated by soluble mediators, but it is also definite that some types of interaction require actual contact between the macrophage and lymphocyte plasma membranes (PM), possibly because a function of the macrophage would b e to “present” the antigen to the lymphocyte (for a review of this topic, see 533,644). I will consider here only the nature of the mechanism by which lymphocyte and macrophage form contacts, and, to remain within the scope of this review, will restrict myself to the importance of cell cortex-plasma membrane interactions in such contact interactions (see also Section IX,B). Early studies of a type of in vitro macrophage-lymphocyte interaction that does not involve immunological specificity (359) showed ( a ) the need for active metabolism of the macrophages, and ( b )the formation of broad contact area between the macrophage and the thymus lymphocyte, sometimes suggestive of a process of phagocytosis. Such nonimmune adherence was essentially studied to determine the requirements shown by the macrophage, and it was shown to be inhibited by cytochalasin B, vinblastine, colchicine, trypsinization, calcium chelation, and low temperature (360). A different picture emerges in the case of antigen dependent macrophage-lymphocyte interactions (443, 544, 671). Morphological studies of the macrophage-lymphocyte clusters obtained in vitro
20
FRANCIS LOOR
show that 10-20 T cells adhere to one central lymphoblast, either a B blast (443, 671) or a T blast (57, 58), which is itself attached to the macrophage. Soon after cluster formation, the central lymphoblast is the only cell engaged in DNA synthesis (58).The central lymphoblast is villous, while the surrounding T cells are smooth except for their uropode, which contains most of the M F of the cell (443).The focalization of most T cell MV toward the central, “activated” lymphoblast suggests that T cell membrane components having recognition, adhesion and/or effector functions are concentrated there to interact with the blast cell membrane (363, 364). This would explain why cytochalasin B blocks such antigen-dependent binding of immune lymphocytes to macrophages (59) and why it inhibits antigentriggered lymphocyte proliferation (544).The drug needs to be added at the initiation of the culture, being less and less inhibitory as its addition is delayed, with practically no effect if added after 1 or 2 hours. Such inhibition happens in conditions where neither antigen handling by macrophage nor mitogen-triggered lymphocyte proliferation are detectably affected (544). This suggests that the drug interferes at the level of the MV on the lymphocytes, perhaps at the phase of recognition of the blasts by the T cells. A direct consequence of such failure of T cells to interact would be a lack of proper activation of the B or T blasts. This assumption may be supported by the fact that in vitro cytochalasin B inhibits release of lymphokines by immune lymphocytes (36, 486, 699). Production of the macrophage migration inhibiting factors (MIF) occurs in lymphocytes, but is dependent upon the presence of macrophages (431). Such MIF production in lymphocytes seems to depend on lymphocyte-macrophage cluster formation, as it shows the same early sensitivity to inhibition by cytochalasin B: the drug needs to be added at the beginning of the culture (36, 486), and it is not effective if added 2 hours later (36). Colchicine and vinblastine do not M and show only a influence MIF production when used at up to slight effect at M (486). Microtubules sensitive to such doses are thus not involved in either the recognition or the effector (secretory) processes. At higher doses (10’’ M ) , colchicine caused an important inhibition of MIF production if added at the beginning of the culture, but not if delayed for 2 hours, suggesting again an effect at the early phase of T cell activation (36); however, this was not documented. It has also been stated that a similar, time-dependent effect could be obtained with N-ethylmaleimide; this SH poisoning reagent has a multiple effect, namely to inhibit capping (372), but it was not said at which concentration the drug was used (36). On the contrary, it was
PLASMA MEMBRANE AND CELL CORTEX JNTERACTIONS
21
stated that ouabain, which is a Na+,K+-ATPase inhibitor without detectable effects on mouse lymphocyte membrane morphology, such as MV and capping (F. Loor, unpublished), did not have any effect on MIF production, but this again was not documented (36). Finally, some agents that increase the endogenous cAMP level of the cell (theophylline, chlorphenesin) can inhibit MIF production, but another such agent (isoproterenol) does not (486), and it is therefore not possible to establish straightforward relationships between cAMP level and MIF production. To extrapolate the effect of cAMP on MV function and lymphocyte-macrophage interactions would be possible but delicate. As far as the macrophage that is the target of MIF is concerned, it has been shown that the integrity of the microtubular cytoskeleton of that macrophage is a prerequisite for MIF-induced inhibition of cellular motility (487). Drugs directed to the MT allow the capping of Con A on niacrophages (362,490) and, most interestingly, M I F partially prevents such cap formation (490). Thus, it has antagonistic effects to MT-disrupting drugs, and there is some evidence that M I F actually works by promoting MT assembly (488, 491). Production of MIF is probably not the best means to monitor macrophage-lymphocyte interactions. In a more direct, morphological approach, Colcemid was found to block cluster formation, if present at the beginning of the cocultivation of lymphocytes and macrophages, at the low dose of 0.1 pg/ml(-2.5 x lop7M ) , but if added after 20 hours of cocultivation it had no effect below 100 pglml (-2.5 x lo4 M ) , at which concentration it reduces the number of clusters by roughly half (59). If such dissociation of clustered cells by a high dose of the drug can be attributable to their well known membrane effects, the interferences of lower doses, such as to lop6M , which are observed on the initiation of the cluster fonnation are probably due to effect of the drug on the M T themselves. Pretreating the lymphocyte with Colcemid did not interfere with cluster formation (59),but this should not suggest that the effects ofthe drug were at the level ofthe macrophage, as most of the binding of the drug to the lymphocyte must be reversible. Still, at whichever level the drug is working, it is difficult to interpret how actually interference with M T organization or function eventually interferes with cluster formation: indeed, such drugs do not detectably affect either antigen uptake or anti-Ig capping on lymphocytes or on macrophages (368, 369), or MV formation [although some effects were observed on their reexpression at 37°C after chilling on ice (368)],and they even favor uropode formation and MV redistribution on the uropode (see Section 111).
22
FRANCIS LOOR
Along the same line, it is easy to accept that inhibition of DNA synthesis (by cytosine arabinoside) does not interfere with cluster formation (59), since DNA synthesis inhibitors do not have any detectable effects on lymphocyte membrane dynamics (125, 363, 438, 570), and that most clusters are formed before DNA synthesis is detected in the central lymphocyte (58): it simply shows that cluster formation precedes, and is not dependent on, DNA synthesis. But it is more delicate to understand why the initiation of cluster formation is prevented by inhibitors of RNA and of protein syntheses (cyclohexiniide, puromycin, actinomycin D) and that established clusters are still sensitive to protein synthesis inhibitors, as they disappear within a few hours (59).This may indeed suggest that a specific protein that would rapidly turn over was required to keep the clustered cells in firm association with each other. The spontaneous dissociation of the clustered cells, which normally happens with time and seems to depend upon the decay of antigen associated with the macrophage (37), might be due to an arrest in the synthesis of such a short-lived protein. One more step in the speculation would be that the synthesis of such a protein actually depends on antigen triggering, a point that must be difficult to test experimentally. Further investigation should be done to determine whether a protein such as is required for the particular type of cell-cell attachment that occurs in the clusters is a membrane protein directly involved in adhesion properties of the cell or a protein essential for appropriate cell cortex-PM interactions (e.g., for MV expression) or a protein with still another function. Inhibitors of protein and RNA syntheses have not been shown to have any action on capping (125,363,438,570),but to the best of my knowledge their possible effects on other cell surface activities have not been investigated. If such a short-lived protein required for cell adhesion were a membrane protein, a clue could be found for the inhibitory effects of colchicine on cluster formation, which were reported above (59).Indeed, although colchicine (as well as cytochalasin B) does not seem to block secretion of immunoglobulins b y plasma cells (476), it inhibits the reexpression of niembraiie immunoglobulin b y lymphocyte after capping removal (645). Such reexpression is also inhibited by protein synthesis inhibitors (570). This suggests that the exposure at the PM level of new protein components requires protein synthesis and intact MT function. Exposure at the membrane level of a labile membrane protein with adhesion function, of the type postulated above, would then be sensitive to both classes of inhibitors, with the known consequence of decrease of cell-cell interaction. This is, however a far-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
23
fetched explanation proposed with the only purpose of trying to make sense of the data. Still, it fits with the observation that the interaction of immune lymphocytes with antigen-pulsed macrophages is abolished b y fixation of either the macrophage or the lymphocytes, or both (358): the adhesive protein may be shed or eluted or denatured by the fixation procedure, although this was mild enough to preserve the antigen-binding capacity of the lymphocyte. This latter observation suggests to me that the constitution of macrophage-lymphocyte complexes cannot be (solely at least) explained by simple antigen-mediated, physical bridging between macrophage receptors and lymphocyte ones. This contention is further supported by the fact that there is also strong inhibition of the macrophage-lymphocyte interactions by inhibitors of cell metabolism (358), in conditions that block capping, but not patching, and would therefore even allow clustering of receptors and cooperative binding interactions to occur. These authors concluded that lateral mobility of the receptors in absence of cellular metabolism is not sufficient to allow specific cell-cell associations. Indeed, the conditions used (20 mh4 2-deoxy-~-glucosetogether with 1 mM 2,Cdinitrophenol) strongly deplete cellular ATP. As seen in a preceding section, ATP depletion results in an increased MV expression by the cell, but such cells showed a severely impaired capacity to adhere. As is the case for the other recognition-adhesion phenomena studied in this review, low temperature was also shown to inhibit the macrophage-lymphocyte interactions(358). I suggest that such inhibition is not d u e only to a reduction of lymphocyte mobility, but also, perhaps principally, because low temperature strongly impairs the capacity of lymphocytes to develop MV (368). Finally, it is important to note that depletion of Ca2+leads at 37°C to the dissociation of already interacting lymphocytes and macrophages, although this would not dissociate antigen from the macrophage, and that no such macrophage-lymphocyte dissociation occurs at 4°C because of removal of calcium (358). In 111y opinion, this is a further demonstration that, after recognition has taken place, the interactions that take place between lymphocyte and macrophage are essentially of “adhesion type” and do not, or no longer, involve the antigen as a possible bridge between the interacting cells.
E. KILLER-TARGET INTERACTIONS This section covers the whole range of killing systems (with or without “immune” specificity, i.e., natural killer cell, antigen- or lectintriggered killer, antibody-dependent cell cytotoxicity), but it deals mainly with results obtained with cytolytic T lymphocytes and with K
24
FRANCIS LOOR
cells (the general name “killer” will be used) (for a review, see 76). Differences exist from one killing system to another, but they will not be considered here in great detail. Both cytotoxic T cells and K cells bind to their target, the killing of which is accompanied by the characteristic membrane blebbing phenomena (zeiosis), whose alterations are quite different from those observed in antibody- and complement-mediated lysis. Two important differences noted between T cell killing and K cell killing are as follows (197).The‘first is that the nature of the killer is different, since the cytotoxic T cell is an immune cell (which results from antigenic stimulation) and that it possesses its own specific receptor for the target antigen, whereas the K cell is also present in nonimmune animals (perhaps even in increased amounts in thymus-deficient ones if natural killer (NK cells) belong to the K cell category), and that it recognizes the target cells only if covered with antibody, via their Fc pieces. The second difference deals with timing of the killing process, since K cells kill their target within 15 minutes of contact, whereas cytotoxic T cells may take very variable times to do it, from seconds to hours, apparently as a random event after contact with the target. A central role for the killer lymphocyte uropode and MV in the establishment and maintenance of stable killer-target contacts was suggested by early morphological studies on the interactions of lymphocytes with their target (3, 24). This seems to be of general value, whether the killer is a T cell or a K cell. Others had not observed such uropode formation or extension of microprojections toward the target cell at any stage of the killer-target interaction (349, 555), but more recently the importance of timing for such membrane phenomena was realized, and it seems clear that the constitution of microprojection by the killer cell is a general feature of early stages of its interaction with the target (197, 557). The difficulty of observing microprojections, which was encountered by some authors, seems to be attributable to the short life span of microprojections and to the fact that they are a very early event of the interaction of the killer with the target: a maximum of such microprojections is found after 5 minutes of contact with the target in the K cell system (197) and after 10 minutes in the T cell system (557). The analysis of various killer-target systems, from the point of view of their requirements, shows that at least two phases should be distinguishable (more likely three phases, see below): the recognition phase (adherence of killer to target monolayers, formation of killer-target conjugates in suspension) and the effector phase (lysis of the target). Thus, the recognition phase requires active metabolism of the killer
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
25
(inhibition b y azide, by dinitrophenol) and shows temperature dependence (669), being better at physiological temperature, though not totally inhibited in the cold (2-22OC) (42). In contrast, the effector phase occurs only at a temperature between 22°C and 37°C and is not affected by inhibition of metabolism (42). Different classes of cytotoxic T killer cells seem to be generated iri uitro at different times of a priniary MLC and of a secondary MLC, some being sensitive to inhibition by 2-deoxy-D-glucose, and some not (379).Although the reason for such differential sensitivity is not yet elucidated, the inhibition brought on some classes of killer is at the level of the recognition phase: their binding to the target is the inhibited step (379). Drugs that affect M T function inhibit the killing of targets (625). Such inhibition, which is irreversible, seems to affect only the effector phase (492), since the drugs have no effect on the recognition phase (i.e., the adsorption of cytotoxic T lymphocytes on fibroblasts) (258). However, in a case of antibody-dependent cellular toxicity (ADCC) in the rabbit, the effector phase does not need intact MT function either (191). On the contrary, the ADCC kind of killing operated by polymorphonuclear leukocytes is sensitive to colchicine and vincristine (94). In a case of secondary cell-mediated lympholysis (CML) in the niouse (which is a fast reaction, completed in 10-20 minutes), a drastic decrease of the number of specific interactions between killer and target was indicated (25); this, however, was obtained at the remarkably high colchicine concentration of 200-500 kg/ml (-5 x lo4 to 1.25 x M ) and therefore is questionable as to its relevance to M T destruction, since colchicine has also purely membranous effects. Further, the number of mobile forms was decreased in the killer population (25) although colchicine in lower doses does normally rather increase the number of mobile forms of lymphocytes (570). Microfilaments are definitely involved in all types of killer-target interactions, as shown by the fact that cytochalasin B drastically inhibits the various systems of target cell killing (25, 75, 191,209,258,298, 492). Such inhibition is reversible, requires the presence of cytochalasin B at the initiation of the reaction and interferes with the actual binding of the killer to the target. At least part of the inhibition depends on M F function rather than on alteration of glucose uptake: killer-target interactions can occur in glucose-free media and can be inhibited by cytochalasin B (68,191);moreover, cytochalasin A, which has little effect on glucose transport, also inhibits killer-target interactions (68, 191,209). There are some discrepancies concerning the effects of cytochalasin A, some authors indicating a reversibility (rabbit ADCC, 191) and others an irreversibility (mouse CML, 28). Local
26
FRANCIS LOOR
anesthetics block the reaction, presumably by affecting M F function (see Section VI); such inhibition is reversible, since preincubation only of killer with anesthetics has no effect (308). The recognition phase should in fact be subdivided into an early stage of specific recognition of target membrane antigens, and a second phase, which follows a few minutes later, called the lethal hit stage (208, 493). While the early recognition stage does not require Ca2+in the medium, the lethal hit stage does (208, 493), which may indicate that the latter stage corresponds to a stronger association of the killer to the target, For later stages Ca2+is not needed, and bivalent cation chelating agents do not block the lytic reaction once it is initiated (75). Recently, all cytochalasins (A-E) were found to show no effect at all on the killing phase, whereas they all inhibit killer-target interactions at the early recognition stage, and all but cytochalasin A also inhibit the CaZ+-dependentlethal hit stage (209,632).This difference will be a useful tool in achieving more detailed understanding of the killer-target interactions. The first result of such analysis seems to b e that structures sensitive to cytochalasin A are required not only for the early recognition phase itself, but also at later stages, because cytochalasin A treatment of preformed killer-target conjugates makes them labile to shear forces (599).Another result of such studies is that differences are being found between the apparent requirements for intact killer MF function shown by the various killer-target assay systems (381, 599). Thus, some of the metabolic requirements for “recognition” are not fixed but “are imposed by the experimental conditions under which recognition is tested” (599). And this should eventually lead to considering that, beyond true recognition, the kiss of death may no longer involve recognition structures, but perhaps exclusively adhesion structures. By use of microcinematography to study killer-target cell interactions, three phases were recognized (396):the recognition phase corresponds to a random crawling of the killer over the target cell surface, which stops as fimi contact is established; the postrecognition phase is characterized by the stability of the association killer-target without gross modification of either cell; and the last phase corresponds to the target cell death, with the classical membrane blebbing. Thus, the general sequence is ( a ) recognition of the killer by the target; ( b ) firm binding of the killer to the target (“postrecognition,” “lethal hit,” “kiss of death” phase); and ( c )target cell death. How does the killer cell cortex mediate these various stages? Transmission and scanning electron microscope analyses of the interactions reveal some similar features of the various killer-target sys-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
27
terns (25,209, 296, 297,298, 349, 551). The principal one is that binding of the killer to the target is achieved through important killer cell “pseudopodes” that penetrate deeply into the target cell. The only detectable components present in such pseudopodes or at their bases are M F (197,297,298,551,557). Early stages of killer-target interactions (conjugates in suspension) show that immediate contact is achieved through MV localized to about the contact area on both killer and targets (296). Later multiple contacts get established between the membrane of the penetrating pseudopodes of the killer cell and the membrane of the target: short point contacts (100-200 nm broad) (296, 551,556) and/or a broad contact area (up to 1 pm) (25,50,551,556).At variance with the microprojections of the killer cell (either K cell or T cell), which contain only fibrillar material, the interdigitating “pseudopodes” of the target cell contain ribosomes (197, 557). Thus, they are of a different nature, and target pseudopodes may not be the result of an active process of the target cell cortex: their “pseudopodal” appearance may come from target plasma membrane infolding caused by the killer cells projections that push deeply into the target cytoplasm and alter all target cell organelles on their way, including the nucleus. In the killer pseudopodes, there are no morphological signs of a secretory activity (such as Golgi elements or vacuoles) (296,297, 298) as would be expected if target cell lysis were due to secretion of a lymphotoxin. Although some cases of membrane “fusion” or “junction” are reported in some systems (25,297),others find no detectable specialization of the membranes in the contact area, such as intercytoplasmic bridges (50,296,349,396). Cases of rupture of the target cell membrane, however, have been reported (326).Deep penetration of killer pseudopodes into the cytoplasm of the target cell through its disrupted membrane have been reported (556). More recently, however, the same authors indicated that no other such case has been observed, and that the general observation, both for T cell killing and for K cell killing, is that the plasma membrane of both the killer cell and the target cell appear to remain intact throughout the killing process (197,557).At zeiosis, the blebs on the target appear all around the cell (not only in the contact area) (76), and there could even be numerous membrane pores (159). It is unclear why colchicine inhibits the killing, for MT are not readily apparent in any of the figures published so far. However, as reported in previous sections, they are not numerous underneath the lymphocyte PM, and they may thus have escaped detection. Drugs that inhibit M T might interfere with the release of “lymphotoxin”
28
FRANCIS LOOR
secreted by the killer, but there is no clear evidence of secretory activity in the killer cells, and thus no support for such an interpretation of colchicine effects on killing. Small vesicles have been described to occur at regions of contact, between the membranes of killer and target cells (326), and transfer of killer lymphocyte vesicles within the cytoplasm of the target cell has been reported (4). This might be the colchicine-sensitive step, but no such transfer was reported elsewhere, and it still awaits confirmation. The involvement of MT itself in the killing process may not be general. Indeed, a recent report indicates that the beige mutation in the mouse selectively impairs natural killer function (532). As seen in a previous section, the beige mutant in the mouse is the equivalent of the Chediah-Higashi syndrome in man; the defect being attributed to an abnormal function of the MT system. Since lytic mechanisms of immune T cells, macrophages, and K cells involved in ADCC were apparently unaffected, a functionally intact M T system does not seem crucial for killing to occur, and this may reinforce the suggestion that the observed colchicine effect may be principally membranous. Unfortunately, controls with lumicolchicine have not been published. T cell killing was also shown to be inhibited when the killers had been pretreated with 25-hydroxycholesterol in order to inhibit the synthesis of cellular sterol. This effect could be abolished if the T cells were given exogenous cholesterol during the pretreatment period (256). This confirms that cholesterol is an important component for the maintenance of appropriate structural and functional features of the PM (363),since the presence of the large amounts of cholesterol present in the mammalian cell membranes reduce or abolish phase separation ofmembrane lipids (102,455),andthat in particular it controls the microviscosity of lymphoid cell membrane (598). Thus, it may control membrane component distribution; receptors or adhesive sites may not be appropriately distributed in cholesterol-depleted T cells. For instance, the association of killer receptors to target membrane glycoproteins may be impaired b y lateral dispersion of normally clustered sites, thus making the appropriate pairing more difficult and the killer-to-target binding less avid. Another consequence of altered membrane fluidity may be to perturb the hydrophobic interactions that appear to take place between the membrane lipids of the killer and target cells at the recognition phase (43). Membrane disorganization could also be the explanation for the inhibitory effects of local anesthetics, rather than on MF, already reported above (308), as well as for the inhibitory effects of high doses of colchicine (25). Finally, the observation that a depletion of cholesterol reduces MV expression in a
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
29
fibroblast line (350)might even provide a clue for the defective killing of cholesterol-depleted T cells, and they will probably show defective capacity to form the MV essential for the recognition phase. Perhaps this should be related to the observation that, in sarcoplasmic reticulum membrane, varying the amount of cholesterol reversibly modulates Ca2+-dependentATPase activity (382); but, for the time being, that could only be a speculation. A quite different interpretation of the effects of cholesterol depletion on killers’ activity may be an abnormal production or release of cytotoxic Iipids. Free fatty acids from lymphocytes have been reported to be highly cytotoxic for tumor cells (315, 454, 640, 641). More recently, the incubation with allogeneic tumor cells of sensitized T lymphocytes was found to result in a decrease of the cholesterol and free fatty acid content of the killers (314). Since ( a ) the killers and the targets come in close contact with each other, ( b )the target tumor cells have usually lower cholesterol content conferring on them higher fluidity (reviewed in 363,438), and (c) transfer of hydrophobic probes from killer to target (43) suggests intimate contact of their membrane lipid bilayers, it does not seem unreasonable to speculate that a possible mechanism of killing is the transfer of some lipids to the target membrane, so that its permeability to ions would be altered, causing osmotic lysis. What could be such toxic lipids? This is unclear, as well as what would be the mechanism of synthesis. But the absence of morphological characteristics suggestive of active secretion in the killer pseudopodes may indicate that killing is rather mediated by the killer cell membrane itself. I would suggest the hypothesis that killing by T cells is only an evolutionary improvement of killing by macrophages, the difference being only at the level of the recognition structures; the K cell would be somewhere between the T killer and the standard macrophage. The hypothesis is thus that the mechanism of killing, once recognition has been completed, may not be much different for T killer, K cells, and phagocytes. Phagocytes can kill in many different ways, principally intracellularly [see review by Roelants (533)],but there are also ways by which they kill extracellularly: these are the production and release of peroxide (114,534), superoxide anions and hydroxyl radicals (e.g., 14, 541), and toxic catabolites of lipid hydroperoxide, namely, malonyldialdehyde (e.g., 624). The aklylating potential of the latter is probably responsible for its toxicity, as revealed, for instance, by the inhibition of fibroblast growth that it causes (518). There is increasing evidence that 0, metabolites are not produced only by phagocytes, where their production does not require previous phagocytosis to occur, and that
30
FRANCIS LOOR
they are produced also at the cell PM level (see, for references, 63, 130, 205). Peroxidation during activation seems to be a universal feature for many cell types, including the fertilized egg (320),and, of more interest for the present review, neutrophils (130, 219, 341, 401), even though superoxide release by neutrophils was recently questioned as being actually physiological (577). A central role for superoxide radicals as the mechanism of target cell killing was recently shown in the case of neutrophil-mediated cytotoxicity (607): Con A- and PHAinduced superoxide radical production correlates with timedependent and dose-dependent cytotoxicity induction over a wide lectin dose range, whereas pokeweed mitogen (PWM) is negative for both 0,- production and cytotoxicity activation. Neutrophils can indeed kill target cells coated with antibody, thus in an ADCC way, by recognizing the target cell-bound antibody via Fc receptors (94, 186,380,702). This looks very much like NK cells and K cell killing. The myeloperoxidase-H202-halide system of polymorphonuclear leukocytes, however, was not found to be responsible for their ADCC type of cytotoxicity (94). Thus, there is yet some controversy; still, it does not seem unreasonable to speculate that K and T lymphocytemediated killing might also involve peroxidative mechanisms with release of 0, metabolites or of toxic peroxidized lipids. It seems to me that this is amenable to actual testing, ( a ) by addition of superoxide dismutase to the medium to destroy superoxide enzymically and of catalase to destroy HzOz produced extracellularly, to try to inhibit killing; ( b ) by inactivating the endogenous catalase by addition of 3-aminotriazole, and the endogenous glutathione peroxidase b y using glucose-free medium during the killing phase, so that both of the two mechanisms allowing H,O, destruction are inactivated and killing should be enhanced. A series of agents that cause an elevation of CAMP in the lymphocyte can inhibit cytotoxicity (259, 347, 626, 681). The level at which such agents are working is unclear, however; a high CAMPlevel is known to lead to higher expression of MV on the cell surface, but, as already discussed, they may not be good for intercellular adhesion. Another agent that could reversibly inhibit killer activity was recently found to be 3-deazaadenosine (704). This was not due to effects on levels of CAMP or of the triphosphonucleotides, and the inhibitory effect of 3-deazaadenosine was attributed to the inhibition of an unidentified but crucial, S-adenosylmethionine-utilizingmethyltransferase within the cytolytic lymphocytes (704).The requirement for such an enzymic activity was considered only in terms of need for methylation of proteins. However, S-adenosylmethionine is also a donor of methyl groups for lipids, namely, for the synthesis of phosphatidylcholine by
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
31
three successive methylations of phosphatidylethanolamine; such a process is activated in the case of the early mitogenic response of lymphocytes to PHA (166, 167). Thus, methylation of proteins may be needed at some stage of the killer activation, and this may be for plasma membrane or cell cortex rearrangements (704), but it may as well be a niethylation of phospholipids. Changes of phospholipid synthesis and/or turnover seem to be a very general feature of “cell activation,” particularly in the case of lymphocyte activation by lectins, of which an early sign is precisely an increased turnover of phosphatidylinositol and of phosphatidylcholine (e.g., 166, 167, 519), but also for other cell types, e.g., macrophages, where similar observations are being made (453). Changes of PM phospholipid composition may in turn alter distribution or interaction of PM components, among themselves and with cell cortex elements, and therefore be responsible in an indirect way for the coupling-uncoupling of recognition structures and of adhesion structures on lymphocyte cell surface. Effective recognition, i.e., recognition followed by adhesion, may thus depend on quite unspecific factors: among these is simply the mobility of the killer lymphocytes. This appears evident when considering recent data (78) showing how much and how fast lymphocytes can crawl on and beneath cultured fibroblasts: up to 20 pm/min. Such a property, which is characteristic of T cells from immunized animals, is not antigen specific, since immune T cells crawl as much and as fast on syngeneic as on allogeneic targets. Crawling is inhibited by inhibition of energy of metabolism, alteration of M F or MT activities, alterations of ionic content of the medium, and so on-but all this is unrelated to the recognition structures themselves. Still, the morphological data analyzed above principally by electron microscopy have shown interesting features of the killer-target interactions: that is, in the various types of killer-target interactions, lymphocyte cortex activities were implicated. The MF/MV system seems to b e involved throughout the interaction even if it is sensitive to inhibition only at early stages of the recognition phase. The remainder of the killing process is still obscure as to its mechanism, it no longer seems to involve MF, and although colchicine inhibits the killing phase, it is unclear whether it is due to its effects on M T or to its membrane effects. VI.
Cell Cortex a n d Plasma M e m b r a n e Functions in the (Mitogenic) Activation of Lymphocytes
In previous sections I have analyzed various phenomena that appear to involve interactions of outer cell cortex elements, such as M F and MT, with specific receptive or with adhesive membrane components,
32
FRANCIS LOOR
or both, exposed on the outside face of the plasma membrane. All such phenomena were of short duration, occurring within minutes, and there was little chance that they could be due to side effects of the various MT- and MF-directed drugs, e.g., phenomena such as capping, MV expression, agglutination, rosette formation, killer-target interaction, and so on are all inhibited instantaneously when MT- and/or MF-directed drugs are present at their initiation. Potential effects of the drugs on protein, RNA, or DNA synthesis do not worry anybody. The situation is dramatically different in the case of the stimulation of lymphocytes by mitogens, where the effect of any drug cannot be evaluated within minutes of its addition. Usually, the test is the capacity of cells that were stimulated with a mitogen at a time zero to incorporate a radioactive precursor of RNA or DNA given 48 hours after the mitogen for the following 24 hours. The effects of classical MF- or MT-directed drugs is tested either while they remain throughout the culture period or are present only at specific periods of mitogenic activation of the cells. The general significance of such methodology and its limits will be briefly discussed. A possible role of the cortical CslCm elements in the mitogenic stimulation of lymphocytes was primarily considered with regard to its possible function as a relay for transmitting a mitogenic signal from the cell membrane to the cell nucleus (139, 141). The effects of drugs affecting M F or M T functions were thus analyzed with regard to various parameters that are supposed to be integral parts of the mitogenic response, i.e., the activation of the cells out of their resting state (Go), into the beginning of the mitotic cycle, the GI phase. Among these parameters, late ones are cell proliferation, DNA synthesis, and blast transformation; early ones are the increased turnover of PM phospholipids, the increased influx of K+ and of Ca2+,the increased uptake of nucleosides, sugars, and amino acids, and alterations of cyclic nucleotide levels (for reviews, see 461,661). Beyond the mitogenic activation, there can be additional differentiation or expression of cellular potentialities, such as antibody synthesis or mediator release or activation of cytotoxic functions. Recent studies on the mechanism of cell division in culture suggest the existence of a “commitment” or “no return” point, somewhere between the Go and the S phase of the cell cycle, i.e., that certain biological events can determine the cell irreversibly to DNA synthesis (reviewed in 472). Triggering to mitosis is thus the whole sequence of events that bring the cell from the resting state to the commitment to mitosis, to the “no return” point, thus to a “disequilibrium” situation. A triggering signal is then the event, biological or not, that introduces a disequilibrium in the functional status of the cell so as to oblige it to
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
33
divide to reach its normal biological equilibrium again. To study what triggering is, or what are the signals for triggering, is not exactly the same thing. It is not my purpose here to discuss what lymphocyte activation may be-I could not. Hence, I would like to point out in the strictly mitogenic response that, before DNA synthesis is initiated, there is a need for previous activation of RNA and protein synthesis, and that lymphocyte stiniulation by a mitogen is accompanied by an increased transport of a number of metabolites, among which are sugars and nucleotides. Thus, alterations of MT or M F functions may modulate lymphocyte activation because of “side effects,” not necessarily because M T or M F or both are at all implicated in the activation signal sequence. On the other hand, some other “factors” (i.e., drugs or experimental conditions) that modulate lymphocyte mitogenic responses and are not recognized MT- or MF-directed drugs may in fact play a role, direct or indirect, at the level of such structures and interfere in such a way with the transmission of a mitogenic signal. Therefore, the mechanism of action of various possible inhibitors that affect cell metabolism, membrane transport mechanism, and/or M F and MT functions will not be understandable in simple terms, and the data are as easy to produce as they are difficult to interpret in a meaningful way. A complication is the lack of synchrony with which the cells enter the mitotic cycle, since some are rapidly stimulated and others take some 20 hours to enter it, with a consequential overlap and blurring of individual biochemical modifications. Moreover, multiple (cell-mitogen) systems have been used, and only a few general lines can be deduced. Although one can clearly distinguish the effects of the classical drugs affecting M T from those affecting MF, there are also alterations of the ‘‘normal’’ mitogenic response of lymphocytes that can be brought about by a number of other drugs or experimental conditions (Section VI, C). Some of them may actually interfere with activation because of their effect on the cortical Cs/Cm elements, directly or indirectly, but perhaps have nothing to do with them. The activation of lymphocytes by anti-Ig is discussed in Section VII, and the importance of clustering of PM components as an activation signal is demonstrated in Section VIII. Finally, in Section IX, a few cases of activation of nonlymphoid cells are reviewed. A. EFFECTSO F CLASSICAL DRUGS THATAFFECT THE
CYTOMUSCULATURE These drugs are the cytochalasins; principally (CB) cytochalasin B has been used. Different results are obtained depending on the dose of mitogen, on the dose of cytochalasin B used, and on the time of any
34
FRANCIS LOOR
addition to culture. When the effect of the drug is tested on cells responding to approximately the optimum mitogenic dose, i.e., the dose of mitogen that gives maximal stimulation in the absence of any drug, one can distinguish two ranges of doses of cytochalasin B giving opposite effects. With doses usually higher than 2 pg/ml(-4 x 1OP6M), increased inhibition of the optimum mitogenic response is obtained by increasing doses of the drug (45, 195, 262, 272, 402, 520, 698-700). Such an inhibition is reversible by simply washing away the drug (262, 402, 520, 700). It is obtained only if the drug is present when the mitogen is added to the cells, but not when added 1 or 2 hours after the mitogen (262,402,520).Thus, the drug interferes with an early step of mitogen stimulation and cannot block the triggering mechanism once turned on. When doses lower than 0.5-1.0 pg/ml ( 1 to 2 x lO+M) were tested, it appeared that the drug increases the optimum mitogenic response of the lymphocytes (45, 46, 262, 358, 402, 460, 700). Such potentiation is reversible by washing the drug before the cells make contact with the mitogen (262); if addition of the drug is delayed for 6 hours after mitogen addition, it is still half as effective, but if delayed 20 hours, it gives almost no potentiation (460). Potentiation is noted as increased synthesis of protein, RNA, and DNA, as well as an increased number of morphologically transformed cells (46). However, in some circumstances potentiation can be obtained with doses of cytochalasin B that would be inhibitory if added at the initiation of a response to an optimum mitogen dose. This happens if the dose of lectin is supraoptimal (higher than the optimal dose) (195),or if addition of the drug is delayed until the last 24 hours of the culture (262). The interpretation of potentiation and of inhibition of mitogen responses by different doses of cytochalasin B is complex. In fact, there are many other described effects of the drug that depend on timing of the addition, shaking of the culture or not, giving the cells serum or not, etc. Those data are as yet isolated findings, not really worth analyzing in the context of this general review. The inhibitory effects of cytochalasin B affect an early step of mitogen stimulation and may be attributed to many causes; these are simply listed here (see discussion in 195, 262). The drug may indeed interfere, through an alteration of the M F system, with the transmission of a signal from the mitogen on the membrane to the inside of the cell, or with various types of possible synergistic interactions between different cells, perhaps between lymphocytes and macrophages. The drug may, however, just as well inhibit the mitogen response through its effects on the membrane itself, by inhibiting the transport of various metabolites, as was recently suggested concerning the transport of
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
35
glucose (272).With regard to this point, it would be worth testing other cytochalasins that affect M F function without having marked effects on glucose transport, such as cytochalasin A, which has little effect and dihydro-cytochalasin B (19) or cytochalasins C, D, and E, which have no detectable effect on glucose transport (19, 294, 354,355). The stimulatory effects of cytochalasin B may also depend on multiple causes (195, 262). They could be interpreted in terms of specific elimination of suppressor cells or of hypersensitization of cells to a mitogenic signal (see below). However, they may be due to apparently trivial, but nevertheless interesting, causes, e.g., the effects of the drug on lectin-induced agglutination of lymphocytes. A dose of lectin lower than the optimum mitogenic dose does not aggregate the cells very much, whereas a dose of lectin higher than the optimum clumps all the cells into large aggregates. On the contrary, the aggregation of the cells by a mitogenic dose of lectin occurs progressively, taking several hours of culture. A low dose of cytochalasin B might then regulate the lectin-mediated agglutination, inasmuch as it would avoid the constitution of massive aggregates of cells, but still allow intercellular interactions to occur, since the cells still form small clusters in its presence. Cells in such clusters may show less inhibition of contact than cells in big aggregates, or have better access to essential metabolites of the medium. No cell aggregation at all occurs at higher doses of cytochalasin B, but here the inhibitory effects (e.g., on glucose transport) of the drug take over. If added at the end of the cell culture or when a supraoptimal dose of lectin was used, the drug competes with the lectin with regard to the size of the cell clusters, and, whenever they can be of appropriate size and the dose of drug not too damaging for the cell metabolism, potentiation will be obtained. Another alternative to interpret the stimulatory effects of low doses of cytochalasin may be found in their capacity to inhibit capping in a reversible way. At doses of lo* M or less, the capping inhibition is incomplete, being more a delay of the capping process than its actual blockage. Such a retardation of ligand removal may in fact be needed for, or at least may favor, interactions of liganded PM components with other PM components that would constitute a step in the triggering of the cell. A too rapid ligand removal would thus impair triggering, as already postulated (363, 366).
B. EFFECTSOF CLASSICAL DRUGSTHATAFFECT THE CYTOSKELETON These drugs definitely inhibit the proliferative response of lymphocytes to the various mitogens; most experiments have been performed with colchicine. If one follows the same type of reasoning as used
36
FRANCIS LOOR
regarding the potentiation of lymphocyte activation by low doses of cytochalasin, the opposite effect of drugs that disrupt MT would be easy to understand: the disruption of the cytoskeleton would fasten the capping removal of liganded PM components, and therefore it would actually impair the succeeding early steps of the sequence of events leading to stimulation. Unfortunately for that speculation, it is absolutely not clear at which stage or stages the inhibition takes place. The effects of the drugs on various early parameters of the mitogenic response are thus the object of many controversial reports (47, 195,220, 229, 375, 402, 521, 594, 619, 633, 659). The activation of lymphocytes seems to take some 20 hours, after which the lectin does no longer seem to be needed and can be removed from the medium and from the lymphocyte surface. This has no effect on the rate of DNA synthesis by the individual cells or on the number of cells that have been committed to synthesize DNA (228). Earlier removal of the mitogen leads to a decrease in both parameters (142). Inhibition of the proliferative response to mitogen can be obtained by use of drugs affecting MT (colchicine, vinblastine) even if their addition is delayed some 16-24 hours after addition of the mitogen (47, 195, 402). This type of inhibition has consequences for later stages of the proliferative response and may be due to various effects of the drugs on multiple targets: metabolic transport, spindle formation, cell cortex organization. There are reports, however, that stimulated lymphocytes that “have passed the commitment point by more than 3 hours” can become refractory to inhibition by the drug (229). Whichever late stages of the proliferative response are sensitive to inhibition by drugs affecting MT, this inhibition is not contingent upon earlier steps that may be related to lymphocyte commitment. Various attempts were made to determine whether intact MT function was required within the first 20 hours of mitogen addition-either by administering the drugs at various intervals after the mitogen, or by trying to analyze early parameters of lymphocyte activation. Approaches of the second kind are more complex, because different early steps of the mitogenic response have been analyzed (see below). In an approach of the first kind on Con A-induced lymphocyte stimulation, the kinetics of inhibition, by colchicine added at different times after mitogen, was found to be similar to the kinetics of inhibition obtained by removing the Con A from the cells by amethyl-D-mannoside (229). This was interpreted as suggesting that both treatments affect the same early step(s) after mitogen binding, perhaps affecting the transmission of the triggering signal itself (229). Other similar experiments had shown that such inhibition of Con A
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
37
stimulation by drugs affecting MT occurred prior to the S phase, that it was not due to mitotic arrest, to cell death, to blockage of D N A synthesis, or to inhibition of thymidine transport (659). There is not total agreement with this, however (see below). When the effect of colchicine is tested on the small proportion of the cells that are activated by a single short pulse of Con A (some 2 hours) and respond in a rather synchronized fashion, two phases of sensitivity to the drug are found (410):a pulse of colchicine given together with Con A markedly inhibits mitogenesis, as does also the presence of the drug for the last 24 hours of culture of Con A pulse-stimulated cells; however, a pulse of colchicine given just after the Con A pulse does not inhibit mitogenesis very much. This suggests that colchicine sensitive structures are necessary at late stages of the mitogenic response but also, and probably for different reasons, at the time of triggering itself or at its very early steps (410). A quite different conclusion was reached by others, however, using colchicine too or another drug, Isoptin, which binds competitively with colchicine to tubulin and depolymerizes M T in uitro: they also abolish D N A synthesis induced by lipopolysaccharides (LPS) or Con A in mouse lymphocytes, when they are left with the mitogen-treated cells thoughout the whole culture period. However, when colchicine or Isoptin are removed 20 hours after the start of the culture, the uptake of D N A precursors by the cells starts without delay (375).It is thus suggested that the disturbance of MT function does not interfere with the commitment of the cells for D N A synthesis; i.e., polymerized MT have no raisoii d’ktre for the early steps of triggering. Before trying to conclude with the first kind of analysis of the mitogeiiic activation, i.e., “late” stages of the response, it seems worth mentioning that the significance of the assay itself, i.e., the uptake of radioactive precursors of DNA, has been rightly put in question as actually being a good parameter of D N A synthesis (619). The blastogeiiic response of lymphocytes to a mitogenic lectin can be evaluated by measuring the growths of cellular and nuclear volumes as monitored by volume spectroscopy arid by measuring the increase of their D N A content (the sign that they have entered the S phase) as monitored by flow cytofluorometry, and these parameters can be compared with the classical [JH]thymidiiie incorporation by the cells. Using such an approach it was found (619) that 10P M colchicine or Colcemid have little effect on nucleus growth and cell growth during the first 2 days of culture, after this, however, cell growth stagnates. Despite the presence of the drugs, a number of cells enter the S phase, since an increase of DNA content is monitored in a number of cells,
38
FRANCIS LOOR
which is about 60% of those that enter S phase in absence of the drug, but D N A synthesis in presence of the drug stops before completion of the S phase. The [3H]thymidine uptake by the cells is strongly suppressed by the drugs (619). Thus, the latter parameter cannot be interpreted as showing that drugs that affect MT inhibit the triggering of lymphocyte by mitogens: they are triggered to enter S phase, but the activation is abortive because of some other factor. What could be the target of action of the drugs that would explain this? It appears that mitogen-induced lymphocyte stimulation is associated in its early stages with an increase in the tubulin content of the cell, which is higher than the total protein increase, and probably due to increased tubulin synthesis (594). The ratio of tubulin to MT remains constant, and there is a proportional increase in MT content, which is at least 2.5-fold, 36 hours after mitogen stimulation. Such an increase may be needed for succeeding stages of cell transformation and division, including the increase of DNA synthesis. The effect of colchicine would be to block polymerization of newly synthesized tubulin into MT (594). A clue for a possible function for M T may be found in electron microscopic observations on the effect of drugs on the mitogen-induced transformation of lymphocytes (633).In the presence of drugs, no MT are seen, but also the mitogen-induced development of many organelles is partly inhibited. The development of the Golgi complex and the dictyosomes particularly seem to b e altered. It is suggested (633) that, in the absence of MT, the Golgi complex and some associated organelles cannot “respond” to nu togenic stimulation, i.e., to permit an increased cell growth (an increase of general cell metabolism and cell enlargement precedes the start of DNA replication). Further, inhibition of GI-associated cell growth would inhibit initiation of DNA synthesis (633). Although exogenous thymidine can b e a source of thymidine for the lymphocyte, as shown by the classical tritiated thymidine assay, DNA synthesis in lymphocytes does not depend on exogenous thymidine; most classical cell culture media do not contain thymidine, and, as seen above, DNA synthesis occurs even when thymidine uptake is practically blocked (619). Therefore, the classical measurement of uptake of tritiated thymidine b y cells is nothing more than a measure of the contribution of exogenous thymidine to total thymidine incorporation into DNA. The relative contributions of endogenous and exogenous pools of thymidine may vary a lot, even at an equal rate of DNA synthesis, as long as the available endogenous thymidine is not a limiting factor. Thus, DNA synthesis may still proceed nonnally, even if w e monitor no incorporation of tritiated thymidine at all, because a drug
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
39
interferes with its uptake. Such interference niay occur either because the drug acts directly on the thymidine transport system and inhibits it, or because it modifies the balance of exogenous and endogenous contributions of thymidine in any other way, so that the cell will preferentially use endogenous thymidine, even though there is no blockage of the transport mechanism for exogenous thymidine. Let us turn to the effects of the drugs that disrupt M T on early parameters of the lymphocyte activation, keeping in mind that their actual significance is not universally acknowledged. Drugs that disrupt MT were found to have no effect on the increased RNA synthesis that normally precedes lectin-induced DNA synthesis, which was inhibited (47, 521). Similarly, they did not affect the increased turnover of membrane phospholipids that occurs soon after stimulation (521). They did not affect the early Ca2+uptake associated with (required for?) lymphocyte activation (220),but they were found to inhibit the early, lectin-induced increase in sodiuni-dependent amino acid transport, as well as the late thymidine uptake (220). Finally, they have been reported to amplify and prolong the early, lectin-induced response of CAMP metabolism (220),perhaps by modifying PM organization. The idea is that PM bound enzymes that allow positive and negative regulation of the cyclic nucleotide metabolism would not be distributed randomly, but, on the contrary, in a high degree of organization dependent on MT. Drugs disrupting MT would disturb such organization and consequently alter the normal function of such “metabolism control units.” Since lumicolchicine (the ultraviolet lightinactivated isomeric form of colchicine) does not have such effects on cyclic nucleotide metabolism, the effects of colchicine are definitely niediated via M T alteration (lumicolchicine is inactive) rather than via membrane effects (lumicolchicine alters membrane organization like colchicine). My only conclusion is that there exists a real controversy as to the involvement of M T as a crucial signal in mitogenic stimuIation by lectin binding on the PM: there are as inany good arguments pro as good arguments contra. With tetravalent Con A, the curve for Con A dose-rnitogenic response was found to be very closely correlated to that for the Con A dose-receptor mobility restriction, and M T structures were also implicated in the mechanism by which the higher Con A doses lead to decreased lymphocyte mitogenesis (141).The situation is complex, for the action of tetravalent Con A is in fact “paradoxical” in that the lectin both stimulates and inhibits, the latter process predominating at high Con A doses. The nature of the inhibitory effect of hyperoptimal doses of tetravalent Con A was unclear. and it could be due to trivial
40
FRANCIS LOOR
reasons of the kind involved in Sections VI,A and B or to cytotoxicity alone (412). Hyperoptimal doses of Con A are indeed inhibitory to both normal lymphocytes and lymphoid cell lines (398). The inhibition does not, however, seem to b e due to some trivial negative effect of cell clumping by the high lectin dose, since the inhibition is maintained when the cells are grown as a suspension in agarose gel (398). Furthermore, the inhibitory eEects of hyperoptimal doses of Con A appear to be reversible (80, 398). The inhibition is not a blockage of stimulatory signals, which appear to be always delivered even by completely nonmitogenic hyperoptimal doses of Con A: the initiation of the blastogenesis process as measured by the capacity to enter the S phase if not blocked; on the contrary, the rate of commitment to enter the S phase increases with the lectin dose even in the hyperoptimal range (398). Doses of Con A found to b e hyperoptimal for exogenous thymidine uptake induce the blastogenic response as detected by growth of cellular and nuclear volume and DNA synthesis, actually in a larger fraction of the cells than the “optimal” mitogenic doses (620). This is the case also for another parameter taken as an early sign of activation: the increased phosphatidylcholine synthesis, which does not occur after treatment with nonmitogenic lectins, is actually proportional to the Con A concentration, showing no falling limb at hyperoptimal “nonmitogenic” doses of the lectin (84). If hyperoptimal doses of Con A deliver increased stirnulatory signals, they do so also for inhibitory signals. Apparently, the cells treated with hyperoptimal doses of Con A can enter the S phase, but their progression through the S phase and the G2 phase stagnates, and only a few cells can complete mitosis (620). The nature of the inhibitory signal(s) is not yet defined (for a discussion, see 620).
c. EFFECTSO F O T H E R “FACTORS” These factors are drugs or experimental conditions that have been shown to modulate the mitogenic activation of lymphocyte and whose mechanism of action may possibly be at the level of the cell cortex. Some of them may affect the capping removal of liganded PM components and thus influence the activation of the cell, even if they are not directly active upon MT or M F activities, e.g., increase of membrane lipid viscosity, chemical cross-linking of PM components. I will consider the following factors: local anesthetics, ions, ionophores, thiols, sulfhydryl reagents, as well as oxidative stimulation. This section cannot, however, be a complete review of the field, and I will try to restrict myself to what may relate the mitogenic activation to PM, MT, MF, capping, etc. With few exceptions, the tests for the
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
41
mitogenic activation of lymphocytes have been limited to the classical tritiated thymidine uptake method. 1 . Loccil Aiiesthetics
Local anesthetics have been shown to block mitogen-induced stimulation, when present from the lieginning of the culture (162, 163, 375). Unfortunately, the interpretation of the effects of local anesthetics, as well as of any membrane-active agent, always remains highly speculative: they do affect the M F and/or M T systems, but is it in a more or less direct way, or as a more or less late consequence of other effects, e.g., on menibrane fluidity, on nienibrane curvature, on disturbance of distribution of PM components, on Ca2+permeation, and so on (363, 438, 570, and see below)? An effect on cell aggregation was invoked to explain the concurrent inhibition by chlorpromazine, of the Con A-induced lymphocyte aggregation and mitogenesis (162): a similar concentration (-5 x 1 O F M ) blocks both the mitogenic response and the cell clustering. Such blockage of mitogenicity was later extended to other niitogens of both B and T cells (Con A, PHA, PWM, LPS) and to another anesthetic, lidocaine, which, however, requires the high concentration of at least -2 x 1 O P 121 to be efficient (163).The inhibitory capacity of high doses of lidocaine on Con A and on LPS stimulation of mouse cells was coilfinned by others (375), but it was also found that low doses of the anesthetic were actually stimulatory, since they gave an enhanced mitogenic response (375).This dual effect of lidocaine recalls the dual effects of cytochalasin, which also potentiate responses at low doses and block them at high ones (see preceding section). It appears to b e very difficult to determine the mechanism by which these tertiary aniines inhibit lectin-induced mitogenesis. Such inhibitory effects are claimed to he readily and totally reversed by anesthetic removal from the culture up to 4 hours after its start (163).This has allowed pulse-type experiments to be perfonned, and it could be shown that the inhibited event occurred rather early after exposure to the mitogen (163). A fast reversibility should not be interpreted as demonstrating a “membrane” effect; it can still be at the cell cortex level, e.g., the effects on capping of many other inhibitors (NaN,1, cytochalasin) are also reversed within a matter of minutes-just the time required to wash away the inhibitor. High doses of lidocaine left for 20 hours with the mitogen-treated cells irreversibly inhibit their mitogenic activation (375). From structural and morphological studies, it is known that local anesthetics actually induce alterations of both the plasma membrane
42
FRANCIS LOOR
and the elements of the cell cortex. At the membrane level, these cationic amphipathic molecules interact with lipids, especially with the anionic phospholipids, and with membrane protein components; they interfere with membranous ATPase activities and with membrane transport systems; they displace Ca2+from the membranes and cause molecular disordering of lipid bilayers (83,471,576).They induce a number of alterations of the function and the morphology of the cell surface, some of which can be related to the MT and M F elements of the cell cortex (for review, see 440). Thus, local anesthetics seem to disrupt both MT and M F organization in fibroblasts (442,499).They inhibit the capping of membrane immunoglobulins on lymphocytes (498,549) and even allow reversion of preformed caps of Ig or Con A (498,568),and finally they allow increased clustering of bound ligands (499):this is best understood if the cell cortex elements that control cell membrane component mobility and distribution lose their attachment to the membrane, i.e., there is “transmembrane receptor uncoupling” (440).Since local anesthetics displace CaZ+from membranes, it has been proposed that the uncoupling results principally from the dissociation of M F from membranous Ca2+-dependent attachment sites and from the dissolution of M T due to the increase in cytoplasmic Ca2+concentrations freed from The disruption of caps with reversiotl to the spot the membrane (440). stage by anesthetics does not require energy, even it is enhanced in the presence of metabolic inhibitors, but it can be counteracted by increasing extracellular Ca2+(568). This suggests that anesthetics work by dissociating clustered PM components from the bound M F bundles and that such association is dependent on membrane-associated Ca2+. With regard to opposite consequences observed in the case of the effects of Ca2+ionophores on similar processes (see below), the interpretation of the actual mechanisms of action of anesthetics on capping remains obscure (see 125,570).Recently, lidocaine was shown to bind directly to MT, and in a competitive fashion with colchicine (326). In conclusion, local anesthetics may alter the mitogen-induced activation of lymphocytes as a result of their effect on MT and MF, but this is not the only possibility, and the M F and MT alterations they cause are not even fully understood. 2. loris and lonophores The mitogenic activation of lymphocytes, and other cells, has regularly been associated with the movements across the plasma membrane of principally two cations: K+ and Ca2+.Can one establish direct
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
43
correlations of such ion fluxes with activities of M T and M F elements of the cell cortex? Maybe. I am not trying to interpret every phenomenon in terms of M T and M F activities, but I only try to see how in such a perspective it would be possible to interpret known data on mitogen stimulations, and I know that multiple alternative explanations are possible. Thus, fluxes of Ca2+and of K+ are apparently required. Potassium ions fluxes have been implicated as early events in lymphocyte stimulation, as there is an increase in the concentration of intracellular K+ in lymphocytes induced by mitogens (PHA, Con A), as a result of increased K+ influx, possibly because of an increase in the number of ion pumping sites on the cell surface (23,302,348,430,508, 509, 579), unless it results from a decrease of the cell water volume, the absolute amount of K+ remaining constant (265). The membraneassociated enzyme Na’, K+-ATPase (involved in the transport of Na+ and K+) has been implicated directly in the activation process of the lymphoid cells (23),and there is indeed some evidence that it plays an important role: mitogen-induced responses of lymphocytes are severely impaired by the addition of nontoxic doses of substances known to inhibit Na+,K+-ATPases, such as polar digitalis glycosides (G strophanthin or ouabain) (506, 507, 558) and cholera toxin (653). Ouabain also inhibits lymphocyte cytotoxicity (336,688)by an inhibition not only of proliferation but also of effector function, and it was suggested that this may be due to depletion of energy production (688).There are, however, cases where ouabain treatment seems to be able to induce ‘cspontaiieous” cytotoxic mechanisms in mononuclear leukocytes in culture (337), but the mechanisms of this curious effect have not been explored. A nonpolar digitalis glycoside (lanatoside C ) was found to be a potent mitogen for mouse B lymphocytes (234). However, there was no direct correlation between the effects of glycoside on the Na+,K+-ATPase activity and the activation of the lymphocytes (235). Potcissiuni ioiiophores at lo-” to lop6M can also impair the mitogen stimulation of lymphocytes. One of them is valinomycin, a neutral cyclic depsipeptide, which forms a lipid soluble complex with K+, highly specific for that monovalent cation (242,502).Thus, it carries K+ through biological membrane by diffusion, and when it is in the presence of lymphocytes, it inhibits their mitogenic activation by lectins ( 1 10,542,558);its inhibitory properties can be reversed by increasing the K’ ion concentration outside the cells (110).The other ionophore is nigericin, a K+-H+ exchanger (502), which gives also a reversible inhibition (112). The inhibited step is an early one, since inhibition is obtained only when the K+ ionophores are added to the cultures within
44
FRANCIS LOOR
the first 24 hours of culture. Thus, loss of intracellular K+ impairs some early step of the triggering mechanism, and an increase of intracellular K+ concentration (either due to arrest of pumping out K+ or to loss of water) appears to be associated with the lymphocyte activation. It is in any case difficult to find a link between such a fact and the MT and M F functions in the cell cortex. It is possible that activation was mediated by a change of membrane potential and has little or nothing to do with MT and MF. However, since the Na+, K+-ATPase controls transport systems essential for the cell metabolism, such as uptake of sugar and amino acids, there may be rapid consequences at the level of cell cortex activities that would modulate its capacity to transmit the mitogenic signal further. Such an indirect effect may be on the transport of other ions, since ouabain, although it does not affect the early Ca2+ion uptake that follows mitogenic stimulation of the cells, was found, however, to inhibit partially Ca2+uptake 12 hours after lectin binding (673, 674). With regard to the capping process, no detectable inhibition was found with ouabain (F. Loor, unpublished data, 1973). This result had been obtained without a prolonged preincubation in the ouabain-containing medium, and it may be necessary to repeat the experiment, since other effects of ouabain have been found to require such preincubation for a few hours (688). Similarly, lo+ M valinomycin had no detectable effect on capping of membrane immunoglobulins and of Con A (497).This is true at least when tested at the classical temperature for capping test, i.e., 37°C. However, at room temperature (2122OC) at which capping is slower, valinomycin was found to be capable of inhibiting the capping of membrane immunoglobulins, detectably even at M ( 111). At M , valinomycin inhibits morphological changes and capping of mouse lymphocytes bound to nylon fibers, also usually observed at room temperature (2lOC) (548). Although no reversibility of the capping inhibition was found in the latter case, in the former, however ( l l l ) ,it was found to be reversible by washing the drug away, and some prevention of the inhibitory effects of the drug could be achieved by increasing the concentration of K+ ions in the external medium (111). The inhibition of capping and of lymphocyte morphological changes by valinomycin may be due to its selective translocation of K+ across the cell PM with a consequent alteration of its electrical properties. This in turn may alter the interaction of Cs or Cm elements of the cell cortex with the PM, resulting in capping inhibition. Another possible cause for such inhibition may be found in that valinomycin can also uncouple oxidative phosphorylation and thus interfere with energy metabolism; this has been shown also to
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
45
impair capping (e.g., with DNP), although usually much higher concentrations are needed (-10-’ to lop3M ) to obtain good inhibition (370, 570). The different results obtained for the effects of valinomycin on capping at 37°C or at room temperature may be due to the general observation that the capping, which is slower at room temperature, is much more sensitive to any inhibitor whatever of metabolism of M F or others, possibly because of alterations of membrane organization. Cnlciuni ions have also been implicated in the mitogenic process, and their effects on the M T and M F of the cell cortex may be more direct than the effects of monovalent ions; this still does not prove that Ca2+ions influence the mitogenic stimulation by means of M T andlor MF. The requirement for the presence of bivalent cations in the medium stems from the effective inhibition of PHA-induced lymphocyte transformation by citrate and EDTA, and it was shown that, although Mg2+, Zn2+,and Fe’+ were important, the most essential cation was Ca2+(8).The presence of 1-1.6 m;zI Ca’+ would enhance PHA stimulation of lymphocytes (246). N o stimulation at all of human or mouse T lymphocytes could be obtained at extracellular Caz+ concentration lower than 3 x 10*M (270).Removal of Ca2+specifically by EGTA up to 16 hours after PHA treatment can inhibit the mitogenic response
(673, 674). Recent data suggest that the Ca’+ ions are specifically needed “for one or more of the very early steps in the mitogenic activation of T lymphocytes” (270).Uptake of Ca“+ions by the lymphocytes was found to show a rapid increase, occurring within 1 hour of the cell-mitogen contact (673, 674). Such a rapid influx of Ca2+was found to occur in lymphoid cells treated by a variety of mitogenic agents, and it has been proposed that early steps of mitogenic stimulation were dependent on Ca’+ fluxes (177,473,512,673,674).Ligand binding, or even mitogen binding to the lymphocyte membrane, does not, however, automatically induce a C i 2 + influx. It is well known that lectins that are mitogenic for T cells bind as well to B as to T cells (362), and it has been shown that such lectins open “Ca’+ gates” on T cells, but do not on B cells (177).Nonmitogenic lectins, such as WGA, bind but do not increase Ca’+ uptake (176). For some reason, the Ca’+ gating phenomenon had not been observed with B cell mitogens (177) until recently (176),but the regularity of the induction of an early Ca’+ uptake seems to be less for B cell mitogens than for T cell mitogens. In particular, different mitogenic LPS preparations influence Ca” uptake in B cells very differently, from no detectable increase to an increase as great as the one obtained with T cell mitogens on T cells (176).The selectivity of the Ca2+gating phenomenon for B and T cells was reassessed (Con A
46
FRANCIS LOOR
and PHA work on T cells only; LPS on B cells only), and, interestingly, using rabbit lymphocytes it was shown that a mitogenic anti-allotype antibody also stimulates an early Ca2+influx in these cells (176). With PHA-stimulated human lymphocytes, a rather long Ca2+ gate opening (24 hours) was observed (673,674), while with mouse spleen cells, Con A or PHA treatment was reported to induce an extremely fast uptake of 45Ca2+: it was detectably increased 45 seconds after addition of the lectin and completed by 1minute (177). Such a fast increase could not be reproduced in a more recent study (176):the initial rise in 45Ca2+influx occurs within 6 minutes addition of Con A, PHA, or active LPS preparations to the mouse spleen cells, the maximum 45Ca2+content of the cells being reached within 6 minutes of PHA or LPS addition and within 30 minutes of Con A, then remaining constant for some 4-6 hours, and finally decreasing during the next 12-20 hours. Ionophorous properties were also identified for another mouse B cell mitogen, keyhole limpet hemocyanin (KLH) (52), and another ionophore excitability-inducing material (EIM) (333)was found to be a B cell mitogen (542). No straightforward correlation exists, however, between ionophorous activity and mitogenicity, even though this is sometimes the case. Attempts to introduce Ca2+through ionophores have been made. “Calcium ionophores” A23187 and X537A can carry Ca2+through the membrane: this allows modulation of the concentration of CaZ+ ions within the cell as a function of their concentration in the extracellular medium. They should not be termed specific Ca2+ionophores, however, particularly when used at high concentration, because they are not strictly specific for Ca2+and can actually transport other bivalent cations across biological membranes (242, 502). In particular, the commonly used A23187 is a monobasic carboxylic antibiotic molecule (made by Streptomyces chartreusensis),two molecules of which combine with the divalent cation as a lipid-soluble complex; although it has a high affinity for Ca2+[it has a 100 times higher affinity for Mn2+ (483)1, it also allows translocation of other divalent cations through membranes quite well, especially Mg2+(516). Furthermore, it can also form lipid-soluble complexes with leucine and other amino acids (268). X537A forms complexes not only with divalent cations, but also with monovalent cations and amines (502). This is not always kept in mind when using them. Nevertheless, A23187 has been shown to be able to trigger, in the presence of extracellular Ca2+,blast transformation, RNA and DNA syntheses, and mitosis in human, rabbit, and porcine lymphocytes (269,290,376,386,522). The appearance of the cells at the optical and
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
47
electron microscope levels was the same as for PHA blasts except that mitochondria showed quite peculiar abnormalities, perhaps due to abnormally high fixation of C i 2 + and alteration of respiration (386). Both the amount of A23187 engaged and the Ca2+concentration in the medium show optimal values and minimum exposure times to obtain triggering (269, 376, 386, 522). Competition for binding to A23187 by La:$+ions, to block Ca2+transport, also results in abrogation of triggering (376). The stimulation of lymphocytes by A23187 could be obtained only in a very narrow range of doses, becoming rapidly toxic at higher doses (376,386),but it does not stimulate mouse lymphocytes at any concentration (542). It is not without effect, however, since in a system where mouse lymphocytes can be stimulated by two pulses of Con A, the first given at 0-3 hours and the second at 15-18 hours, the first Con A pulse can be replaced by a pulse of A23187 (638).Thus, the early Ca2+uptake may be needed for triggering, but in the case of mouse lymphocytes, either the higher Ca‘+ concentration in the cell becomes toxic or the Ca2+uptake is not sufficient for the development of later stages of mitogenesis. With regard to the early Ca’+ uptake, it was shown, using 45Ca2+,that human and mouse lymphocytes incubated with a low, “nontoxic” concentration of A23187 (-60 mM) show a rapid Ca2+influx within 1 minute of treatment, with the maximalization of the celliilar Ca2+content within 1 hour and its stability over the next 4-6 hours (176, 290). The early kinetics is thus not dramatically different from that obtained with Con A, but at variance with the latter, quality proven, mitogen, the increase in cellular Ca2+content brought about b y A23187 is much higher (sixfold instead of twofold with mitogens) (176). Whether it is introduced in the cell by ionophorous carriers or by gates open in the membrane by mitogens, Ca2+would passively diffuse through the membrane in the cell, its influx being independent on energy metabolism, while its efflux does require a Ca’+-dependent ATPase. At the steady state there would be equilibrium of influx and efflux mechanisms. The efflux would take over 12 hours after mitogen contact as a result of the closing of the Ca2+gates and/or of the increase in the number of metabolic units engaged in Ca2+efflux. Regarding the possible involvement of MT and/or M F in these phenomena, a few preliminary experiments seem to indicate that drugs that affect the structure and function of M T and MF, as well as those that modulate the intracellular levels of cyclic nucleotides, can all modulate the opening and closing of Ca’+ gates by mitogenic lectins (176), but interpretation of the data is difficult given the undissociable side effects of such drugs and the irregularity of a given effect [e.g.,
48
FRANCIS LOOR
colchicine does it, lumicolchicine (no effect on MT, but same effects on membrane as colchicine) does not do it, but unfortunately, for the simplicity of the interpretation, vinblastine does not do it well either]. Finally, it is still difficult to know ( a ) the actual distribution of the Ca2+that is taken up early after stimulation, namely, how much of it does not become sequestered in mitochondria and can therefore interfere with Ca2+-dependentactivities or stabilities of M T and M F systems; and ( b )what can be the consequences for M T and M F of the other alterations of cellular metabolism that the increase in intracellular Ca2+provokes. The effects of modulating intracellular Ca‘+ concentration have been directly tested on cell surface activities. Divalent cation ionophores A23187 and X537A have a definite effect on the capping process, even though controversial, probably because of differences in the experimental conditions (497, 569). Depletion of bivalent cations by EDTA or citrate or of Ca2+more specifically by EGTA do not detectably affect capping (363,570). In the absence of extracellular Caz+,the A23187 and X537A ionophores do not have any effect either (497,569). In the presence of extracellular Ca2+(2 x 10P to 4 x lop4M ) in the culture medium, some found still no inhibition of mIg capping by A23187 and X537A, but on the contrary relief from Coil A-induced inhibition of mIg capping (497).This may be interpreted as due to MT disruption by higher Ca2+concentration inside the cell. Others, however, obtained inhibition of capping of mIg by A23187 in presence of extracellular Ca2+, and even by a 20-100 times lower dose of ionophore than the first group (569). Furthermore, preformed caps were disrupted, reversed to spots, by introduction of Ca2+inside the cells b y A23187; however, at variance with the disruption obtained with local anesthetics (568),the disruption by A23187 required simultaneous inhibition of energy metabolism (569). At variance with the mechanism suggested for capping inhibition and cap disruption b y local anesthetics, i.e., displacement of Ca’+ from the membrane with disengagement of M F because of loss of anchorage to the PM, the requirement of energy metabolism blockage to disrupt cap by A23187 and Ca2+,suggest that the association of M F with the PM is maintained and that M F remain contracted in the cap and can be relaxed only b y inhibition of energy (=equivalent to “growth of MV, M F bundles,” by ATP depletion). Why does A23187 + Caz+inhibit capping, in the absence of the inhibition of energy metabolism, when its reversion requires such a condition? One could suggest that the ionophore, although it introduces Ca2+in the cell, lets its Mg2+ leak out, and that
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
49
depletion of Mg’+ in the cell cortex would lead to a malfunction of Mg’+-dependent ATPase of actomyosin. Some batches of A23187 at least can transport Mg’+ quite efficiently (e.g., 238). This may not have happened with the A23187 batch used b y the group who could not obtain direct capping inhibition. There are other alternative explanations for both the mechanism of stimulation b y the A23187 ionophore and its effect on membrane dynamics-that is, that the ionophore molecules “perturb” the membrane organization. This may be suggested by the interesting observation that, although this requires Ci2+in the medium, rabbit lymphocyte activation by A23187 is not actually dependent on Ci2+translocation inside the cell: there can be activation in conditions where the A23187 is saturated with M$+ or Mn2+ ions and does not transport Ca2+ (522). The suggestion is that the ionophore that intercalates into the lipid phase of the membrane would initiate “the same changes in the membrane itself that are induced b y mitogens that bind to surface receptors,” i.e., increase the turnover of‘membrane phospholipids, as actually observed (522). Lithium ions have recently been shown to modulate lectin-induced mitogenesis of lymphocytes. The concentration of Li+ ions in human beings and animals modulates a number of physiological functions; the mechanism of this modulation is, at the molecular level, being attributed to known effects of‘ Li+ on the activity of many enzymes, among which are adenylate cyclase and membranous Na+,K+-ATPases. It would therefore affect, by complex and ill understood ways, energy metabolism, ion distribution, cyclic nucleotide metabolism, and so on (for review, see 288, 292). Side effects of pharniacological treatments with Li+ have been noted as alterations of the immune system, with a concentration of Li+ in the cortex of the thymus and the involution of the organ (432, 480). 1n uitro both enhancement and inhibition of mitogenic responses have been described, and Li+ ions alone, in absence of lectin, have no reported effect on the proliferation of lymphocytes (192, 245, 247). Thus, Li+, at 5 x lop3M , was reported to increase PHA-induced human lymphocyte stimulation (192); at 1 to 10 x M , Lif greatly enhances the hamster lymphocyte proliferation induced by optimal and suboptimal (threshold) concentrations of PHA, the effect being specific to Lit monovalent cations and being obtained only if it is added within the first 24 hours of culture (245). With Con A as a mitogen, modulation of the proliferative response by Li+ ions was obtained only for thymocytes, but not for lymphocytes from other
50
FRANCIS LOOR
sources; potentiation of the response to suboptimal (threshold) doses of Con A is obtained by 1 to 25 x M Li+ ions, an effect specific to that ion and not competed for by other ions. However, there is inhibition of the response to optimal and supraoptimal doses of Con A, an effect specific to Li+ ions, which can be competed for by K+ ions, but not by Na+, Ca2+,Mg2+,or a combination of them (247). Li+ ions may have an effect on activities of the cell cortex, since preincubation of human lymphocytes with 5 x 10" M for 5 minutes keeps intact their capacity to form E rosette (human T cell marker) in the presence of drugs that elevate their endogenous CAMP level, and which alone abrogate E rosette-forming capacity (192). Thus, Li+ may modulate CAMP level; since the latter has been implied in the control of the degree of M T polymerization, it is possible to build up a relation between Li+, CAMP, MT, lectin binding to lymphocyte PM components and triggering, but I leave to each the responsibility of building up such a relation in his personal way. Rubidium ions were recently shown to inhibit lymphocyte mitogenesis induced by a variety of mitogens, among which are LPS, PHA, Con A, PWM, and purified protein derivative, when used at 10 to 50 x 10" M (236),confirming an earlier observation (558), at which concentration Rb+ was a Na+, K+-ATPase inhibitor (330). Its effects may thus be similar to those of other inhibitors of the enzyme, which already have been discussed. The biudent cations MnZ+,Coz+, CdZ+,Cu2+,and Ni2+ all inhibit DNA synthesis of normal human lymphocytes at concentrations as low as lop7M (40).This toxic effect may be due to competition with other physiologically important metals. Trace metal ions such as Pb2+,Cd", and C?+ have variable effects on the ability of lymphocytes to be transformed by mitogens, both T (593) and B (187) lymphocytes: they can be toxic, but each of those metal ions can also be mitogenic b y itself. Mn2+ ions have been reported to enhance mitogen stimulation of M , but to inhibit it at lo4 M . Further, they lymphocytes at lop6to have differential effects 011 different cell types, as shown by the facts that responses to T cell mitogens are affected but those to B cell mitogens are not, and that lymphocytes from different sources show different susceptibilities to be protected from Mn2+inhibitors by other bivalent cations (due to difference of lymphocyte maturity?). Whatever complex this is, the inhibition by Mn2+can be obtained only if added within the first 4-16 hours after stimulation, thus it inhibits an early step of the mitogenic activation (243). M 8 + ions would have an enhancing effect on PHA stimulation, at 1
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
51
to 10 x M (246).No enhancing effect had been found for Ca2+and Mg'+ on the DNA synthesis of normal lymphocytes (40). Zinc ions have effects that deserve more comment. Zn2+ions have been shown to stimulate DNA synthesis and blast transformation (at the optimal concentration of lo4 M ) in human lymphocytes (40, 317, 546,547) and animal lymphocytes (244).Zn2+ions appear to be essential for mitogenesis (8),and indeed chelation by o-phenanthroline blocked thymidine uptake by PHA-triggered lymphocytes and blocked their activation, the nonchelating analog rn-phenanthroline being without effect, and the inhibition by o-phenanthroline of lectin-mediated activation could be reversed by addition of Zn2+ or prevented by Zn" or NiZ+(677). Differential effects on B and T cell activation by LPS and Con A would be mediated by 1 to 5 x M Zn2+, since B cell activation is enhanced while T cell activation reM , Zn2+ inhibits mains unchanged (244). However, at 1 to 5 x PHA-induced stimulation of human lymphocytes (40, 230), and the stimulation of rat thymocytes by succinyl-Con A (390)to make RNA and DNA and to transform into blasts. The step of stimulation that is inhibited b y Zn2+is a very early one, the strongest inhibition being obtained when Zn" was present during the first few hours of mitogentreated culture, and no effect of Zn2+being detectable when added later on, when DNA synthesis is taking place (390). If Zn2+ ions can either stimulate or be toxic, they might also be essential for some stage of lymphocyte activation by lectins, since addition of Zn'+ to the medium is the means to reverse the inhibitory effects caused by chelators such as a-phenanthroline (677) and EDTA (85).More recently, a sulfhydryl metal chelator with a high affinity for Cuz+,Ni2+,and Zn'+, sodium diethyldithiocarbamate, was shown to have inhibitor effects on the PHA stimulation of human lymphocytes; its toxicity was found to be biphasic, with a fast phase at 2.5 x M and a second at doses higher than 2.5 x M , while no toxicity was found at intermediate doses, such as 2.5 x lop4M . As partial reversion M Zn2+and of the fast phase toxicity could be achieved by 2.5 x the second one by lo-' M Cu'+, it is suggested that the drug inhibits the function, in the first case, of a Zn-metalloenzyme and, in the second case, of a Cu-metalloenzyme (526). Zinc is a trace metal whose deficiency, either inherited or acquired (e.g., induced by a special diet), has dramatic consequences on development; this is not unexpected because it is an essential element of many metalloenzymes (among which are DNA and RNA polymerases) involved in nucleic acids synthesis, in protein synthesis and degradation, and in energy metabolism. Both in human beings and in animals,
52
FRANCIS LOOR
zinc deficiency, either natural or induced, results in various immunodeficiency states, among which is thymus atrophy (329), most of which can be corrected by administration of Zn (66, 164, 500). The effects of Zn2+ on lymphocytes do not seem to require its interaction with cytoplasmic proteins (230), and stimulation of lymphocytes can be obtained by a nonpenneable probe, such as ZnZ+ complexed to transferrin (454), and as Zn2+chelated with hydroxyquinoline (86, 87). This suggests that the effects of Zn2+ might thus be at the membrane level, and indeed lymphocytes have receptors for transferrin (484),Zn2+is an inhibitor of membrane-bound Na+, K+-ATPase (87), and it is an antagonist of Ca2+for red cell membrane (137, 138). It has been suggested that Zn2+would interfere with stimulation at the level of the membrane or of the cell cortical constituents by some modification of cell surface receptor number and mobility (390). There are indeed many possible ways by which ZnZ+could interfere with membrane dynamics and organization, from the reaction with intrinsic PM components to the inactivation of energy-providing enzymes (e.g., see 86,87). While Con A or PHA stimulation of human lymphocytes is inhibited by lo4 M Zn2+when the cells come from aged donors, no effect or a slight enhancement is obtained for similarly treated cells coming from young persons (514);the differential effect of Zn2+on the two groups might then be related to the alterations of cell membrane structure that occur with aging (223), but it seems to me difficult to establish which relation exactly. At 3 x lo-” M , Zn2+inhibits E rosette capping on human lymphoM it restricts, though weakly, Con A cytes (400), and at 5 x receptor mobility (390) and delays membrane immunoglobulin capping on B cells (389). There are reports that Zn2+,as well as Cd2+, stabilizes membranes (199,435),producing larger fragments to which most cytomuscular proteins of the cell cortex remain bound: actin, myosin, a-actinin as well as another material (spectrin-like, filaminlike, or actin-binding protein-like) (415). From such membrane fragments all cytomuscular proteins but myosin can be extracted by EDTA, thus indicating that they are peripheral proteins except for myosin, which may be an integral membrane protein. Since both the interaction of myosin with actin and the actomyosin function require ATPase activities, and these have essential SH in their active sites and Zn2+ is a SH blocking reagent, one may speculate on inhibition of capping as due to Zn2+inhibition at the level of interaction of actin MF with membrane-bound myosin. On the other hand, Zn2+promotes in vitro a very special kind of polymerization of tubulin (also an SHcontaining protein): formation of sheets of tubulin rather than MT;
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
53
although the process of polymerization is not known, either it is not related to the SH blocking capacity of Zn”, or it is specific to this ion, since other SH reagents do not induce similar tubulin sheets (335).It is so far not known whether or not those effects of Zn2+on tubulin in vitro can be extrapolated to intact cells, but hyperpolymerization of tubulin would probably restrict mobility of the plasma membrane and of the whole cell. I would favor a Zn2+effect at the level of M F since it enhances E rosette formation (which depends on M V structures) while leaving unaltered the complement-dependent rosettes (which do not depend on MV, see Section V,B) (400). Mercury ions also deserve special comment, as they have been shown to stimulate DNA synthesis and blast transformation in human lymphocytes, the optimal concentration of Hg2+being -lOP5M (40,41, 565). Particularly interesting are the studies performed with organomercurials, where Hg2+ions are complexed with organic material to make various polyvalent, bivalent, and monovalent mercury derivatives (41); while a monovalent mercury derivative and a multivalent one showed only the Hg toxic effects, a divalent mercury derivative is a potent stimulator of lymphocyte mitogenesis. The interpretation is that only the divalent one is a bifunctional sulfhydryl reagent capable of appropriately cross-linking some unidentified PM component(s), the monovalent being unable to cross-link and the multivalent inducing a too extreme cross-linking of cell surface sulfhydryl-containing components. The analogy with the effects of antibodies and lectins are evident (see Section VIII). Far from making things simpler to understand, there are data on the interplay of several of those ions on the mitogenic responses: varying the molarities of Li+ and Ca2+ions and looking at the mitogenic properties of Zn2+ and Hg2+ ions (246, 247). Thus, 1 mM Ca2+ 10 mM Li+supplemented media potentiate the mitogenic effects of both Zn2+and Hg”, though differently with regard to the characteristics of the dose response curve. Not much interpretation is provided, however. (How could it be?) Channel-former type ionophores allow translocation of ions through membranes by forming channels through it, thus by a mechanism different from that of the valinomycin, A23187, and X537A ionophores (which form complexes with the ions and diffuse through the membrane). A series of them have been used ( a ) to try to stimulate lymphocyte to mitosis; and ( b ) to test their effects on mitogen-induced lymphocytes (233,542).These channel-former ionophores were gramicidin A and excitability-inducing material (EIM), which make membranes permeable to all cations; the Streptomyces-derived poIyene antibiotics nystatin and amphotericin B, which show anion prefer-
+
54
FRANCIS LOOR
ence; and alamethicin, which makes them permeable to both cations and anions. These substances also show a different requirement for permeation (394, 542). In one study, only one channel-former ionophore (EIM) was found to be mitogenic for mouse lymphocytes by itself. Mitogenesis induced by PHA or LPS was strongly inhibited by low doses of gramicidin A, but slightly enhanced by similar low doses of alamethicin and nystatin. These were inhibitory at high doses, though not toxic for the cells (543). In the other study, both nystatin and amphotericin B were found to induce DNA synthesis and polyclonal antibody production in murine B cells (233). The simple induction of ion flows through the cell membrane might not provide by itself a signal that would be sufficient for activation. The lack of selectivity of all these channel-forming ionophores makes it difficult to draw firm conclusions, as practically every ion can leak through the membrane when some of them are used. More demonstrative of a lack of correlation of the ionophorous properties and of the mitogenic properties are other experiments, using heat-denatured EIM and KLH: this destroys their ionophorous activity but preserves their mitogenicity (542). Thus, ionophorous activity may be a side effect of some mitogen, and not only would not be sufficient, but also would not even be needed for mitogenicity. Finally, the crucial importance of polyvalency of mitogens was stressed, since both EIM and KLH are polymers of high molecular weight, and their mitogenicity, as well as their ionophorous capacity, are lost upon dissociation into monomeric units (542). This suggests again that clustering of some specific membrane components should play a crucial role in the activation process.
3 . Modulating Membrane Lipid Composition Various alterations of the membrane lipid composition have resulted in modulation of the mitogenic responses of lymphocytes to lectins (e.g., 18, 82, 378, 407, 433, 434, 478, 501, 550, 672). A number of nonexclusive interpretations can be given to such a modulation, some of which are trivial. It is evident that changes of membrane fluidity can alter, for instance, the early fluxes of ions across the PM that follow stimulation (see Section VI,C,2), and more generally the permeability to, or the transport of, all essential metabolites. But this is not always the case in the examples given here. A disturbance of the normal lipid composition of the membrane may alter some of the changes of fluidity perhaps needed for further steps of the triggering, such as the early, transient increase of fluidity, the decrease of viscosity reported to occur within some 30 minutes of
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
55
lectin binding or with its return to normal, which follows within 1hour (28, 636). It may interfere with some unknown function of the enhanced phospholipid turnover that follows lectin binding (e.g., 501, 519).An alteration of membrane fluidity may also alter the capacity of a ligand to aggregate the membranous determinants to which it binds; as will be seen later, such aggregates seem to be crucial for the first triggering step to take place, even if not sufficient. Alteration of membrane lipids composition may also interfere with triggering because of altered interaction of the liganded PMC with the cell cortical MF/MT structures, and this may be a more subtle way of action. Indeed, an altered membrane lipid composition may result in altered distribution of PM components; for instance, there could be segregation of these components into more or less fluid menibraiie domains leading to nonuniform distribution of the components that are recognized by the ligands; especially they may be absent from some menibrane domains that would show interactions with the cell cortex MF/MT structures, or inversely they would redistribute to membrane domains that do not show interactions with such structures. Thus, the binding of a mitogenic ligand would still result in appropriate aggregation of the corresponding PM components, but this would not be followed by the interaction with the MF/MT structures, since aggregated PM components and component-associated MF/MT are in different domains. The consequence would be no transmission of the signal, or too long opening of the Ca2+gates because of too slow removal of the clustered conipohents, or anything else resulting in a lack of cellular activation or into toxic effects. Which ways have been used to modify lipid composition of lymphocyte membranes, and what were the effects of such alterations? Some are reviewed herein, but within the limits of this review I will not comment on them very much. The lectin activation of lymphocytes could be modulated when the cells were grown in culture medium containing saturated and unsaturated fatty acid (407,478,672).If cells are treated in culture with avidin to impair their own fatty acid biosynthetic capacities, they will use the exogenously supplied fatty acids (e.g., from the serum). Using medium containing lipid-depleted serum but supplemented with specific fatty acids, one can then influence greatly the eventual lipid composition of the cell niembraiie (287, 679). Such an approach has used, as an exogenous supply, either oleate (monounsaturated fatty acid, with a cis double bond) or elaidate (trans isomer of oleate). Oleate (melting point 13.4"C) should lower, whereas elaidate (melting point 45°C) should raise, the liquid crystalline transition temperature of the membrane. This results in changes of the
56
FRANCIS LOOR
agglutination properties of fibroblasts by lectins (267). No important effect was found, at 3TC, on the stimulation of lymphocytes by Con A. In contrast, around 30"C, oleate-enriched cells were still as well activated as at 3TC, whereas elaidate enriched cells were much less so (about 2.5 times less), the control cells being intermediate (378). A different type of approach consists in modulating lymphocyte membrane composition by treatment with liposomes of controlled composition. All experiments mentioned hereafter show satisfactory controls that viability of the cells was not drastically altered. Thus, the treatment of lymphocytes with lecithin-cholesterol liposomes decreases the mitogenic response to various lectins (530, 637). Incubation of the lymphocytes with pure lecithin liposomes also results in a suppression of the induction phase of lectin-mediated activation, in a reversible way (530). While lecithin-cholesterol liposomes can b e used to increase the level of cell membrane cholesterol relative to phospholipids, pure lecithin liposomes have the opposite effect, yet both depress lymphocyte activation. Thus, it would appear that the nature of the added lipids is less important than the perturbation of the lymphocyte membrane fluidity or organization they provoke. Along the same lines, it has been observed that egg lecithin can abrogate human lymphocyte blastogenic response to a series of T cell and B cell mitogens (433). This latter inhibition occurs best if the lecithin is added at the same time as the lectin to the lymphocytes and less and less with time thereafter, being without effect after 48 hours. Sonicated phosphatidylserine, phosphatidylethanolamine, or lecithin are all similarly potent inhibitors of blastogenesis, whereas unsonicated lecithin is less so and sphingomyelin, which gives higher viscosity, is not potent (433,434). The differential effect of sonication is interesting inasmuch as it may be related to the observation that sonicated egg lecithin liposomes induce the expression of MV on the surface of fibroblasts, whereas unsonicated egg lecithin liposomes and sonicated dipalmitoyllecithin do not (showing the importance of the chemical nature, but also of the physical state, of the liposome lipids for the induction of MV formation) (388).Where a lecithin liposome fuses with the membrane, one can expect the constitution of a domain of higher fluidity; thus this may cause further segregation of other membrane components and "disconnection" of PM components having primary recognition function from components having the function to relay the signal and to transmit it inside the cell. The differences of inhibitory effects caused by lecithin, whether it is sonicated or not, may b e related to its capacity to actually fuse with the membrane, since sonicated lecithin liposomes
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
57
are small single-layered ones (with lipids in the liquid state), whereas unsonicated ones are multilayered ones, and the latter enter the cells mostly by endocytosis rather than by fusion with the PM, as also do liposomes whose lipids are in the gel state, such as the dipalmitoyllecithin lipids (391,470). Although some membrane properties such as MV expression may thus appear to be altered by liposomes (388), not every parameter allows detection of changes, such as cell aggregation by PHA, which can remain unchanged (433). Important in the present context is the observation that polymorphonuclear leukocytes treated with liposomes composed of phosphatidylcholine, cholesterol, and dicetylphosphate, although they keep intact a number of functions, have some others altered, in particular a reduced capacity to phagocytose bacteria and a decreased random mobility (109,621); thus, the exogenously incorporated membrane lipids can affect functions that are definitely related with cell motilty-that is, functions of cell cortex MF/MT elements. Another means of changing the lipid composition of cellular membranes is to interfere with their sterol synthesis, and inhibition of sterol synthesis by the presence of oxygenated sterols in the medium leads to decreased blast transformation and mitogenic activation of lymphocytes by lectins: 20u-hydroxycholesterol and human lymphocytes (501), 25-hydroxycholesterol and niouse and human lympho- . cytes (18, 82). The mechanism of the inhibition remains unclear: it may alter the response at many stages, either the increase of synthesis of sterol and phospholipid that follows stimulation or the early change of membrane viscosity that follows Iectin binding, although these changes were also found to occur with other nonmitogenic lectins that agglutinate the lymphocytes (501), and they may not be actual components, but only pennissive components of the triggering sequence. In this context, it is recalled that depletion of cholesterol content in the lymphocyte membrane alters the recognition capacity of killer cells (256; see above), and that may be related to the impaired expression of MV by fibroblasts depleted of cholesterol by similar procedures (350). Depletion of cholesterol and enrichment in lecithin would have opposite effects on MV expression (even though expression of more MV cannot be autoinatically linked to higher cell cortex activities, particularly if such higher expression results in a disturbance of PM component organization). Those effects of cholesterol content in the lymphocyte membrane on their immunological reactivity recall the hypothetized bioregulator function of cholesterol (277) and suggest that the alterations induced in vitro may have physiological in uiuo equivalents.
58
FRANCIS LOOR
Finally, a last manipulation of in vitro responses of lymphocytes makes use of gangliosides; these sialoglycosphingolipids are common integral membrane components of the mammalian cell; and in the case of lymphocytes some are found almost exclusively on the T cells. Gangliosides are amphipathic molecules, both water soluble and liposoluble; this physicochemical property would enable gangliosides, unlike phospholipids, to migrate passively between adjacent cell surfaces through aqueous boundaries (419).A series of brain gangliosides were shown to be able to suppress the LPS activation of mouse B cells, and one of them, on the contrary, was strongly mitogenic by itself for mouse spleen cells (550).This may indeed suggest that T cell-derived B cell-active enhancing and suppressing factors were of ganglioside nature, in which case ganglioside translocation would constitute a new, unsuspected, non-antigen-specific physiological modulation of immune responses.
4 . Thiols und Sulfnydryl Reugeizts A series of low molecular weight sulfhydryl containing compounds have been shown to be consistently capable of enhancing various immunological responses of lymphocytes, although the mechanism of such enhancement has received little attention thus far. In contrast, a series of sulfhydryl binding reagents have been shown to inhibit lymphocyte responses. The possible cellular targets for thiols and sulfhydry1 reagents are multiple, i.e., as many as all the SH-containing molecules of the cells, namely, all the proteins. However, in the context of this review and because of some recent data from my laboratory, I shall restrict the SH molecules we are concerned with to only a few: membrane proteins, tubulin, and myosin. That other interpretations of the effects of thiols and S H reagents are possible must be kept in mind. The presence of some thiols in the culture medium appears to favor the proliferation of some cell lines (among which are lymphoid ones), their disulfide being also effective (64,65). Thiols are also required for the clonal proliferation of single lymphocytes cultured in gelified media (408). Thiols may functionally replace macrophages, at least for some types of immunological responses and in some culture conditions (81, 95,discussed in 567). More specifically, thiols enhance the transformation and DNA synthetic responses of lymphocytes to a variety of anti-Ig and lectin reagents (65,157,309,469,F.Loor, unpublished). Thiols also enhance both the proliferative responses of T cells to allogeneic cells in vitro [increased mixed lymphocyte reaction (MLR)] (48,253,309)and the generation of the cytotoxic T cells [increased cell-mediated lympholysis (CML)] (48,77,147).
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
59
The thiol compounds that are used include mainly cysteine, glutathione, cysteamine, 2-mercaptoethanol, dithiothreitol, a-thioglycerol, thioglycolate, and also sulfite. However, when looking at the data more closely, it appears that important differences exist from one assay system to the next with regard to the nature of the thiols that are effective, or are not, to the effective concentrations to be used optimally, and also to the obtention, under some conditions, of inhibition instead of potentiation. This excludes any simple interpretation to be made, particularly because of the relative absence of systematic studies concerning the kinetics of action of thiols, their chemical modifications, the analysis of various cell activation parameters, not just one, for each assay system and so on. A few examples of such effects of the thiols are worth mentioning. The presence of thiols in the culture medium seems to be essential; for instance, in protein-free medium a unidirectional mouse MLR could be obtained if the medium contained some reducing agents, such as 2-mercaptoethanol (at optimum: 2.5 x lop5M ) , L-cysteine (2.5 x M ) , or dithiothreitol (5 x lo4 M ) , but reduced glutathione did not show any “permissive” activity over the to lod3M range (253). At variance with this, however, glutathione potentiates the stimulation of rabbit lymphocytes b y PHA or antibody reagents (157). For the same cell population, the dose of reducing agents that gives maximal potentiation depends on the nature of the reducing agent, on the nature of the ligand, and on its dose: for example, the stimulation of rabbit lymphocytes by various anti-Ig reagents is best potentiated by 5 x 10P M reduced glutathione, lop3M cysteine, or M sulfite; 5 x 1W3hi’ cysteine does not enhance rabbit lymphocyte stimulation by anti-Fab, anti-Fc, or anti-allotype reagents as well as does lW3M cysteine, but it is the optimum concentration for potentiating their stimulation by PHA; finally, the best cysteine concentration to enhance PHA stimulation of human lymphocytes is 5 x lop3M when PHA is at 50 pglnil and lo-* M when PHA is at 20 pg/ml (all examples from ref. 157). Similar observations were made in the other studies quoted above. In all cases, it was shown that the thiol group was important for the effect to be obtained, since the non-SH analogs (such as ethanol for mercaptoethanol) or their S-alkylated derivatives (such as S-methyl-L-cysteine) were not effective (e.g., 65, 253). Higher doses of thiols may become inhibitory; for example, M cysteine inhibits the stimulation of rabbit lymphocytes by anti-Ig reagents, which may be explained by the reduction of part of the antibody since their monovalent Fab and Fab’ do not stimulate (157, 158).While disulfides from active thiols can be enhancing too, they may also become inhibitory, depending on
60
FRANCIS LOOR
the conditions (e.g., 65). Whereas L-cysteine can enhance lectininduced lymphocyte transformation, D-cysteine does not do so and can even strongly inhibit it (309). Generally, it has been shown that the thiol compounds did not have any stimulatory effects by themselves (e.g., 65, 157, 309), although some researchers claim that 2-mercaptoethanol and a-thioglycerol were mitogenic by themselves, for a small proportion of lymphocytes (210-213, 276, 342). Thus, it is evident that thiols do something, but it is still difficult to determine what they do and particularly when they do it, since in most studies the thiols were present throughout the culture period and could be effective at many stages of cell transformation. It has been reported that potentiation of the response requires the presence of thiol during the first 24 hours of culture (253). However, under the cell culture conditions, the added thiols are present for only short periods in the reduced state. Further, it is not clear whether the physiologically active form is the reduced form or the oxidized one. If disulfides enter the cells, they will be readily reduced given the redox potentials of DPN+ and flavoproteins. At least in the case of cysteine, it appears that the reduced form is taken more readily by the lymphocyte than its disulfide cystine, and that cysteine becomes readily incorporated into glutathione (y-L-glutamyl-L-cysteinylglycine):some 16% within 3 minutes of incubation and some 30% after 30 minutes (563). Therefore, a high cysteine content of the medium may result in a high glutathione content in the cells. When other reduced forms of thiols are used, they may interfere with the atmospheric oxidation of the cysteine that flows out of the cells in the reduced state, and this favors its reuptake by the cells. This simple example of a type of alteration that can be induced by thiols in the medium shows how complex will be the interpretation of their effects at the molecular or cellular level. To test for the involvement of SH groups in lymphocyte activation, a series of reagents have been used that react with SH groups either by alkylating them (N-ethylmaleimide and iodoacetamide), b y oxidizing them [diazinedicarboxylic acid bis(N,N’-dimethylamide) or diamide], or b y blocking them [ p-hydroxymercuriphenylsulfonic acid, an organic mercurial, and 5,5’-dithiobis(2-nitrobenzoic acid)]. In contrast to thiols, which had potentiating activities on the lectin-induced lymphocyte activation and were inhibitory only at very high doses (lo+ M ) , all the SH reagents provoke important degrees of inhibition of the PHA- or Con A-stimulated thymidine or uridine uptake by lymphocytes, even at very low doses (for some of them down to lop6M ) , when added together with the lectin or at any time thereafter (79). Diamide was particularly interesting inasmuch as it is much more inhibitory
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
61
when added either within 30-60 minutes of lectin or when added after 24 hours than between these two periods. The intermediate “refractory” period is all relative, however, and it should not be interpreted as showing a full resistance of the lectin-treated cells to diamide, but only as a period of lower sensitivity, as actually appears from the data; for example, Fig. 3 of Chaplin and Wedner (79) shows only partial resistance, since in the “refractory” period there is anyhow still +SO% of inhibition compared to the controls in absence of diamide. The short half-life of diamide in culture (328) would allow distinguishing the early and late events in the activation sequence, whereas the other SH reagents would persist long enough in culture to alter the cell S H group all along (79). When disulfide linkage is induced, as is the case with diamide, cellular reductases may correct the disulfide bridging and the cell may “ recover,” but what happens with other reagents, and what is the meaning of the inhibition of lymphocyte transformation that they cause? It is actually remarkable that low doses of them can block the latter process, since the medium for lymphocyte culture contains 15% of fetal serum (79), i.e., proteins that, as sulfhydryl compounds are potential targets for the SH reagents, during the preparation of the media, before any contact with the lymphocyte membrane proteins. All the SH reagents used so far were in fact selected by biochemists because of their high reactivity for SH radicals, not as site-specific reagents. To give just two examples-diamide had been synthesized as an oxidant specific for glutathione, but it is not and reacts with a number of other nonprotein SHs and proteins, including tubulin (404);and N-ethylmaleimide is a very active alkylating reagent, which must form quite stable serum protein derivatives (529). One may wonder how much an SH reagent actually acts on the cell, and if the inhibitory component of the medium is not an altered serum component. Among these are not only the proteins, but also the smaller molecules, such as cysteine, cystine, cystinylglycine, and glutathione, whose extracellular pool may be depleted to some extent by the SH reagent. The cystine content of serum from various species (e.g., human beings, M , and its cystinylglycine content is cattle, rats) is about 4 to 7 x about 1 x M (15). Thus, SH reagents as well as thiols may modify the total content of the cysteine that is available to be taken up by the cells, and this might alter the lectin-induced transformation. There are many other possible targets for the action of those simple molecules, which have been referred to as having their effects on a “cascade of molecular interactions” or as “short-circuiting the chain of transmission for activating
62
FRANCIS LOOR
stimuli perceived at the cell surface en route to the nucleus.” A less vague, but still quite general statement, could be that thiols may chelate heavy metal ions whose traces would inhibit SH-requiring enzymes, and that on the contrary SH reagents would block the functions of such enzymes by binding to their essential SH group. Thiols and SH reagents may interfere either directly with the activation of such SH enzymes-and they are a multitude-or indirectly with the activation of enzymes that require a specific SH cofactor by competing with it or by blocking it (see discussion in 366,372). Binding does not necessarily mean inactivation; for example, in the case of dynein, a myosinlike protein obtained from sea urchin sperm flagella, and having three different essential SHs, their progressive reaction with N-ethylmaleimide can lead to very different results: reaction of the first one (most reactive) leads to blockage of a capacity to associate with another protein; of the second, to an enhancement of ATPase activity; and of the third, to a loss of catalytic activity (597). Thus, not only are the target molecules in the cell for reducing agents and for SH binding reagents multiple, but, even at the level of a single molecule, their action may have very different consequences. To return to the surface (of the cell), it has been suggested that the early sensitivity of lectin-induced transformation to diamide shows that the affected step would occur at or near the PM, principally because the early biochemical alterations related to cell activation are largely confined to that cellular compartment (79). There are so far few, if any, published data regarding the effects of SH reagents, or of thiols, on such early parameters of lymphocyte activation. Among thiols, cysteine and glutathione would increase the magnitude of the cAMP rise and the initial rate of cAMP accumulation, whereas 2-mercaptoethanol and dithiothreitol would increase only the magnitude of the cAMP response (662). Among other possibilities, one can mention that SH reagents do affect ATPase activities and permeability to ions, in the case of erythrocyte membrane at least (517), and, as was discussed in a preceding section, membrane ATPase activities and ion fluxes may be important for activation. Another possible target is adenylate cyclase, which is also inactivated by SH reagents (397) and whose activation seems to be linked to lymphocyte activation (661). Thiols and SH reagents may also interfere with the extensive disulfide bonding that would exist at the level of the surface of mammalian cells (275). Such an effect has been shown to occur for cysteine, which could increase the number of titratable SH groups on the lymphocyte surface, a maximum being obtained within 10 minutes of cysteine addition, but, for unknown reasons, not obtained with
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
63
glutathione and dithiothreitol(552). This may result in an alteration of any membranous function that would depend on a supramolecular organization of membrane components mediated via SH bonding. If, as suggested (365, 367), the membrane components are segregated into domains having specialized functions of recognition and adhesion or of metabolism control, one can easily imagine that thiols and SH reagents will disorganize either the first domains or the second ones or the passage of information from the first to the second. What we know from experiments performed in our laboratory is that thiols and SH reagents affect gross membrane dynamic changes, such as capping, and, at least for thiols, this was rather unexpected. Indeed, simple thiol compounds, such as cysteine, thioglycolate, and glutathione (both reduced and oxidized), as well as sulfite, do inhibit the process of capping of lymphocyte membrane components such as B cell inimunoglobulin and Con A binding sites (366). This inhibition occurs in a range of doses around lop3M , which is similar to the doses required to obtain potentiation of mitogenic responses. For unclear reasons, we have not yet obtained equally marked capping inhibition with mercaptoethanol, a-thioglycerol, and dithiothreitol. The thiolmediated inhibition of capping is a delay of the capping process rather than its blockage: it is evident after 10 minutes at 3TC, becomes less and less evident with time, and is hardly significant after 60 minutes. At least in the case of cysteine, it appears that atmospheric oxidation of the cysteine-containing medium is a cause for loss of activity, since little inhibitory capacity is found when the medium is left for 2 hours in the open air, whereas most of it is kept if the medium is left for 2 hours under nitrogen. All the SH reagents tested so far, which include all those indicated above and a few others, can also inhibit capping, but this is a blockage of the process that is not easily reversible (as is, for instance, the inhibition caused by 10’ M NaN3, 70-80% of which is reversible by washing away the drug) (372). It had been suggested that “SH interactions may be an obligatory event in the activation sequence” and that “the maintenance of free SH groups is important during the early induction phase of lymphocyte proliferation” (79). Such an early step may be “capping”; however there is no simple correlation to be made: capping is only partly inhibited by doses of SH reagents that are quite sufficient to block either lymphocyte proliferation itself or only its triggering (79, 372). Although there are differences from one SH reagent to another, it appears that higher doses of SH reagents are required to obtain a full inhibition of capping, usually in the M range (372): thus, capping appears to be about 100 times more resistant to SH poisoning than lymphocyte trig-
64
FRANCIS LOOR
gering and proliferation, and capping is probably not the step that is affected by SH reagents in the proliferation assay. What can be the explanation for the effects of thiols on capping, and is it related to their enhancing effects on lymphocyte activation? The fact here is that there is a rough correlation of the doses that give both effects. The lack of detectable inhibition obtained with some thiols may be due to differences in the assay systems. Indeed, for capping studies, to reduce to a minimum the unknown variables of the medium and increase our’chances of understanding we use a simple medium made of sodium phosphate buffer supplemented with 0.5% bovine albumin; and experiments on lymphocyte activation are made in culture medium containing fetal calf serum, then, as indicated above, a number of sulfhydryl compounds, among them cystine and cystinylglycine, which may be readily reduced by stronger reducing agents, such as 2-mercaptoethanol, dithiothreitol and a-thioglycerol to give cysteine. Even when used in the M range and up to 5 x M , the thiols do not significantly interfere with the clustering of mIg by anti-Ig, although at the higher doses a few more cells are stained as rings rather than as spots (366).The dramatic effect of the thiols is thus not on the clustering of the PM components by the ligand, but on their removal by capping. We have suggested already that delayed capping may allow better stimulation of the lymphocyte because there is better transmission of the signal through the membrane (364). Indeed, if the triggering of a cell is a chain of reactions that requires the cooperation of several plasma membrane components, some of them rare, a rapid removal by capping, endocytosis, or shedding of clustered receptors will not favor such cooperation and will impair triggering. Usually capping of mIg (370) or of low, mitogenic doses of lectins (362) is fast. We suggest that it is too fast: the clustered PM components are removed from the cell surface before the clustering of receptors can be “felt” by the cells. Therefore, any physiological condition that will slow down the capping, without interfering with the clustering, should favor the early membranous sequencies of triggering and potentiate the signal. The evidence that clustering is required will be analyzed later. In the particular case, it remains to speculate how thiols can delay the capping process. Besides the general considerations on their possible effects at the level of SS bridging of membrane proteins and of SH-enzyme activities, thiols may interfere with activities of the cell cortex, both at the microfilament (MF) and at the microtubule (MT) levels. As far as the cytoskeleton is concerned, we have seen above that the presence of thiols in the medium may eventually lead to an increase in
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
65
cellular glutathione in the reduced state GSH. This will alter the tubulin-MT equilibrium: SH groups of tubulin are involved in the in vitro polymerization process (404), and, according to a recent view, M T would be unstable in viuo if too little glutathione is in the reduced form (69).Thus, at high levels of GSH, most tubulin molecules may be expected to be in the polymerized form and there would be many MT in the cell; since MT appear to interfere with M F activities leading to capping, this is a first way by which thiols may inhibit capping. A second, nonexclusive, way, is an action at the level of the cytoinusculature. Thiols may alter the normal function of the myosin ATPases needed for the MF-mediated dragging of patches of membranous material to the cap-either the ATPase required for association of myosin with actin or the ATPase whose function depends on previous actin binding; thus, for instance, either myosin and actin do not interact or they do so, but there is no M F sliding. It is not difficult to link, at least theoretically, the effects of thiols on capping and on activation with cell cortex activities; it seems clear, however, that the solution of the problem will not come from studies with such compounds with undefined sites of action. Still related to thiols and SH reagents are the effects on lymphocyte stimulation of larger SH compounds and of inorganic and organic mercurials. It has been shown that protease (usually trypsin) inhibitors can enhance the activation of small lymphocytes b y niitogenic lectins (e.g., 240, 24 1, 554). Various trypsin inhibitors show important homologies of large sections of amino acid sequences, particularly an unusually high content of cysteine-in one such case, of up to 19 mol% (239). It was shown that such cysteine-rich peptides can counteract the autoxidation of thiols of the medium by molecular oxygen, a reaction that is normally catalyzed by metal ions, such as Cu2+,present as trace contaminants in the media: 1 mM cysteine is autoxidized within 15 minutes in the presence of 1 p.R.1 Cu" (241). Better potentiation of lyniphocyte activation was obtained if the cysteine-rich peptides were in the reduced form, probably because they can thus protect essential SH groups of membrane from oxidation by chelating the catalytic metal ions. When such cysteine-rich peptides are in the oxidized form, they may still protect the membranes against proteolytic enzyme digestion (241). Among the thiol compounds, particular attention was given to the antihelminthic drug levamisole (~-2,3,5,6-tetrahydro-6-phenyliniidazo[2,1-h Ithiazole hydrochloride). Levamisole has been reported to stimulate the in vitro proliferation of mouse lymphocytes exposed to T cell mitogens or allogeneic cells in both the mouse (231, 406, 686, 687) and human beings (231, 345, 655). In a recent study (469),
66
FRANCIS LOOR
the stimulatory effects of levamisole on Con A-induced proliferation of mouse cells were confirmed, but they were also shown to be very dependent on the presence or the absence of other thiols in the medium; for instance, mercaptoethanol or cysteamine can, alone, potentiate the response, as does also levamisole, but when mixed together there is a complete suppression of cell proliferation. A few other observations of that kind have been made, but they d o not help to clarify the mechanism of action of thiols. As analyzed in Section VI,C,2, divalent mercury ions can stimulate lymphocytes to proliferate. A possible mechanism of such triggering is through interaction of such Hg2+ ions with free sulfhydryl groups, since such groups are present on the cell surface (2). A divalent organomercurial (1,4-bismercuri-3,4-dihydroxybutane) can, as does divalent mercury, stimulate DNA synthesis in lymphocytes, whereas monovalent methylmercury does not (41),and it has been suggested that organomercurials can stimulate lymphocytes only if they are able to cross-link protein sulfhydryl groups and to bring them into close proximity on the cell surface (41).A polyvalent reagent made of mercury-substituted dextran lacks stimulatory activity, perhaps because of too important cross-linking leading to cell surface freezing, or perhaps because of a too fast capping removal of clustered components. The point should be analyzed. Finally, another combined effect of bivalent cations and thiols is that when they are present together there may be generation of peroxide, which has definite membrane effects. It seems difficult to obtain media that do not contain traces of heavy cations, principally copper. It has been shown that micromolar quantities of Cu2+can catalyze within minutes the atmospheric autoxidation of cysteine leading not only to cysteine depletion in the medium, but also to generation of H202(161). Peroxide production is a mechanism used by phagocytic cells to kill bacteria, and there is also release of measurable amounts of H20zin the outside medium. One can evaluate that in extreme conditions in uitro, the amount of HzOz produced is measurable in terms of nanomoles per minute per lo6 cells (e.g., 541).Chick embryo fibroM H 2 0 2(482). At a conblasts are killed when exposed to -5 x centration of -low5 M , H 2 0 2blocks capping of membrane immunoglobulins (F. Loor, unpublished). It can be expected that lower concentrations of locally produced H20z,i.e., in the vicinity of Cu2+ions bound to the cell surface, will affect neighboring membrane components, in particular there will be oxidation of SH groups of membrane proteins, with consequent alterations offreedom of component distribution. HzOz has also been implied, like oxidized glutathione and other
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
67
endogenous oxidants, as possibly playing a role in the control of microtubule polymerization in normal cells, as it would oxidize tubulin SH and inhibit their polymerization, thus, a role opposite to the one played by reduced glutathione (69, 458). It is presently not possible, however, to know at which level(s) a role is played by the reduction of S-S b y thiols and by the oxidation of other SH groups by the H 2 0 2 generated by those thiols and Cu'+: it can be on the surface membrane proteins, on the cell cortex elements, or anywhere else in the cell. HzOzrelease was recently identified (when monitored by generation of chemiluminescence by the cells in the presence of luminol) as a very early event following lymphocyte activation by mitogen or A23187 (689). The function of such H z 0 2 production is not known; perhaps it could be related, in the case of killer cells, to generation of cytotoxic lipids by formation of aldehydes by peroxidation of polyunsaturated fatty acids as the possible mechanism of killing (see Section V,E). Such peroxidation of the unsaturated fatty acids of the phospholipid moiety of the membrane leads to membrane damage. H,O, treatment of lipids in natural and model membranes results in loss of polyunsaturated lipids, such as phosphatidylethanolaniine, by peroxidation, concomitantly with important alterations of membrane permeability and viscosity (see, e.g., 135, 616). There would be correlations between membrane viscosity and levels of tissue antioxidants in uiuo (207, quoted in 135). I n vitro HzO, production may be a regular event of cell culture, and it may become toxic if the proper balance of prooxidant and antioxidant is not appropriate. In the course of studies on the cytotoxicity to various cultured fibroblasts of ascorbate and other reducing, antioxidant agents (cysteine, glutathione, and others), the toxic compound was identified as being HzOz(482).H 2 0 2was not fomied by oxidation of ascorbate or thiols in the medium but probably arose from oxidative reactions taking place either intracellularly or at the level of the cell surface; addition of catalase could prevent the toxic effects of glutathione or ascorbate. Thus, although antioxidants are required, they may become toxic; principally it would appear that many types of cultured cells have too low intracellular levels of catalase to degrade completely the amount of H z 0 2formed when the concentration of thiols is higher than the physiological level. It is interesting to recall here that ( a ) Zn'+ ions, with which we dealt in a previous section, inhibit lipid peroxidation both in uivo and in vitro (88-90); ( b ) inhibition is also achieved by some other antioxidants, among which are selenium (132,628), i.e., another compound found to be essential for the culture of lymphocytes (226); and ( c ) transferrin, which binds Fe3+but also complexes ZnZ+and binds to lymphocyte
68
FRANCIS LOOR
membranes (see preceding section), is another essential medium compound for lymphocyte culture (278). Besides catalase, the other mammalian enzyme capable of degrading H 2 0 2 is the glutathione peroxidase, whose function requires NADPH, which is produced by oxidation of glucose via the hexose monophosphate shunt. It is thus remarkable that the best means to obtain a marked inhibition of the capping process, i.e., preincubation of the cells in PBS with 10 mM NaN3, will impair the ability of the cells to destroy H,O,; indeed, the lack of glucose should impair glutathione peroxidase function and the azide should inhibit the catalase. One may therefore wonder if at least some of the effects of NaN3 on cell surface morphology (for example, decreased capacity to cap, increased expression of microvilli) were not due to H202-mediated effects rather than to a direct inhibition of metabolism. Finally, a major role for H 2 0 2in the mitogenic activation of lymphocytes might possibly be found at the level of their interactions with macrophages or other phagocytes, adherent or accessory cells of the immune system. Activation of macrophages, for instance, leads to HzOzproduction and release in the external medium (see Section IX,B). Still, in cell cultures, dividing lymphocytes are frequently seen to be in close contact with the macrophages (clusters, see preceding section). One can expect that part of their membrane at the contact level must be altered in some way by the HzOzresulting from surface activities of the macrophage. A similar situation may exist in vivo, where besides the antigen contact and antigen priming of the lymphocytes, the phagocytes are also present and release all kinds of mediators, probably including H,O,. Thus, in summary, it seems possible to speculate that if general exposure of the whole cell membrane to HsOz is probably toxic, local exposure of limited areas of the cell membrane to locally produced H z 0 2may induce surface alterations (of fluidity, of permeability) that, if controlled, may be an essential function in activation. The existence of those “trace impurities” of the medium and similar other “complications” of the cell culture system should not be disregarded or overlooked, however, when trying to understand mechanisms of lymphocyte triggering i n uitro. 5 . Miscellaneous Factors Not only can some lymphocytes be triggered to mitogenesis by lectins, antibodies, and small molecules such as ions and thiols, but also they can be triggered by a chemical reaction-that is, a mild oxidation of membrane sialyl residues by periodate (448).The aldehyde moiety that results from the oxidation is presumed to be essential for the
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
69
transformation to occur. Such a periodate-induced transformation of lymphocytes can indeed be blocked with aldehyde blocking agents (134). In such a system the concentration of cysteine and other thiols available to the cells will modulate the response, depending on the proportion of aldehyde groups the thiols will block. In relation to this so called “oxidative stimulation” of lymphocytes, some treatments or substances were found to be mitogenic only after modification of the lymphocyte surface with iieuraminidase (which removes the terniinal sialyl residues of the complex carbohydrates of membrane glycoproteins and glycolipids). Cells with such altered surface can then be triggered to transform by galactose oxidase (449);by soybean lectin, which binds to galactosyl and N-acetylgalactosamine residues (450); and by peanut lectin, which binds to galactosyl residues (451). Although there are species differences, these studies show that galactosyl residues are probably involved in this type of stimulation. This is further substantiated by the marked decrease of mitogenic response obtained with these stimulants when the cells were pretreated with Pgalactosidase (451). P-Linked galactosyl residues are thus involved in lymphocyte triggering by all these mitogens (447). Since these enzymic modifications of the lymphocyte surface do not affect the mitogenic response to other lectins, such as Con A, the membrane site of action of the latter mitogen is different, and the enzymic treatment does not affect a late phase of cell triggering but rather an early one, specific for galactose oxidase, soybean, and peanut lectin (presumably, for the latter, binding to or redistribution of membrane components). Proteolytic activities and proteolytic inhibitors can also modulate the state of responsiveness of the lymphocytes. Thus, trypsin can stimulate lymphocytes to incorporate thymidine, and this is blocked by simultaneous addition of soybean trypsin inhibitor to the cells (248, 651). Chymotrypsin could also stimulate (248). A series of protease inhibitors were shown to inhibit lymphocyte proliferation induced by PHA and PWM (113) and b y Con A and LPS (249), thus c o n h i n g earlier studies of depression of PHA stimulation of lymphocyte by inhibitors of proteolysis (261). The maximal suppressive effect was found when the protease inhibitors were added shortly after the mitogenic lectin ( 1 13, 248, 249, 261). Although the cells are still viable, it is remarkable that they become “spherical and motionless,” that is, they do not show the typical uropode of cells in culture (113). Among various causes, one may suggest that the cell cortex structures required for motility have been influenced by the protease inhibitors, and it would be interesting to look at the structural organization of MT
70
FRANCIS LOOR
and MF in protease inhibitor-treated cells. If a proteolytic event may thus be involved in an early step of the activation of lymphocyte, still the different cell subsets, the different mitogens, show various sensitivities to the protease inhibitors, and it is difficult to find out their mechanisms of action; for example, some protease inhibitor blockage can be reversed by glutathione or cysteine in the case of one mitogenic activation system, but cannot in another (249). Last, but not least, it seems worth mentioning here that the mitogenic stimulation of lymphocytes by C3b is presumably due to activation of the alternative complement pathway (a sequence of proteolytic enzyme actions) by lymphocyte membrane-bound C3b (250),thus giving some support to a hypothesis formulated earlier (136).The complement sequence with its various proteolytic steps had also been suggested earlier as a source of many possible different transmembrane signals, leading to activation or suppression of lymphocyte activities (363, 364, 370), but the idea was that such a sequence was operated by membrane proteins with complement component-like activities rather than by membrane-adsorbed serum complement components. It is known that various proteolytic treatments of cell surface induce redistribution of membrane components, but whether such rearrangement can specifically lead to triggering, and is followed by rearrangement of cell cortex elements, is simply not known at all. It remains that proteases and protease inhibitors from the serum and proteases present on the membrane of cells may play a role in the activation of lymphocytes or modulate it, and therefore it is a potential source of variability in the results on the mitogenic activation of lymphocytes as a function of the presence of serum, of accessory cells, of thiols, of cations, and so on. Finally, possibly in correlation with the protease inhibitors, one should also mention that inhibition of lectin-induced mitogenesis in lymphocytes can be obtained by treating them with chalones (aqueous extracts of lymphoid tissues) (20, 527). For instance, a chalone extracted from bovine thymus strongly inhibited DNA synthesis in activated human peripheral lymphocytes, bone marrow cells, thymocytes, and lyrnphoblastoid cells, in a dose-dependent fashion. It had no effect on the proliferation of nonlymphoid cell lines, and it was considered to be “lymphocyte specific” (527). Some chalones would be T or B specific but work across species (20). They would be proteins contained in the crude ultrafiltrate of the lymphoid cells themselves, not made by other cells of the lymphoid organs. Chalone-mediated inhibition of lymphocyte activation is not due to trivial causes such as competition for lectin binding, but their mechanism of action is not known
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
71
(20). Attempts to characterize the precise biochemical nature of chalones are in progress (for recent references, see 395). The lectin-induced activation of lymphocytes also depends on a number of other small molecules, some of which are endogenously produced and can indeed be physiological regulators of the lymphocyte responsiveness. Among these are some hormones, histamine, interferon, and others, all of which can modulate lymphocyte mitogenesis, but the mechanisms of their reaction have not yet been much explored (214). Another factor that may play a role in lymphocyte activation is the possibility for intercellular interaction, namely, the requirement for the presence of plastic adherent cells. But this requirement does not show constant features: although activation of lymphocytes by oxidation with NaIO, and by Con A and PHA does not require physical contact between lymphocytes and adherent cells, such contact is required for their activation by PWM or by ZnZ+ions (183).There could be a number of speculations on the mechanisms involved; let us consider but a few. For instance, for Con A stimulation, one now thinks in terms of a dual effect: the Con A should on the one hand induce the in .situ production of T cell growth factors and, on the other hand, render resting T cells sensitive to the mitogenic activity of these growth factors, but the mere binding of Con A to the surface membrane of T cells would not at all trigger the cells to go through the mitotic cycle (334).Do the membrane effects of Con A disclose growth factor receptors on the T cells? Do MT-directed and MF-directed drugs then interfere with stimulation by modulating the expression of such growth factor receptors? Growth factors are usually found in sufficient concentration in the “conditioning medium” after a day or two of culture. But a number of other “factors” appear in the medium as well, i.e., the lymphokines whose extracellular appearance is stimulated by the lectin. Such lymphokines have been shown to be capable of enhancing synthesis and secretion of Ig by B cells (273,274) or to inhibit it (524). Some of the lymphokines are factors regulating mitogenesis as well, but their mechanisms of action are obscure. Some factors may also be released b y some lymphocytes within seconds of the binding of the ligands to their cell surface, as was recently observed in the case of rabbit thymocytes treated by Con A: proteins of nonimmunoglobulin nature appear in increased amounts in the extracellular medium virtually without delay following Con A binding (206). What are these proteins? Are they important? Would the stimulation be the same if they were washed away before culturing the cells? There is still much to do in the field of the factors.
72
FRANCIS LOOR
When cell aggregation seems to be required for efficient lectininduced mitogenesis in the lymphocytes, various speculations can be put forward-from the simple cooperation models, such as an appropriate “presentation” of clustered lectins bound to the PM, to the more elaborated “cell communication” models, such as the constitution of gap junctions between the cells, possibly to allow flows of ions or other substances. There is some evidence for the existence of such interactions, in at least some cases (e.g., 299). VII. Activation of Lymphocytes by Anti-Ig Antibodies
The effects of factors that influence the mitogenic activation of lymphocytes have been analyzed so far only in the particular case of lectins and other polyclonal activators as mitogens. One does not know whether activation of lymphocytes via a “mitogen” receptor is a rule or an exception, the rule being activation through the antigen receptor. It seems therefore mandatory to analyze as well the parameters that influence the response of the lymphocytes to the ligands they should normally meet, i.e., at least the antigens, if not the anti-idiotypes born out of network theories. At the time being, the persisting uncertainty about the molecular species of the T cell receptors makes our conclusions limited to activation of B cells only. Specific antigens, or anti-idiotypes, cannot be used, as they would be only oligoclonal activators; the activation of a few clones per millions of lymphocytes would not emerge from the background noise. Thus, the most logical B cell polyclonal activator to be used is the anti-immunoglobulin antibody, and since the B cells generally show various restrictions of expression of their batch of Ig genes, various anti-isotype and anti-allotype antibody preparations should be more restricted polyclonal activators. Much work has been done in this field, but the effects of anti-Ig sera on B cell function are nevertheless not solved. A concluding remark of Moller, in his masterly review of the topic, summarizes well what we know (413): “Presently the only safe conclusion to be drawn from the totality of the experiments dealing with the effect of anti-Ig sera on B cells is that no safe conclusion can be drawn.” And it would be very tempting to plagiarize both content and conclusions of Moller, as I could not better summarize the data and I share his conclusions. There have been a number of studies performed on the effects of anti-Ig antibodies on specific immune responses in vitro, and they have given all possible results-from enhancement to suppression, in-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
73
cluding induction (see 413, for review). Those systems are too complex to be suitable attempts to correlate, in molecular terms, the effects of the binding of the ligands to the lymphocyte membrane, to the consequent triggering, potentiation, or depression of DNA synthesis. I will briefly review only the effects of anti-Ig either alone as polyclonal activators or in conjunction with other polyclonal B cell activators. In a number of systems, anti-Ig can lead to a polyclonal activation of B cells to make DNA, although they do not activate Ig synthesis or plasma cell differentiation. This has been obtained with lymphocytes of rabbits, human beings, chickens, rats, pigs, and guinea pigs, but not with mouse lymphocytes. The latter case will be analyzed below, but one should not oversimplify the conclusion about the other species: even though stimulation was obtained, there is no absolute consensus as to which specific anti-Ig antibody works and which does not. Thus, in guinea pig, a slight stimulation was obtained with an anti-y2globulin, but a much weaker one than with ALS, and no stimulation was obtained with anti-F(ab’)2(169);in the rat, a rabbit anti-rat L chain stimulates (324), as does a rabbit anti-pig Ig in the pig (387). With chicken lymphocytes, stimulation was obtained by an anti-IgG (not y specific) (10, 316), thus equivalent to the “multispecific” anti-Ig used by others (613), but class-specific anti-p and anti-y also stimulate, and even better when used together or in sequence (613). With human lymphocytes a weak stimulation only was first obtained by a polyspecific anti-Ig and also with specific antibody-anti-IgG and anti-IgM being better than anti-IgA, anti-rc, and anti-A (462). Good stimulation indexes were obtained with an anti-F(ab’), (182) and with almost all specific antisera tested (217): anti-IgG, anti-IgM, anti-IgA, anti-rc, anti-A; but weak stimulation, if any, was obtained with antiIgD. In more recent studies, good stimulation was reported for classspecific anti-y, anti-a, and anti-p (190) and with purified anti-rc, but not anti-A (528). In all these cases, the percentages of cells that are transfornied usually are in the range of 5-10%, while PHA transforms up to 80% of the cells. There seems to be an age dependency of the results, which has been discovered only very recently (668): indeed peripheral blood lymphocytes from elderly human beings (average age of 78 years) respond better to anti-Ig stimulation than lymphocytes from young adults (average age of 29 years). As lymphocytes from the two groups show an inverse sensitivity to activation by PHA, and because of some other data not discussed here, it is suggested that the greater stimulation of lymphocytes from old individuals by anti-Ig were due to the loss of thymic influence that occurs with age (668).A difference
74
FRANCIS LOOR
between mature and immature B cells in humans may be that the mature cells can regenerate mIg after capping, whereas the immature ones cannot (21). With rabbit lymphocytes, the most remarkable fact is the ease with which anti-Ig can stimulate lymphocyte transformation (see later). It is also remarkable how high is the proportion of lymphocytes that are transformed-of the order of 80% (see, e.g., 582,583);however, as can be expected, there are differences from one organ to the other (108).In the rabbit, practically every anti-Ig antibody that was used worked: anti-total Ig, anti-Fab, anti-Fcy, various specific anti-H and anti-L chains, and almost the entire series of the various anti-allotypes that are available and are specific ligands for the different parts of the different chains of Ig(33,156,158,193,319,583,584,585,and others). Particularly, thanks to the use of such specific anti-allotypes, it was possible to show that the stimulation depends on the binding of antibody to membrane immunoglobulin of endogenous origin (193), and interesting “summation” effects were obtained by use of antibodies to the different Tg allotypes expressed in the animal (193). I pointed out in an earlier review (364) the interesting observations that the complexes of anti-Ig with mIg persist for several hours on the lymphocytes from this species (427); indeed, a lack of total and rapid removal by capping of the clusters might be the reason for the high mitogenicity of anti-rabbit Ig sera (364). Further, since anti-Ig of practically any specificity works, whether it is anti-L chain or anti-Fc, it would seem unlikely that triggering was due to very subtle conformational changes in the mIg receptor molecules, and it would appear that on the contrary the only important step for triggering is the formation of clusters, and their persistence for some time on the membrane; e.g., to constitute calcium gates and enough Ca2+entry. That clustering is required is demonstrated by the lack of stimulation by monovalent Fab, while F(ab‘), are active (l58), as will be analyzed further in Section VIII. In the case of mouse lymphocytes, earlier studies indicated that little or no stimulation could be obtained by various anti-Ig antibody preparations (13,146,217,566,652).What was most usually found was that the anti-Ig antibody treatment of a suspension of mouse lymphocytes was depressing or suppressing their response to other polyclonal activators (13,566).A case of direct stimulation of B cell to proliferate was reported (474) using an anti-Ig antibody (predominantly of anti+ specificity) covalently coupled to the surface of polyacrylamide beads, but in the soluble form the same antibody did not stimulate. Activation to DNA synthesis was obtained, however, within only a narrow range of density of the anti-Ig on the polyacrylamide beads, and it could not
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
75
be obtained with anti-Ig-coated Sephadex beads (474) or with antiIg-coated Sepharose beads (216). Still, in a more recent paper (475) it appears that sustained cell division (and differentiation to Ig secretion of these insoluble anti-Ig-activated B lymphocytes) could be obtained only in the presence of T cell factors, whereas in their absence only limited proliferation was obtained. Only recently has the capacity of anti-Ig to induce mitosis in mouse lymphocyte been a regular finding. This major breakthrough was obtained by three different groups; it depends on a critical condition for the assay to be appropriate for anti-Ig triggering (603, 605, 665), but unfortunately there is no agreement as to what that crucial condition is. For one group (665) the key factor seems to be the age of the mouse: anti-Ig, both a specific anti-IgM and unrestricted anti-Ig, regularly stimulates DNA synthesis in lymphocytes from mice that are 5-7 months old at least, but not, or only rarely, in younger animals. In a subsequent paper, however (664), the requirement for anti-p was indicated, since only anti-p works, not anti-a or anti-y. For the second group (605),even young adult mice (-8 weeks old) have lymphocytes that can be activated by anti-Ig, but the key trick is to use purified antibody rather than the whole antiserum; anti-p works best, but anti-rc and anti-y are not too good. Finally, for the third group (603), lymphocytes from mice of all ages can proliferate in response to anti-IgM, but this requires the presence of a 65,000 MW cofactor that is generated from the serum by 2-mercaptoethanol. This does not make it easy to present an integrated view. There is some agreement, however, that lymphocytes from old mice really respond better to anti-Ig reagents (though their proliferative response to LPS does not change with aging, e.g., 426): thus, neonataI mouse Iymphocytes do not show any response at all to anti-IgM (604, 606), and the anti-p responsiveness does not appear until 4 weeks of age (606) and does not reach its maximal levels until 8 weeks of age (606) or even later (604). This is still very different from the 7 months required in the first group’s system (665).Similar are the observations that anti-p are the best stimulators, but anti-a, anti-y, anti-rc, and anti-Fab are not so good (604,605, 664),although the responsiveness to anti-y, anti+ may be attributable to a subset of B cells that appear much later and whose appearance depends on thymus control (606). Cross-linkage is needed since F(ab’), works but Fab does not (604, 605, 664). The Fc piece may even be responsible for an inhibitory effect since F(ab’)* anti-IgM were reported to be better stimulators than intact antibody (604). But definitely, since F(ab’)zanti-Ig reagents can stimulate, it shows that the Fc is not involved in delivering the activation signal either directly on the
76
FRANCIS LOOR
B cell or via other cells with Fc receptors. Thus, the antibody-binding capacity of the reagent is needed, as well as cross-linking of receptors, but it is not known for what exactly. In a search to determine why lymphocytes from old mice respond better to anti-Ig, it was found that F(ab')2anti-Ig does actually stimulate lymphocytes from young mice quite well, but that such stimulation was severely depressed by intact anti-Ig, while stimulation of old mouse lymphocytes by F(ab'), anti-Ig was only marginally sensitive to the presence of intact anti-Ig (574). Anti-Ig stimulation of B cells could also be enhanced in young mice, provided they had been deprived of thymic influence b y neonatal thymectomy (666), but further experiments (574) exclude that the normally low responsiveness of young mouse B cells was due to suppressor T cells. Thus, the better stimulatory capacity of F(ab'), anti-Ig reagents (574, 604) seems to be not a trivial fact; nevertheless, the nature of the age-related change in the sensitivity of B cells to the inhibitory effect of the Fc of anti-Ig antibody is unknown. A possible function of macrophages cannot be excluded, since B cells would require macrophages to respond to intact anti-IgM, but not to F(ab')z anti-IgM (252). The two other groups, however, reported that the stimulation by intact anti-Ig does not need accessory cells (604, 606). Another alternative is that with aging, possibly as a result of thymus influence, different B cell subsets appear that show different sensitivity to the Fc of anti-Ig, the reason for which could be found at the level of the B cells themselves, particularly of the molecular relationships that exist in their membrane between mIg and FcR (Fc receptors), which may differ in different B cell subsets. The association mIg-FcR found in some mouse B cells (1, 172) may not be extrapolable to every B cell subset; for instance, a recent study indicates a loss of such association after B cell activation (573). Thus, generalizations are not permitted: membrane components that are in association in one cell type may not be so in another cell type-that is, control of molecular interactions between membrane components may indeed regulate the transmission of signals through the membrane. Such controls may find their origin in membrane composition-many examples have been seen already, such as different lipids or different proteins, but the controls of molecular interactions of membrane components may also be of cell cortex origin. This brings us back to the fate of the anti-Ig bound to the B cells. It is known that after capping the mouse B cell can resynthesize and reexpress new mIg (146, 363, 364, 370, 570), but it is also known that only mature B cells can do that, whereas immature B cells do not resynthesize mIg after capping (513,601). An additional requirement
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
77
to stimulate the proliferation of mouse B cells by anti-Ig is continuous exposure of the cells to the antibody for at least 48 hours (604,667); its removal after an earlier time is without effect, thus confirming that just a cycle of binding, clustering, capping is not sufficient to activate the cells to DNA synthesis (146).When mouse lymphocytes were cultured in excess of anti-Ig reagent, microprecipitates of mIg-anti-Ig were found to persist on the surface of the cells (370), a finding also reported for rabbit cells (427) and confirmed in the mouse, but only for cells derived from young animals (667). When cells from old animals are used, there is a much stronger persistence of the rabbit antibody on the surface of the cells, even after 2 or 3 days of culture, when intact antibody is used, but not when F(ab‘), fragments are used (667).Thus, there is no simple correlation to be made between extent of anti-Ig persistence on the cell surface and triggering, as we had temptingly suggested (363,364),or other factors play also a role that have not yet been identified. The conditioning of the medium, the presence of niacrophages, accessory cells, the presence of cofactor generated by thiol treatment of the serum, the role of the Fc portion of the antibody, the function of T cells-all are variables that have been analyzed already, but perhaps not yet systematically enough, for they niay have no role to play for the activation of some B cell subsets and still be important for others. In this whole series of experiments, it is difficult to relate the events that happen at the cell surface with activities of the cell cortex and with aotivation, simply because the proper experiments have not yet been done. Definitely the stiinulatory ligands are bivalent and thus induce clustering and capping, this being followed by endocytosis and/or shedding. But this is not sufficient, for one cycle of capping of Con A or PHA could not stimulate the cells either. To the best of my knowledge, there is only one series of experiments that can be related to the cell cortex activity, and that is with rabbit lymphocytes (319). Rabbit lymphocytes can be stimulated to proliferate when exposed to anti-Ig antibody for 24 hours, and if one then supplements the culture medium with an “enhancing soluble factor” (a T cell factor), there is also IgG production. Mitotic inhibitors do not affect such IgG production when present in the first stage of the culture, and cytochalasin B even enhances it; but if added at the second stage of the culture, they all block IgG production (319).Thus, the possible involvement of the cell cortex in lymphocyte activation by anti-Ig reagents is as yet unclear. Ifanything, one could say that inactivation ofthe M F at the early stages actually favors the triggering, perhaps as a result of the delayed capping, but that is all.
78
FRANCIS LOOR
No more information comes from their study as to the molecular mechanisms of action of anti-Ig as suppressor of effects of polyclonal B cell activation, such as LPS. The suppression of LPS-induced Ig synthesis by B cells (13), and of LPS-induced proliferation of some unidentified B cell subsets (403), has not been studied in relation to possible mechanisms in molecular terms. This is mainly because the success of suppression by anti-Ig reagents of LPS-induced proliferation of B cells has been variably reported in the past, even by the same authors (601-604). Even if it can be obtained not only in neonatal mice, but also in young adult ones, it remains that it is easiest to induce in the youngest animals. All anti-Ig antibodies (unrestricted anti-Ig, specific anti-IgM, specific anti-IgD) can block proliferation of lymphocytes from mice of all ages, and only anti-IgM can stimulate, but this requires continuous exposure of the cells to the antibody in presence of serum cofactors (generated by serum incubation with 2-ME followed by dialysis; stable, alkylation-insensitive factors) (604). If this is of general value, then we may have something to learn from the analysis of such a serum cofactor, in particular of its effect, if any, on the cell surface dynamics, VIII. To Cluster but Not to Cap-Is
That What Triggers?
When one considers the entire series of experiments performed on the stimulation of lymphocytes, it appears that for all the ligands that have been assayed for mitogenicity (antibodies, lectins, and others), the capacity of a ligand to trigger mitogenesis invariably depends on its capacity to cross-link membrane components: it has to be at least bivalent, and monovalent ligands do not stimulate. There are a few exceptions to this rule, which in my opinion may not be real (see later). Thus, there is tremendous support for the generalization that clustering of PM components is a prerequisite for triggering. That is, in all cases where a ligand is found to be mitogenic, it is a hi- or polyvalent ligand, and its monovalent derivatives are no longer mitogenic. On the contrary, there is not one case where a ligand was mitogenic when monovalent but not when bi- or polyvalent. Even more conclusive is that monovalent, nonmitogenic ligands can be made mitogenic by polymerization. Thus, cross-linking of membrane sites not only does not impair triggering, but is even required for it. There now follow examples; the exceptions come later. To my knowledge, the first example was obtained with an antihuman lymphocyte serum, whose IgG were mitogenic for human lymphocytes, as well as their F(ab'),, but not Fab' (685). This was con-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
79
finned shortly afterward for an anti-mouse thymocyte antibody (525) and later for a burro anti-rabbit lymphocyte serum (586).This was different, however, from results based on transformation of rabbit 1ymphocyte b y anti-Ig antisera, where transformation was obtained both with intact IgG, F(ab’), (pepsin digested), and also Fab (papain digested) either from a polyspecific sheep anti-rabbit Ig (582,586) or from a specific rabbit anti-rc chain allotype (586).This was the first exception, to be discussed later, for the rule of a lack of activation by Fab holds true also for anti-Ig antibodies. Unlike the intact IgG, the Fab anti-L chain is unable to stimulate human lymphocyte to transform (218). Similar data were reported then in various species. Thus, rabbit lymphocytes are stimulated to transform by intact anti-Ig IgG antibodies and b y their F(ab’),, but not by their Fab’ or Fab; however, if the monovalent fragments are further cross-linked by an anti-rabbit Ig, blast transformation is obtained (158).Similarly, though an intact anti-L chain is sthulatory for rat lymphocytes, its Fab fragments are not, unless they are further cross-linked by an anti-Fab (324);a similar result was obtained with pig lymphocytes (387).Where anti-Ig stimulation of mouse lymphocyte to proliferate works, it is only with bivalent antibody or F(ab’)2,but not at all with Fab or Fab’ (604,664). Not only proliferation of cells can be taken as a criterion of activation: with most cells there are other parameters (to be seen later), but also with lymphocytes there are other signs of cell activation. Thus, a secondary in vitro anti-hapten IgG response can be induced in rabbit B cells by anti-Ig plus an enhancing soluble factor: this is obtained with intact IgG and with F(ab’),, but not with Fab’ (318). Similarly, not only sheep or horse anti-human lymphocytes can stimulate uridine uptake in the treated lymphocytes, but they also induce interferon synthesis: this property is kept by the F(ab’), but lost by Fab’ and Fab (155). The importance of the valency of the antibody reagent is best demonstrated by Fanger’s experiments (156):a univalent Fab’ fragment of goat antibody directed to rabbit Fab is nonstiniulatory, but it becomes a good stimulator of rabbit lymphocytes after it is polymerized by glutaraldehyde; furthermore, when the goat IgG with anti-rabbit Fab antibody activity is itself polymerized, it better stimulates DNA synthesis at lower concentrations than does the untreated antibody. It is thus tempting to suggest that the results found by one group (582,586) that univalent Fab stimulates as well as F(ab’)* might be due to the presence of some undigested IgG or bivalent F(ab’), or even of sinall amounts of aggregated antibody or antibody fragments. The method used for the separation of such Fab (molecular sieving on
80
FRANCIS LOOR
G-200) does not allow 100% confidence in the monovalency of the isolated material, nor the total lack of some complement components. The methods used to show that their material was univalent could only show absence of important contamination b y bivalent material: ( a ) the lack of microprecipitation in gel is not sensitive enough; ( b )the lack of agglutination of the cells is well known to depend on many variables, namely on the class of antibody, some of them being nonagglutinating though being bivalent. Finally, although their Fab could inhibit agglutination induced by intact antibody, and was anyhow still mitogenic, it only shows that strong agglutination of the lymphocytes was not needed for their activation. But this does not exclude that clustering is not needed. I have never been able to inhibit clustering induced b y F(ab’)zby excess Fab (as can be followed easily when they are differently labeled with fluorochromes): even when going down to the limit of detection of the rhodamine F(ab’), anti-mouse Ig bound to the cells in an excess of fluorescein Fab antibody of the same origin, one can observe that the F(ab’), will form clusters and cap (F. Loor, unpublished). Carbohydrate binding reagents, principally lectins, generally show similar requirements for achieving lymphocyte stimulation; that is, they need to be polyvalent, although it is not fully established that cross-linking is the key factor. Thus, Con A, a strongly niitogenic lectin, is a tetramer made up of identical subunits of molecular weight 26,000, each subunit having a saccharide binding site, with specificity for mannoside and glucoside residues (104). Under physiological conditions for mammalian cells (around 37°C and pH 7.0), Con A essentially consists of tetravalent tetramers, whereas below p H 6.0 it forms divaleiit dimers (140),and it is almost completely in a dimeric form at or near 0°C (215, 271). By alkylation it is possible to stabilize Con A in the dimeric form, and both succinyl-Con A and acetyl-Con A derivatives were shown to be stable bivalent dimers (227).The native form of the monomer is unstable under physiological conditions (104), but two ways have been used to obtain monovalent Con A: the one is use of a succinylated Con A dimer, one of the two active sites of which is blocked by photoaffinity labeling with p-azidophenyl-a-D-niannopyranoside (175), and the other is by a monomer obtained by limited proteolysis (658), the monovalent molecules being in both cases selected b y affinity chromatography. When compared with tetravalent tetramers for mitogenicity on mouse lymphocytes, both bivalent dimers are found to b e niitogenic, at low concentration up to 5 pg/ml, in terms of the total amount of their thymidine uptake. While slightly higher concentrations (-20 pg/ml) of
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
81
tetravalent Con A rapidly become toxic for the lymphocytes, concentrations of bivalent succinylated derivative, which are much larger (2100 pg/ml) than the optimal ones, still give only slightly decreased mitogenic stimulation. However, the bivalent acetylated derivative behaves like tetravaIent Con A in that it gives the same mitogen dose-mitogen response curve (227). Still, the apparent properties of the two bivalent dimers are similar in that, in contrast to tetravalent Con A, they do not induce strong cell agglutination, they do not cap the Con A-recognized lymphocyte membrane carbohydrates, and they do not inhibit mIg capping, unless further cross-linked by bivalent antiCon A antibody (227). Since there are no dramatic differences in the binding capacity of bivalent and tetravalent Con A to the cells, bivalency is enough to keep the capacity to stimulate mitogenesis, even though it confers restricted ability to induce a series of changes of cell surfhce organization and activities that probably result from crosslinkage of cell surface carbohydrate. What about monovalent derivatives of Con A? Monovalent dimers behave just like their divalent honiologs (both are succinylated dimers) with regard to both their cell membrane-binding effects and their mitogenicity for lymphocytes ( 175). In contrast, monovalent monomers, although they bind almost as well as tetravalent Con A, are only weakly mitogenic; at low doses they do not stimulate blast transformation or lymphocyte-mediated cytotoxicity, and cells precoated with monovalent monomers show impaired binding of, and decreased mitogenic inducibility by, tetravalent Con A (658). Some mitogenic responses are observed at high doses, but they may be due to aggregation or contamination with intact Con A. Thus, this suggests that bridging is required to induce lymphocyte activation. The different results obtained with monovalent dinieric Con A can be reconciled with this concept because of a fact, usually forgotten by reviewers, including myself (364), but of crucial importance in the present context. Although its preparation and purification is very elegant, it remains that the monovalent dimer is not stable as such; that is, when kept in solution for 12 hours at 37"C, it gives a mixture of 70%monovalent dimers, 15%inactive dimers, and 15%bivalent dimers, probably by virtue of the reversible association of the monomers, allowing subunit interchange (175).The bivalent diniers must show a much higher avidity of binding to the lymphocyte surface, and such bound dimers will not b e as easily rearranged by subunit exchange as the soluble ones; thus, the cell membrane will absorb selectively the bivalent dimers and provoke a shift in the equilibrium of monovalent dimers toward the mixture of bivalent ones and of inactive ones. This suggests that even
82
FRANCIS LOOR
when cells are treated by monovalent dimeric Con A, the lectin molecules that trigger the cells to mitogenesis may well be divalent dimeric ones, which lead to some critical degree of cross-linkage of glycoproteins and glycolipids of the membrane of the lymphocytes. A similar conclusion seems to be valid for their triggering by another lectin, soybean agglutinin (SBA). This is made up of four subunits, but it displays only two valences, with specificity for N acetyl-D-galactosamine or D-galactose. It stimulates lymphocyte mitogenesis, but only after neuraminidase treatment of the cells, at least for most species (447, 451). To give triggering, the native lectin needs to be at a high concentration (100-2000 pglnil), but glutaraldehyde cross-linked lectin is much more mitogenic (373). Aggregatefree lectin does not stimulate, thus suggesting that bivalent lectin is not stimulatory and that the mitogenic properties of the native lectin are due to the multivalent lectin aggregates only; as recently shown, tetrad e n t soybean agglutinin complexes give optimal stimulation at concentrations as low as 10 Fg/ml (559). Since multivalent soybean lectin stimulates despite the presence of large amounts of divalent lectin, it also shows the selective advantage given by higher valence: for lectins with binding sites of low intrinsic affinity, the small amounts of highly polymerized lectin present, should bind highly selectively to the cell surface, with practically no competition by native lectin. For various antigen-antibody or lectin systems, enhancement values (i.e., ratio avidityhntrinsic affinity) of lo4to more than lo6 can be obtained when comparing the natural polyvalent reagent (IgC or IgM antibody, polymeric lectin) to its monovalent monomers (301). Phaseolus vulgaris PHA is another polyvalent, mitogenic lectin, with four valences specifically inhibitable b y N-acetylgalactosamine (589). Studies with derivatives having different valences have not been done. However, kinetics studies on the binding of PHA to lymphocytes have revealed the presence of less than 10% of high affinity sites on the lymphocytes and also that maximal stimulation is achieved when less than 20% of the total binding sites on the lymphocytes are bound with PHA (481). This lectin patches and caps on the lymphocytes (362). Some of the observed characteristics of the kinetics of PHA binding could not be related directly to mitogenesis, but may involve some requirement for receptor cross-linkage, thus requiring polyvalence (481). It is also clear that though cross-linking was similarly induced b y PHA on B and T cells, only T cells were stimulated (362), thus forbidding straightforward correlations. A similar conclusion could be reached from the data known for Con A (362).There are also species differences that preclude generalization. Thus, another
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
83
lectin that binds to related carbohydrate residues is peanut agglutinin (PNA), which binds to galactosyl residues exposed only after treatment with neuraminidase, at least for rat, mouse, guinea pig, and human lymphocytes. However, only rat and human lymphocytes are stimulated; mouse lymphocytes are not, even though they show lectin binding similar to that of rat lymphocytes (451).Not yet published is the state of polynierization of the peanut lectin, and what is the valence of the monomer, but polyvalence would be required for stimulation of mouse lymphocytes (504).In all cases where binding of SBA, PNA, and PHA resulted in lymphocyte stimulation, the binding characteristics showed positive cooperativity (503).This indicates that the functional affinity, the avidity of lectin-receptor interactions, increases as the extent of receptor site occupancy increases. A possible interpretation is that as cross-linkage of glycosylated receptors by lectins proceeds, redistribution phenomena make the free glycosylated residues on those or other receptors become more avid ligands for the lectinsthat is, either because of clustering and polyvalence, or of unmasking of cryptic but better fitting receptor sites. Multivalence of lectin would thus be involved to obtain receptor clustering, either because the clusters themselves are the stimulator units or because clustering is required to unmask other molecules that would be the actual stimulator units or growth factor receptors. I n the latter alternative, the differences of mitogenicity of different lectins, with regard to the nature of the lectin itself and that of the cell type, could not b e due to the capacity, for any lectin-cell system, to obtain clustering or not, but to lead to proper exposure of the “stimulator units” for their activation (again, the lectin may or may not bind to them). In other words, ifwe consider that most of the cell glycocalyx is a kind of buffer coat, mitogenic lectins should be able not only to remove such a coat to uncover cryptic stimulator sites, but also to allow the activation to proceed through such sites. The specificity of binding of SBA, PHA, and PNA to galactosyl residues has been taken as a further argument to assign an important function in lymphocyte stimulation to such residues, principally because oxidative stimulation works through them and a series of other mitogenic lectins bind closely related carbohydrates (451).Such a generalization is at least risky, inasmuch as another lectin, H e l i x pomcitiu A agglutinin (237),is not at all mitogenic, although it has an affinity of binding to N-acetyl-D-galactosanline very similar to the one displayed by the mitogenic SBA (257).Such a lack of mitogenicity does not depend on the valence of H e l i x poniatici A lectin, since it is found both for the naturally occurring hexa-
84
FRANCIS LOOR
valent molecule, its divalent derivative, and a highly polymerized (insolubilized) one (131). Still this nonmitogenic lectin was shown to bind either to the same membrane components as the mitogenic lectins or to different but physically associated ones (257). This was done principally by competition studies and by the cocapping methodology, and we have already discussed what the limitations of such methods are (363,364).There can still be subtle differences of membrane sites that are recognized by the different lectins, even if they similarly bind the same sugar in the soluble form. Besides the two classes of mitogenic lectins that can be inhibited either by N-acetylgalactosamine or by a-methyl mannoside, a few other lectins were found to be mitogenic and inhibitable by galactose or lactose. Among these are the Abrus and Ricinus agglutinins: they are divalent lectins, made of four peptide chains, with specificity for lactose, and their mitogenicity can be demonstrated after they are freed from their toxins (abrin and ricin, respectively). The nontoxic ricin B chains, which bind only one sugar, were found to be weakly mitogenic (96). However, the stimulation varied from one B chain preparation to another, as it was due to the small, variable amounts of dimers and polymers ofthe B chain present in the preparation (96).This again favors cross-linkage as a crucial step of lymphocyte triggering. Besides the lectins of plant and invertebrate origins, carbohydratereactive antibodies have been tried with varying success. A series of antibodies were raised in rabbits against sugar azoproteins and purified by affinity chromatography by elution with different monosaccharides; they were shown to react with terminal, nonreducing sugars on mammalian cell surface, Not one of them was found to be mitogenic for mouse, rat, or human lymphocytes, whether or not they were pretreated with neuraminidase (51). Further, some of them could even inhibit Con A-induced mitogenesis, an effect that was obtained also with the monovalent Fab fragments. Thus, it was due not to a possible cytotoxicity of their Fc (e.g., allowing recognition by any killer cell), nor to removal of the mitogenic sites by a clustering process, but more probably to competition of the Fab anti-carbohydrate for some specific Con A binding sites (bulk binding of the Con A was not significantly competed for b y those Fab) (51). In sharp contrast with this lack of mitogenicity of antibodies resulting from an immunization procedure, others (580) reported in normal serum the presence of antibodies that are reactive with cell surface carbohydrates and are weakly mitogenic. Such antibodies were purified by affinity chromatography on fetuin, a glycoprotein whose
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
85
carbohydrate moiety shows some similarity of structure with that of IgG; they induced patching and capping of lymphocyte membrane components. Although present in the serum of a number of species, only the rabbit and chicken antibodies were found to be mitogenic for mouse lymphocytes, the shape of the dose response curve resembling the one obtained with bivalent Con A, but the degree of thymidine uptake being an order of magnitude lower (580). Later, further data were presented showing that the putatively monovalent Fab from the fetuin carbohydrate-specific, natural antibody of chicken origin, were even slightly more mitogenic than the intact bivalent IgG molecule
(581). This was presented as the evidence that “for saccharide-specific residues, receptor cross-linkage may not be a stringent requirement for lymphocyte stimulation” (581). Since then, unfortunately, there has been no further substantiation of such an important observation; this is worrying inasmuch as the monovalency of the chicken Fab preparation, which is the crucial point, does not appear quite evident. Apparently, they were prepared by pepsin digestion of chicken Ig with no further separation, on the basis of a statement in a review article from 1969 that “unlike mammalian yG, pepsin digestion results in a univalent Fab’ rather than a bivalent F(ab’),” (221, p. 69). Such a statement was not documented by any references, and, despite an extensive review of the literature, I could not find any appropriate reference before 1969. Although limited, later published information shows that pepsin digestion of avian immunoglobulins can lead to fragments of various valences and of various sizes; that Ig from different avian species show different susceptibilities to pepsin; that some classes of Ig are more susceptible and some are more resistant to degradation; and that the conditions of the pepsin digestion, principally the pH, do dramatically influence the yields of Fab’, F(ab’)z,undigested Ig, and small peptides for a given class of Ig (343, see also 344). The pepsin treatment of the anti-fetuin carbohydrate molecules was apparently not made in strictly controlled conditions, and the gel electrophoresis analysis of the pepsin digests (581) do not, in my opinion, demonstrate the monovalence of the preparations, but, at best, only that pepsin had degraded part of the Ig. The only argument in favor of a lack of bivalent fragments is the lack of agglutination of pig erythrocytes by the pepsin preparation, which would show that less than 0.1% of bivalent material was present (581).Agglutination is a sensitive method, but it is a secondary phenomenon, not a direct consequence of bivalent antibody binding; for instance, some classes of antibody agglutinate better
86
FRANCIS LOOR
than others, although all are bivalent, and it is therefore possible that a partial degradation of an agglutinating antibody could abolish its agglutinating activity without touching its valence. Finally, it is regrettable that this major exception to the rule that cross-linkage is a crucial step of triggering was not better documented. One would have liked to see controls that stimulation was not due to contaminating complement components or to some low level of remaining proteolytic activity, and it would have been worth showing that stimulation by the pepsin-digested, unpurified fragments of anti-fetuin carbohydrate could indeed be competed for by free carbohydrate molecules. The importance of cell surface carbohydrates and of cross-linkage for triggering is further supported by another series of experiments. Various ligands (haptens such as dinitrophenol, arsanilic acid, biotin) can be chemically coupled to different functional groups of the lymphocyte membrane components, e.g., of proteins (amino groups, sulfhydryl groups, tyrosyl residues, either directly or via a spacer) or of carbohydrate (aldehyde groups formed after mild periodate oxidation of sialyl residues). The mitogenicity of the anti-ligands (i.e., the respective antibodies, the avidin) can then be tested. When the ligands were attached to cell surface proteins, the anti-ligand was not mitogenic (rat spleen cells), but when the ligands were linked to cell surface carbohydrate, the anti-ligand induced the blast transformation of the lymphocyte (690). Using such a system (dinitrophenylated thyrnocytes with the DNP bound to carbohydrate), F(ab’), fragments anti-DNP were found to be an even more potent mitogen than the intact antibody, but the Fab fragments were not mitogenic at all (515), thus supporting again the cross-linking concept. The higher mitogenicity of F(ab’), over intact IgG was interpreted as being possibly d u e to greater flexibility of divalent fragments or to an escape from cellular cytotoxicity mechanisms mediated via Fc recognition. Finally, the importance of cross-linkage for lymphocyte stimulation is also stressed by the observation that divalent mercury cations are mitogenic, whereas monovalent methyl mercury is not, and that a divalent organomercurial (1,4-bimercury-3,4-dihydroxybutane) is also mitogenic (41). However, with a polyvalent organomercurial, a mercury-substituted dextran, where the distance between two mercury atoms was large, no stimulation was obtained, and it was suggested that this may indicate that cross-linking itself is not sufficient, but that it has to occur between proteins (sulfhydryl groups) and bring them in close proximity (41). To conclude with the activation of the lymphocytes, let us recall that certain polymeric proteins such as KLH and “excitability inducing
PLASMA MEMBRANE AND CELL CORTEX JNTERACTIONS
87
material” were active (both for ionophorous and mitogenic activities) only when they were in the polymerized state, not as monomers (542). IX. About Activation of Nonlymphoid Cells
Other cell types show similar characteristic requirements for activation. A. FIBROBLASTS It is possible to stimulate mouse L cells to incorporate exogenously supplied nucleosides and to divide by rabbit antibody directed against the mouse L cell membrane. While both IgG and F(ab’), are active, Fab’ and Fab are not. The divalent reagent aggregates the cells and caps on their membrane (590). The lack of growth observed when aggregation is forbidden (by suspending the cells in an agar gel) might not show the requirement of aggregation for stimulation (590)but perhaps indicates that growth of those cells is anchorage dependent or may be due to the anticomplementarity of the agar. Indeed, the complement seems to play a crucial role in the stimulation of L cells by antibody (591):the response is much increased b y use of fresh serum, and, by using various complement-deficient sera, it would appear that activation was through C 3 by either the classical or alternative pathways.
B. MACROPHAGESAND POLYMOWHONUCLEAR NEUTROPHILS These inflammatory cells are in many respects related to the lymphocyte compartment of the immune system, with which they cooperate. The term “activation” in their case is not used to define a stimulation to mitosis, but rather a stimulation of the aggressor or scavenger capacities of neutrophils and macrophages. To obtain such functional activation requires conditions that recall those needed for lymphocyte activation or, as will be seen, basophil-mast cell activation. Activation, in the case of macrophages and neutrophils, is monitored by a number of parameters that go from the general morphology of the cells to more refined biochemical activities that develop (for reviews, see 100, 300,446,533,564,644).Activation can in all cases be measured by the comparative biochemical analysis of the cellular components of “resting” and “activated” cells. Both for macrophages and neutrophils, activation is manifested by increased motility and by a release of a variety of enzymes and metabolites in the medium around the cells (30,100,300,561,600).Similar panels of stimuli can activate macrophages and neutrophils, both particulate ones and soluble ones,
88
FRANCIS LOOR
and they are extremely varied. Both activation by phagocytosable material, with or without actual phagocytosis, and activation by surfacereactive stimuli leads to an activation of oxidative metabolism in the cells; this is characterized by increased 0, consumption and production of peroxide and superoxide radicals, with their release in the mediuni, thus allowing the use of simple monitoring methods of activation, such as increment of cheniiluminescence in presence of luiiiinol (114, 130, 205, 219, 401, 428, 534, 537, 538, 541, and others). Thus, such activation is obtained by phagocytosable material, even if not digestible, such as opsoiiized zymosan particles (130, 204, 260, 293,401,564),and inore conventionally by immune complexes or IgG aggregates (204,260,293),or by digestible antigens, such as sheep red blood cells (564). It seems that activation of macrophages can be closely related with activation of hexose monophosphate shunt (HMS) activity, and that phagocytosis does not act as a stimulus if not accompanied by HMS stimulation, as was shown to be the case for phagocytosis of latex particles (564). However, nonphagocytosable material such as methylene blue can stimulate both HMS and release of enzymes and metabolites in the mediuni, and the tendency is now to consider that a triggering ligand, either particulate or soluble, is a ligand that gives a rise in HMS activity. This is supported by a number of observations made with a variety of surface reagents, many of which strongly recall the factors causing or modulating “lymphocyte activation” that were discussed in previous sections. This will be considered hereafter, without much development, as it is not possible here to review such a field in any extensive or critical manner. Thus, an activation of phagocytes and neutrophils (e.g., activation of spontaneous mobility, of enzyme secretion or release, of oxidative metabolism, of response to chemotactic factors or to MIF) or a modulation of such activation could be obtained using ( a )A23187 ionophore, in a Ca2+ion dependent way (e.g., 149, 401, 536, 537, 540, 560, 675) and perhaps in an M$+ dependent way (238);( b )La”+ions (536); (c) Zn2+ ions (91-93), (a!) amphipathic molecules, among which are fatty acids (295,701); ( e )membranous phospholipid splitting by phospholipase C (477);(f)some lectins, principally Con A (143,194,200-202,224,225, 263,424,535,538,607,650,697)(see also further); ( g ) phorbol myristate acetate (115, 130, 341, 401, 650) which is mitogenic for some T cells too (543, 635); and also by some other surface active agents, not yet tried so far as I know, on lymphocyte activation, such as ( h )digitoiiin (99); ( i ) chsmotactic agents, namely, the C5a fragment from complement, as well as a series of small synthetic N-formylmethionyl peptides (30, 31, 149, 189, 204, 251, 341, 561, 600, 622, 675), which are
PLASMA MEMBRANE AND CELL CORTEX LVTERACTIONS
89
chemotactic agents requiring Ca2+and Mg2+ions for optimal activity, i.e., two ions being needed for actomyosin function [It is also remarkable that the latter chemotactic agents can also stimulate basophils to release anaphylaxis mediators in a Ca’+ dependent way (266,610)];( j ) various as yet ill-defined lymphokines (321,339,429,648); ( k )agents modulating the CAMP level (251, 325, 422, 424, 485, 489, 588, 649), which itself can be modulated b y some lectins, among which is Con A (194, 224, 225), as was also the case with lymphocytes (e.g., 98, 232,
377). Clustering of membranous components seems to be an important requirement for stimulation by lectins, as best indicated by studies performed with Con A derivatives. This concerns both the activation of oxidative metabolism (538,697) and the modulation of responsiveness to hormonal stimuli that act through the CAMP messenger (194, 225). Thus, the activation of oxidative metabolism of neutrophils (monitored essentially by O2 metabolite release) was shown (538) to be more pronounced after treatment of the cells with native, tetravalent Con A, than after similar treatment with its succinyl, divalent, derivative; however, greater response was obtained if the latter ligand was further cross-linked b y anti-Con A antibodies. With peritoneal macrophages, differences were also obtained for tetravalent and divalent Con A, but they may not be due to the activation step itself (697):indeed, Con A, but not succinyl-Con A, could cause a significant 02-relecise in the medium, but both could trigger the enhancement of consumption of O2 by the cells and of intracellular 02-production. In both neutrophils and macrophages, actual binding of the Con A to surface receptors was shown to be needed for the activation to proceed, as removal of the lectin by excess sugar could block it (538, 697). However, the data on the macrophages seem to indicate that, though binding of Con A to the receptors would be sufficient for 0, consumption and intracellular superoxide production, clustering of the Con A receptor is further required for the subsequent high rate of superoxide release outside the cell (697). Few other lectins have been used to look for oxidative metabolism stimulation [e.g., PHA also stimulates, but PWM does not (607)],but a number have been tried in the case of the facilitation of the responsiveness of peritoneal macrophages to adenylate cyclase stimulators: neither PWM, SBA, WGA, the lectins from Lotus tetraglonobus and liina bean, or LPS can stimulate the latter response as does Con A (194, 225). There are differences that may be due to methodology or to different animal species; for example, PHA was found to stimulate guinea pig macrophages (225) but not rat macrophages (194).The len-
90
FRANCIS LOOR
ti1 lectin would activate as does Con A (225). Thus, not any lectin or other ligand binding to the macrophage membrane can enhance the sensitivity of the cells to the adenylate cyclase stimulators, but further, for those lectins that activate, mere binding is not sufficient and clustering of the membrane receptors seems to be mandatory (225). Indeed, only native tetravalent Con A can activate, but succinyl, divalent, Con A or Con A insolubilized on a solid matrix (Sepharose beads) cannot. This recalls similar observations performed with lymphocytes (98,232,377).The intracellular concentration of CAMP, whether it is a second messenger of activation or not, has been shown to modulate most of the parameters of activation of the macrophages that were listed above (see 251, 422, 485, 489, 588, 649). Are MT, MF, and general metabolism involved in “activation” of macrophages and neutrophils? Probably yes, but it is difficult to determine how, In the absence of any other ligand, such as Con A, preincubation of guinea pig macrophages with lop6M colchicine enhances their sensitivity to adenylate cyclase stimulators (224). However, lumicolchicine does the same, as well as colchicine in conjunction with D,O to stabilize the MT (194), which suggests that such activation is due not to MT disruption but merely to membranous effects of the alkaloids. Pretreatment with cytochalasin B has no effect on the Con A-induced facilitation of adenylate cyclase stimulation (225),thus showing that clustering is more crucial than capping, shedding, or endocytosis of the clusters, for that type of response at least. In the case of the activated oxidative metabolism response, there is some controversy. With peritoneal exudate macrophages there was no detectable effect, on the amount of superoxide formed intracellularly, of pretreatment with inhibitors of electron transport (rotenone M, or M ) , thus sugsodium malonate 10” M ) or with colchicine (up to gesting no essential role for MT or metabolism in the activation of 0,formation (697).However, the release of 0,- by peripheral blood neutrophils was found to be potentiated by agents that stabilize MT (D,O, Con A) and to be inhibited by agents that destabilize them (colchicine, vinblastine) (424). Thus, to monitor “activation” by measuring 0,release may lead to artifacts. Furthermore, the interpretation of the latter set of data is difficult, since they deal with an 0,- release induced by a cytochalasin and may therefore not represent an actual activation model. There have been many attempts to use cytochalasin to study the involvement of M F and the requirement for phagocytosis for activation. Cytochalasin B treated cells are still capable of generating and releasing 0,- when stimulated with phagocytosable particles or surface
PLASMA MEMBWIVE AND CELL CORTEX INTERACTIONS
91
active reagents, although phagocytosis itself is strongly inhibited (105, 204). This happens despite the fact that cytochalasin B-treated cells exhibit a significant reduction in 0, consumption (e.g., 539), as well as a decreased phagocytotic capacity (703). It would seem that most of the 0,- release occurs from oxidative metabolism happening at the PM level, and only the latter would take place in the presence of cytochalasin B, with no 0,- formation intracellularly in the phagocytosis vesicles. The actual origin of both 0,- and H z 0 2released outside the cell is, however, still difficult to define exactly (e.g., 130, 205, 541). Treatment with cytochalasin B would even enhance 0,release by cells stimulated with opsonized zymosan, C5a, IgG aggregates (204); it would have no effect on 0,- release by neutrophils stimulated with phorbol myristate acetate, but it would enhance severalfold the 02-release obtained by their treatment with the formylmethionyl peptides (341), as well as their release of lysosomal enzymes. Cytochalasin E, which also inhibits phagocytosis, could even by itself stimulate an 0,- release similar to or even better than that obtained by particle phagocytosis (424,425). However, those authors found no stimulation by other cytochalasins (A and B) (425) and even a defective capacity to release 0,- in response to cytochalasin E stimulation in cells that had been pretreated with cytochalasin A or B (424). It is thus difficult to conclude on the involvement or not of M F structures in this type of activation or in its expression. The different cytochalasins have some common and some unique high affinity binding sites on the mammalian cell membrane (19, 294, 354, 355), and this may explain their differences in behavior. It seems that one is dealing more with effects due to modulation of the membrane itself than with effects at the level of M F activities. But the issue is still confusing. Any firm conclusion is in any case obviated by the impossibility of defining what activation actually is. In a recent paper (453), surface active agents and phagocytosable material that activate the niacrophage were shown to stimulate a burst of turnover of phosphatidylinositol in the PM, whereas nonactivating particles, although phagocytosable, were not doing this. Does this also depend on HMS activation? Can it be correlated to the similar early membrane changes following lymphocyte activation? These are difficult questions to answer. But also, is it fair to consider “activation” globally, as done here, from the first triggering of membrane events to the eventual release of aggressive O2metabolites and lysosomal enzymes? It would seem so. Indeed, in a recent paper (341), the same formylmethionyl peptides were shown to be chemotactic and to be able to stimulate oxidative metabolism, 0,- release, and lysosomal enzyme release by
92
FRANCIS LOOR
neutrophils, but the latter activation required concentrations of peptides 10-50 times higher than that required for chemotaxis stimulation. Thus, there may be “unique” stimulants for “different” activation responses, and this opens up an interesting possibility: that is, that chemotactic factors that are generated at sites of infection and diffuse as a gradient in the tissue will attract the killer and scavenger neutrophils and macrophages from very far in the tissue, but will stimulate the aggressor properties of those cells only when they have reached the sites of infection where the cells meet a much greater dose of “chemotactic” factor. Still, d o all these types of activation, from chemotaxis to enzyme release, depend for their initiation on exactly similar membrane component changes, such as “clustering”? What are C5a, phorbol myristate acetate, formylmethionyl peptides, and other surface active agents actually doing to the cell membrane? There are a few morphological and biochemical analyses of their effects (e.g., 188, 423, 452, 676), a number of membranous activities appear altered (transmembrane potential, fluxes of ions, membrane spreading and ruffling, pseudopode formation, cell swelling, cell aggregation, and others), but whether this can be related to induction of membrane component clustering, or mimicked by it, cannot be decided at present.
C. BASOPHILSAND MAST CELLS There is abundant evidence that the activation of basophils and mast cells to release mediators of anaphylaxis makes use of mechanisms very similar to the ones used by lymphocytes, as cross-linkage of receptors is also required, and that many of the factors now found to influence lymphocyte activation were actually long known to influence basophils and mast cell activation. These cells are the mediators of anaphylaxis, and the mechanism of their activation has been the subject of recent reviews (282,304,409).Some of the membrane dynamic aspects of their stimulation have been discussed in a previous review (364).These cells have membranous IgE of exogenous origin, to which binding of allergen results in activation of the cell, i.e., release of its mediators of anaphylaxis (histamine, serotonin, slow reacting substance of anaphylaxis, and others). Various tests can be performed in vivo or i n vitro for basophil and mast cell activation, which will not be taken account of here, as they routinely gave similar information on the mechanisms of activation when both in vitro and i n vivo tests of a particular condition were done, and in vitro results fit with i n vivo ones. The analysis of the factors or conditions that affect such basophil-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
93
mast cell activation shows striking similarities with what was found for the mitogenic activation of lymphocytes, the principal condition being that membrane IgE should be clustered to activate the cell (e.g., by allergen or anti-Ig). This was first shown by obtaining reversed-type erythema wheal reactions by reduced and alkylated anti-IgE and by its F(ab‘)z but not by Fab’, implying that bridging of IgE was needed (279). The binding of IgE to the basophils was shown to be due to its Fc rather than through the antibody active sites, since no passive sensitization could be done by the F(ab’)zof IgE (280). Whereas the Fc monomer could not do anything else than bind to the cells, bisdiazotized benzidine-aggregated Fc could stimulate mediator release both in uiuo and in uitro, the similarly treated F(ab’)z of IgE being inactive (281). This was interpreted in terms of conformational changes in the Fc due to aggregation, but, as will appear below, it actually showed that activation is produced by some clustering of the membrane receptor sites for the Fc of IgE (see below). From the morphological point of view, the treatment of basophils by anti-IgE was shown to induce patching and capping of mIgE (165, 627), but very soon it was also shown that no simple correlation could be shown between mediator release and gross redistribution of mIgE (32) because ( a ) cells that release histamine in uitro were indistinguishable from those that cannot, as far as redistribution of mIgE was concerned (same anti-IgE dose, time and temperature dependence); ( b )optimal mediator release occurred, at low anti-IgE doses, before or in the absence of redistribution; (c) redistribution could occur at high anti-IgE doses in the absence of mediator release. Thus, although bridging is needed, only a certain kind of mIgE cross-linking effectively stimulates the basophils (32). The natural IgE distribution was shown to be essentially uniform, and patching required cross-linking, e.g., use of F(ab’), anti-IgE (73, 338). While capping was definitely obtained on basophils (32,73, 165,627),mast cells failed to cap well, for some still unknown reason (338, 405). Basophil capping requires active metabolism, as it is inhibitable by NaN3, and also intact MF function, as it is inhibitable by cytochalasin B (73). Important for the mechanism of activation are the observations that the maximum size of mIgE cluster that could be necessary for signaling seems to be less than 10 IgE molecules (at least for mast cells) (338) and that the degranulation reaction is not an all-or-none response, but can be localized to the area where the mIgE is clustered (338, 511). This is a particularity that cannot be found in the case of mitogenic activation of lymphocytes. But there are several other parallels with lymphocyte activation, to
94
FRANCIS LOOR
be listed hereafter. Like lymphocyte activation, IgE-bearing cell activation requires energy; for instance, it is inhibitable by the metabolic antagonist 2-deoxyglucose (346). Sulfhydryl binding reagents, such as N-ethylmaleimide, iodoacetic acid, and p-chloromercuribenzoate, strongly or totally inhibit mediator release (22, 144, 145, 414), the phase of the activation inhibited by them being apparently available prior to the allergen-IgE interaction (22). Various thiol compounds, such as cysteine, thioglycolate, glutathione, and sulfite, although they inhibit mediator release in a reversible way when used at high concentration (144, 145, 479), can on the contrary potentiate such release when used at lower concentrations (145, 465467,479). While colchicine has an inhibitory effect on the release of slow reacting substance of anaphylaxis (SRSA), though only when M (463), both cytochalasins A used at the very high dose of and B, in a more classical (lower) dose range, appear to effect a dosedependent inhibition of SRSA release (464). At high doses they also inhibit histamine release (cytochalasin A being irreversible, B being mostly reversible), but at low doses cytochalasins A and B both give enhancement of such a response (464).The cytochalasins appear to be more effective when present during the antigen-dependent, Ca”independent phase of the activation of mediator release, thus at a very early stage of activation (464), as was also the case for lymphocyte activation. Modulation of the cAMP content also affects the stimulation; for instance, dibutyryl cAMP and agents that increase endogenous cAMP content inhibit the early stages of allergen stimulation
(346). As for lymphocyte activation (see preceding section) and for the activation of many other cell types (e.g., macrophages), IgE-bearing cell activation is accompanied by marked and selective changes in the turnover of phospholipids (310), and it is possible to modulate their secretory response by the addition of phospholipids: phosphatidylserine enhances the response (393, 609), several of its N-substituted derivatives being inactive and even competing with phosphatidylserine enhancement, and other diacyl phospholipids were found to be ineffective (393). Finally, lysophosphatidylserine is much more active than phosphatidylserine, the latter working on the cells in a micellar state and the former being active well below its critical micelle concentration (392). A similar approach to lymphocyte stimulation has not yet been made. IgE-bearing cell activation can be performed not only with the specific allergen, but also with IgE-directed ligands, such as anti-IgE
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
95
antibody (see later), and also with lectins, such as Con A (e.g., 97, 310, 392, 393, 511, 611). Activation by Con A occurs presumably via the binding of the lectin to IgE carbohydrate groups, since, on a molar basis, IgE was 4000 times more active than a-methyl-D-mannoside in inhibiting Con A stimulation (611).The divalent succinylated Con A is not as good a stimulator as the tetravalent native Con A. As for antiIgE, the binding of the Con A can be dissociated from the stimulation (histamine release), and the cells can be desensitized by prolonged incubation in the absence of cations or in an excess of Con A, this making the cells insensitive to activation b y allergen or anti-IgE (611). Activation of mIgE-bearing cells can also be obtained by enzymic attack of the membrane, e.g., by chymotrypsin (see 97). Chymotrypsin inhibitors have long been known to inhibit histamine release (22). The stimulation seems to be linked to Ca’+ fluxes through the niembrane (511). No stimulation can be obtained in the absence of Ca2+ (though redistribution is not affected), and it is indeed the way to desensitize the cells (e.g., 32). As expected, the bivalent cation ionophore A23187 can induce histamine release in a Ca2+-dependent way (310, 346, 467). A “blockage” of Ca’+ gates by lop7M La3+completely inhibits both A23187 and allergen-dependent stimulation (346). Stimulation can also be obtained by Ca2+(but not Mg’+ or K+) entrapped within phospholipid vesicles whose constitution (unilamellar, fluid, and negatively charged) makes them capable of fusing with the cell membrane (631).Possibly related to Ca2+gating is the fact that no stimulation can be obtained in the absence of phosphatidylserine (609).Although the exact mode of action of the lipid is not known, it is postulated that it modulates the changes of Ca2+gates in the PM or that it is required for activation of a Na+,K+-ATPase essential for mediator release (171). Probably relating to Ca2+gating are the long known protective effects of Znz+on the disruption of mast cells by the allergen-mIgE reaction or other means (compound 48/80, see later) (264): Zn2+ inhibits histamine release at concentrations too low significantly to chelate histamine within the cells [a former hypothesis on the effect of Zn2+, due to the selective accumulation of Zn2+ within mast cells (307)] (305). In fact, much stronger than with ZnC1, is the inhibition obtained by the use of unsaturated equimolar zinc-8hydroxyquinoline complex, known to be unable to penetrate into the cells and to act only on the mast cell membrane (306). There are also a number of studies on the kinds of clustering of mIgE on cells with IgE receptors that lead to stimulation. Such studies have not been performed with lymphocytes. It was indeed suspected
96
FRANCIS LOOR
that only certain kinds of cross-linking of IgE could effectively stimulate the basophils (32); extensive clustering was not the best way to stimulate. Morphological ultrastructural studies fit this concept (338). First, a series of elegant studies were performed using monovalent, bivalent, and multivalent haptens differing in molecular size, degree of substitution, and rigidity, some being homologous and others heterologous (e.g., there are two different haptens, H, and He, separated by a small carbonated chain). Thus, using dinitrophenol (DNP) as hapten, monovalent haptens were found to be inactive, while at equimolar doses the bivalent cr-N-(e-N-DNP-aminocaproyl-)-A‘DNP-L-lysine was as effective as the multivalent DNP-albumin (420). Identical results were obtained using benzylpenicilloyl (BPO) as hapten, BPO-NH-(CH,),,NH-BPO as homologous bivalent haptens, and BP09-Lys,, as homologous multivalent hapten (612).The flexibility of the segment of the molecule separating the two haptens is important, as exemplified by the lower activity shown in the above systems by, respectively, the more rigid di-DNP-gramicidin J (420) and di-BPO hapten separated by rigid diphenyl bridges or diphenylsulfone (612).The distance separating the two haptens is also important with, in the above systems, much less activity shown by a, E-N-bis DNP-lysine (420) and in the above di-BPO hapten system an optimum when the number ( 1 % ) of CH, (flexible chain) is 6 to 9, much less activity being found for ri = 3 and n = 12 (612). No difference of activity due to chain length is found for bivalent haptens separated by a rigid spacer (612). Thus, bridging is needed, and internal bridging between the two active sites of a single IgE is probably not the mechanism leading to histamine release. That there is a critical size clustering for stimulation to occur is also suggested from experiments (385) showing that histamine release stimulated by a suboptimal anti-IgE concentration is inhibited by Fab anti-IgE, while the one stimulated by a supraoptimal anti-IgE concentration is on the contrary enhanced by Fab anti-IgE treatment of the cells. It is interesting to note that a high dose of anti-IgE leads to unresponsiveness to the allergen, whereas a similarly high dose of Fab anti-IgE has no effect on the allergen-induced stimulation (385). The probable explanation is that the anti-IgE not only binds to IgE but clusters it, while the Fab anti-IgE only binds-the active sites of the IgE remaining available for allergen binding and allergen-induced critical” clustering. The “critical size clustering” hypothesis seems to be the inescapable conclusion from mathematical models as well (see later). Studies performed with heterologous bivalent haptens and cells sensitized to “
PLASMA hfEMBRANE AND CELL CORTEX INTERACTIONS
97
either one or both haptens have shown that ( u ) in doubly sensitized animals, the IgE of both specificities are on the majority of the basophils; ( b ) they are not grouped as a function of their allergen specificity; and ( c ) the bridging must occur between two IgE of the same cell in order to stimulate (384, 420, 612). Thus, after suitable in v i w sensitization with anti-DNP und anti-ABA (p-azobenzenearsonate), the heterologous divalent hapten eN-DNP-a-N-[(4-hydroxy3-ABA)-phenacetyl]-~-lysine could elicit the reaginic reaction (Prausnitz-Kustner skin reaction); it did not work after sensitization with only one of those IgE antibodies, and complete blockage of the reaction could be obtained by equimolar doses of monovalent hapten derivatives (420). Using BPO and DNP for either mono- or bihaptenation of octamethylenediamine, either for homologous or for heterologous bihaptens, similar results were obtained in vitro with singly or doubly sensitized rabbit basophils (384) and confirmed using hexaniethylenediamine as spacer (612): (u ) with doubly sensitized basophils, all bivalent haptens, both homologous and heterologous are activating; I!(. ) with singly sensitized basophils only the corresponding homologous bivalent hapten can activate; (c) a mixture of the two singly sensitized basophils is not stimulated by the heterologous bivalent hapten. Another elegant approach to determine the minimal degree of cross-linking required to stimulate degranulation of mast cells consists of treating them with monomers, dimers, trimers, and higher polymers of IgE (appropriately separated after treatment with a cross-linking reagent (dimethyl suberimidate): preformed dimers were capable of triggering the cells, thus suggesting that the bridging of two IgE molecules only is sufficient to generate a “unit signal,” the summation of such signals leading to the degranulation (578). A number of mathematical models have been presented to interpret the requirement for receptor cross-linking for cell activation, in general (e.g., 34, 35, 116, 119), and basophil-mast cell activation in particular (117, 120). Such mathematical models have been tested with specifically sensitized mast cells using homologous bivalent hapten, varying its concentration and measuring the amount of mediator release (117, 118, 121, 122). The observations generally fit the predictions of the models. They are consistent with the concepts that (a ) the amount of cross-linked IgE molecules on the cell surface is what controls the amount of mediator release (117,121);(b) there is a threshold number of clusters to obtain release and only a small fraction of clusters can efficiently transmit signals (117); ( c ) cross-linking is required throughout the course of mediator release, as there is reequilibration
98
FRANCIS LOOR
within seconds of any perturbation of the cross-linking level (122); ( d ) the number of clusters is more important than their size (1 17). Some data (117) suggest “a high degree of cooperativity in the biochemical sequelae to clustering” (117) and can indeed be interpreted as showing that clustered IgE have to interact with some other membrane molecule in order to give an effective triggering signal; this is in conflict with an earlier statistical analysis (383) of such a reaction, which had suggested that the correlation of the amount of mediator release by basophils and the mean number of activations, was more compatible with a one-hit activation (i.e., the condition and geometrical configuration necessary to elicit the response are fulfilled by a single molecule), than with multihit processes. Various deviations to the optimum (allergen dose-mediator release) response curves suggest that both the signal for release and the signal for desensitization are “clustering” in nature and increase as the cluster concentration increases, and that such signals reach a plateau at large cluster numbers (118). As a complement to the studies on the critical cluster size leading to activation, a number of approaches have been done to determine the valence of the receptor of IgE, present in the basophil-mast cell membrane. These receptors have a high affinity for the IgE Fc, but there always remain a number of free receptors on the immediate hypersensitivity cells. It was found that “empty” receptors would comigrate with receptors to which IgE was bound (cocapping by antiIgE), suggesting that on basophils a group of receptor sites are associated with each other on the cell membrane or that receptor molecules are multivalent with respect to the combining sites for IgE (283). This observation has not been substantiated more recently, on the contrary. On mast cells coated by fluorescein-labeled IgE and rhodamine-labeled IgE, anti-IgE coredistributes both, but antifluorescein redistributes only very little of the rhodamine-IgE (405). This suggests the monovalence of the receptor for IgE or a very low oligovalence. Monovalence of this receptor was also concluded from data from fluorescence photobleaching recovery, which is a method used to measure the macroscopic lateral motion of fluoresceinated cell surface components (562). This method consists of bleaching a very small region of the membrane fluorescence by a laser beam and by measuring the rate of fluorescence recovery by diffusion of fluorochrome from the adjacent parts of the cell. It also gives much information on the alteration of mobility of the receptor after various treatments: reduction after extensive aggregation in conditions that inhibit mediator release; reduction by doses of c9tochalasin B that are inhibitory; no detectable effect of colchicine; no effect of the limited cross-
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
99
linking by anti-fluorochrome as adequate for degranulation (562). The effect of cytochalasin B may indicate that the receptor mobility depends on M F activity, but this should require further analysis. The receptor for IgE would be a glycoprotein of 58,000 MW in gel electrophoresis (285),although a value of 45,0O0-5O70O0was obtained after purification by repetitive affinity chromatography on IgE (331). The properties of the isolated receptors seem to fit the ones they show in situ: binding of IgE is specific as far as animal species, Ig class, and lack of binding after IgE denaturation are concerned (332). The problem of the monovalence of the receptor comes back, since at least the solubilized, purified receptors would exist in monomeric and multimeric forms, all capable of binding IgE (332).Antibody preparations directed against the receptor for IgE have been obtained by using as antigen either the gel electrophoresis band material temptingly identified as the Fce receptor of mast cells (285) or the complexes of Fce receptors, IgE and anti-IgE, isolated by detergent lysis of basophils (284), the antisera being made specific by absorptions. Both antibody reagents bind to cell-bound FCEreceptors and can block IgE binding to the mast cell-basophil membrane. In the presence of excess IgE, to saturate all FCEreceptor, neither antibody preparation can activate mediator release. However, the preparations do it when not all Fce receptors are loaded with IgE (284, 285). Such an activation by the anti-Fce receptor requires Caz+(285) and cross-linking, since whole antibody and F(ab’)qactivate whereas Fab’ do not (284,285).Fab’ antiFce receptor followed by a divalent anti-Fab also stimulates (284). Finally, the release of mediators can be stimulated by anti-Fce receptor antibody on mastocytoma cells lacking completely IgE (285). The conclusion from both groups is thus that the aggregation of the receptors for IgE provides the critical signal for cell activation. In this context, and though it is unclear that one is not dealing with contaminating antibodies, it is worth recalling that mediator release has also been obtained with antibody directed against major histocompatibility complex antigens of mice (107) and rats (421).In the mouse system, allogeneic anti-H2 were used, their Fab were able to compete with allergen binding to the cells, not all classes of antibody could activate (only IgG,, and not IgG, and IgM), and the Fc portion of the activating antibody was needed, though this was not due to complement dependence (107).This is thus a rather different type of activation than that obtained by anti-Fce and anti-IgE where F(ab’)zcan activate. Morphologically, the degranulation induced by anti-H2 looks like that induced by immune complexes (340),also a situation where a critical size clustering of identical determinants is probably achieved.
100
FRANCIS LOOR
It would appear that the requirement for the Fc portion of the IgG, with anti-H2 activity is required for interaction with an Fc receptor on the same cell to which the antibody is bound through its Fab portions (106). In conclusion, the basophil-mast cell activation system is the one where requirements for clustering have been best demonstrated. There is only little information on how the FCEreceptors interact with the cell corte? elements, but the fact that cytochalasin B controls the mobility of these receptors suggests that they would not be freely diffusible ones and must belong to the class of membrane components directly associated with MF. There would be many experimental testable models of lymphocyte activation to be found in the presently known characteristics of IgE-bearing cell activation. This might also be valid for their inactivation. With the latter cells, in uitro,the conditions to desensitize are easy to provide: for instance, one simply gives the IgE-directed ligand, either the allergen itself or an anti-IgE antibody or a lectin, but in the absence of Ca2+or of phosphatidylserine (e.g., 609). It may be worth considering that conditions that lead to basophil-mast cell desensitization would similarly lead to lymphocyte desensitization or paralysis. X. Concluding Remarks
Essentially two kinds of interactions of the plasma membrane and of the cell cortex have been considered, one happening as an active process of effective recognition performed by the lymphocyte itself, and the other happening as a passive recognition of the lymphocyte membrane b y various ligands. In the first case, the lymphocyte had decided to move in a given direction, or to associate with another cell, either for cooperation or for killing, or otherwise. The elements of the cell cortex and the plasma membrane were definitely involved in the effective recognition process, as best evidenced by the formation of microprojections. Furthermore, as shown for the B lymphocyte, the membranous immunoglobulins were more concentrated on such microvilli; that is, there is a strong deviation from random distribution of components that have definite capacity and functions for recognition. The reason why those receptor molecules tend to accumulate there is unclear, for, if we consider that expression of microvilli is already a sign of an “activated” state, what is the purpose then of having mIg concentrated on microvilli? Is it for better shedding? Is it for better recognition and complete a&ivation? The development of many microprojections by a cell
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
101
may not be a sign of profound activation, but only that the cell has been shifted from the resting state to a highly receptive one, e.g., by a lymphokine, by some unspecific factors. The second case, passive recognition, would be what is happening when the cells are exposed to ligands that bind to or react on their membrane and either put them in a receptive state or push them throughout the whole activation sequence. Here, a number of factors were considered. Most of the factors that had some effect were “membrane active agents,” showing the importance of membrane structural organization. The role of the cell cortex was much less evident: of course, the cells do not feel too well with their microtubules and/or microfilaments damaged, but these structures do not seem to play crucial roles in the sequence of signals leading to full cell activation. Where a role might be played by the cell cortex elements, it is rather at the primary level of the modulation of plasma membrane organization, not at the secondary one of transmitters of signals. All factors that could interfere in one or another way with membrane organization could modulate cell activation, but the most remarkable one was the stringent requirement for clustering shown by lectins and antibodies. Particularly in the case of the B cell, the receptor immunoglobulin is there on the membrane to inform the cytoplasm of what is happening outside, at the cell surface. That receptor has its built-in requirement for clustering: it is bivalent and symmetrical, both sites being of same specificity. What could have been the selective pressure for bivalence of a membrane receptor if not the clustering? It cannot be so much the affinity of binding, since those receptors are not soluble (in a three-dimensional sense), but only diffusible in the membrane (soluble in a two-dimensional sense) and without fipflop. At equal densities of surface diffusible receptors, monovalent ones and bivalent ones should provide to the surface a roughly similar avidity of binding: the cellular membrane itself constitutes the polymerized antibody. Thus the probable raison d’6tre for bivalence of mIg is the capacity for clustering. Not every clustering lectin or antibody could stimulate, but among the stimulating ones, the capacity to stimulate was lost as the clustering capacity was lost. Experiments with basophils-mast cells were the most beautiful. What about simpler molecules, the so-called membrane active agents? Do they also induce clustering? The answer is unclear. Some might do it, but some might allow bypassing the clustering step, the initial “perturbation” of the membrane that is felt by the cell as first signal, and might deliver directly the second signal. What could this be? Perturbing the ion fluxes, the internal pH, the cyclic nucleotide
102
FRANCIS LOOR
metabolism, the calcium concentration, the phospholipid turnover, or what else? There is no answer yet, but probably it is of general biological vaIue. After the clustering induced by antigen, the mIg has no further role to play, the rest of the sequence is no longer a specific immunological process. The whole series of factors that were capable of modulating lymphocyte function and activation could play exactly the same tricks on the activation of other cell types. And many of those assays recalled similar attempts, performed years ago, to try to modulate the division of fertilized eggs or to try to kick its parthenogenetic development.
REFERENCES Abbas, A. L., and Unanue, E. R. (1975).J. Zmmunol. 115, 1665. Abbott, R. E., and Schachter, D. (1976).J.B i o l . Chem. 251, 7176. Able, M. E., Lee, J. C., and Rosenau, W. (1970).Am. J . Pathol. 60, 421. Adelstein, E. H., Barrett, B. A. L., and Senhauser, D. A. (1976). Cancer Res. 36, 302. 5. Albertini, D. F., and Anderson, E. (1977).J.Cell Biol. 73, 111. 6. Albertini, D. F., Berlin, R. D., and Oliver, J. M. (1977).J . Cell Sci. 26,57. 7. Alderson, J. C. E., and Green, C. (1978). Exp. CeZZ Res. 114,475. 8. Alford, R. H. (1970).J.Zmmunol. 104, 698. 9. Allwood, G., Asherson, G. L., Davey, M. J., and Goodford, P. J. (1971).Zmmunology 21,509. 10. Alm, G. V., and Peterson, R. D. A. (1969).]. E x p . Med. 129, 1247. 11. Anderson, A. O., and Anderson, N. D. (1976). Zmmunology 31, 731. 12. Anderson, A. O., Anderson, N. D., Hoffman, R. D., and White, J. D. (1978). Fed. Proc., Fed. Am. Soc. E x p . Biol. 37, 591. 13. Anderson, J., Bullock, W. W., and Melchers, F. (1974). Eur. J . Zmmunol. 4, 715. 14. Ando, M., Suga, M., Sugimoto, M., and Tokuomi, H. (1979). Infect. Immun. 24, 404. 15. Armstrong, M. D. (1979). Biochim. Biophys. Actu 584, 542. 16. Ash, J. F., and Singer, S. J. (1976). Proc. Natl. Acad. Sci. U . S. A. 73, 4575. 17. Ash, J. F., Louvard, D., and Singer, S. J. (1977). Proc. Natl. Acad. Sci. U. S. A. 74, 5584. 18. Astruc, M., Laporte, M., Tababik, C., and Crastes de Paulet, A. (1978). Biochem. Biophys. Res. Commun. 85, 691. 19. Atlas, S . J., and Lin, S. (1978).J.Cell Biol. 76, 360. 20. Attallah, A. M., Sunshine, G. H., Hunt, C. V., and Houck, J. C. (1975).E x p . Cell Res. 93,283. 21. Auk, K. A., and Unanue, E. R. (1977).J.Immunol. 119,327. 22. Austen, K. F., and Brocklehurst, W. E. (1961).J.E x p . Med. 113, 521. 23. Averdunk, R., and Lauf, P. K. (1975). Exp. Cell Res. 93, 331. 24. Ax, W., Malchow, H., Zeiss, I., and Fischer, H. (1968). Exp. Cell Res. 53, 108. 25. Barber, T. A., and Alter, B. J. (1978). Scand. J . Zmmunol. 7, 57. 26. Barber, B. H., and Crumpton, M. J. (1976). F E B S Lett. 66,215. 27. Barber, B. H., and Delovitch, T. L. (1979).J . Zmmunol. 122,320. 28. Barnett, R. E., Scott, R. E., Furcht, L.T., and Kersey, J. H. (1974).Nature (London) 249, 465. 29. Baur, P. S., Thurman, G. B., and Goldstein, A. L. (1975).J.Immunol. 115, 1375. 1. 2. 3. 4.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
103
30. Becker, E. L. (1976). Am. J. Pathol. 85,385. 31. Becker, E. L., and Showell, H. J. (1972). Z. Zmmunitaetsforsch., E x p . Klin. Z m munol. 143, 466. 32. Becker, K. E., Ishizaka, T., Metzger, H., Ishizaka, K., and Grimley, P. M. (1973).J. E x p . Med. 138, 394. 33. Bell, C., and Wigzell, H. (1977). Eur. J. Zmmunol. 7, 726. 34. Bell, G. I. (1974). Nature (London) 248,430. 35. Bell, G. I., and D e Lisi, C. (1974). Cell. Zmmunol. 10, 415. 36. Ben-Sasson, S. Z., and Rosenthal, A. S. (1975).J. Zmmunol. 115, 1140. 37. Ben-Sasson, S. Z., Lipscomh, M. F., Tucker, T. F., and Uhr, J. W. (1978).J. Zmmunol. 120, 1902. 38. Ben-$haul, Y., and Moscona, A. A. (1975). E x p . Cell Res. 95, 191. 39. Bentwich, Z., Douglas, S. D., Siegal, F. P., and Kunkel, H. G. (1973). Clin. Zmmunol. Immunopathol. 1, 511. 40. Berger, N. A,, and Skinner, A. M. (1974).J. Cell Biol. 61, 45. 41. Berger, N. A., Kociolek, K., and Pitha, J. (1979). Biochem. Ciophys. Res. Commun. 86, 1234. 42. Berke, G., and Gahison, D. (1975).E u r . J . Zmmuno!. 5,671. 43. Berke, G., Tzur, R., and Inbar, M. (1978).J. Zmmunol. 120, 1378. 44. Berlin, R. D., and Oliver, J. M. (1978).J. Cell B i d . 77, 789. 45. Bernard, D. P., and Waksman, B. H. (1975). Ann. Zmmunol. (Paris) 126c, 121. 46. Bernard, D. P., Carboni, J. M., and Waksman, B. H. (1975).Ann. Zmmunol. (Paris) 126c, 107. 47. Betel, I., and Martynse, J. (1976). Nature (London) 261, 318. 48. Bevan, M. J., Epstein, R., and Cohn, M. (1974).J. E x p . Med. 139, 1025. 49. Biberfeld, P. (1971). E x p . Cell Res. 66,433. 50. Biherfeld, P., and Johansson, A. (1975). E x p . Cell Res. 94, 79. 51. Bloch, R., Maccecchini, M. L., Jumhlatt, J., Buttrick, P., and Burger, M. M. (1977). Eur. /. Biochem. 80, 261. 52. Blumenthal, R. (1975). Ann. N.Y. Acad. Sci. 264, 476. 53. Booth, A. G., and Kenny, A. J. (1976). Biochem. J. 159, 395. 54. Bourguignon, L. Y. W., and Singer, S. J. (1977). Proc. Natl. Acad. Sci. U. S . A. 74, 5031. 55. Bourguignon, L. Y. W., Hyman, R., Trowhridge, I., and Singer, S. J . (1978). Proc. N u t / . Acad. Sci. U . S. A. 75,2406. 56. Bourguignon, L. Y. W., Tokuyasu, K. T., and Singer, S. J., unpublished data, quoted in Bourguignon and Singer (54). 57. Braendstrup O., and Werdelin, 0. (1977). Cell. Zmmunol. 32,263. 58. Braendstrup, O., Andersen, V., and Werdelin, 0. (1976). Cell. Zmmunol. 25, 207. 59. Braendstrup, O., Anderson, P., Jensen, H., and Werdelin, 0. (1977). Cell. Zminunol. 32,274. 60. Brain, P., Gordon, J., and Willets, W. A. (1970). Clin.E x p . fmmunot. 6,681. 61. Bretscher, A,, and Weber, K. (1978). E x p . Cell Res. 116, 397. 62. Bretscher, A., and Weher, K. (1978).j. Cell Biol. 79, 839. 6:3. Briggs, R. T., Drath, D. B., Kamovsky, M. L., and Kamovsky, M. J . (1975).J. Cell Biol. 67, 566. 64. Broome, J. D., and Jeng, M. W. (1972).J. Natl. Cancer Znst. 49, 579. 65. Broome, J. D., and Jeng, M. W. (1973).J. E x p . Med. 138,574. 66. Brown, G. J., and Prasad, A. S., eds. (1977). “Zinc Metabolism: Current Aspects in Health and Disease.” Alan R. Liss, Inc. New York. 67. Brown, S., and Revel, J . P. (1976).J. Cell Biol. 68, 629.
104
FRANCIS LOOR
68. Bubbers, J. E., and Henney, C. S. (1975).J. Immunol. 115, 145. 69. Burchill, B. R., Oliver, J. M., Pearson, C. B., Leinbach, E. D., and Berlin, R. D. (1978).J. Cell Biol. 76,439. 70. Burger, M. (1973).Fed. Proc., Fed. Am. SOC. Exp. Biol. 32,91. 71. Butcher, E., Scollay, R., and Weissman, I. (1979).Nature (London) 280, 496. 72. Carlsson, L., Nystrom, L.-E., Sundkvist, I., Markey, F., and Lindberg, U. (1977).J. Mol. B i d . 115, 465. 73. Carson, D. A., and Metzger, H. (1974).J. Immunol. 113, 1271. 74. Celis, J. E., Small, J. V., Andersen, P., and Celis, A. (1978).Exp. CeZl Res. 114,335. 75. Cerottini, J.-C., and Brunner, K. T. (1972).Nature (London),New B i d . 237,272. 76. Cerottini, J.-C., and Brunner, K. T. (1977).In “B and T Cells in Immune Recognition” (F.Loor and G . Roelants, eds.), p. 319. Wiley, New York. 77. Cerottini, J.-C., Engers, H. D., MacDonald, H. R., and Brunner, K. T. (1974).J. Exp. Med. 140, 703. 78. Chang, T. W., Celis, E., Eisen, H. N., and Solomon, F. (1979).Proc. Natl. Acad. Sci. U . S. A . 76,2917. 79. Chaplin, D. D., and Wedner, H. J. (1978).Cell. Immunol. 36, 303. 80. Chauvenet, A. R., and Scott, D. W. (1975).J. Immunol. 114,470. 81. Chen, C., and Hirsch, J. G . (1972).]. E x p . Med. 136,604. 82. Chen, H. W., Heiniger, H. J,, and Kandutsch, A. A. (1975).Proc. Natl. Acad. Sci. U . S . A. 72, 1950. 83. Chen, S. S. (1974).J . Physiol. (London) 238, 313. 84. Chen, S. S. (1979).E x p . Cell Res. 121,283. 85. Chesters, J. K. (1972).Biochem. J. 130, 133. 86. Chvapil, M. (1973).Lije Sci. 13, 1041. 87. Chvapil, M. (1976).Med. Clin. North A m . 60, 799. 88. Chvapil, M., Ryan, J., and Zukoski, C. (1972).Proc. SOC.Exp. B i d . Med. 141,150. 89. Chvapil, M., Ryan, J., Elias, S., and Peng, Y. M. (1973).E x p . Mol. Pathol. 19, 186. 90. Chvapil, M., Peng, Y. M., Aronson, A. L., and Zukoski, C. (1974).]. Nutr. 104,434. 91. Chvapil, M., Stankova, L., Zukoski, C., IV, and Zukoski, C., I11 (1977).J.Lab. Clin. Med. 89, 135. 92. Chvapil, M., Stankova, L., Bemhard, D. S., Weldy, P. L., Carlson, E. C., and Campbell, J. B. (1977).Infect. f m m u n . 16, 367. 93. Chvapil, M., Stankova, L., Bartos, Z., Cox, T., and Nichols, W. (1979).J . Reticuloendothelial SOC. 25,345. 94. Clark, R. A., and KlebanoE, S. J. (1977).J. Immunol. 119, 1413. 95. Click, R. E., Benck, L., and Alter, B. J. (1972).Cell. Immunol. 3, 155. 96. Closs, O., Saltvedt, E., and Olsnes, S. (1975).J. Immunol. 115, 1045. 97. Cockcroft, S., and Gomperts, B. D. (1979).Biochem. J. 178, 681. 98. Coffey, R. G . , Hadden, E. M., and Hadden, J. W. (1977).J. Immunol. 119, 1387. 99. Cohen, H. J., and Chovaniec, M. E. (1978).J. Clin.Inuest. 61, 1088. 100. Cohn, Z. A. (1978).J. Immunol. 121, 813. 101. Cohnen, C., Fischer, K., and Brittinger, C. (1975).Immunology 29,337. 102. Comell, B. A., Sacre, M. M., Peel, W. E . , and Chapman, D. (1978).F E B S Lett. 90, 29. 103. Criswell, B. S., Rich, R. R., Dardano, J., and Kimzey, S. L. (1975).Cell. Irnmunol. 19, 336. 104. Cunningham, B. A., Wang, J. L., Pflumm, M. N., and Edelman, G . M. (1972). Biochemistry 11, 3233.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
105
105. Cumutte, J. T., and Babior, B. M. (1975). B l o o d 45, 851. 106. Daeron, M., and Voisin, G. A. (1978). Cell. Ztnmunol. 37, 467. 107. Daeron, M., Duc, H. T., Kanellopoulos, J., Le Bouteiller, P., Kinsky, P., and Voisin, G. A. (1975). Cell. Zminunol. 20, 133. 108. Daguillard, F., and Richter, M . (1969).J. E x p . Med. 130, 1187. 109. Dahlgren, C., Kihlstrom, E., Magnusson, K. E., Stendahl, O., and Tagesson, C. (1977).E x p . Cell Res. 108, 175. 110. Daniele, R. P., and Holian, S. K. (1976). Proc. Natl. Acud. Sci. U . S. A . 73, 3599. 111. Daniele, R. P., and Holian, S. K. (1976). Cell. Znzrtiuitol. 26, 21. 112. Daniele, R. P., Holian, S. K., and Nowell, P. C. (1978).J. E x p . Med. 147, 571. 113. Darzynkiewicz, Z., and Arnason, B. G. W. (1974). E r p . Cell Re.9. 85, 95. 114. DeChatelet, L. R. (1975).J. Ztifect. Dis. 131, 295. 115. DeChatelet, L. R., Shirley, P. S., and Johiiston, R. B., Jr. (1976). B l o o d 47, 545. 116. De Lisi, C., and Perelson, A. (1976).J. Theor. B i o l . W2, 159. 117. De Lisi, C., and Siraganian, R. P. (1979).J . Zntinuitol. 122, 2286. 118. De Lisi, C., and Siraganian, R. P. (1979).J . Zniniunol. 122, 2293. 119. De Lisi, C., and Thakur, A. K. (1977). Cell. Zirimunol, 28, 416. 120. Dembo, M., and Goldstein, B. (1978).J. Zrnniunol. 121, 345. 121. Dembo, M., Goldstein, B., Sohotka, A. K., and Lichtenstein, L. bl. (1978). J . Z f l l m u n o l . 121, 354. 121. Dernbo, M., Goldstein, B., Sobotka, A. K., and Lichtenstein, L. M. (1979).J . Zrrtrnunol. 122, 518. 123. de Petris, S. (1974). Nuture (London) 250, 54. 124. de Petris, S. (1975).J . Cell B i o l . 65, 123. 125. de Petris, S. (1977). Cell Surf. Reo. 3, 643. Elsevier, Amsterdam. 126. de Petris, S. (1977). B u s e l Z i t s t . Znirnuiiol., Atmu. Rep. p. 57. 127. de Petris, S. (1978).J. Cell B i o l . 79, 235. 128. de Petris, S. (1978). Nottrre (Londori)272, 66. 129. de Petris, S., RaK, M. C., and Mallucci, L. (1973).Nuture (London), N e u B i d . 244, 275. 130. Dewald, B., Baggiolini, bl., Cumette, J. T., and Babior, B. M. (1979). J . Cliit. Zt1t;gst. 63, 21. 131. Dillner, M.-L., Hanimarstrom, S., and Perlmann, P. (1975). E.rp. Cell Res. 96,374. 132. Di Luzio, N. R. (1973). F e d . Proc., Fed. Atn. Soc. E x p . B w l . 32, 1875. 133. Di Pauli, G., and Brdiczka, D. (1974). Biochim. Biophys. A c t u 352, 252. 134. Dixon, J. F. P., O’Brien, R. L., and Parker, J. W. (1975). Exp. Cell Res. 96, 383. 135. Dobretsov, G. E. , Borschevskaya, T. A., Petrov, V. A., and Vladimirov, Y. A. (1977). F E B S Lett. 84, 125. 136. Dukor, P., and Hartmann, K.-U. (1973). Cell. Znirnunol. 7, 349. 137. Eaton, J. W., Berger, E., White, J. G., and Jacob, H. S. (1977).Prog. Cliti. B i d . Res. 14, 275. 138. Eaton, J. W., Berger, E., White, J. G., and Jacob, H. S. (1977).I n “Zinc Metabolism: Current Aspects in Health and Disease” (G. J. Breuer and A. S. Prasad, eds.), p. 275. Alan R. Liss, Inc., New York. 139. Edelman, G. M. (1976). Science 192, 218. 140. Edelman, C . M., Cunningham, B. A,, Reeke, G. N. Jr., Becker, J. W., Waxdal, M. J., and Wang, J. L. (1972).Proc. Natl. Acud. Sci. U . S. A. 69, 2580. 141. Edelman, G. M . , Yahara, I., and Wang, J. L. (1973). Proc. N a t l . Acud. S c i . U . S. A . 70, 1442.
106
FRANCIS LOOR
142. Edelman, G. M., Wang, J. L . , and Yahara, I. (1976).Cold Spring Harbor Con$ Cell Proliferation 3,305. 143. Edelson, P. J. , and Cohn, Z. A. (1974).J. Exp. Med. 140, 1364. 144. Edman, K. A. P., and Mongar, J. L. (1961).J. Physiol. (London) 157,40. 145. Edman, K. A. P., Mongar, J. L., and Schild, H. 0. (1964).J. Physiol. (London) 170, 124. 146. Elson, C. J. , Singh, J., and Taylor, R. B. (1973).Scand. J. Zmmunol. 2, 143. 147. Engers, H. D., MacDonald, H. R., Cerottini, J. C., and Brunner, K. T. (1975).Eur. J . Immunol. 5, 223. 148. Enlander, D., Tobey, R. A., and Scott, T. (1975).Erp. Cell Res. 95, 396. 149. Estensen, R. D., Reusch, M. E., Epstein, M. L., and Hill, H. R. (1976). Infect. Zmmun. 13, 146. 150. Fagraeus, A., and Norberg, R. (1978).Cum. Top. Microbiol. Zmmunol. 82, 1. 151. Fagraeus, A., Lidman, K., and Biberfeld, G. (1974).Nature (London) 252,246. 152. Fagraeus, A., Lidman, K., and Norberg, R. (1975).Clin. E x p . Zntmunol. 20,469. 153. Fagraeus, A., Nilsson, K., Lidman, K., and Norberg, R. (1975).J . Natl. Cancer Inst. 55, 783. 154. Fagraeus, A., Norberg, R.,and Biberfeld, G. (1978).Ann. Immunol. (Paris) 129c, 245. 155. Falcott, R., Oriol, R., and Iscaki, S. (1972).Eur. J. Zmmunol. 2,476. 156. Fanger, M. (1972).Immunol. Commun. 1, 315. 157. Fanger, M. W., Hart, D. A,, Wells, J. V., and Nisonoff, A. (1970).J. Zmmunol. 105, 1043. 158. Fanger, M. V., Hart, D. A., Wells, V. J., and Nisonoff, A. (1970).J. Zmmunol. 105, 1484. 159. Farber, P. A., and Friedman, H. (1979).Experientia 35, 832. 160. Fechheimer, M., and Cebra, J. J. (1979).J . Immunol. 122,2590. 161. Fedorcsak, I., Harms-Ringdahl, M., and Ehrenberg, L. (1977).E x p . Cell Res. 108, 331. 162. Ferguson, R. M., Schmidtke, J. R., and Simmons, R. L. (1975).Nature (London) 256, 744. 163. Ferguson, R. M., Schmidtke, J . R., and Simmons, R. L. (19761.J. Zmmunol. 116, 627. 164. Femandes, G., Nair, M., Onoe, K., Tanaka, T., Floyd, R., and Good, R. A. (1979). Proc. Nat!. Acad. Sci. U . S . A. 76,457. 165. Ferrarini, M., Munro, A., and Wilson, A. B. (1973).Eur. J . Zmmunol. 3, 364. 166. Fisher, D. B., and Mueller, G. C. (1969).Biochim. Biophys. Acta 176,316. 167. Fisher, D. B., and Mueller, G. C. (1969).Nature (London) 221,566. 168. Flanagan, J., and Koch, G. L. E. (1978).Nature (London) 273, 278. 169. Foerster, J., Lamelin, J. P., Green, I., and Benacerraf, B. (19691.J. Exp. Med. 129, 295. 170. Ford, W. L., Sedgley, M., Sparshott, S. M., and Smith, M. E. (1976).Cell Tissue Kinet. 9, 351. 171. Foreman, J. C., and Garland, L. G . (1974).J. Physiol. (London) 239, 381. 172. Fonii, L., and Pemis, B. (1975). In “Membrane Receptors of Lymphocytes” (M. Seligmann, J. L. Preud’homme, and F.M. Kourilski, eds.), p. 193. Am. Elsevier, New York. 173. Fram, R. J., Sidman, C. L., and Unanue, E . R. (1976).J. Zmmunol. 117, 1456. 174. Frankel, F. R., Tucker, R. W., Bruce, J., and Stenberg, R. (1978).J. Cell B i d . 79, 401.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
107
175. Fraser, A. R., Hemperly, J. J., Wang, J. L., and Edelman, G. M. (1976).Proc. Natl. Acad. Sci. U . S . A . 73, 790. 176. Freedman, M. H. (1979).Cell. Immunol. 44,290. 177. Freedman, M. H., Raff, M. C., and Gomperts, B. (1975).Nature (London ) 255,378. 178. Freitas, A. A., and Bognacki, J. (1979).Immunology 36,247. 179. Freitas, A. A., and de Sousa, M. A. B. (1975).Eur. J. Immunol. 5, 831. 180. Freitas, A. A., and de Sousa, M. (1976).Eur. J. Immunol. 6, 703. 181. Freitas, A. A,, Rocha, B., Chiao, J. W., and de Sousa, M. (1978).Cell. Imm*inol. 35, 59. 182. Froland, S. S., and Natvig, J. B. (1970).I n t . Arch. Allergy Appl. Immunol. 39, 121. 183. Frost, A. F., Monahan, T. M., and Abell, C. W. (1978).Immunol. Commun. 7,251. 184. Fujiware, K., and Pollard, T. D. (1978).J. Cell Biol. 77, 182. 185. Gabbiani, G., Chaponnier, C., Zumbe, A., and Vassalli, P. (1977).Nature (London) 269, 697. 186. Gale, €3. P., and Zighelboim, J. (1975).J. Irnmunol. 114, 1047. 187. Gallagher, K., Matarazzo, W. J., and Gray, I. (1979).Clin. Itnmunol. Zmrnunopnthol. 13, 369. 188. Gallin, E. K., and Gallin, J. I. (1977).J. Cell Biol. 75, 277. 189. Gallin, J. I., and Rosenthal, A. S. (1974).J. Cell Biol. 62, 594. 190. Gausset, O., Delespesse, G., Hubert, C., Kennes, B., and Govaerts, A. (1976).J. Immunol. 116,446. 191. Gelfand, E. W., Morris, S . A., and Resch, K. (1975).J. ImmunoE. 114,919. 192. Gelfand, E. W., Dosch, H.-M., Hastings, D., and Shore, A. (1979).Science 203,365. 193. Cell, P. G. H., and Sell, S. (1965).J. E x p . Med. 112, 813. 194. Gemsa, D., Steggemann, L., Till, G., and Resch, K. (1977).J. Immuiiol. 119, 524. 195. Gery, I., and Eidinger, D. (1977).Cell. Immunol. 30, 147. 196. Gilbert, D. (1978).Nature (London)272, 577. 197. Glauert, A. M., and Sanderson, C. J. (1979).J. Cell Sci. 35, 355. 198. Glenney, J. R., Jr., Hixson, D. C., and Walhorg, E. F., Jr. (1979).Exp. Cell Res. 118, 353. 199. Click, M. C., and Warren, L. (1969).Proc. N n t l . Acnd. Sci. U . S. A . 63, 563. 200. Goldman, R. (1976).Exp. Cell Res. 99, 385. 201. Goldman, R., and Raz, A. (1975).E.sp. Cell Res. 96, 393. 202. Goldman, R., Sharon, N., and Lotan, R. (1976).Exp. Cell Res. 99, 408. 203. Goldman, R., Pollard, T., and Rosenbaum, J., eds. (1976). “Cell Motility,” Cold Spring Harbor Lab., Cold Spring Harbor, New York. Book A-C. 204. Goldstein, I. M., Roos, D., Kaplan, H . B., and Weissman, G. (1975).J.Clin. Ziiuest. 56, 1155. 205. Goldstein, I. M., Cerqueira, M., Lind, S., and Kaplan, H. B. (1977).J. Clin. Inuest. 59, 249. 206. Gollapudi, S. V. S., and Kern, M. (1979). Riochem. Biophys. Res. Commurr. 88, 804. 207. Goloschapov, A. N. (1976). Dissertation Chernogolovka, quoted by Dobretsov (135). 208. Golstein, P., and Smith, E. T. (1976).Eur. J. Immunol. 6, 31. 209. Golstein, P., Foa, C., and MacLennan, I. C. M. (1978).Eur. J. Zmmuttol. 8, 302. 210. Goodman, M. G., and Weigle, W. 0. (1977).J. Exp. Med. 145, 473. 211. Goodman, M. G., and Weigle, W. 0. (1978).J. Immunol. 121, 1453. 212. Goodman, M. G., Fidler, J. M., and Weigle, W. 0. (1978).J. Inamunol. 121, 1899. 213. Goodman, M. G., Fidler, J. M., and Weigle, W. 0. (1978).J. Immunol. 121, 1905.
108
FRANClS LOOR
214. Goodwin, J. S., Messner, R. P., and Williams, R. G., Jr. (1979). Cell. Immunol. 45, 303. 215. Gordon, J. A., and Marquardt, M. D. (1974). Biochim. Biophys. A C ~ Q 332, 136. 216. Greaves, M., Janossy, G., Feldman, M., and Doenhoff, M. (1974).In “The Immune System: Genes, Receptors and Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 271. Academic Press, New York. 217. Greaves, M. F. (1970). Transplant. Reu. 5,45. 218. Greaves, M. F., Torrigiani, G., and Roitt, I. M. (1969).Nature (London) 222,885. 219. Green, M. R., Hill, H. A. O., Okolow-Kubkowska, M. J., and Segal, A. W. (1979). FEBS Lett. 100,23. 220. Greene, W. C., Parker, C. M., and Parker, C. W. (1976).]. Immunol. 117, 1015. 221. Grey, H. M. (1969).Adu Zmmunol. 10,51. 222. Griffith, L. M., and Pollard, T. D. (1978).J . Cell Biol. 78, 958. 223. Grinna, L. S . (1977). Gerontology 23,452. 224. Grunspan-Swirsky, A., and Pick, E. (1978). fmmunopharmacology 1, 71. 225. Grunspan-Swirsky, A., and Pick, E. (1979). Cell. Zmmunol. 45, 415. 226. Guilbert, L. J., and Iscove, N. N. (1976).Nature (London) 263,594. 227. Gunther, G. R., Wang, J. L., Yahara, I., Cunningham, B. A., and Edelman, G . M. (1973). Proc. Natl. Acad. Sci. U. S. A. 70, 1012. 228. Gunther, G. R., Wang, J. L., and Edelman, G. M. (1974).]. Cell Biol. 62,366. 229. Gunther, G. R., Wang, J. L., and Edelman, G. M. (1976).E x p . Cell Res. 98, 15. 230. Gunther, T., Averdunk, R.,and Riihl, H. (1979).Z. lmmunitaetsforsch. 155,269. 231. Hadden, J. W.,Coffey, R. G., Hadden, E. M., Lopez-Corrales, E., and Sunshine, G. H. (1975). Cell. Immunol. 20, 98. 232. Hadden, J. W., Hadden, E. M., Sadlik, J. R., and Coffey, R. G. (1976).Proc. Natl. Acad. Sci. U . S . A. 73, 1717. 233. Hammarstrom, L., and Smith, E. (1976). Scand. J. Zmmunol. 5,37. 234. Hammarstrom, L., and Smith, C. I. E. (1978).Cell. Immunol. 36, 377. 235. Hammarstrom, L., and Smith, C. I. E. (1979).Immunology 37,389. 236. Hammarstrom, L., and Smith, C. I. E. (1979).Esp. Cell Res. 119, 343. 237. Hammarstrom, S. (1973). Scand. J. lmmunol. 2,53. 238. Hand, N. L., King, N. L., Johnson, J. J., and Lowe, D. A. (1977).Nature (London) 265,543. 239. Harms-Ringdahl, M., and Jomvall, H. (1974).Eur. J. Biochem. 48,541. 240. Harms-Ringdahl, M., Fedorcsak, I., and Ehrenberg, L. (1973). Proc. Natl. A d . Sci. U. S. A . 70, 569. 241. Harms-Ringdahl, M., Forsberg, J., Fedorcsal, I., and Ehrenberg, L. (1979). Biochem. Biophys. Res. Commun. 86,492. 242. Harold, F. M., Altendorf, K. H., and Hirata, H. (1974).Ann. N . Y. Acad. Sci. 235, 149. 243. Hart, D. A. (1978).E x p . Cell Res. 113, 139. 244. Hart, D. A. (1978).Infect. Immun. 19,457. 245. Hart, D. A. (1979).E x p . Cell Res. 119, 47. 246. Hart, D. A. (1979).E x p . Cell Res. 121,419. 247. Hart, D. A. (1979). Cell. Zmmunol. 43, 113. 248. Hart, D. A., and Streilein, J. S. (1976).E x p . Cell Res. 102, 246. 249. Hart, D. A., and Streilein, J. S. (1976).Exp. Cell. Res. 102, 253. 250. Hartmann, K.-U., and Bokisch, V. A. (1975).]. E x p . Med. 142,600. 251. Hatch, G. E., Nichols, W. K., and Hill, H. R. (1977).J. Zmmunol. 119,450.
PLASMA MEMBRANE AND C E L L CORTEX INTERACTIONS
109
252. Hattori, M., Scribner, D. J., and Moorhead, J. W. (1979). Nature (London) 280, 316. 253. Heber-Katz, E., and Click, R. E. (1972).Cell. Zmmuiiol. 5 , 410. 254. Heggeness, M. H., and Ash, J. F. (1977).J. Cell B i d . 73, 783. 255. Heggeness, M. H., Wang, K., and Singer, S . J. (1977).Proc. Natl. Acad. Sci. U . S. A . 74, 3883. 256. Heiniger, H.-J., Brunner, K. T., and Cerottini, J.-C.(1978).Proc. NatE. Acad. Sci. U. S. A . 75, 5683. 257. Hellstrom, U., Dillner, M.-L., Hammarstrom, S., and Perlmann, P. (1976).Scund.J. Zmmunol. 5,45. 258. Henney, C. S., and Bubbers, J. E. (1973).J. Zmmunol. 111, 85. 259. Henney, C. S., and Lichtenstein, L. M. (1971).J.Immunol. 107, 610. 260. Henson, P. M., and Oades, Z. G . (1975).J. C h i . Znuest. 56, 105.3. 261. Hirschom, R., Corossman, J., Troll, W., and Weissman, G. (1971).J.Clin. Inoest. 50, 1206. 262. Hoffman, R., Ferguson, R., and Simmons, R. L. (1977).J.Immunol. 118, 1472. 263. Hoffstein, S., Soberman, R., Goldstein, I., and Weissnian, G . (1976).J.Cell B i d . 68, 781. 264. Hogberg, B., and Uvnas, B. (1960).Acta Physiol. Scand. 48, 133. 265. Holian, A., Deutsch, C. J., Holian, S. K., Daniele, R. P., and Wilson, D. F. (1979).J. Cell. Pliysiol. 98, 137. 266. Hook, W. A., Schiffmann, E., Aswanikumar, S., and Siraganian, R. P. (1976).J. Immunol. 117, 594. 267. Honvitz, A. F., Hatten, M. B., and Burger, M. M. (1974).Proc. Nutl. Acud. Sci. U . S. A. 71, 3115. 268. Hovi, T., Williams, S. C. , and Allison, A. C. (1975).Nature (London) 256, 70. 269. Hovi, T., Allison, A. C., and Williams, S. C. (1976). E x ~ JCell . Res. 97, 92. 270. Hovi, T., Allison, A. C. , and Williams, S. C. (1979). Riochem E i o p h y s . Res. Conimuti. 88, 1337. 271. Huet, C., Lonchampt, M., Huet, M., and Bernadac, A. (1974). Riochim. Biophys. Acta 365,28. 272. Hume, D. A. , Hansen, K., Weidemann, M . J., and Ferber, E. (1978). Nature (Lotidoti ) 272, 359. 27.3. Hiinig, T. H., Schinipl, A., and Wecker, E. (1977).J.E x p . bled. 145, 1216. 274. Hiinig, T. H., Schimpl, A., and Wecker, E. (1977).J.E x p . Med. 145, 1228. 275. Hynes, R. O., and Destree, A. (1977). Proc. Natl. Acad. Sci. U. S. A . 74, 2855. 276. Igarashi, T., Kada, M. O., Kishimoto, T., and Yamaniura, Y. (1977).J. Zmmunol. 118, 1697. 277. Inbar, M., and Shinitzky, M. (1974).Proc. N u t / . Acad. Sci. U . S. A . 71, 4229. 278. Iscove, N. N., and Melchers, F. (1978).J . E x p . itfed. 147, 923. 279. Ishizaka, K., and Ishizaka, T. (1969).J. Imniutiol. 103, 588. 280. Ishizaka, K., Tomioka, H., and Ishizaka, T. (1970).J . immunol. 105, 1459. 281. Ishizaka, K., Ishizaka, T., and Lee, E. H. (1970). Ini,)iuiiochetnistry 7, 687. 282. Ishizaka, T., and Ishizaka, K. (1975).Prog. Al/ergy 19, 60. 283. Ishizaka, T., and Ishizaka, K. (1975).Afiii. N . Y. A c u ~Sci. . 254, 462. 284. Ishizaka, T., and Ishizaka, K. (1978).J. Zrnmwiol. 120, 800. 285. Izerski, C., Taurog, J. D., Poy, G . , and Metzger, H. (1978).J . Immuiiol. 121, 549. 286. Jacobs, M. (1975).Nature (London ) 258, 659. 287. Jacobs, R. A., and Majerus, P. W. (1973).J. B i d . Chem. 248, 8392.
110
FRANCIS LOOR
288. Jefferson, J. W., and Greist, J . H., eds. (1977). “Primer of Lithium Therapy,” Williams & Wilkins, Baltimore, Maryland. 289. Jensen, P., and Rasmussen, H. (1977). Biochim. Biophys. Acta 468, 146. 290. Jensen, P., Wiri;<er,L., Rasmussen, H., and Nowell, P. (1977). Biochim. Biophys. Acta 496, 374. 291. Jerrells, T. R., and Hinrichs, D. J. (1975).Res,J. Reticuloendothelial. SOC. 17, 84. 292. Johnson, F. M., and Johnson, S., eds. (1977). “Lithium and Medical Practice,” Univ. Park Press, Baltimore, Maryland. 293. Johnston, R. B., Jr., Lehmeyer, J . E., and Guthrie, L. A. (1976).]. E x p . Med. 143, 1551. 294. Jung, C. Y., and Rampal, A. L. (1977).J. Biol. Chem. 252,5456. 295. Kakinuma, K. (1974). Biochim. Biophys. Acta 348, 76. 296. Kalina, M., and Berke, G. (1976). Cell. Zmmunol. 25, 41. 297. Kalina, M., and Ginsburg, H. (1975). Proc. SOC. E x p . Biol. Med. 149, 796. 298. Kalina, M., and Hollander, N. (1975). Imniunology 29, 709. 299. Kapsenberg, M. L., and Leene, W. (1979). Exp. Cell Res. 120,211. 300. Karnovsky, M. L., and Lazdins, J. K. (1978).J. Zmmunol. 121, 809. 301. Karush, F. (1976). Contemp. Top. Mol. Zmmunol. 5,217. 302. Kay, J. E. (1972).E x p . Cell Res. 71, 245. 303. Kay, M. M. B. (1975).Nature (London ) 254,424. 304. Kazimierczak, W., and Diamant, B. (1978).Prog. Allergy 24, 295. 305. Kazimierczak, W., and Maslinski, C. (1974).Agents Actions 4, 1. 306. Kazimierczak, W., and Maslinski, C. (1974).Agents Actions 4, 203. 307. Keller, R., and Sorkin, E. (1970).Experientia 26, 30. 308. Kemp, A. S., and Berke, G. (1973).Eur. J. Zmmunol. 3,674. 309. Kendall, P. A., and Hutchins, D. (1978). Immunology .35, 189. 310. Kennerly, D. A., Sullivan, T. J., and Parker, C. W. (1979).J. Zmmunol. 122, 152. 311. Kersey, J. H., Horn, D. J., and Buttrik, P. (1974).J. lmmunol. 112, 862. 312. Kieda, C., Roche, A.-C., Delmotte, F., and Monsigny, M. (1979). FEBS Lett. 99, 329. 313. Kieda, C. M. T., Bowles, D. J., Ravid, A., and Sharon, N. (1978).F E B S Lett. 94,391. 314. Kigoshi, S. (1979).Experientia 35,836. 315. Kigoshi, S., and Ito, R. (1973).Experientia 29, 1408. 316. Kirchner, H., and Oppenheim, J. J. (1972).Cell. Immunol. 3,695. 317. Kirchner, H., and Riihl, H. (1970).E x p . Cell Res. 61,229. 318. Kishimoto, T., and Ishizaka, K. (1975).J. Zmmunol. 114, 585. 319. Kishimoto,T., Miyake, T., Nishizawa, Y., Watanabe, T., and Yamamura, Y. (1975).J. Zmmunol. 115, 1179. 320. Klebanoff, S. J., Foerder, C. A., Eddy, E. M., and Shapiro, B. M. (1979).J. E x p . Med. 149, 938. 321. Klimetzek, V., and Sorg, C. (1977). Eur. J. lmmunol. 7, 185. 322. Knutton, S. (1976).Exp. Cell Res. 102, 109. 323. Koch, C. L. E., and Smith, M. J. (1978).Nature (London) 273,274. 324. Koch, K., and Nielsen, H. E. (1973). Scand. J. Zmmunol. 2, 1. 325. Koopman, W. J., Gillis, M. H., and David, J. R. (1973).J. lmmunol. 110, 1609. 326. Koren, H. S., Ax, W., and Freund-Moelbert, E. (1973).Eur. J . Zmmunol. 3, 32. 327. Kom, E. D. (1978). Proc. Natl. Acad. Sci. U.S. A . 75, 588. 328. Kosower, E. M., and Kosower, N. S. (1969).Nature (London) 224, 117. 329. Kroneman, J., Van der Mey, G. J. W., and Helder, A. (1975). Zentralbl. Veterinaermed., Reine A 22,201.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
111
330. Ku, D., Akera, T., Tobin, T., and Brody, T. M. (1975).Arch. Pharmocol. 290, 113. 331. Kulczycki, A., Jr., and Parker, C. W.(1979).J . Biol. Chem. 254, 3187. 332. Kulczycki, A., Jr., Hempstead, B. L., Hofmann, S. L., Wood, E W., and Parker, C. W. (1979).J. Biol. Chem. 254, 3194. 333. Kushnir, L. D. (1968).Biochim. Biophys. Actu 150, 285. 334. Larsson, E.-L., and Coutinho, A. (1979).Nature (London) 280, 239. 335. Larsson, H., Wallin, M., and Edstrom, A. (1976).E x p . Cell Res. 100, 104. 336. Lawrence, E. C., Muchmore, A. V., Dooley, N. J., and Blaese, H. M. (1979).Cell. Zmmunol. 46, 100. 337. Lawrence, E. C., Muchmore, A. V., Decker, J. M., Dooley, N. J., and Blaese, H. M. (1979).Cell. Zmmunol. 46, 110. 338. Lawson, D., Fewtrell, C., Gomperts, B., and Raff, M. C. (1975).J.E x p . Med. 142, 391. 339. Lazdins, J. K., Kiihner, A. L., David, J. R., and Kamovsky, M. L. (1978).J . Exp. bled. 148, 746. 340. Le Bouteiller, P., Daeron, M., Duc, N. T., Highenzi, S., and Voisin, G. A. (1976). E u r . J. Zmmunol. 6, 326. 341. Lehmeyer, J. E., Snyderman, H., and Johnston, H. B., Jr. (1979).Blood 54, 35. 342. Lemki, A., and Optiz, H.-G. (1976).J. Zmmunol. 117, 388. 343. Leslie, G. A., and Benedict, A. A. (1970).J. Zmmunol. 104, 810. 344. Leslie, G. A., and Clem, L. W. (197q.J. fmmunol. 108, 1656. 345. Lichtenfeld, J. L., Desner, M., Mardiney, M. H., Jr., and Wiemik, P. (1976).Cancer Treat. Rep. 60, 571. 346. Lichtenstein, L. M. (1975).J. Zmmunol. 114, 1692. 347. Lichtenstein, L. M., Bourne, H. H., Henney, C. S., and Greenough, W. B., 111 (1973).J . Clin. Itruest. 52, 691. 348. Lichtman, M. A,, Jackson, A. H., and Peck, W. A. (1972).J . Cell. Physiol. 80, 383. 349. Liepins, A., Faanes, H. B., Lifter, J., Choi, Y. S., and de Harven, E. (1977).Cell. Zmmunol. 28, 109. 350. Lin, P.-S., and Chen, H. W. (1979).Cell Bwl.I n t . Rep. 3, 51. 351. Lin, P.-S., and Wallach, D. F. H. (1974).Science 184, 1300. 352. Lin, P.-S., Cooper, A. G., and Wortis, H. H. (1973).N. E n g l . J. Med. 289, 548. 353. Lin, P.-S., Wallach, D. F. H., and Tsai, S. (1973).Proc. Natl. Acad. Sci. U . S . A. 70, 2492. 354. Lin, S., and Snyder, C. E., Jr. (1977).J. Bwl. Chem. 252, 5464. 355. Lin, S., Lin, D. C., and Flanagan, M. D. (1978).Proc. Natl. Acad. Sci. U . S. A . 75, 329. 356. Linthicum, D. S., and Sell, S. (1975).J. Ultrastruct. Res. 51, 55. 357. Lipscomb, M. F., Holmes, K. V, Vitetta, E. S., Himmeding, U., and Uhr, J. W. (1975).E u r . J. Immunol. 5,255. 358. Lipscomb, M. F., Ben-Sasson, S. Z., Tucker, T. F., and Uhr, J. W. (1979).E u r . J . Zmmunol. 9, 119. 359. Lipsky, P. E., and Rosenthal, A. S. (1973).J . E x p . Med. 138, 900. 360. Lipsky, P. E., and Hosenthal A. S. (1975).J. Immunol. 115,440. 361. Loor, F. (1973).Exp. Cell Res. 82,415. 362. Loor, F. (1974).Eur. J. Zmmunol. 4, 210. 363. Loor, F. (1977).Prog. Allergy 23, 1. 364. Loor, F. (1977). In “B and T Cells in Immune Recognition’’ (F. Loor and G. Hoelants, eds.), p. 153, Wiley, New York.
112
FRANCIS LOOR
365. Loor, F. (1979).In “Cell Biology and Immunology of Leukocyte Function” (M. R. Quastel, ed.), p. 3. Academic Press, New York. 366. Loor, F. (1980).Eur. J . Zrnmunol. 10, 53. 367. Loor, F. (1980). Cell Surf. Rev. 7, in press. 368. Loor, F., and Higg, L. B. (1975). Eur. J. Zmmunol. 5, 854. 369. Loor, F., and Roelants, G. E. (1974). Eur. J. Immunol. 4, 649. 370. Loor, F., Fomi, L., and Pemis, B. (1972).Eur. J. Zmmunol. 2, 203. 371. Loor, F., Block, N., and Little, J. R. (1975).Cell. Zmtnunol. 17, 351. 372. Loor, F., Martin-Pelissard, C., and Angman, L. (1980). Cell Immunol., in press. 373. Lotan, R., Lis, H., Rosenwasser, A,, Novogrodsky, A., and Sharon, N. (1973). Bwchetn. Biophys. Res. Commun. 55, 1347. 374. Lotem, J., Vlodavsky, I., and Sachs, L. (1976).Exp. Cell Res. 101, 323. 375. Lother, H., Blitstein-Willinger, E., and Diamantstein, T. (1979). Z. Zmmunitaetsfor.sch. 155,346. 376. Luckasen, J. R., White, J. G., and Kersey, J. H. (1974).Proc. Natl. Acad. Sci. U . S. A . 71, 5088. 377. Lyle, L. R., and Parker, C. W. (1974). Biochemistry 13,5415. 378. Maccecchini, M.-L, and Burger, M. M. (1977).Biochim. Biophys. Acta 469,33. 379. MacDonald, H. R., and Cerottini, J.-C. (1979).J.Zmmunol. 122, 1067. 380. MacDonald, H. R., Bonnard, G. D., Sordat, B., and Zawodnik, S. A. (1975).Scand. I. Immunol. 4, 487. 381. MacLennan, I. C. M., and Golstein, P. (1978).J. Zmmunol. 121, 2542. 382. Madden, T. D., Chapman, D., and Quinn, P. J. (1979).Nature (London) 279, 538. 383. Magro, A. M. (1975). Immunochemistry 12, 389. 384. Magro, A. M., and Alexander, A. (1974).J. Immunol. 112, 1757. 385. Magro, A. M., and Alexander, A. (1974).J. Immunol. 112, 1762. 386. Maino, V. C., Green, N. M., and Crumpton, M. J. (1974).Nature (London)251,324. 387. Maino, V. C., Hayman, M. J., and Crumpton, M. J . (1975).Biochem. J. 146, 247. 388. Margolis, L. B., and Bergelson, L. D. (1979).E x p . Cell Res. 119, 145. 389. Maro, B., and Bomens, M. (1979). F E B S Lett. 97, 116. 390. Maro, B., Avrameas, S., and Bomens, M. (1979).Exp. Cell Res. 118, 85. 391. Martin, F. J., and MacDonald, R. C. (1976).J.Cell Biol. 70, 515. 392. Martin, T. W., and Lagunoff, D. (1979). Nature (London) 279,250. 393. Martin, T. W., and Lagunoff, D. (1979). Science 204, 631. 394. Marty, A., and Finkelstein, A. (1975).J. Gen. Physiol. 65, 515. 395. Maschler, R., and Maurer, H. R. (1979).Hoppe-Seyler’s2. Physiol. Chem. 360,735. 396. Matter, A. (1979). Znamunology 36, 179. 397. Mavier, P., and Hanoune, J . (1975).Eur. J. Biochem. 59, 593. 398. McClain, D. A,, and Edelman, G. M. (1976).J. Exp. Med. 144, 1494. 399. McFarland, W. (1969). Science 163, 818. 400. McMahon, L. J., Montgomery, J., Guschewsky, A., Woods, A. H., and Zukosli, C. F. (1976). Zmmunol. Commun. 5, 53. 401. McPhail, L. C., De Chatelet, L. R., and Johnston, R. B., Jr. (1979).j. Clin.Inuest. 63, 648. 402. Medrano, E., Piras, R., and Mordoh, J. (1974).Exp. Cell Res. 86, 295. 403. Melchers, F., von Boehmer, H., and Phillips, R.A. (1975).Transplant. Reu. 25,26. 404. Mellon, M. G., and Rebhun, L. I. (1976).J. Cell Biol. 70, 226. 405. Mendoza, G., and Metzger, H. (1976).Nature (London) 264, 548. 406. Merluzzi, M. J., Badger, A. M., Kaiser, C. W., and Cooperland, S. R. (1975).Clin. Exp. Immunol. 22, 486.
PLASMA MEMBRANE AND C E L L CORTEX INTERACTIONS
113
Mertin, J., and Hughes, D. (1975). Znt. Arch. Allergy A p p l . inamunol. 48, 203. Metcalf, D. (1976).J. Zrnmunol. 116, 635. Metzger, H. (1977). Recept. Recognition, Ser. A 4, 75. Milner, S. M. (1977). Nature (London) 268, 441. Molday, R. S. , Dreyer, W. J., Rembaum, A., and Yen, S. P. S. (1975).J. Cell Biol. 64, 75. 412. Moller, G. (1976). Scand. J . Imniunol. 5, 583. 413. Moller, G. (1978). Scand. J . Inimuriol. 8, 469. 414. Mongar, J. L., and Schild, H. 0. (1962). Physiol. Reo. 42, 226. 415. Moore, P. B., Anderson, D. R., Huggins, J. W., and Carraway, K. L. (1976). Biochein. Bi0phy.P. Res. Conimun. 72, 288. 416. Mooseker, S. (1976).J. Cell B i d . 71, 417. 417. Mooseker, S., and Tilney, L. G. (1975).J. Cell B i d . 67, 725. 418. Mooseker, M. S., Pollard, T. D., and Fujiwara, K. (1978).J. Cell B i d . 79, 444. 419. Moss, J., Fishman, P. H., Manganiello, V. C., Vanghan, M., and Brady, R. 0. (1976). Proc. Nntl. Acrid. Sci. U . S . A . 73, 1034. 420. Mossmann, H., Meyer-Delius, M., Vortisch, U., Kickhofen, B., and Hammer, D. K. (1974).]. E s p . Med. 140, 1468. 421. hlossmann, H., Lauer, W., Vortisch, U.,Giinther, E., Bazin, H., and Hammer, D. K. (1976). Eur. J . Inimunol. 6, 591. 422. Muschel, P. J., Rosen, N., Rosen, 0. M., and Bloom, B. R. (1977).J. Inimunol. 119, 1813. 423. Naccache, P. H., Showell, H. J., Becker, E. L., and Sha'afi, R. I. (1977).J. Cell Biol. 75,635. 424. Nakagawara, A., and Minakami, S. (1979). Biochim. Biophys. Actci 584, 143. 425. Nakagawara, A,, Shibata, Y., Takeshige, K., and Minakami, S. (1976).E . Y Cell ~ Res. 101, 225. 426. Nariuchi, H., and Adler, W. H. (1979). Cell. Inimunol. 45, 295. 427. Nash, A. A., and Ling, N . R. (1976). Nnture (London) 264, 255. 428. Nathan, C., Nogueira, N., Juangbhanich, C., Ellis, J., and Cohn, Z. (1979).J. E x p . ,$led. 149, 1056. 429. Nathan, C. F., Karnovsky, M. L., and David, J. R. (1971).J, E x p . Med. 133, 1356. 430. Negendank, W. G., and Collier, C. R. (1976). E x p . Cell Res. 101, 31. 431. Nelson, R. D., and Leu, R. W. (1975).J. Imniunol. 114, 606. 432. Nelson, S . C., Herman, M. M., Bensch, K. G., Sher, R., and Barchas, J. D. (1976). E x p . Mol. Pnthol. 25, 38. 433. Ng, M. H., Ng, W. S., Ho, W. K. K., Fung, K. P., and Lamelin, J. P. (1978). E x p . Cell Re.7. 116, 387. 434. Ng, W. S., Ng, M. H., and Ho, W. K. K. (1979). E . Y ~Cell . Res. 121, 347. 435. Nickolson, V. J., and Vddstra, H. (1972). FEBS Lett. 23, 309. 436. Nicolson, G . L. (1972). Nature (London),Nerc Biol. 239, 193. 437. Nicolson, G. L. (197.3).Ser. Haenicitol. 6, 275. 438. Nicolson, G. L. (1976). Biochini. B i o p h y . ~Acta . 457, 57. 439. Nicolson, G. L. (1976). Biochim Biophys. Acta 458, 1. 440. Nicolson, G. L., and Poste, G . (1976).J . Supramol. Struct. 5, 65. 441. Nicolson, G. L., and Poste, G . (1979). Biochim. Biophys. Acta 554, 520. 442. Nicolson, G . L., Smith, J. R., and Poste, G. (1976).J. Cell B i d . 68, 395. 443. Nielsen, M. H., Jensen, H., Braendstrup, O., and Werdelin, 0. (1974).J E x p . Med. 140, 1260. 444. Norberg, R., Fagraeus, A., and Lidman, K. (1975). Clin. E x p . Zmmutiol. 21, 284. 407. 408. 409. 410. 411.
114
FRQNCIS LOOR
445. Norberg, R., Lidman, K., and Fagraeus, A. (1975).Cell 6,507. 446. North, B. J. (1978).J . Immunol. 121, 806. 447. Novogrodsky, A. (1976). In “Mitogens in Immunobiology” (J. J. Oppenheim and D. L. Rosenstreich, eds.), p. 43. Raven Press, New York. 448. Novogrodsky, A., and Katchalsky, E. (1972).Proc. Natl. Acad. Sci. U . S. A. 69,3207. 449. Novogrodsky,A., and Katchalsky, E. (1973).Proc. Natl. Acad. Sci. U . S. A . 70,1824. 450. Novogrodsky, A., and Katchalsky, E. (1973).Proc. Natl. Acnd. Sci. U . S . A. 70,2515. 451. Novogrodsky, A,, Lotan, R., Ravid, A., and Sharon, N. (1975).J. Immunol. 115, 1243. 452. O’Flaherty, J. T., Kreutzer, D. L., and Ward, P. A. (1977).J. Immunol. 119,232. 453. Ogmundsddttir, H. M., and Weir, D. M. (1979).Immunology 37, 689. 454. Okudaira, H., Kataoka, T., Okada, H., Furuse-hie, R., Kawachi, S., Nojima, S., and Nishioka, K. (1970).J. Biochem (Tokyo) 68,379. 455. Oldfield, E., and Chapman, D. (1972). FEBS Lett. 23,285. 456. Oliver, J. M., Ukena, T. E., and Berlin, R. D. (1974).Proc. Natl. Acad. Sci. U . S. A. 71, 394. 457. Oliver, J. M., Zurier, R. B., and Berlin, R. D. (1975).Nature (London) 253, 471. 458. Oliver, J. M., Albertini, D. F., and Berlin, R. D. (1976).J. Cell B i d . 71, 921. 459. Oliver, J. M., Lalchandani, R., and Becker, E. L. (1977).J . Reticuloendothel. Soc. 21, 359. 460. Ono, M., and Hozumi, M. (1973). Biocheni. Biophys. Res. Commun. 53, 342. 461. Oppenheim, J. J., and Rosenstreich, D. L. (1976).Prog. Allergy 20, 65. 462. Oppenheim, J. J., Rogentine, G. N., and Terry, W. D. (1969).Immunology 16,123. 463. Orange, R. P. (1974).Prog. Immunol., Proc. Int. Congr., and, 1974 Vol. I I 4, p. 29. 464. Orange, R. P. (1975).J. Immunol. 114, 182. 465. Orange, R. P. (1978).In “Immediate Hypersensitivity, Modern Concepts and Developments” (M. K. Bach, ed.), Immunol. Ser. Val. 7, p. 589. Dekker, New York. 466. Orange, R. P., and Chang, P. L. (1975).J. Imtnunol. 115, 1072. 467. Orange, R. P., and Moore, E. G. (1976).]. Immunol. 116, 392. 468. Ostlund, R. E., and Pastan, I. (1976). Biochim. Biophys. Acta 453, 37. 469. Ottemess, I. G., Torchia, A. J., and Bliven, M. L. (1979).Cell. Immunol. 43, 62. 470. Papahadjopoulos, D., Poste, G., and Mayhew, E. (1974). Biochim. Biophys. Acta 863, 404. 471. Papahadjopoulos, D., Jacobson, K., Poste, G., and Sheperd, G. (1975). Biochim. Biophys. Actn 394, 504. 472. Pardee, A. B., Dubrow, R., Hamlin, J. L., and KIetzien, R. F. (1978).Annu. Reo. Riochem. 47, 715. 473. Parker, C. W. (1974). Biochem. Biophys. Res. Commun. 61, 1180. 474. Parker, D. C. (1975).Nature (London) 258, 361. 475. Parker, D. G., Fothergill, J. J.. and Wadsworth, D. C. (1979).j. lmmunol. 123,931. 476. Parkhouse, R. M., and Allison, A. C. (1972).Nature (London)New Biol. 235,220. 477. Patriarca, P., Cramer, R., Marussi, M., Moncalvo, S., and Rossi, F. (1971). J. Reticuloendothel. Soc. 10, 251. 478. Paty, D. W., Cousin, H. K., and McDonald, L. E. (1975). Lancet 1, 1197. 479. Perera, B. A. V., and Mongar, J. L. (1965). lmmunology 8, 519. 480. Perez-Cruet, J., and Dancey, J. T. (1977).Experientia 33, 646. 481. Perks, B., Flanagan, M. T., Auger, J., and Crumpton, M. J. (1977). Eur. J . Immunol. 7, 613. 482. Peterkofsky, B., and Prather, W. (1977).]. Cell. Phy.yiol. 90,61. 483. Pfeiffer, D. R., and Lardy, H. A. (1976). Biochemistry 15, 935. 484. Phillips, J. L., and Azari, P. (1974). Cell. Immunol. 10, 31.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
485. 486. 487. 488. 489. 490. 491. 492. 493. 494. 495. 496. 497. 498. 499. 500. 501. 502. 503. 504. 505’. 506. 507. 508. 509. 510.
511. 512. 513.
514. 515. 516. 517. 518. 519.
115
Pick, E (1972). Nature (London),New B i d . 238, 176. Pick, E. (1974). Zmmutiology 26, 649. Pick, E., and Abrahamer, H. (1973). I r i t . Arch. Allergy Appf. lmnunol. 44,215. Pick, E., and Grunspan-Swirsky, A. (1977).Cell. Zmutzol. 32, 340. Pick, E., and Manheimer, S. (1974). Cell. Itnmuiiol. 11, 30. Pick, E., and Wilner, I. (1979). Exp. Cell Res. 118, 151. Pick, E., Honig, S., Griffel, B., and Grunspan-Swirsky, A. (1976). Cell. Inzmutiol. 27, 339. Plaut, M., Lichtenstein, L. M., and Henney, C. S. (1973).J. Zmmunol. 110, 771. Plaut, M., Bubbers, J. E., and Henney, C. S. (1976).J. Zmnuriol. 116, 150. Polliack, A., and de Harven, E. (1975). Cliti. Zmmunol. Zmmunopathol. 3,412. Polliack, A,, Fu, S. M., Douglas, S . D., Bentwich, Z., Lampen, N., and d e Harven, E. (1974).1.E x p . Med. 140, 146. Porter, K., Prescott, D., and Frye, J. (1973).]. Cell Bwl. 57, 815. Poste, G., and Nicolson, G. L. (1976). Biochim. Bi0phy.p. Acta 426, 148. Poste, G., Papahadjopoulos, D., Jacobson, K., and Vail, W. J. (1975). Nuture (London ) 253, 552. Poste, G., Papahadjopoulos, D., and Nicolson, G. L. (1975). Proc. N u t / . Acud. Sci. U . S. A. 72,4430. Prasad, A. S., ed. (1976). “Trace Elements in Human Health and Disease.” Academic Press, New York. Pratt, H. P. M., Fitzgerald, P. A., and Saxon, A. (1977). Cell. Zmniunol. 32, 160. Pressman, B. C. (1973).Fed. Proc., Fed. Am. Soc. E x p . B i d . 32, 1698. Prnjansky, A,, Ravid, A., and Sharon, N. (1978). Biochim. Biophys. Act0 508, 137. Prujansky, A., Ravid, A., Lis, H., and Sharon, N. (1977).Quoted in Prujansky et (11. (503, p. 144). Pugh-Humphreys, R. G. P., and Sinclair, W. J. (1970).J Cell Sci. 6, 477. Quastel, M. R., and Kaplan, J. G. (1968). Notrire (London) 219, 198. Quastel, M. R., and Kaplan, J . G. (1970). Exp. Cell Res. 62, 407. Quastel, M. R., and Kaplan, J. G. (1970). E x p . Cell Res. 63,230. Quastel, M. R., and Kaplan, J. R. (1975). E x p . Cell Res. 94,351. Raff, M. C., and de Petris, S. (1974). I n “The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 247. Academic Press, New York. Raff, M. C., and Lawson, D. (1978). In “Immune System: Genetics and Regulation” (E. E. Sercarz, I. A. Herzenberg, and C. F. Fox, eds.), ICN-UCLA Symp. Mol. Cell. Biol. Vol. 6, p. 669. Academic Press, New York. Raff, M. C., Freedman, M., and Gomperts, B. (1975). In “Membrane Receptors of Lymphocytes” (M. Seligmann, J.-L. Preud’homme, and F. M. Kourilsky, eds.), p. 393. Am. Elsevier, New York. Raff, M. C., Owen, J. J. T., Cooper, M. D., Lawton, A. R., 111, Megson, M., and Gathings, W. E. (1975).J.Exp. Med. 142, 1052. Rao, K. M. K., Schwartz, S. A., and Good, R. A. (1979). Cell. Ininzutiol. 42,270. Ravid, A., Novogrodsky, A., and Wilchek, M. (1978). Eur. J . Znlmuiiol. 8, 289. Reed, P. W., and Lardy, H. A. (1972).]. Biol. Chem. 247, 6970. Rega, A. F., Rothstein, A., and Weed, R. I. (1967).]. Cell. Physiol. 70, 45. Reiss, U., Tappel, A. L., and Chio, K. S. (1972). Bic~chem.Biophys. Res. Commun. 48, 921. Resch, K., Gelfand, E. W., Hansen, K., and Ferber, E. (1972). Eur. J , Ztnmutzol. 2, 598.
116
FRANCIS LOOR
520. Resch, K., Prester, M., Ferber, E., and Gelfand, E. W. (1976).J. Zmmunol. 117, 1705. 521. Resch, K., Bouillon, D., Gemsa, D., and Averdunk, R. (1977). Nature (London) 265, 349. 522. Resch, K., Bouillon, D., and Gemsa, D. (1978).J. Zmmurwl. 120, 1514. 523. Reyes, F., Lejonc, J. L., Gourdin, M. F., Mannoni, P., and Dreyfus, B. (1975).J.Exp. Med. 141, 392. 524. Rich, R. R., and Pierce, C. W. (1974).]. Zmmunol. 112, 1360. 525. Riethmiiller, G., Riethmiiller, D., Stein, M., and Hausen, P. (1968).J. Zmmunol. 100,969. 526. Rigas, D. A., Eginitis-Rigas, C., and Head, C. (1979). Biochem. Biophys. Res. Commun. 88,373. 527. Rijke, E. O., and Ballieux, R. E. (1979). Cell Tissue Kinet. 12, 435. 528. Ringden, O., and Moller, W. (1975). Scand. J. Zmmunol. 4, 171. 529. Riordan, J. F., and Vallee, B. L. (1972). In “Methods in Enzymology” (C. H. W. Hirs, ed.), Vol. 25, Part B, p. 449. Academic Press, New York. 530. Rivnay, B., Globerson, A,, and Shinitzky, M. (1978). Eur. J. Zmmunol. 8, 185. 531. Roath, S., Newell, D., Polliack, A., Alexander, E., and Lin, P.-S. (1978).Nature (London) 273, 15. 532. Roder, J., and Duwe, A. (1979).Nature (London)278,451. 533. Roelants, G. E. (1977). In “B and T Cells in Immune Recognition” (F. Loor and G. E. Roelants, eds.), p. 103. Wiley, New York. 534. Romeo, D., and Rossi, F. (1973).Acta Vitaminol. Enzymol. 27, 145. 535. Romeo, D., Zabucchi, G., and Rossi, F. (1973). Nature (London),New B i d . 243, 111. 536. Romeo, D., Zabucchi, G., Miani, N., and Rossi, F. (1975). Nature (London) 253, 542. 537. Romeo, D., Zabucchi, G., and Rossi, F. (1977). In “Movement, Metabolism and Bactericidal Mechanisms of Phagocytes” (F. Rossi, P. Patriarca, and D. Romeo, eds.), p. 153. Piccin Medical Books, Padova. 538. Romeo, D., Zabucchi, G., Berton, G., and Schneider, C. (1978)J. Membr. Biol. 44, 321. 539. Roos, D., Hamon-Muller, J. W. T., and Weening, R. S. (1976). Biochem. Biophys. Res. Commun. 68,43. 540. Root, R. K., and Metcalf, J. (1976). Clin. Res. 24, 318. 541. Root, R. K., and Metcalf, J. A. (1977).J. Clin. Znuest. 60, 1266. 542. Rosenstreich, D. L., and Blumenthal, R. (1977).]. Zmmunol. 118, 129. 543. Rosenstreich, D. L., and Mizel, S. B. (1979).J. Zmmunol. 123, 1749. 544. Rosenthal, A. S., Blake, J. T., and Lipsky, P. E. (1975).]. Zmmunol. 115, 1135. 545. Rottmann, W. L., Walther, B. T., Hellerqvist, C. G., Umbreit, J., and Roseman, S. (1974).J. Biol. Chem. 249,373. 546. Riihl, H., Kirchner, H., and Bochert, G. (1971). Proc. SOC. E x p . Biol. Med. 137, 1089. 547. Riihl, H., Scholle, H., and Kirchner, H. (1971).Acta Haematol. 46, 326. 548. Rutishauser, U., Yahara, I., and Edelman, G. M. (1974). Proc. Natl. Acad. Sci. U . S . A. 71, 1149. 549. Ryan, G. B., Unanue, E. R., and Kamofsky, M. J. (1974).Nature (London)250,56. 550. Ryan, J. L., and Shinitzky, M. (1979).Eur. J . Zmmunol. 9, 171. 551. Ryser, J.-E., Sordat, B., Cerottini, J.-C., and Brunner, K. T. (1977).Eur.1. Zmmunol. 7, 110.
PLASMA MEMBRANE AND CELL CORTEX INTERACTIONS
117
552. Ryzewski, J., Roszkowski-Sliz, W., and Kryzystyniak, R. (1976).Immunology 31, 145. 553. Sahyoun, N., Hollenberg, M. D., Bennett, V., and Cuatrecasas, P. (1977). Proc. Natl. Acud. Sci. U . S . A. 74,2860. 554. Saito, M., Aoyagi, T., Umezawa, H., and Nagai, Y. (1977).Biochem. Biophys. Res. Commun. 76,526. 555. Sanderson, C. J. (1976).Proc. R. Soc. London, Ser. B 192, 241. 556. Sanderson, C. J., and Glauert, A. M. (1977).Proc. R. SOC. London, Ser. B 198,315. 557. Sanderson, C. J., and Glauert, A. M. (1979).Immunology 36, 119. 558. Santalo, R. G. (1975).Exp. Celt Res. 96, 429. 559. Schechter, B., Lis, H., Lotan, R., Novogrodsky, A., and Sharon, N. (1976).Eur. J. Immunol. 6, 145. 560. Schell-Frederick, E. (1974).F E E S Lett. 48, 37. 561. Schiffmann, E., Corcoran, B. A., and Wahl, S. M. (1975).Proc. Natl. Acad. Sci. U.S.A. 72, 1059. 562. Schlessinger, J., Webb, W. W., Elson, E. L., and Metzger, H. (1976).Nature (London) 264,550. 563. Schneider, J. A., Bradley, K. H., and Seegmiller, J . E. (1968).Pediatr. Res. 2,441. 564. Schnyder, J., and Baggiolini, M. (1978).]. E x p . Med. 148, 1449. 565. Schopf, E., Schulz, K. H., and Gromm, H. (1967).Naturwissenschaften 54, 568. 566. Schrader, J. W. (1975).J. Immunol. 115,323. 567. Schreier, M. H., and Nordin, A. A. (1977).In “B and T Cells in Immune Recognition” (F. Loor and G. E . Roelants, eds.), p. 127. Wiley, New York. 568. Schreiner, G. F., and Unanue, E. R. (1976).Clin. Immunol. Immunoputhol. 6,264. 569. Schreiner, G. F., and Unanue, E. R. (1976)../.Exp. Med. 143, 15. 570. Schreiner, G. F., and Unanue, E. R. (1976).Adu. Immunol. 24,37. 571. Schreiner, G. F., Braun, J., and Unanue, E. R. (1976).J. E x p . Med. 144, 1683. 572. Schreiner, G. F., Fujiwara, K., Pollard, T. D., and Unanue, E. R. (1977).j . Exp. Med. 145, 1393. 573. Scribner, D. J., Weiner, H. L., and Moorhead, J. W. ( 1 9 7 7 ) Immunol. ~ 119,2084. 574. Scribner, D. J., Weiner, H. L., and Moorhead, J. W. (1978).J. Zmmunol. 121,377. 575. Sedgley, M., and Ford, W. L. (1976).Cell Tissue Kinet. 9, 231. 576. Seeman, P. (1972).Phannacol. Rev. 24,583. 577. Segal, A. W., and Meshularn, T. (1979).FEES Lett. 100, 27. 578. Segal, D. M., Taurog, J. D., and Metzger, H. (1977).Proc. Natl. Acad. Sci. U . S. A. 74, 2993. 579. Segel, G. B., and Lichtman, M. A. (1976).]. Clin. Inuest. 58, 1358. 580. Sela, B. A,, Wang, J. L., and Edelman, G. M. (1975).Proc. Natl. Acad. Sci. U . S . A . 72, 1127. 581. Sela, B. A., Wang, J. L., and Edelman, G.M. (1976).J. E x p . Med. 143,665. 582. Sell, S. (1967).J. Immunol. 98, 786. 583. Sell, S. (1967).J. E x p . Med. 125, 289. 584. Sell, S. (1970).Trunsplant. Rev. 5, 19. 585. Sell, S., and Gell, P. C . H. (1965)../.E x p . Med. 122, 423. 586. Sell, S., Mascari, R. A., and Hughes, S. J. (1970).J . Immunol. 105, 1400. 587. Senda, N., Tamura, H., Shibata, N., Yoshitake, J., Kondo, K., and Tanaka, K. (1975). E x p . Cell Res. 91, 393. 588. Seyberth, H. W., Schmidt-Gayk, J., Jacobs, K. H., and Hackenthal, E. (1973).]. Cell Biol. 57, 567. 589. Sharon, N., and Lis, H. (1972).Science 177, 949.
118
FRANCIS LOOR
590. Shearer, W. T., and Parker, C. W. (1975).J. Zmmunol. 115, 613. 591. Shearer, W. T., Atkinson, J. P., Frank, M. M., and Parker, C. W. (1975).J.E x p . Med. 141, 736. 592. Sheetz, M. P., Painter, R. G., and Singer, S. J. (197q.J. Cell B i d . 70, 193. 593. Shenker, B. J., Matarazzo, W. J., Hirsch, R. L., and Gray, I. (1977).Cell. Immunol. 34, 19. 594. Sherline, P., and Mundy, G. R. (1977).J. Cell Biol. 74, 371. 595. Sherline, P., and Schiavone, K. (1977).Science 198, 1038. 596. Sheterline, P. (1978).E x p . Cell Res. 115, 460. 597. Shimizu, T., and Kimura, I. (1974).J. Biochem. (Tokyo)76, 1001. 598. Shinitzky, M., and Inbar, M. (1974).J.Mol. Bwl. 85,603. 599. Shortman, K., and Golstein, P. (1979).J. Zmmunol. 123, 833. 600. Showell, H. J., Freer, R. J., Zigmond, S. H., Schiffmann, E., Aswanikumar, S., Corcoran, B., and Becker, E. L. (1976).J. E x p . Med. 143, 1154. 601. Sidman, C. L., and Unanue, E. R. (1975).Nature (London)257, 149. 602. Sidman, C. L., and Unanue, E. R. (1976).]. E x p . Med. 144,882. 603. Sidman, C. L., and Unanue, E. R. (1978).Proc. Natt. Acad. Sci. U.S . A. 75,2401. 604. Sidman, C. L., and Unanue, E. R. (1979).J. Zrnrnunol. 122,406. 605. Sieckmann, D. G., Asofsky, R., Mosier, D. E., Zitron, I. M., and Paul, W. E. (1978). 1.E x p . Med. 147, 814. 606. Sieckmann, D. G., Scher, I., Asofsky, R., Mosier, D. E., and Paul, W. E. (19781.J. E x p . Med. 148, 1628. 607. Simchowitz, L., and Spilberg, I. (1979).Immunology 37, 301. 608. Singer, S. J. (1974).Annu. Reu. Biochem. 43,805. 609. Siraganian, R. P., and Hazard, K. A. (1979).J. Immunol. 122, 1719. 610. Siraganian, R. P., and Hook, W. A. (1977).J. Zmmunol. 119,2078. 611. Siraganian, R. P., and Siraganian, P. A. (1975).J. Irnmunol. 114, 886. 612. Siraganian, R. P., Hook, W. A., and Levine, B. B. (1975).Zmmunochemistry12,149. 613. Skamene, E., and Ivanyi, J. (1969).Nature (London)221,681. 614. Smith, M. E., Sparshott, S. M., and Ford, W. L. (1977).E x p . Cell Biol. 45, 9. 615. Smith, S. B., and Revel, J.-P. (1972).Deu. B i d . 27, 434. 616. Smolen, J. E., and Shohet, S. B. (1974).J. Lipid Res. 15,273. 617. Soni, S. L., Kalnins, V. I., and Haggis, G. H. (1975).Nature (London) 255, 717. 618. Stamper, H. B., Jr., and Woodruff, J. J. (1977).J. Immunol. 119, 772. 619. Steen, H. B., and Lindmo, T. (1978).Eur. J. Zmmutiol. 8,667. 620. Steen, H. B., and Lindmo, T. (1979).E u r . J. Zmmunol. 9, 434. 621. Stendahl, O., and Tagesson, C. (1977).E x p . Cell Res. 108, 167. 622. Stossel, T. P. (1973).J. Cell Biol. 58, 346. 623. Stossel, T. P., and Pollard, T. D. (1973).J. Bwl. Chern. 248, 8288. 624. Stossel, T. P., Mason, R. J., and Smith, A. L. (1974)J. Clin. Znoest. 54, 638. 625. Strom, T. B., Garovoy, M. R., Carpenter, C. B., and-Merrill, J. P. (1973).Science 181. 171. 626. Strom, T. B., Carpenter, C. B., Cragoe, E. J., Norris, S., Devlin, R., and Perper, R. J. (1977).Transplant. Proc. 9, 1075. 627. Sullivan, A. L., Grimley, P. M., and Metzger, H. (1971).]. E x p . Med. 134, 1403. 628. Tappel, A. L. (1973).Fed. Proc., Fed. Am. Soc. E x p . Biol. 32, 1870. 629. Taylor, A. C. (1966).J . Cell Biol. 28, 155. 630. Taylor, R. B., Duffus, P. H., Raff, M. C., and de Petris, S. (1971).Nature (London), New Biol. 233, 225. 631. Theoharides, T. C., and Douglas, W. W. (1978).Science 201, 1143.
PLASMA MEMBRANE AND C E L L CORTEX INTERACTIONS
632. 633. 634. 635.
119
Thorn, R. M., and Henney, C. S. (1976). Nature (London) 262, 75. Thyberg, J., Moskalewski, S., and Friberg, U. (1977).J. Cell Sci. 27, 183. Toh, B. H., and Hard, G. C. (1977). Nature (London) 269, 695. Touraine, J.-L., Hadden, J. W., Touraine, F., Hadden, E. M., Estensen, R., and Good, R. A. (1977). J . E x p . Med. 145,460. 636. Toyoshima, S., and Osawa, T. (1975).J. Biol. Cherrr. 250, 1655. 637. Toyoshima, S., and Osawa, T. (1976). Exp. Cell Re.v. 102, 438. 638. Toyoshima, S., Iwata, M., and Osawa, P. (1976). Nuture (London) 264,447. 639. Tsokos, G. C., and Papamichail, M . G. (1979). E x p . Cell B i d . 47, 183. 640. Turnell, R. W., and Burton, A. P. (1974). Cnncer Res. 34, 39. 641. Turnell, R. W., Clarke, L., and Burton, A. F. (1973). Cuncer Res. %3, 203. 642. Ukena, T. E., and Berlin, R. D. (1972).J. E x p . Med. 136, 1. 643. Ukena, T. E., and Kamovsky, M. (1977). Err,. Cell Res. 106, 309. 644. Unanue, E. R. (1972). Ado. Zmrnunol. 15, 95. 645. Unanue, E. R. (1974).A m . J . Puthol. 77,2. 646. Unanue, E. R., and Kaniovsky, M. J. (1974).J. Exp. Med. 140, 1207. 647. Van Ewijk, W., Brons, N. H. C, and Rozing, J. (1975). Cell. Znmutiol. 19, 245. 648. Vassalli, J. D., and Reich, E. (1977).J. E x p . Med. 145, 429. 649. Vassalli, J. D., Hamilton, J., and Reich, E. (1976). Cell 8, 271. 650. Vassalli, J. D., Hamilton, J., and Reich, E. (1977). Cell 11, 695. 651. Vischer, T. L. (1974).J. Inirnunol. 113, 58. 652. Vischer, T. L., and Jaquet, C. (1972). Znmaunology 22, 259. 653. Vischer, T. L., and LoSpalluto, J. J . (1975). Iniinunology 29, 275. 654. Vitetta, E. S., and Uhr, J. W. (1975). Biochitn. Biophys. Acta 415, 253. 655. Wachi, K. K., Kimura, L. H., Perreira, S., Hokama, Y., Perri, S., and Palumbo, N. (1974). Res. Conimun. Chent. Puthol. Phur?nucol. 8, 681. 656. Walsh, F. S., and Crunipton, M. J. (1977). Nature (London)269, 307. 657. Walther, B. T. (1976). 1tr “Coucanavalin A as a Tool” (H. Bittiger and H. P. Schnebli, eds.), p. 231. Wiley, New York. 658. Wands, J. R., Podolsky, D. K., and Isselhacher, K . J. (1976). Proc. N u t / . Accrd. Sci. U . S. A. 73, 2118. 659. Waiig, J. L., Gunther, G. R., and Edelnian, G. M. (1975).]. Cell B i d . 66, 128. 660. Watanabe, T., Kishimoto, T., Miyake, T., Nishizawa, Y., Inoue, H., Takeda, Y., and Yamamura, Y. (1975).J. Zmrnunol. 115, 1185. 661. Wedner, H. J., and Parker, C. W. (1976). Prog. Allergy 20, 195. 662. Wedner, H. J., Chaplin, D. D., and Parker, C. W. (1976). Cell. Initnuno1. 27, 338. 663. Weeds, A. (1978). Nature (London) 274, 417. 664. Weiner, H. L., Moorhead, J. W., Yamaga, K., and Kubo, R. T. (1976).J. Ztnrtiunol. 117, 1527. 665. Weiner, H. L., h4oorliead, J. W., and Claman, H. N. (1976). J . ~nrrtzroio!.116, 1656. 666. Weiner, H. L., Scribner, D. J., and Moorhead, J. W. (1977). Cell. Z t n n i u t d . 31, 77. 667. Weiner, H. L., Scribner, D. J., arid Moorhead, J. W. (1978).J. In~rnutiol.120, 1907. 668. Weiner, H. L., Scribner, D. J., Schocket, A. L., and Moorhead, J. W. (1978). C l i i i . Ztnrnutiol. Zna7ritr1io~~citho/. 9, 356. 669. Wekerle, H., Lonai, P., and Feldman, M. (1972).Proc. Nutl. Acud. Sci. U . S. A . 69, 1620. 670. Weller, N. K. (1974).1. Cell. B i d . 63,699. 671. Werdelin, O., Braendstrup. O., and Pedersen, E. (1974). J . E x p . Med. 140, 1245. 672. Weyman, C., Morgan, S. J., Belin, J., and Smith, A. D. (1977). Biochim. Biophys. Acta 496, 155.
120
FRANCIS LOOR
Whitney, R. B., iind Sutherland, R. M. (1972).J. Cell. Phy.siol. 80,329. Whitney, R. R., and Sutherland, R. M . (1972). Cell. ~ l l Z ? J f U l L O / .5, 137. Wilkinson, P. G. (1975). Eap. Cell Res. 93, 420. Williams, L. T., Snyderman, R., Pike, M. C . ,and Lefkowitz, R. J. (1977).Proc. N n t l . Acrid. S c i . U . S. A. 74, 1204. 677. Willimis, R. O., and Loeb, L. A. (1973).J . Cell Biol. 58, 594. 678. Willinghain, hl. C., and Pastan, I. (1975).Proc. N u t / . Acrid. Sci. U . S. A. 72, 126.3. 679. Wisiiieski, R. J., Williams, R. E., and Fox, C. F. (197.3). Proc. N u t / . Acrid. Sci. U . S. A . 70, 3669. 680. Wofsy, L., Baker, P. C., Thompson, K., Goodman, J., Kiniura, J., and Henry, C. (1974).J. E s p . hletf. 140, 523. 681. Wolberg, G., Zimmennan, T. P., Hienistra, K., Winston, M., and Chu, L.-C. (1975). Sciertce 187, 957. 682. Woociruff, J. J., and Gesner, B. hl. (1968). Science 161, 176. 683. Woodruff, J. J., and Gesner, €3. M. (1969).J. E.xp. Aled. 129, 551. 684. Woodruff, J. J., Kntz, I. M . , Lucas, L. E., and Stamper, H. B., Jr. (1977).J. I ~ ~ n n t i ~ t o / . 119, 1603. 685. Woodmff, ht. F. A., Reid, B., uid James, K. (1967). Nntrire ( L o d o n )215, S9l. 686. Woocls, W. A , , Siege], hi. J., and Chirigos, M. A. (1974). Cell. Ztnmumd. 14, 327. 687. Woods, W. A,, Fliegelman, bl. J., and Chirigos, hl. A. (1975). Proc. Soc. E s ~ JR. i d . Jled. 148, 1048. 688. Wright, P., and Demarchi, M . (1979). Cell. Z m n i i d n o l . 45, 318. 689. Wrowinann, K., Weideinann, M.J., Peskar, B. A,, Staudiiiger, H., Rietschel, E. T., iind Fischer, H , (1978). Errr. J. Z i ~ t i i i ~ ~ i ~8,r i749. l. 690. Wynne, D., Wilchek, M., and Novogrodsky, A. (1976). Biochern. Biophys. Res. ~ f > J l i l J l I L l l 68, . 730. (591. Yahara, I., arid Edelnian, G. M. (1973). E.xp. Cell Res. 81, 143. 692. Yahara, I., and Edelman, G. M. (1975). E x p . Cell Res. 91, 125. 69.3. Yahara, I., and Kakirnoto-Sanieshinia, F. (1977). Proc. Nritl. Acod. Sci. U . S. A. 74, 451 1. 694. Yahara, I., arid Kakinioto-Sameshirna, F. (1978). Cell 15, 251. 695. Yahara, I., and Kakirnoto-Sameshima, F. (1979). E x p . Cell Res. 119, 237. 696. Yamada, K. M., Ohanian, S. H., and Pastan, I. (1976).Cell 9, 241. 697. Yasaka, T., and Kambara, T. (1979). Biochini. B i < i p h ! / ,Actci ~ . 585, 229. 698. Yoshinaga, M., Waksman, B. H., and Malawista, S. E. (1972). Transplant. Proc. 4! ( 3 ) ,325. 699. Yoshinaga, M., Waksnian, B. H., and Malawista, S. E. (1972). Science 176, 1147. 700. Yoshinaga, M., Yoshinaga, A., and Waksman, B. H. (1972). Proc. Nrztl. Acrld. Sci. U . S. A. 69, 3251. 701. Zatti, M., and Rossi, F. (1967). Biochini. Biophys. Actrz 148, 553. 702. Zighelboim, J., Gale, R. P., and Kedar, E. (1976). Trcinsplontcltioti 21, 524. 703. Zigmond, S. H., and Hirsch, J. G. (1972). EX^. Cell Res. 73, 383. 704. Zimmerman, T. P., Wolberg, G . , and Duncan, G. S. (1978). Proc. N a t l . Acrid. Sci. U . S. A. 75, 6220.
673. 674. 675. 876.
ADVANCES I N IMSiIlNOLOC:Y, VOL.
‘31)
Control of Experimental Contact Sensitivity’ HENRY N. CLAMAN, STEPHEN D. MILLER, PAUL J. CONLON, AND JOHN W. MOORHEAd Deportmenfs of Medicine ond of Microbiology ond Immunology, University of Colorodo Medico/ School, Denver, Colorodo
I. Introduction ... 11. Contact Sensi I1 a ................. .4.Chemical Requirements for Induction of Contact Sensitivity B. The Afferent Limb of Contact Sensitization . . . . . . . . . . C. The Efferent Limb of Contact Sensitization: The Elieitation Reaction.. .. D. Genetic Control of Contact Sensitivity . . . E. Antibody and Contact Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Control of Contact Sensitivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Influence of the Major Histocompatibility Complex (MHC) . B. Anti-Idiotype Antibodies Regulating the Duration of Contact Sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Tolerance to Contact Sensitivity ........ A. General Features . . . . . . . . . B. Tolerance in the Absence of Any Demonstrable Suppressor T Cells (“Clone Inhibition”) C. Suppressive Mechani D. Suppressor Factors . . E. Summary . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
122
125 128 128 130 1.30 131 133
15.3
I. Introduction
Historically, the concept that eczematous, inflammatory reactions in the skin to various chemicals was due to an increased sensitivity to the substances rather than to their toxic properties was initiated b y the studies of Joseph Jadassohn (1895). Since then, contact hypersensitivity has been extensively studied in man and experimental animals and has been clearly established as a fonn of antigen-specific delayed hypersensitivity. Once considered to be a relatively simple model of delayed hypersensitivity, studies in the last 10 years have shown contact sensitivity to be an extremely complex phenomenon. Furthermore, the increasing interest in the endogenous regulation of immune processes has prompted detailed studies of the ways in which contact sensitivity is controlled. This paper will review the work that has been done in our laboraSupported in part by NIH grants AI-12685, AI-12993, and AI-14913. Recipient of NIH Research Career Development Award K04-AI-00125. 121 Copyright @ 1980 b y .4cc.~demic Press, Inc. All right5 of reprodxloction in m y torn1 reserved. ISBN 0-12-022430-5
122
HENRY N. CLAMAN ET AL.
tory, and others, on contact hypersensitivity to dinitrofluorobenzene (DNFB) and related chemicals using experimental animals. We will be primarily concerned with regulatory and suppressive mechanisms that have been shown to control the induction, magnitude, expression, or duration of the contact allergic reaction. Owing to space limitations, we have not discussed studies of contact allergy in man. This material may be found in several articles by Levis and his colleagues (Miller and Levis, 1973; Levis et a/., 1975, 1978). II. Contact Sensitivity: Induction a n d Elicitation
REQUIREMENTSFOR INDUCTION OF CONTACT SENSITIVITY We reviewed this subject previously (Claman and Moorhead, 1976), and the basic information remains as follows. The classic studies of Eisen et ul. (1952)showed that the ability of a contactant to sensitize or elicit sensitivity when applied epicutuneously is proportional to its ability to couple covalently to protein. Highly reactive compounds such as DNFB are good couplers and good sensitizers; moderately reactive compounds such as DNBSOBare fair couplers and weak sensitizers; nonreactive congeners such as dinitrophenyl (DNP)-lysine do not couple and do not sensitize to DNFB. We also pointed out that this relation between sensitization and the ability to couple covalently is perturbed when the allergen is not painted on the skin but rather is injected, particularly with Freund’s adjuvant. Godfrey and Baer (1971) showed that adjuvant can convert a poor coupler into a good sensitizer. Furthermore, incorporation into complete Freund’s adjuvant can convert a tolerogen to a sensitizer (Baer et al., 1970). These experiments indicate that injection of allergen with adjuvant changes antigen “processing” in the host. The correlation between the ability of a contact allergen to couple to protein and its ability to sensitize when painted on the skin suggests that the “real” allergen is fomied i n vivo and is “hapten-coupled-toself.” There has been no agreement about the nature of the self carrier. Some data (reviewed by d e Weck, 1977) indicate that haptens coupled to skin proteins can sensitize (when injected with Freund‘s adjuvant). This may indicate that the carrier is skin protein(s). However, other experiments showed that DNP coupled to autologous lymphoid cells would sensitize when injected intraperitoneally (without adjuvant); DNP coupled to erythrocytes (RBC), serum proteins, or killed lymphocytes did not sensitize (Baurngarten and Geczy, 1970). Similar reA.
CHEMICAL
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
123
sults have been seen in contact sensitivity to beryllium (Jones and Amos, 1975). We believe that the dilemma concerning the nature of the self carrier for contact allergens will be resolved when we consider the role of Langerhans cells in the sensitization process.
B. THE AFFERENT LIMBOF CONTACTSENSITIZATION It is convenient to separate the afferent and efferent limbs of the contact sensitization reaction. The afferent limb includes the events following antigen presentation to immunologocally naive animals, and we believe that this is best analyzed when antigen is given epicutaneously. The afferent limb is complete when the animal is sensitized and capable of giving a positive elicitation reaction. In this sense, the afferent limb is activated over a period of 3-7 days. The efferent limb is activated during the elicitation (challenge) phase, and this requires 24-72 hours in most cases. In terms of the current models for delayed hypersensitivity, the afferent limb comprises three links: the forniation of hapten-self carrier; the recognition of this antigen by clonally restricted set(s) of T lymphocytes, which then proliferate; and the dissemination throughout the lymphoid system of these “sensitized” or immune T cells (TDH for T cells mediating delayed hypersensitivity). 1. The Langerhans cell and the formation ofhapten-selfcarrier. The current opinion among cellular immunologists is that, for activation, T lymphocytes best recognize antigen on the membranes of presenting cells that carry components of the major histocompatibility complex (MHC) (Doherty et al., 1976a). Efficient antigen-presenting cells appear to be macrophages and for the T cells of delayed hypersensitivity, the I region (immune response) region of the MHC is the critical associated membrane component (Shevach and Rosenthal, 1973). The “rediscovery” of the Langerhans cell in the skin and elsewhere has come as a breakthrough in our understanding of contact allergy (reviewed by Silberberg-Sinakin et al., 1978). After languishing in “the literature” for almost a century, the Langerhans cell has emerged as a major link in the afferent limb of the contact sensitivity response. Langerhans cells have the following important characteristics: ( a )they are dendritic cells that are particularly plentiful in the epidermis but are also found in gastrointestinal mucous membranes, the dermis, lymph node, tonsils, and thymus; ( b) their identification requires either electron microscopy or special stains for the light microscope, such as ATPase; (c) they are phagocytic and can take up clinically important contact allergens (Shelley and Juhlin, 1976); ( d ) after elicitation of contact sensitivity by skin-painting of sensitized animals,
124
HENRY N. CLAMAN ET AL.
Langerhans cells are frequently seen in close opposition to lymphocytes at the challenge site. These features strongly suggested to Silberberget al. (1974)that, in fact, Langerhans cells (which hitherto had no definitely recognized function) were antigen-presenting cells similar to the macrophages of the lymphoid tissues. The appropriateness of this conclusion was reinforced by the subsequent finding that Langerhans cells carried Ia molecules (Klareskog et d.,1977). Therefore, to put the previous facts together, the best current formulation is that in contact sensitivity the afferent limb begins when epicutaneously applied allergen is associated with Langerhans cells in the epidermis (and perhaps the dermis). It is the complex of hapten coupled to an Ia-positive Langerhans cell membrane that is recognized by the appropriate T lymphocytes. 2. Antigen recognition by T lymphocytes: The question of “peripheral sensitization.” The next link in the afferent limb is the meeting of T lymphocytes with the hapten-Langerhans cell complex. There has long been a controversy over whether the antigen moves from the skin to the draining lymph node, where it meets the T cell (afferent sensitization), or whether T cells from the circulation meet the antigen in the skin and then travel to the draining lymph node (peripheral sensitization). The fact that contact sensitization is accompanied by enlargement and increased cell proliferation in the draining nodes (Turk and Stone, 1963)does not settle the issue. Such a finding is consistent with afferent sensitization, but it is also consistent with peripheral sensitization if T lymphocytes complete their clonal expansion in the node after they arrive from the skin. A number of interesting experiments have been done to answer the question: Does sensitization of lymphocytes in contact allergy begin peripherally in the skin or more proximally in the draining lymph node? Frey and Wenk (1957) showed that guinea pigs could not be contact-sensitized on alymphatic skin islands. This experiment established that the lymphatic pathways are essential, but it did not answer the question. By cutting off ears at various times from guinea pigs injected with trinitrochlorobenzene (TNCB) or dinitrochlorobenzene (DNCB), Macher and Chase (1969) came to the conclusion that the sensitizing antigen was that part which remained in the skin and so supported the peripheral route of sensitization. Friedlander and Baer (1972), in contrast to Frey and Wenk, were able to sensitize guinea pigs via alymphatic skin islands. They believed that although sensitization occurred in lymph nodes, the allergen could arrive there via the bloodstream. [This opinion is at variance with a number of experiments showing that the intravenous route favors tolerance rather than sensitization (see Section IV).] Perhaps the most important experiments
CONTROL OF EXPERIMENTAL
CONTACT SENSITIVITY
125
were done by McFarlin and Balfour (1973) in the pig. This animal is a good choice for such studies because it easily develops contact sensitivity to DNFB and because it is large enough so that the lymphatics draining the site of contact sensitization (usually a limb) can be isolated and cannulated. Lymph can be removed for analysis and/or injection into lymphatics of a naive recipient to test for sensitization potential of the draining lymph. The results of these and other studies (Balfour et al., 1974) showed that lymph draining the skin that has been painted with DNFB contained several sensitizing agents including DNFB itself, DNP-conjugated proteins, and DNP-labeled cells. It seems entirely likely, therefore, that contact sensitization may begin in more than one site. After epicutaneous painting of allergen, some will bind to Langerhans cells in the skin. The Langerhans cell will then “present” it to wandering T cells, and the latter will travel via lymphatics to the draining node (Silberberg-Sinakin et al., 1978). Another likely pathway is that Langerhans cells will bind allergen in the skin and the complex will then move via lymphatics to draining nodes, where they will then encounter T cells. A third pathway is for allergen-protein conjugates or free allergen (e.g., DNFB) to move from the skin to the regional nodes and there bind to Langerhans cells or macrophages prior to activation of T cells. None of these pathways excludes either of the other two. 3. Nature and dissemination of T cells in contact sensitivity. The essential experiment that establishes the nature of a “classic” delayed hypersensitivity state is that sensitivity can be transferred to naive recipients with T lymphocytes, not with serum. Such experiments have indicated that the TDHwhich are generated by the end of the afferent limb of sensitization carry the Lyt-1 lymphocyte marker (Vadas et al., 1976). These cells almost certainly derive directly from the draining lymph node, as splenectomy does not affect contact sensitization in the mouse (Sy et al., 1977b). The effect of adult thymectomy prior to sensitization is not clear. Erard et a2. (1979)found that thymectomy several weeks before sensitization led to a diminution of contact sensitivity, whereas our laboratory did not (H. N. Claman, unpublished). In any case, the fact that animals contact-sensitized on one area of the skin are able to show an elicitation reaction on remote parts of the skin just a few days later indicates that TDH may be raised locally but are soon available systemically (Asherson and Zembala, 1973).
C. THEEFFERENTLIMBOF CONTACTSENSITIZATION: THE ELICITATIONREACTION There is general agreement that in contact sensitivity the elicitation reaction is carried out by skin painting (almost by definition). In a
126
HENRY N. CLAMAN ET AL.
“
pure” delayed hypersensitivity reaction (without significant IgE or Arthus reactivity) the tested area shows few if any changes for several hours, has a peak response around 24 hours, and then is fading by 48 hours. We h d it extremely intriguing that there are significant differences in the anatomic location of optimal sites for elicitation. In the guinea pig, virtually any area of the skin may be used, but the flank is preferred. In the mouse, by contrast, while it is possible to elicit contact reactions on trunk areas (Crowle and Crowle, 1961) one gets much larger reactions if the eliciting dose is applied to the ears (Asherson and Ptak, 1968).The reasons for this are not clear. They may be related to the provocative (and perhaps related) findings that the ears have very high concentrations of histamine (Riley and West, 1956), serotonin (Gershon et al., 1975),and Langerhans cells (Silberberg-Sinakin et
al., 1978). 1 . Measurement of the Elicitation Reaction This measurement has undergone modifications recently. The classic method has been to assess semiquantitatively the degree of erythema and induration in guinea pig skin following the patch test. In fowl, the traditional method for assessing delayed hypersensitivity is measurement of thickening of the injected wattle. Asherson and Ptak (1968) introduced the method of contact elicitation in the mouse by painting contactant on the ear and measuring ear swelling with a micrometer. More recently several other quantitative assays were introduced for delayed hypersensitivity. In a comparison of four of these, Robinson and Naysmith (1976) used mice immunized with protein antigens in Freund’s complete adjuvant. When antigen was injected for challenge, they found that ear swelling was more sensitive and reliable than ( a )footpad swelling; ( h )arrival of SICr-labeledsyngeneic cells; or (c)the accumulation of l T U d R at the challenge site in the ear. In contrast, Sabbadini et al. (1974) found that, in contact sensitivity, measurement of arrival of 51Cr-labeled cells at the ear challenge site was more sensitive than the ear swelling test.
2 . Elicitation Reaction in Vivo and in Vitro De Weck has recently summarized this subject (1977); a detailed exposition of this topic is beyond the scope of this review. The reaction appears to start in the dermis (Kerl et al., 1974), but the epidermis is also involved (Silberberg et al., 1974). There follows an accumulation of mononuclear phagocytic cells (of hematogenous origin) and lymphocytes, but basophils also occur (Dvorak et al., 1970). The cellular accumulation is caused by the release of lymphokines at the challenge
CONTROL OF EXPEFUMENTAL CONTACT SENSITIVITY
127
site. These substances attract and activate circulating mononuclear cells, resulting in the cellular infiltrate mentioned above. The best studied lymphokine is migration inhibition factor (MIF), whose measurement and characteristics have been well described (David and David, 1972). The production of this substance is usually assayed in uitro by reexposing lymphoid cells of sensitized subjects to the sensitizing antigen (or a conjugate) and measuring the inhibition of migration of macrophages. Recent data indicate that such lymphokines are probably active in uiuo. Geczy et a / . (1975)prepared an antiserum against partially purified lymphokines, and this could inhibit delayed-type hypersensitivity (DTH) tests i t i uiuo. Such studies have not been done in contact allergy, but there is no reason to suspect that the mechanism is significantly different. Animals sensitized for contact allergy have cells that can proliferate specifically in uitro upon reexposure to the contactant or a congener (Geczy and Baumgarten, 1970; Barnes, 1973; Phanuphak et al., 1974a). As such a proliferative response is seen with cells of sensitized subjects but not with cells of normal or tolerant subjects (Phanuphaket u1., 1974a,b), it is reasonable to consider the proliferative response as an in uitro correlate of the contact-sensitized state. This is not strictly true, however, as Moorhead (1978) has recently shown that there are two separable populations of antigen-reactive T cells in the nodes of DNFB-sensitized mice. One population can passively transfer contact sensitivity, is nylon wool nonadherent, Ia negative, and does not proliferate i n uitro upon reexposure to antigen. The other T cell population proliferates in uitro, but cannot passively transfer contact sensitivity; it is nylon wool adherent and Ia positive. Similar results have been obtained recently by Shirley and Little (1979).Thus, while contactsensitized animals have cells that can passively transfer contact allergy and can respond by proliferation to antigen in uitro, these functions appear to be carried out by separate T cell populations.
3 . Role of Basophils, Must Cells, und Vusoactiue Amities in Contact Allergy Recent work indicates that basophils and mast cells are important components of “delayed-in-time” hypersensitivity reactions. Several of these have been grouped together under the term “cutaneous basophil hypersensitivity” (CBH). The subject has been reviewed by Askenase (1977) and is beyond the scope of this paper. Nevertheless, it should be noted that reagents that inhibit vasoactive anlines (e.g., reserpine used as a serotonin antagonist in the mouse) inhibit the elicitation of the contact reaction to DNFB (Gershon et UZ., 1975). The
128
HENRY N. CLAMAN ET AL.
precise role of such vasoactive amines is not yet clear, but it is thought that sensitized T cells, upon reaction with antigen, might cause their release. The subsequent local vasodilation would facilitate entry of lymphoid and moiionuclear phagocytic cells into the site of antigen challenge. D. GENETICCONTROLOF CONTACT SENSITIVITY It is now clearly established that immune responses are controlled b y genes, most of which lie in the MHC. Most of these studies have been done in situations in which the antigenic stimulus has been limited either by a low dose, a restricted complexity of antigenic epitopes, or very specific “readouts.” The data with regard to genetic control of contact allergy are as yet incomplete. Early work with guinea pigs showed that inbred strains 2 and 13 differed in their responses to injected TNCB but the relative responsiveness of a given strain depended strongly upon the method used to induce the sensitized state (Chase, 1967; Polak and Turk, 1969).Similar results were reported b y Polak et 01. (1974). Early studies in the mouse (now the favorite species for genetic analysis of immune responses) indicate that while MHC regions may be important, other regions may also be involved (Schultz and Bailey, 1975; Fachet and Ando, 1977). Both of these studies used powerful sensitizers (TNCB and oxazolone), and the role of antibody was not investigated b y transfer experiments.
E. ANTIBODY AND CONTACT SENSITIVITY The relation of antibody to contact sensitivity has been a vexed topic for some time. We believe that sufficient data have been gathered to shed light on the problem. We will approach the situation by asking three related questions: Can antibody be formed during contact sensitization? Can contact sensitization occur in the absence of antibody? Are there contact reactions that require antibody? 1. Can antibody be formed during contact sensitization? Many studies have shown that circulating antibodies are produced during developnieiit of contact allergy. As contact sensitization is often induced by injection of sensitizer in Freund’s adjuvants (which are powerful inducers of antibody formation), this result is hardly surprising. Nevertheless, antibody can also be produced by epicutaneous sensitization aloiie (Chase, 1967; Taylor and Iverson, 1971). Careful studies of this phenomenon by Takahaski et c i l . (1977) showed that a single painting of mice with DNFB induced contact allergy but no increase in anti-
CONTROL O F EXPERIMENTAL CONTACT SENSITIVITY
1%
DNP plaque-forming cells in the lymph nodes. More intense epicutaneous sensitization (either multiple paintings closely spaced or two paintings 10 days apart) did induce considerable antibody formation. Epicutaneous sensitization (particularly at high antigen doses) induces hapten-specific antibodies that can transfer passive cutaneous anaphylaxis (PCA) and other antibodies that cause hemagglutination of hapten-coated cells; the latter contain both 2-mercaptoethanol sensitive and resistant agglutinins (Chase, 1967; Thomas et iiZ., 1978). Thus antibodies of the IgM, IgG, and reaginic (IgE-like) classes may appear during contact sensitization. 2 . Can contact sensitization occur in the absence of antibody? From the philosophical standpoint, this question is difficult to answer definitively because any experiment purporting to show contact sensitivity zi)itlzout antibody is open to the objections that there wus antibody present but it was not detected because (a ) the assay was not sensitive enough or ( b ) antibody was not sought in the right place-e.g., it is cell-bound and not circulating. Leaving this problem aside, there are a number of observations that strongly indicate (but do not prove) that contact sensitivity reactions do not require antibody. The classical studies of Landsteiner and Chase (Chase, 1967) showed that contact sensitivity induced by injection of sensitizer can be passively transferred with cells but not with serum. More recent studies of epicutaneously sensitized mice have shown that lymphoid cells are efficient in the passive transfer of contact allergy and that contact sensitivity is T cell dependent (Parrott et al., 1970). Strictly speaking, this shows only that T celIs are required for contact allergy; it is possible that they are needed as helpers for B cells in the inoculum. This is unlikely, as contact sensitivity is transferred with purified T cells (with 1% B cell contamination) but not with B-enriched cell populatioiis with 84% B cells (Moorhead, 1978). This seems to provide unequivocal evidence that contact sensitivity can exist in the absence of antibody. 3. Are there some contact reactions that require antibody? Again, this brings us to the area of CBH. There is now no doubt that contact sensitization, usually induced with antigens injected with FCA, but also with powerful epicutaneous sensitizers such as oxazolone, can be transferred with serum (Askenase, 1977). These important studies indicate that not all “delayed-in-time” cutaneous reactions are solely T cell functions. To summarize the role of antibody in contact allergy, the following statements can be made.
130
HENRY N. CLAMAN ET AL.
1. “Pure” contact sensitivity may be expressed by T cells in the absence of antibody. 2. Many contact systems involve the coexistence of both immune T cells and antibody, which may be responsible for different manifestations, e.g., classic delayed contact responses and anaphylactic reactions. 3. Some contact reactions may be expressed via antibody alone (although T cells are doubtless needed for the production of that antibody). 111. Control of Contact Sensitivity
A. INFLUENCE OF THE MAJORHISTOCOMPATIBILITY COMPLEX (MHC) As mentioned earlier, immune responses involving T cells are controlled by genes that map in the MHC (Paul and Benacerraf, 1977; J. F. A. P. Miller, 1979). Activation of most T cells involves recognition of antigen in association with certain gene products of the MHC. This so-called “H-2 restriction” has been well documented for T cells involved in helper functions (Erb and Feldmann, 1975; Katz and Benacerraf, 1975; Sprent, 1978b), in vitro proliferation in response to allogeneic cells (Alter et at., 1973), or antigen-pulsed macrophages (Paul et al., 1977a), cytotoxicity against virus-infected (Zinkemagel and Doherty, 1975) or hapten-modified (Shearer et al., 1975) target cells, and delayed hypersensitivity to protein (Miller et al., 1975) or viral antigens (Zinkernagel, 1976). In general, I region gene products, in association with antigen, control the activation of helper T cells and T cells mediating delayed hypersensitivity, whereas gene products of the K and D regions control effector functions of cytotoxic T cells. Few studies have been done exploring the MHC control of contact sensitivity. Vadas et at. (1977) using the in vivo *251UdRassay, have reported the successful transfer of contact sensitivity to DNFB between donors and recipients that share either H-2K, H-2D, or Z region genes. Recent studies carried out in our laboratories, both in vivo and in uitro, have failed to confirm these findings (J. W. Moorhead, unpublished). Using ear swelling as a measure of contact sensitivity, our studies have shown that successful transfer of contact sensitivity to DNFB occurs only when the donor and recipient mice share Z region genes [the precise subregion(s) involved in this transfer have not yet been identified]. N o transfer of immunity is seen with donor-recipient combinations sharing only H-2K and/or H-2D regions. Similar results
CONTROL OF EXPERIMENTAL
CONTACT SENSITIVITY
131
have been obtained in uitro using production of M I F as an indicator of T cell activation. Production of M I F occurs only when purified DNFB-immune lymph node (LN) T cells are cocultured with Z region-compatible, DNP-labeled splenic macrophages. The precise explanation for the difference between our results and those of Vadas is not clear. However, it should be pointed out that the Iz5IUdR assay is simply a measure of cell arrival at the skin test site rather than a measure of effector cell functions. Thus, it is possible that, in addition to delayed hypersensitivity T cells, this assay is also detecting the arrival of cytotoxic T cells that see antigen associated with H-2K or H-2D molecules. Cytotoxic T cells are induced by painting contact sensitizers on the skin of mice (Tagart et al., 1978).In contrast, ear swelling (edema, induration, etc.) and in uitro production of lymphokines such as MIF are dependent on effector cell functions, and these procedures may be detecting only those T cells mediating the delayed reaction. Nevertheless, both sets of data clearly show that T cells that mediate contact sensitivity reactions are controlled by genes within the MHC. B. ANTI-IDIOTYPE ANTIBODIES REGULATING THE DURATIONOF CONTACTSENSITIVITY Primary delayed hypersensitivity (contact sensitivity) is generally thought to be long lived (Crowle, 1975).In the guinea pig, the primary sensitized state may last for months and transferred immunity has been shown to last for at least a year (Polak, 1977). The situation in the mouse, however, appears to be quite different. In mice, optimal sensitization to DNFB results in peak responsiveness 4-6 days after sensitization as measured by ( u ) increase in ear swelling; ( h )ability of the lymph node cells to transfer immunity; and ( c )i n vitro proliferation of antigen stimulated T cells (Phanuphak et al., 1974a,b; Sy, 1979). By days 9-12, significant immunity is no longer detectable (Sy, 1979).This rapid decline in immunity has been shown to be partly due to the activation of suppressor cells. Zembala et al. (1976) found that, in inice contact-sensitized with DNFB or TNCB, lymph node cells taken at 4 days transferred sensitivity wheaeas LN cells taken at 7-9 days transferred suppression. The suppressor cells were identified as B cells. We have confirmed these findings. Furthermore, we have found that pretreating mice with cyclophosphamide prolongs contact sensitivity measured both iti vivo and i i i vitro (Sy et ul., (1977a; Sy, 1979). Since cyclophosphaniide is known to eliminate B cells and precursors or inducers of suppressor T cells, we concluded that the transient nature of the contact sensitivity response in mice is due
132
HENRY N. CLAMAN E T AL.
mainly to the activation of a regulatory process and thus represents a form of active suppression. Two experimental findings have suggested that this regulation may be due to anti-DNP antibodies. First, B cells taken from mice primed with DNP-keyhole limpet hemocyanin (KLH) in adjuvant are able to transfer suppression of contact sensitivity (Sy, 1979). Second, contact sensitivity is significantly prolonged in mice tolerized with the B cellspecific tolerogen, DNP-isologous IgG2,. In spite of these findings however, we have not been able to detect either anti-DNP serum antibody or anti-DNP plaque-forming cells in the LN of DNFBsensitized mice. Nevertheless, serum taken from animals 9-15 days after primary sensitization with DNFB blocks the ability of DNFB immune LN cells to passively transfer immunity (Sy et id.,197%). This suppressive immune serum (SIS) is antigen specific but lacks strain specificity in inhibiting passive transfer. The suppressive materials in the serum are immunoglobulins that do not have anti-DNP activity and are not associated with DNP as antigen-antibody complexes. Rather, these suppressive antibodies have affinity for purified mouse anti-DNP antibodies, indicating that they are anti-idiotypic in nature. We believe these anti-idiotypic antibodies to be induced by the expanded clone of DNFB-specific TDHcells and negatively regulate the duration of the contact sensitivity response by interacting with the idiotype of the T cell receptor. Several years ago, Jerne (1974) proposed a network theory for regulation of the immune response. Based on the earlier observations of Oudin and Michel (1963, 1969) that antibody combining sites (idiotypes) are themselves antigenic and can induce the formation of specific antibodies (anti-idiotypes) in the same species, Jerne suggested that the production of anti-idiotype antibodies (against either serum antibodies or membrane antigen receptors on T cells) was a normal event that regulated the intensity and duration of an immune response. Our finding that there are anti-idiotype antibodies in the serum of animals sensitized with DNFB and that these antibodies down-regulate the immune T cells is in agreement with, and provides rather direct evidence for, the existence of such an immunological network. Since Jerne’s proposal, a large number of experiments have been done to determine whether exogenous administration of antiidiotypic antibodies can modulate immune responses either in a positive (helper) or negative (suppressor) way. Space limitations here do not permit a detailed discussion of the various experimental findings, and the reader is directed to several recent comprehensive reviews of the subject (Binz and Wigzell, 1977; Rajewsky and Eichmann, 1977;
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
133
Eichmann, 1978). Suffice it to say that it is now well established that anti-idiotypic antibodies, when administered exogenously, can suppress or stimulate B and/or T lymphocytes that bear the relevant idiotypes. It has also been demonstrated that the antibody response to phosphorylcholine in BALB/c mice (a response restricted to a single clone expressing the TEPC 15 idiotype) is accompanied by the coproduction of anti-idiotypic antibodies (Cosenza, 1976). Thus, idiotype-bearing and complementary anti-idiotypic reactive lymphocytes do indeed coexist in the immune system, and both sets are activated when the system is perturbed by the relevant antigen. IV. Tolerance to Contact Sensitivity
A. GENERALFEATURES Experimental tolerance (unresponsiveness) dates back to work in delayed hypersensitivity. More than 50 years ago, Sulzberger (1929) showed that if guinea pigs were treated via the intracardiac route with the drug neoarsphenamine (NEO), it was then more difficult to sensitize them to NEO. Chase (1946) began many important studies on tolerance by showing that prolonged feeding of DNCB to guinea pigs induced a state of antigen-specific Clnresponsiveness. De Weck et ul. (1964) induced tolerance to contact sensitization by prior injection of various tolerogens. Lowney (1967) and Macher and Chase (1969) investigated the mutual interrelations between sensitivity and tolerance in contact allergy. When Asherson and Ptak (1968) extended quantitative measurements of contact sensitivity to the mouse, it was easier to explore the model further. Another advance was the recognition that unresponsiveness in the contact sensitivity may be mediated by active suppression (Zembala and Asherson, 1973). We have reviewed this topic, both in general (Claman and Moorhead, 1976)and with regard to work in our own laboratory (Claman et al., 1980). This section will concern the induction of tolerance prior to and during sensitization. The development of negative immunoregulation of the contact response itself has been discussed in Section III,B, on the role of antireceptor (idiotype) antibody. Deserisitizatiorz following the establishment of the contact allergic state will not be discussed here (see Phanuphak et al., 1975; Polak and Rinck, 1978). A number of generalizatioiis can b e made about the induction of specific tolerance to contact sensitivity.
1. Nature of the tolerogen. The ability of various contact allergen congeners to induce tolerance is roughly proportional to their chemical
134
HENRY N. CLAMAN ET AL.
reactivity; i.e., analogs of DNFB that are highly reactive in forming covalent bonds (e.g., DNFB itself) are more potent than weakly reactive molecules (e.g., DNBSOJ, and nonreactive molecules (e.g., DNP-lysine or DNP-mouse proteins) are not tolerogenic (de Weck and Frey, 1966; Claman, 1976; Claman et al., 1977a). This implies that the effective tolerogen formed in vivo is the contactant coupled to some self-component. Battisto and Bloom (1966) showed that antigens coupled to autologous cells were powerful tolerogens for delayed hypersensitivity reactions, and we have confirmed this (Miller and Claman, 1976). These facts suggest that unresponsiveness in contact sensitivity is best induced by allergen-autologous membrane conjugates. This is compatible with current views that T cells recognize antigens on membranes (particularly in the context of MHC components). 2. Route of the tolerogen. The route of presentation of contact allergen is crucial in determining whether the outcome will be sensitization or unresponsiveness. This is clearly shown with DNFB, where epicutaneous presentation leads to sensitivity and intravenous induction leads to tolerance (Claman, 1976).This implies that “host processing” of antigen within various microenvironments of the recipient will determine the immunologic result. However, even epicutaneous presentation can lead to unresponsiveness if supraoptimal doses of the antigen are used. 3. Pathways of tolerance. The demonstrations of acquired specific immunologic tolerance prior to the 1950s were quite intriguing but lacked a firni basis for mechanistic interpretations. With the development of the clonal selection theory and the explosion of cellular immunology, such data were usually believed to mean that tolerance occurred when clonally reactive immunocompetent cells (T or B or both) were removed in some way after interaction with tolerogen. This concept has been applied with success to B cells in terms of “clonal abortion” in the neonatal period (Nossal and Pike, 1978). The discovery of active suppressor cell mechanisms (Gershon, 1974), however, has led to a variety of experiments that have unequivocally shown that specific unresponsiveness may be caused by suppressor T cells. In T cell-mediated reactions, including contact sensitivity, this approach has been so successful that other mechanisms of tolerance have lacked attention. The following sections will explore the topic of specific unresponsiveness in contact allergy. Evidence will b e reviewed to indicate that at least two pathways of tolerance can exist simultaneously. One pathway involves suppressor T cells. The other is indicative of inacti-
CONTROL OF EXPERIMENTAL
CONTACT SENSITIVITY
135
vation of TDHcells without demonstrable T suppressor cells; we have called the latter phenomenon “clone inhibition.” B. TOLERANCE IN THE ABSENCE O F ANY DEMONSTRABLE T CELLS(“CLONEINHIBITION”) SUPPRESSOR Intravenous injection of free hapten or i n vitro hapten-modified syiigeneic cells results in the production of tolerance to a normally immunogenic regimen (Chase, 1946; Battisto and Bloom, 1966; Asherson and Zembala, 1974; Miller and Claman, 1976; Sherr et al., 1979). This unresponsive state has been reported to be due to at least two mechanisms-the generation of suppressor T cells (Ts) (see Section IV,C) and the production of clone inhibition (Claman et al., 1976). A similar distinction between suppression (Ts mediated) and tolerance (non-Ts mediated) has been shown using proteins coupled to syngeneic cells (Miller et al., 1979; Sherr et nl., 1979). After the first report by Gershon and Kondo (1970) that tolerance could be transferred by T cells from tolerant donors, numerous investigators reported similar findings (Asherson et uZ., 1971; Phanuphak et ul., 1 9 7 4 ~ )In. DNFB contact sensitivity, the ability to transfer tolerance by T cells is a transient phenomenon lasting from 4 to 21 days after tolerization (Miller et al., 1977a). Yet the tolerant state lasts for approximately 6-8 weeks. Thus, the existence of a tolerant state in the absence of any demonstrable Ts cells argues for another mechanism of tolerogenesis in contact sensitivity. There is more suggestive evidence for tolerance without demonstrable Ts. Regimens exist that inactivate Ts precursors or inducers, including adult thymectomy (Baker, 1975), splenectomy (Pierce and Kapp, 1976; Sy et al., 1977b), and cyclophosphaniide pretreatment (Zenibala and Asherson, 1976; Miller et nl., 1977a). Such regimens abolish the induction of Ts (as measured by the failure to transfer tolerance adoptively), but do not alter the existing tolerant state within the donor (Miller et al., 1977a). Thus, in contact sensitivity, another mechanism of tolerance seems to exist that is independent of the generation of Ts (non-Ts mediated). Although the exact lesion of the non-Ts mediated mechanism or clone inhibition is not known, it is antigen-specific and long-lasting (Miller et d., 1977a). Furthermore, this state cannot be transferred to naive recipients by either serum or cells from tolerant donors (Miller et al., 1977a; Claman et al., 1977a). Injection of hapten-modified cells has been shown to depress both contact sensitivity and antibody production (Asherson and Zembala, 1974; Miller and Claman, 1976; Ptak and Rozycka, 1977; Long and
136
HENRY N. CLAMAN ET AL.
Scott, 1977). DNFB-modified spleen cells (DNP-SC) effectively tolerize both in viuo and in oitro cell-mediated responses to DNFB. This unresponsiveness is similar to that produced by the free haptens DNFB or DNBS, except for one interesting finding. Rather than requiring 4-7 days to develop tolerance (as with the free haptens), DNP-SC induces tolerance immediately. This state of unresponsiveness is not dependent on the generation of Ts, as judged by several criteria including (a ) the inability of such treated animals to transfer Ts to normal recipients; ( b )the inability to inhibit the expression of immune-lymphocytes adoptively transferred from DNFB-sensitized donors; and (c) that regimens that abrogate Ts development do not affect the ability to produce tolerance immediately. Thus, immediate tolerance induction appears to be independent of Ts and may reflect inhibition of T cell clones required for the expression of contact sensitivity. The efficient induction of immediate tolerance to DNFB requires that the donor of the tolerogen bear a haptenated Iu-bearing cell which is Z-region compatible (when KID regions are compatible) with the recipient (Coiilon el al., 1979). Moreover, both splenic B cells or plastic adherent (macrophage-like) cells are able to induce immediate tolerance, whereas theta-bearing (T cells) or red blood cells are inefficient (Miller and Claman, 1976; Conlon et al., 1980a). This suggested that the Ia structure itself was recognized regardless of the cellular origin (Benacerraf, 1979). Somewhat different results have been reported in other experimental systems. For example, with TNBSOs modified cells, only theta-bearing mature (T)cells are able to induce a rapid state of unresponsiveness to contact sensitivity (Asherson and Zembala, 1974). While this apparent inconsistency is not readily explained, it may be due to differences in the in vitro modification by the various haptens. However, workers using TNBSO, reported that adherent cells (macrophages) were also capable of inducing a rapid state of tolerance (Polak and Rozycka, 1977). Nevertheless, if a T cell required for expression of sensitivity to DNFB is inactivated by antigen in association with self Ia antigens, then several predictions can be made. First, since hapten-modified cells present antigen plus l a gene products to the pertinent T cell clones, that interaction alone cannot, a priori, be expected to result in immunity. Rather, another signal or event may be required to produce immunity. In its absence, tolerance ensues. This theory, first proposed by Dresser (1962) for activation of all immunocompetent cells, was put forth by Bretscher and Cohn (1970) for B cell activation. Experimental evidence that such a theory is plausible in T cell responses has re-
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
137
cently been reported (Cleveland and Claman, 1980; Claman, 1979). Animals given the tolerogeii DNP-SC intravenously becaiiie sensitized (not tolerized) if treated with the mitogen concanavalin A (Coii A) shortly after tolerization. Thus, in the absence of any further sensitization with DNFB, the recipients could respond specifically to ear challenge with DNFB. The Coii A apparently provided a second, nonspecific sigiial or factor (signal 2), which, when given along with tolerogeii (signal l), resulted in activation of the antigen-specific clones. Similar experiments have been reported for B cell responses, i.e., converting a tolerogeii (bovine serum albumin or deaggregated human y-globulin), to an immunogen b y concomitant injection of the mitogen lipopolysaccharide (Chiller and Weigle, 1973). However, the ability to rescue tolerant €3 cells was felt to be T cell independent (Chiller aiid Weigle, 1973). Thus, a potent tolerogen can be converted to an immuriogen by addition of a second nonspecific signal in both B ant1 T cell responses. Another prediction coiicerns the dual recognition niechaiiisins in T cells. Recent experiments have shown that within a F, (A x B) animal, there exist separate clones of aiitigen-reactive T cells. One set of cells recognizes antigen in the context of one parental MHC, e.g., Ag + MHC-A (Paul et ol., 1977h; Miller, 1978; Sprent, 1978a,b), and aiiother set recognizes Ag + MHC-B. If rapidly induced tolerance involves inhibition of antigeii-reactive T cells, then it should be possible to tolerize each clone separately. Recent experiments show that (A x B) F, mice tolerized with DNP-MHC-A did not have T cells responsive to DNP-A but had no impaimieiit of those T cells recognizing DNP-B. Furthermore, immunity was not conferred on either parental strain recipient (A or B) if the F, donor mice were tolerized with DNP-labeled spleen cells from F, animals (Conlon et al., 198011). These experiments also indicate that there are no Ts cells present in this system. As discussed separately (see Section IV,C), Ts induced b y the injection of syngeiieic DNP-SC suppress the effector phase of DNFB-immune lyiiiphocytes in a genetically nonrestricted fashion. Thus, if Ts were induced by parental DNP-SC in the F,, they should suppress the transfer of sensitivity to both parental strain recipients and one would not find the selective T cell tolerance as reported here. Therefore, if any suppressive effects exist in immediate tolerance, they must act in a genetically restricted nianner. Although the exact mechanism of tolerogenesis in clone inhibition is nnknown, it is thought that the intravenous route allows for direct interaction of DNP-SC with antigen-specific T cells without appropriate host processing. This thereby tolerizes the T cells required for
138
HENRY N. CLAMAN ET AL.
sensitivity to DNFB skin painting. Whether these T cell(s)clones exist in the host (clone inhibition) or are actually eliminated (clone deletion) is not known. It should be noted, however, that tolerance to DNFB can be converted to sensitivity by Con A 2 days after injection of DNP-SC. This suggests that the inactive T cell clones are still present in the host 2 days after tolerization, but for how long is still not certain. C. SUPPRESSIVE MECHANISMSIN CONTACTSENSITIVITY In the last decade the role of suppressor cells in the induction and maintenance of immunologic tolerance to both humoral and cellmediated immune responses has been the subject of intensive investigation (reviewed in Gershon, 1974; MoIler, 1975; Singhal and Sinclair, 1975; Pierce and Kapp, 1976; Asherson and Zembala, 1976; Claman et ul, 1980).After the development in the early 1970s of the concept that immunologic tolerance may be due in part to active suppression, i.e., infectious tolerance,” by suppressor cells (Gershon and Kondo, 1970), it was recognized that there are a variety of suppressor cell systems that play a critical role in immune regulation. Essentially, all the major cell types involved in a positive immune response have also been demonstrated to be capable of functioning as suppressor cellsthese include B cells (Neta and Salvin, 1974; Katz et ul., 1974; Zembala et ul., 1976), macrophages (Folch and Waksman, 1973; Kirchner et ul., 1974), and especially T cells (Gershon, 1974; Asherson and Zembala, 1976; Miller and Claman, 1976). This section will deal with suppressive mechanisms involved in the regulation of contact sensitivity. It will concern the role of suppressor T cells (Ts) in tolerance to and regulation of contact sensitivity, their mechanisms of action, and genetic constraints on both their induction and expression. These discussions will concern Ts induced by the intravenous injection of free hapten or hapten-modified lymphoid cells, and those induced in a somewhat more physiologic way, i.e., by the epicutaneous application of supraoptimal doses of DNFB. ‘I
I . Induction of Suppressor T Cells ( T s ) with Free Reactive Hupteiis The tolerogenic effects of free hapten introduced via intravenous or intragastric routes on the subsequent development of contact sensitivity has been recognized since the pioneering work of Sulzberger (1929). Chase (1946) was the first to show that oral administration of DNCB to guinea pigs specifically prevented subsequent antibody and DTH responses induced by repeated intracutaneous injections of the contactant. Asherson et al. (1977) subsequently showed that mice fed repeated doses of contactants such as TNCB or oxazolone were
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
139
specifically unresponsive and contained suppressor cells of both the B and T cell lineage. De Weck et al. (1964) induced tolerance to DNCB contact sensitivity by the prior intravenous injection of various DNPcongeners, including DNBSO,, in guinea pigs. In the mouse, tolerance to contact sensitization to TNCB can be induced by injection of TNBSO, (Asherson and Ptak, 1970) and tolerance to contact sensitization to DNFB can be induced by injection of DNBSO, (Phanuphak et al., 1974a,b). With DNBSO,, tolerance required an induction period of 4-7 days to reach maximum levels. Zembala and Asherson (1973) were the first directly to demonstrate that mice made unresponsive to contact sensitization by the prior intravenous injection of hapten (TNBSO,) contained active suppressor cells. These cells were shown to be T cells and acted on the efferent limb of the contact sensitivity response as they inhibited the passive transfer of immunity by TNCB-immune T cells (Asherson and Zembala, 1974). Polak and Turk (1974) indirectly provided evidence for a role for Ts in unresponsiveness to DNCB in guinea pigs. They showed that unresponsiveness induced by the intravenous injection of DNBSO, could be reversed by the administration of cyclophosphamide 3 days before sensitization with DNCB. These data were interpreted to mean that the effect of cyclophosphamide was to eliminate a short-lived, rapidly proliferating suppressor cell that was functioning to inhibit specifically contact sensitization. Tolerance to DNFB induced by injection of DNBSO, is also mediated in part by an antigen-specific Ts (Phanuphak et d., 1974c), but these Ts are quite different from those induced by TNBSO,. The precursors of DNBS03-induced Ts appear to reside primarily in the spleen (Sy et ul., 1977b), as no demonstrable Ts were found in the lymph nodes of tolerized mice that have been previously splenectotnized. The spleen must be present for at least 3 days after tolerization for significant levels of Ts to be present in the peripheral lymph nodes. In contrast to these findings, Asherson et nl. (1976) have shown that TNBS0,-induced Ts are resistant to splenectomy, but sensitive to short-term adult thymectomy. After DNBSO, injection, Ts are first detected 3-4 days post-tolerization, peak at 7 days, and are no longer demonstrable 14 days post-tolerization although tolerance lasts 5-6 weeks. Mature Ts are demonstrable in both lymph node and spleen, ) ~opposed to Ts but not in the thymus (Phanuphak et al., 1 9 7 4 ~as induced with TNBSO, (Zembala and Asherson, 1973). The precursors or inducers of Ts are also sensitive to pretreatment with cyclophosphamide (200 mg/kg). These Ts appear to be Ia- and are not depleted as long as 20 weeks after adult thymectomy, indicating that they are a long-lived population of T cells (J. W. Moorhead, unpublished).
140
HENRY N. CLAMAN ET AL.
A further difference between Ts induced by TNBS03 and Ts induced by DNBS03 concerns the target of suppressive activity. TNBS03-induced Ts inhibit the efferent (elicitation) phase of contact sensitivity (Asherson and Zembala, 1974). In contrast, Ts induced by DNBS03 impair the afferent limb as shown by their ability to block the cell proliferation that occurs in regional lymph nodes after sensitization with DNFB (Moorhead, 1976). These targets are specific in the sense that TNBSO, induced Ts do not block the afferent limb, and DNBS03-induced Ts do not block the efferent limb. This discrepancy may be in some way related to the amount of hapten injected, or to the relative ability of DNBSO, vs TNBS03 to couple to proteins in uiuo. Thomas et al. (1979) have recently shown that multiple injections of TNBSO, lead to the induction ofTs active on both afferent and efferent sensitivity. Interestingly, it has been shown that pretreatment of DNBS03induced Ts with hyperimmune rabbit anti-TNP serum eliminates their ability to confer adoptive tolerance in recipient mice (Moorhead and Scott, 1977).These data were interpreted to mean that LN cell populations containing DNBS03-induced Ts have accessible DNP associated with their membrane, which is necessary for adoptive tolerance. It has also been shown that Ts induced by DNBS03are MHC unrestricted at the effector stage. However, a soluble suppressor factor liberated by LN cell populations containing the MHC-unrestricted Ts is restricted by the H-2K and/or H-2D regions of the MHC (see Section IV,D). 2 . Induction of Ts with Hapten-Modijied Lymphoid Cells (LC) The use of antigen-modified cells or proteins as tolerogens has been an important step forward (Battisto and Bloom, 1966). This approach has been exploited in many tolerance systems using defined cell populations (Asherson and Zembala, 1974; Miller and Claman, 1976; Ptak and Rozycka, 1977; Long and Scott, 1977; Greene et al., 1978; Conlon et al., 1980a), serum proteins (Golan and Borel, 1971; Borel, 1976) or cellular components (Greene et al., 1978; Miller et al., 1980a). When cells or protein are coupled with tolerogen in uitro, one is better able to control both the extent of substitution and the moieties that are modified. Moreover, far smaller quantities of tolerogen are effective whFn they are coupled to cells (Miller and Claman, 1976; Long and Scott, i977dSherr et al., 1979; Miller et al., 1979). a. Eflerent Blocking Ts Induced by Syngeneic DNP-LC i. Characteristics and measurement. As discussed previously, the intravenous injection of DNP-modified syngeneic lymphoid cells induces profound and efficient unresponsiveness to DNFB contact sen-
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
141
sitivity (Miller and Claman, 1976). The induction of tolerance with antigen linked to cells has been found in a variety of contact sensitivity systems (Battisto and Bloom, 1966; Greene et al., 1978; Bach et al., 1978) and recently in delayed hypersensitivity to protein antigens as well (Miller et ul., 1979; Sherr et ul., 1979).The induction of suppressor T cells is an integral part of the unresponsiveness, although, as discussed previously, a nonsuppressor cell-mediated pathway of clone inhibition is also involved. Although donor tolerance develops within 1day and lasts 6-8 weeks, antigen-specific suppressor cells from tolerant mice are demonstrable only from 4-7 to 14-21 days after tolerization with DNP-LC. As with DNBS0,-induced Ts, suppressor cells induced by snygeneic DNP-LC are found in the spleen and LN, but not the thymus of tolerant donors. These suppressor cells have been shown to be T cells, as their activity is abolished by treatment with anti-theta-serum and complement and they cannot be raised in T cell-deprived mice (Miller et a]., 1977a). These cells have been termed syninduced Ts. Precursors or inducers of Ts have been found to be sensitive to 150-200 mg of cyclophosphamide per kilogram (Mil1977a) and appear to arise primarily from the spleen (Sy et ler et d., d.,1977b), as splenectomized mice are tolerant, but contain no demonstrable Ts. Syninduced Ts have been shown by several methods to impair the efferent limb of the contact sensitivity response (Miller et al., 1978a). First, syninduced Ts raised in another animal specifically suppress previously sensitized recipients if transferred 1 day before ear challenge. Second, cotransfer of syninduced Ts with DNFB-immune TDH cells to normal recipients blocks the passive transfer of contact sensitivity. Third, in reverse transfer experiments where immune TI)”are transferred into tolerant recipients, only those recipients containing active syninduced Ts (e.g., 7 days after but not 1 day after tolerization with DNP-LC) can inhibit the expression of sensitivity in the immune recipient. Last, syninduced Ts have been found to inhibit the production and/or liberation of MIF by immune TDH i n vitro (J. W. Moorhead, unpublished observation). In contrast, syninduced Ts have no affect on the afferent limb of the response, as they do not inhibit cell proliferation in the recipient’s draining LN. The efferent locus of suppression was most clearly shown in experiments where mice received syninduced Ts at the same time that they were sensitized. These mice are themselves unresponsive to ear challenge, but contain active T D H able to express sensitivity after transfer to normal recipients. These data were interpreted to mean that the suppressive environment, mediated by syninduced Ts, allows the development of a nonnal complement of
142
HENRY N . CLAMAN ET AL.
TDH cells after skin painting (afferent limb), but inhibits the expression of this sensitivity (efferent limb). These data also suggest that the interaction of syninduced Ts with T,, cells is not cytotoxic. Thus, syninduced Ts induced by DNP-LC mediate suppression at the efferent limb similar to TNBS03-induced Ts (Asherson and Zembala, 1974), but in opposition to DNBS03-induced Ts (Moorhead, 1976). ii. Genetic restrictions on induction and expression. In immunoregulation, antigen in association with I region gene products controls the activation of Lyt I+T cells mediating primarily helper and delayed hypersensitivity functions, whereas antigen in association with KID region gene products controls effector functions of cytotoxic T cells. Contact sensitivity models using haptenated lymphoid cells as tolerogens provided an excellent tool for use in determining whether MHC restrictions played a part in Ts induction or expression. Several observations indicated that this may be the case. First, as in tolerance induction (see above), only those DNP-congeners that covalently couple to self proteins are efficient inducers of Ts-e.g., DNFB, DNBS03, and i n vitro modified DNP-LC or DNP-RBC, but not DNPlysine (Phanuphak et al., 1974c; Miller and Claman, 1976). Second, potent DNP-specific B cell tolerogens, such as DNP-y2aor DNP-D-G, L (Katz et al., 1971; Golan and Borel, 1971), do not stimulate Ts or block their subsequent induction with DNP-LC. These data indicate that Ts can be induced in this system only by DNP in association with self membranes. It has been directly shown (Miller et al., 1980a), using soluble DNP-LC membranes, that DNP-MHC determinants are required for Ts induction. Studies with congenic resistant mice indicated that recognition of DNP-modified MHC determinants on the tolerogenic DNP-LC was, in fact, essential for the induction of syninduced Ts (Miller et al., 1977b, 1978b). Thus, tolerization of BALB/c mice ( H - 2 d ) with either syngeneic DNP-BALB/cLC or with H-2-compatible DNP-BlO.D2 LC is sufficient for induction of Ts that can suppress BALB/c recipients. However, tolerization of BlO.D2 ( H - 2 d )mice with DNP-B 10.BR ( H - 2 k ) LC (compatible only outside the MHC) led to no measurable Ts. Mapping studies indicated that induction of syninduced Ts is associated with recognition of DNP-modified determinants in the H-2D region. Tolerization of donor mice with DNP-LC that are H-2D region compatible is both sufficient and required for the induction of syninduced Ts. Injection of DNP-LC which are compatible at the H - 2 K and/or I regions and incompatible at the H-2D region do not induce efferentblocking syninduced Ts (Miller et al., 1978b). In contrast to the requirement for H-2D region compatibility for
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
143
syninduced Ts induction, these Ts were found to be non-MHC restricted in their expression (suppressive action). For example, Ts induced by the tolerization of BALBic mice with BALB/c DNP-LC could transfer suppression to recipient mice of any MHC haplotype (Claman et al., 1977a). Also, this non-MHC restriction was demonstrated when syninduced Ts were found to suppress DNFB-immune TDHfrom different MHC haplotypes using the efferent suppression assay (Miller et al., 1978a). In examining the non-MHC restriction of efferent suppression, attempts were made to determine the target structure(s) of syninduced Ts. This was especially interesting because of recent reports suggesting Ts active in carrier-specific suppression of antibody responses recognize antigen alone (Okamura et ut., 1977; Taniguchi and Miller, 1977). With Ts induced by DNP-LC, blocking experiments (using various DNP-LC cell lysates and membranes) showed that the apparently MHC-nonrestricted syninduced Ts consist of a polyclonal wave of suppressor activity composed of distinct MHC-restricted Ts, some clones of which are specific for DNP-syngeneic determinants and other clones of which recognize DNP-allogeneic determinants (S. D. Miller, 1979). iii. Role of epitope density on M H C restrictions. It should be emphasized that the above-described polyclonal MHC-unrestricted syninduced Ts were induced by the injection of LC modified with a high epitope density of DNFB (0.5 mM). Recent experiments have shown that somewhat different restrictions of induction and expression are obtained if Ts are induced by DNP-LC labeled with 1/100 x (0.005 mM) concentration of DNFB (S. D. Miller, unpublished). As opposed to Ts induced with high epitope density labeled DNP-LC, both the induction and expression of Ts induced by 1/100 x labeled DNP-LC is restricted by MHC products encoded for by either the H-2D and/or H-2K subregions. Pierres et al. (1980) have made similar observations on Ts induced by TNP-modified lymphoid cells. iv. Two-receptor models to explain M H C restrictions on Ts induction and expression. To explain MHC restrictions on Ts induction and expression at the two epitope densities, a model has been proposed for a dual receptor model of T cell antigen recognition similar to those advanced by Janeway et at. (1976) and Doherty et al. (1976b).According to this model, T cells possess two receptors: No. 1 recognizing antigen (i.e., DNP) with low affinity, and No. 2 recognizing self MHC determinants with low to moderate affinity and in some instances able to cross-react with allogeneic H-2 determinants with high affinity. It was proposed that at limiting DNP epitopes only those clones recog-
144
HENRY N. CLAMAN ET AL.
nizing self with moderate affinity (and not cross-reacting with allogeiieic MHC determinants) are triggered (Miller et ul., 1980b). Thus, the Ts induced would be MHC restricted and suppress only H-2K and/or H-2D compatible target TDH cells. In contrast, at high DNP epitopes the former clones and those recognizing self with low affinity (but that also recognize allogeneic MHC determinants with high affinity) are triggered. The activation of these allosuppressive clones is greatly facilitated by the presence of a large number of DNP epitopes on the DNP-LC tolerogen, which may enhance the interaction between the pre-Ts and the tolerogen, leading to their activation. Activation of the pre-Ts clones may be mediated through nonmacrophage presenting pathways, which has been suggested by i n uitro experiments (Feldmann and Kontiainen, 1976; Pierres and Germain, 1978).These allosuppressive clones can thus be specifically and efficiently inhibited only b y pretreatment with DNP-LC membranes that are MHC-compatible with the target T D H . A prediction of the model would be that allosuppressive clones induced by high epitope density-labeled DNP-LC would be fewer in number and/or have higher affinity receptors for DNP-allogeneic determinants. Results showing that 100-fold higher concentrations of syngeneic DNP-LC membranes than allogeneic DNP-LC membranes are needed to inhibit suppression of syngeneic vs allogeneic TDH,respectively, would seem to support this (S. D. Miller, 1979). Thus, Ts induction with low epitope density-labeled DNP-LC can be thought of as monoclonal, whereas induction with high epitope density-labeled DNP-LC appears to be polyclonal and composed of subsets of distinct MHCrestricted Ts. In addition, one could propose that lack of DNP-H-%K reactive clones in Ts populations induced b y high epitope densitylabeled DNP-LC could be due to the fact that H-2K determinants at that concentration are derivatized to such an extent that they are not recognized by pre-Ts clones, while derivatized H-2D determinants would still be recognized. Thus, there is considerable evidence indicating MHC restrictions on Ts induction and expression. The H-BKID restriction on efferentblocking syninduced Ts activated by low epitope density-labeled DNP-LC correlates well with restrictions shown for efferent-acting suppressor factor in the DNFB contact sensitivity system (see Section IV,D) and with the MHC-restricted efferent-blocking Ts induced b y epicutaneous application of supraoptimal doses of DNFB, which will be discussed later. Although Ts induced by the injection of high doses of protein antigens seem to be able to see native antigen not associated with MHC products (Okumura et al., 1977; Taniguchi and Miller,
CONTROL OF EXPERIMENTAL
CONTACT SENSITIVITY
145
19773, other investigations have revealed Ts-associated MHC restrictions. Pang and Blanden (1976) have shown that specific Ts froin ectronielia virus-infected mice could suppress the response of recipient mice to the virus if the donor and recipient were H-2D region conipatible. Kuinar a n d Bennett (1977) have shown that the interaction between Ts a i d niitogen-responsive cells in cultures infected with Friend leukemia virus required identity at the H-2D subregion. More recently, Epstein aiid Cohn (1978) reported that suppression of an i t i uitro antibody response to SRBC by T cells with specificity for histocompatibility antigens is H-2K and/or H-2D restricted. Also, Pierres et ril. (1980) have shown that Ts induced by 1.0 mM TNP-coupled spleen cells, a s opposed to 10 nlnl coupled cells, exhibited MHC restrictions. Considering the data on epitope density of the DNP-LC tolerogeii, it is possible that identification of MHC restrictions on Ts incluction or expression may critically depend on the dose of antigen used to induce the Ts. ~ 7 .
Requiremerit f o r
(iii
(iiisiliciry .suppressor T cell (Ts-cius)iri effer-
eiit ,yiippre,yiotL. Recent experiments indicate that the expression of
efferent-acting Ts in inhibiting TI,, requires another cell. This has been termed an auxiliary Ts ( T s - ~ u x It ) . is present nonilally in populations from DNFB-sensitized mice and is required for Ts to act on T I ) H . This auxiliary cell is sensitive to cyclophosphaiiiide (whereas ‘ r D H is not) aiid requires antigen activation for development. This cell is sensitive to shnrt-term adult thymectoniy, is killed by anti-theta serum + C‘, and hears 1-1 tleterniiiiants (Sy et ti!., 1979a).The precise interrelations between TI)H, Ts-~~LIx, and efferent-acting Ts are not known, but they may be similar to the feedback suppression system of Eardley et ( I / . (1979). These esperinients showing T-T interactions in the mediation of efferent suppression are similar to other studies involving T-T iiiteractions in either the induction or expression of suppression (Eardley et ( I / . , 1978; Turkiii and Sercarz, 1978; Feldniann and Kontiainen, 1976; Gerniain aiid Reiiacerraf, 1978), although the precise localization of the interaction (i.e., induction or expression) in these systems is not clear. Our observations most closely correlate with the results of Tada et rrl. (1978),who found that, in order for a T cell-derived suppressor factor to act on T helper cells, another T cell derived from antigen primed to mice is required. This cell appears to “accept” the suppressor fktor, and only then can it suppress the helper T cells. This acceptor cell is Lyt 1,2,3+aiid I-J’ similar to our Ts-aux cell. 12. Affereiit-Blockitig Ts ltzducecl by Allogetieic DNP-LC. The intravenous injection of allogerieic DNP-LC is extremely effective in
146
HENRY N . CLAMAN ET AL.
tolerizing for DNFB contact sensitivity (Claman et al., 1977b). This form of tolerance, however, appears not to involve active suppression, as no Ts active in strains MHC compatible with the donor animal are demonstrable. That is, LN and/or spleen cells from BALB/c mice tolerized with CBA DNP-LC are not able to suppress BALB/c recipients. However, these same mice do contain Ts that are suppressive in the allogeneic strain providing the DNP-LC (i.e., CBA recipients in this instance) (Claman et al., 1977b; Miller et al., 1977b). These alloinduced Ts thus appear to be genetically restricted. Characterization of the cells showed that they were theta+, Ia- and do not require the spleen for their induction. Mapping studies indicated that tolerization of donor mice with DNP-LC that were incompatible only at the H-2D region was sufficient for alloinduced Ts induction, and the Ts suppressed recipient mice only if they shared the H-2D region, but not H-2K or I region, with the strain providing the DNP-LC tolerogen (Miller et uZ., 197813). Thus, similar to syninduced Ts, the induction and expression of alloinduced Ts is restricted by MHC determinants. Pierres et al. (1980) have recently shown that the injection of allogeneic TNP-LC modified with 10 mM TNBSOs leads to the induction of MHC nonrestricted Ts active in suppressing TNCB contact sensitivity, whereas injection of allogeneic TNP-LC modified with 1 mM TNBSO, leads to the induction of Ts active only in the strain providing the TNP-LC. Thus, epitope density may also be important in MHC restrictions on induction and expression of alloinduced Ts. Alloinduced Ts act only during the afferent (induction) phase of contact sensitivity, as opposed to syninduced Ts (Miller et d., 1978~). Thus, the target of alloinduced Ts, like DNBS0,-induced Ts, seems to be cell proliferation in the recipients. The failure of recipients of alloinduced Ts to generate DNFB-immune TDH capable of transferring sensitivity also inducates that these Ts act by preventing the development of an expanded clone of mature immune TDHcells. It should be noted that the target of Ts in other systems also seems to be cell proliferation (Baker et al., 1974; Baker, 1975; Gershon et al., 1975; Folch and Waksman, 1974; Moorhead, 1976).
3. Inductioiz of Ts by Supraoptimal Sewsitizution a. Description of the Phenomenon. It has long been recognized that the magnitude of the immune response depends critically on the dose of antigen used to invoke it. The epicutaneous painting of mice with supraoptimal doses of DNFB led to a significantly depressed level of sensitivity. Transfer of cells from such animals caused antigen-specific
C O N T R O L OF EXPERIMENTAL CONTACT SENSITIVITY
147
suppression in recipients (Sy et al., 1977a) indicating a role for active suppression. Pretreatment of mice with cyclophosphamide prior to supraoptimal sensitization led to an increased level of sensitivity. Normal mice sensitized with high doses of DNFB could adoptively transfer suppression but not sensitivity, whereas cyclophosphamidepretreated mice after supraoptimal sensitization failed to transfer suppression but could transfer immunity (Sy, 1979). These data indicated that the cyclophosphamide pretreatment eliminated suppressor cell precursors or inducers, thus allowing the full development of DNFBimmune TDHcells. In addition, the fact that cyclophosphamide pretreatment completely reversed the suppression induced by excess antigen suggests that the phenomenon was mediated exclusively by suppressor cells as opposed to induction of tolerance with DNBSO, or DNP-LC where clone inhibition is an important pathway. The observation that cyclophosphamide heightened the response of supraoptimally, but not optimally, sensitized mice is in some respects similar to that of other workers, who showed that the effect of cyclophosphamide on DTH responses depends on the antigen dose (Askenase et d.,1975; Schwartz et ul., 1978). This observation is also relevant to the enhancing effect of cyclophosphamide on DTH observed by other workers (Maguire and Ettore, 1967; Mitsuoka et d.,1976; Polak and Turk, 1974), where supraoptimal doses of antigen may have been employed. The mechanism(s) by which supraoptimal doses of antigen induce Ts is not clear. One could speculate that excess DNFB may leak into the circulation activating Ts as does reactive DNFB or DNBS03, or conjugate to cells other than macrophages inducing Ts as do syngeneic DNP-LC. Kinetic studies showed that Ts induced with supraoptimal doses of DNFB cannot be shown in adoptive transfer until 3 days after sensitization, and activity is maximal at 5 days and no longer detectable by 11 days. Thus, these Ts are like those induced with DNBSO, or DNP-LC, having a time requirement for induction and a short lifespan. b. ldentijcution of Two Distinct Populutioiis of Ts. It was shown that these suppressor cells were of the T cell lineage (Sy, 1979). In investigating whether Ts induced with supraoptimal doses of DNFB inhibit afferent or efferent sensitivity, it was found that they blocked both limbs of the immune response. More detailed analysis revealed that two distinct populations of Ts are induced with supraoptimal doses of DNFB. One population blocks the afferent limb of the response (Ts-am, is insensitive to splenectomy and short-term adult thymectomy, and is Ia-. These afferent blocking Ts do not exhibit MHC restrictions, as they can transfer suppression to allogeneic mouse
148
HENRY N. CLAMAN ET AL.
strains. Thus, this population appears to be similar in phenotype and mode of action to DNBS0,-induced Ts described earlier. The second population of Ts was found to block the elicitation phase of the response (Ts-em. This population is sensitive to splenectomy and shortterm adult thymectomy. It is also sensitive to lysis by anti-Ia and anti-I-J serum plus complement, and appears to be an I-J' cell. As opposed to the Ts-aff, the Ts-eff induced by supraoptimal doses of DNFB shows strain specificity. The Ts-eff induced in BALB/c mice block the passive transfer of immunity by BALBlc TDHcells (Sy, 1979). Thus, this Ts-eff population directly parallels the phenotype and M HC restrictions exhibited by efferent-blocking syninduced Ts activated by the intravenous injection of lymphoid cell derivatized with low epitopes of DNFB. D. SUPPRESSORFACTORS It is abundantly clear from the preceding sections that suppressor T cells have an important role in the regulation of contact hypersensitivity reactions. The precise mechanism(s) by which these cells suppress precursors of TDHor the effector functions of immune TDHremain to be clearly established. In this regard, however, there is a large body of evidence that indicates that, at least for some Ts, suppression is mediated by soluble macromolecular factors released by or extracted from the cells (Okumura and Tada, 1980). In some cases the factors mimic the action of the suppressor cells, but in others they appear to induce additional suppressor cells (Zembala and Asherson, 1974; Tada et al., 1978; Feldmann et al., 1977; Greene et al., 1977). In this section, we will describe the characteristics and properties of a soluble suppressor factor (SSF) that regulates contact sensitivity to DNFB. Space does not permit a detailed discussion of similar suppressor factors that have been shown to regulate humoral immunity and other types of cell-mediated responses. The reader is referred to recent comprehensive reviews on the subject (Okumura and Tada, 1980; Germain and Benacerraf, 1980).
1. Znduction of Soluble Suppressor Factor ( S S F ) The SSF is obtained from 48-hour culture supernatants of LN cells taken from DNFB-tolerant mice. The factor is released by the cells during the culture period and does not require extraction by sonication. T cells that produce SSF are induced by injecting mice with the free hapten DNBS (750 mgkg). However, triggering of SSF production requires additional stimulation of the cells, which is achieved by epicutaneous application of DNFB 5 days after tolerization
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
149
(Moorhead, 1977a). This secondary stimulation is obligatory for SSF production and is antigen specific. A similar protocol is used to prepare a suppressor factor that regulates contact sensitivity to TNCB (Zembala and Asherson, 1974).The precise nature of this secondary triggering stimulus is unknown, although it may be mediated by the T cells recognizing DNP in association with certain membrane antigens (MHC antigens?). 2. Characteristics, Specificity, and Mode of Action of S S F Suppression by SSF is antigen- and hapten-specific (Moorhead, 1977a). The factor suppresses contact sensitivity to DNFB but not TNCB. Furthermore, the factor is adsorbed by, and can be recovered from, a DNP-KLH affinity column. Columns conjugated with TNPKLH or KLH alone do not adsorb the factor. Thus, specific suppression of contact sensitivity to DNFB by SSF is in part based on the factors’ ability to discriminate between DNP and TNP, suggesting that recognition of hapten is an essential step in the factor-mediated suppression (see below). Soluble suppressor factor has been partially characterized by affinity and column chromatography. The factor appears to have a molecular weight of between 35,000 and 60,000. Whether it exists as a single molecule or has subunit structure is not known. The factor does not bear antigenic determinants of immunoglobulin heavy or light chain constant regions, nor is the hapten DNP associated with it (Moorhead, 1977a). Rather, the factor expresses determinants encoded by genes in the MHC as anti-H-2 antibodies are able to adsorb the suppressor activity. These determinants appear to be coded for by genes in the Z-C subregion. Other suppressor factors have also been shown to carry determinants encoded by Z region genes (Tada et al., 1976; Kapp et al., 1976; Theze et al., 1977; Taniguchi and Miller, 1978; Perry et nl., 1978; Rich et nl., 1979). Of particular relevance here is the factor that regulates contact sensitivity to TNCB (Zembala and Asherson, 1974), which has been shown to carry determinants encoded by the Z-J subregion (Greene et al., 1977). Although the TNCB factor and SSF are prepared in essentially the same way, the finding that each bear separate Z region determinants together with the hapten specificity of SSF distinguishes the two as distinct regulatory entities. Regulation by SSF is limited to suppressing effector functions of DNFB-immune TDH;that is, incubation of DNFB-immune TDHwith SSF iit uitro suppresses the ability of the cells passively to transfer contact sensitivity in uiuo (Moorhead, 1977a) or to produce MIF i n uitro (J. W. Moorhead, unpublished). The SSF has no effect on the
150
HENRY N. CLAMAN ET AL.
induction of immune T cells, nor does repeated injection of SSF into normal mice induce suppressor T cells. This is in contrast to the findings of Greene et al. (1977) that showed that injection of TNCB suppressor factor induces suppressor T cells that block effector functions of TNCB T D H . As indicated previously, SSF-producing T cells are induced by intravenous injection of DNBS. This same tolerogen induces suppressor T cells that suppress precursors of TDHcells by inhibiting cell proliferation (Moorhead, 1976) (see Section IV,C,l). Since SSF-mediated suppression is restricted to the effector limb of the contact reaction, it appears that DNBS induces two populatioils of suppressor cells. One, the afferent suppressor, is apparently activated directly by the DNBS. The second suppressor population is primed or induced by DNBS but requires further stimulation (epicutaneous DNFB) for activation resulting in the synthesis and release of the suppressor factor. The two populations caii be distinguished by their sensitivity to adult thyniectomy and anti-Ia serum. T cells producing SSF are lost within 5 weeks after adult thymectomy and are killed by treatment with anti-Ia serum plus complement. In contrast, the Ts that inhibit afferent sensitivity caii still be induced 20 weeks after thymectomy and are insensitive to treatment with anti-la serum.
3 . Genetic Restrictions for SSF-Mediated Suppression As mentioned earlier, many suppressor factors have been shown to express determinants encoded by genes in the I region of the MHC. Some of these factors are under further genetic control to effect suppression; that is, certain regions of the MHC must be shared between the factor-producing strain and the target cells for suppression to occur (Okumura and Tada, 1980; Germain and Benacerraf, 1980).It has generally been found that the I subregion that codes for the determinants on the factor must also be present in the target cell haplotype. We have shown that SSF-mediated suppression is also genetically restricted by genes that map in the MHC (Moorhead, 1977b).However, in contrast to other factors requiring I region identity with their target cells, suppression by SSF requires identity at the H-2K and/or H-2D region. Identity at only one of the two regions is necessary for effective suppression to occur. Thus, although SSF is a product of the Z region of the MHC, it suppresses DNFB-immune LN cells b y interacting with products ofgenes encoded within the H-JK orH-2D loci. It follows then that these gene products serve as target molecules for the suppressor factor. This point was established by adsorbing the supernatants with a
CONTROL O F EXPERIMENTAL CONTACT SENSITIVITY
151
variety of noniial or immune LN cell populations. It is clear that the target or acceptor molecules for SSF are expressed only on DNFBininiuiie T cells (Moorhead, 1979). Noniial T cells, niacrophages, T cell-depleted DNFB-irnmune LN cells or TNCB-immune LN cells clo not adsorb the suppressor factor and thus do not express the target molecules. This is in contrast to the TNCB suppressor factor, which has been shown to interact with nornial macrophages inducing these cells to produce nonspecific suppressor materials (Ptak et [il., 1978a,1)). In order for DNFB-immune T cells to adsorb, they must share with the factor-producing strain either the H-2K or H-2D region of the MHC (Moorhead, 1979). The essential role of these histoconipatilility antigens for SSF-mediated suppression was directly shown b y antiserum blocking studies. Antibodies specific for H-2K or H-2D determinants blocks adsorption of the factor, whereas treating the cells with anti-Ia antibodies has no effect (Moorhead, 1979). This blocking is highly specific. F, DNFB-immune LN cells treated with aiiti-H-2 serum specific for the PI (parent-1) haplotype do not adsorb SSF specific for that haplotype. However, F, LN cells treated with anti-H-2 serum specific for the P, haplotype are unaffected in their ahility to adsorl, the P, suppressor factor. Previous re s u 1t s had shown , ho we ver, that hist ocom pa t ill i 1it y anti gens alone are not the target molecules for SSF. As mentioned above, the suppressor factor is riot adsorbed b y syiigeneic populations of nornial LN T cells, all of which express the relevant histocompatibilit!, antigens. Thus, there appeared to be something unique ahout the histocompatibility antigens expressed on DNFB-immune T cells. We found that this uniqueness is due to the presence of the hapten DNP, as pretreating immune T cells with purified anti-DNP aiitilmclies blocks their al)ility to adsorl, the factor (hloorhead, 1979). Thus, the target niolecules on DNFB-imiiiune T cells for SSF seeni to be coniprised of histocoi~i),citibilitllaritigeiis ti.ssocicitet1 w i t h the lztipteri DNP. It appears, therefore, that SSF recognizes determinants similar to those recognized b y cytolytic T cells induced i t i uitro to haptenmodified autologous lymphocytes (Schnnitt-Verhulst et d . , 1976; Burakoff et [ I / . , 1976). Precisely what this similarity means with respect to the mechanism of suppression b y SSF remains to be deterni i iied . Exactly how the fhctor interacts with the DNP-H2K/H-:'D complex is not known. The fact that either anti-H-2 or anti-DNP antibodies alone block adsorption indicates that the factor must bind or recognize both moieties simultaneously. This may indicate that the factor is nionovalen t and recognizes a ii eoan t i gen produced by D N P coin p l e xing with the H-2 antigens, i. e., altered self. Alternatively, the factor
152
HENRY N . CLAMAN ET AL.
may be bivalent (or multivalent) and recognize the two determinants as independent structures. It seems unlikely that the factor is directed against a neoantigen, since it is adsorbed b y DNP affinity columns, where such a determinant would not be present. Rather, we believe the factor is at least bivalent, with one receptor site specific for DNP and one site specific for H-2. Still a discrepancy exists in that DNP affinity columns will adsorb the factor whereas DNFB-immune allogeneic LN cells or DNP-modified allogeneic cells, both of which have DNP on their surface, do not adsorb the factor (Moorhead, 1979). This difference may be related both to the affinity of the receptor for DNP and to the difference in the hapten density on the affinity column vs the cell membrane. Final resolution of this problem must await biochemical analysis of the factor.
E. SUMMARY Experimental contact sensitivity is a highly regulated immune response. Sensitization depends .upon application of reactive (haptenic) antigens that couple covalently to self tissues. It now seems likely that both the induction (afferent limb) and elicitation (efferent limb) of contact sensitivity involve reactions between the sensitizer and Iabearing Langerhans cells in the skin. The effector cells in contact allergy are T lymphocytes that are capable of synthesizing and releasing lymphokines. Antibody (anti-antigen) may be produced in contact reactions, but its role in modulating the response is unclear. The duration and magnitude of contact sensitivity are regulated. In the mouse, contact sensitivity wanes concomitant with the production of anti-idiotypic antibodies. The magnitude of the sensitive state also depends upon the dose of antigen, as supraoptimal doses of contactant induce suppressor cells. Tolerance (unresponsiveness) to contact allergy is readily produced b y exposure to the antigen prior to sensitization. The ability to produce tolerance (like the ability to produce sensitization) is proportional to the ability of the haptenic antigen to form covalent bonds with tissue components, especially membranes. The route of presentation is crucial in determining tolerance or sensitivity. Epicutaneous or subcutaneous application favors sensitization whereas intravenous or oral routes favor tolerance. The cellular explanations for these facts are not known, but probably involve differences in host processing. Several nonsuppressive and suppressive mechanisms have been described in tolerance to contact sensitivity. Tolerance can be seen in the absence of demonstrable suppressor mechanisms (“clone inhibition”). Tolerance can be induced by haptenated cells and exists ( a ) before suppressor cells are found; ( b ) after suppressor cells are no longer
CONTROL OF EXPERIMENTAL CONTACT SENSITIVITY
153
found; ( c )when suppressor cells or their inducers have been removed; and ( d ) where suppressor cells are present but are genetically not capable of suppressing the tolerant host itself. Suppressive mechanisms found in tolerance to contact sensitivity include suppressor cells and suppressor factors. Suppressor cells may be induced with free reactive haptens such as DNBSO, or TNBSO, or with haptenated cells. Suppressor cells raised in this fashion are T lymphocytes (Ts). Some Ts act on the afferent limb of the contact response, and others act on the elicitation phase. Genetic restrictions exist in both induction and expression of Ts and map to K and D ends of the major histocompatibility coniplex (MHC). The restrictions observed depend upon epitope density of the haptenated tolerogen. The genetic restrictions of Ts have been analyzed in terms of two receptors on T cells-one for DNP and the other for syngeneic or cross-reactive allogeneic MHC products. Soluble suppressor factors are also induced by tolerogenic regimens. These are nonimmunoglobulin, antigen-specific, and involve MHC products. The targets of such factors appear to be hapten and products of the MHC.
REFERENCES Alter, B. J., Schendel, D. J., Bach, M. L., Bach, F. H., Klein, J., and StirnpHing, J. H. (1973).J.Erp. Med. 120, 1183. Asherson, G. L., and Ptak, W. (1968).1t)ittiutwlogy 15, 405. Asherson, G. L., and Ptak, W. (1970).lintnunology 18, 99. Asherson, G. L, and Zembala, M. (1973).Nnture (Lotadoti),Net; R i o l . 244, 176. Asherson, G. L., and Zembala, M. (1974).Proc. R . Soc. Lotidoti, Ser. B 187, 320. Asherson, G. L., and Zembala, M. (1976).Curr. Top. Microbial. Zmniutwl. 72, 56. Asherson, G. L., Zembala, M., and Barnes, R. M. R. (1971).Cliti. Erp. Z~ritiwraol.9, 111. Asherson, G. L., Zemhala, M., Mayhew, B., and Goldstein, A. (1976).E u r . / . Zmrnunol. 6, 699. Asherson, G. L., Zenibala, M., Perera, M. A. C . C., Mayhew, B., and Thomas, W. R. (1977).Cell. Zrnniunol. 33, 145. Askenase, P. W. (1977).Prog. Allergy 23, 199. Askenase, P. W., Hayden, B. J,, and Gershon, R. K. (1975).J.E x p . Med. 141, 697. Bach, B. A., Sherman, L., Benacerraf, B., and Greene, M. I. (1978).J.Ztiamunol. 121, 1460. Baer, H., Dawson, C. R., Byck, J. S., and Kurtz, A. P. (1970).]. Zrnniutiol. 104, 178. Baker, P. J. (1975). Tronsplant. Rer;. 26, 3. Baker, P. J., Stashak, P. W., Amsbaugh, D. F., and Prescott, B. (1974).]. Zrnniunol. 112, 104. Balfour, B. M., McFarlin, D. E., Sumerska, J., and Parker, D. (1974).Ztr “Contact Hypersensitivity in Experimental Animals” (D. Parker and J. L. Turk, eds.), p. 27. Karger, Basel. Barnes, R. M. R. (1973).Eur. /. Znimunol. 3 , 605. Battisto, J. R., and Bloom, B. R. (1966).Nature (Lotidoti)212, 156. Baumgarten, A,, and Geczy, A. F. (1970).Znrtnutiology 19, 205.
154
HENRY N. CLAMAN ET AL.
Benacerraf, B. (1979).J . Imrnunol. 120, 1809. Binz, H., and Wigzell, H. (1977). Prog. Allergy 23, 154. Rorel, Y. (1976). Trarisplcznt. Rev. 31, 3. Bretscher, P., and Cohn, M. (1970). Science 169, 1042. Burakoff, S. J., Gerniain, R. D., Dorf, M. E., and Benacerraf, B. (1976). Proc. N a t l . Accid. Sci. U . S . A . 73, 625. Chase, M. W. (1946). Proc. Soc. Exp. Biol. &led. 61,257. Chase, M. W. (1967). Htrroey Lect. 61, 169. Chiller, J. M., and Weigle, W. 0. (1973).j.E x p . Med. 137, 740. Claman, H. N. (1976).J. Inzmunol. 116, 704. Claman, H. N. (1979). Cell. I n i n c u t t o / . 48, 201. Claman, H. N., and Moorhead, J. W. (1976). Cointemp. TO),. Immunobiol. 5, 211. Claman, H. N., Miller, S. D., and Moorhead, J. W. (1977a). Cold Spring Hurhor Syrnp. Qucitit. B i d . 41, 105. Claman, H. N., Miller, S. D., and Sy, M. S. (1977b).J . Ex)?. Med. 146, 19. Claman, H. M., Miller, S. D., Sy, M. S., and Moorhead, J. W. (1980). Inzrnunol. Rec. 50 (in press). Cleveland, R. P., and Claman, H. N. (1980).J. Irnmunol. 124,474. 278, 257. Conlon, P. J., Moorhead, J . W., and Claman, H. M. (1979). Notrrre (Lond~in) Conlon, P. J., Moorhead, J. W., and Claman, H. N. (1980a). Cell. Irnniunol. 51, 360. Med. Conlon, P. J., Moorhead, J. W., Miller, S. D., and Claman, H. N. (1980b).J. E.K~J. 151, 959. Cosenza, H. (1976). Eur. J . Zmmztno/. 6, 114. Crowle, A. J. (1975).Arlo. Immutzol. 70, 197. Crowle, A. J., and Crowle, C. M. (196l).J. Allergy 32, 302. David, J . R., and David, R. A. (1972). Prog. Allergy 16, 300. de Weck, A. L. (1977). Fed. Proc., Fed. Am. Soc. E x p . B i d . 36, 1742. de Weck, A. L., and Frey, J. R. (1966). “Iminunotolerance to Simple Chemicals.” Karger, Basel. d e Weck, A. L., Frey, J. R., and Geleick, H. (1964). I t i t . Arch. Allergy A p p l . Ittimuno/. 24, 63. Doherty, P. C., Blanden, R. V., and Zinkernagel, R. M. (1976a). Trmsplnnt. RPD. 29,89. Doherty, P. C., Gotze, D., Trinchieri, G., and Zinkernagel, R. M. (1976b). Inmunogenetics 3, 517. Dresser, D. W. (1962). Immunology 5, 378. Dvorak, H. F., Dvorak, A. M., Simpsen, B. A., Richerson, H. B., Leskowitz, S., and Kamovsky, M. J. (1970).J. Exp. Med. 132, 558. Eardley, D. D., Hugenberger, J., McVay-Boudreau, L., Shen, F. W., Gershon, R. K., and Cantor, H. (1978). J . E x p . Med. 147, 1106. Eardley, D. D., Shen, F. W., Cantor, H., and Gershon, R. K. (1979).J. Exp. Med. 150,44. Eichmann, K. (1978).Ado. Immunol. 26, 195. Eisen, H. N., Orris, L., and Belman, S. (1952).J. Exp. Med. 95,473. Epstein, R., and Cohn, M. (1978). Cell. Imrnunol. 39, 110. Erard, D., Charreire, J., Auffredou, M. T., Galanaud, P., and Bach, J. F. (1979).J . Zmmuno/. 123, 1573. Erb, P., and Feldrnann, M. (1975).J. E x p . Med. 142,460. Fachet, J., and Ando, I. (1977). Eur. J. Zrnrnunol. 7, 223. Feldmann, M., and Kontiainen, S. (1976). Eur. J. Zmmrtnol. 6, 302. Feldmann, M., Baltz, M., Erb, P., Howie, S., Kontiainen, S., Woody, J., and Zwaifler, N. (1977). Prog. Irnrnunol. 3, 331.
CONTROL OF EXPERIMENTAL
CONTACT SENSITIVITY
155
Folch, H., and Waksman, B. H. (1973). Cell. I t t i t n u t t o / . 9, 12. Folch, H., and Waksman, B. H. (1974).J. Itnrtiutiol. 113, 140. Frey, J. R., and Wenk, P. (1957). Z t t t . Arch. '4/lerg!/ Appl. I n t t t i r o t o l . 11, 81. Friedlander, M. H., and Baer, H. (1972).J. Z m t t u t i o l . 104, 1122. Geczy, A. F., and Baumgarten, A. (1970). Ztiunutiology 19, 189. Geczy, A. F., Geczy, C. L., and de Weck, A. L. (1975). Eur. J . Z t t i t t i u t i o / . 5, 711. Gemiain, R. N., and Benacerraf, B. (1978).J . I t r i t t i r r t t d , 121, 602. Gennain, R. N . , and Benacermf, B. (1980). Sprittger Seniirt. Inirriutio~,tztliol.3, 93. Gershon, R. K. (1974). Cotiternp. Top,. I r t t t t t u t d > i d . 3, 1. Gershon, R. K., and Kondo, K. (1970). Ininiuttology 18, 723. Gershon, R. K., Askenase, P. W., and Gershon, hl. D. (1975)./. E q i . Med. 142, 732. Godfey, H. P., and Baer, H. (1971)./. Imrnurid. 106, 431. Golan, D. T., and Borel, Y. (1971).]. E.qi. Med. 134, 1046. Greene, M .I., Pierres, A., Dorf, M . E., and Benacerraf, B. (1977).J. E x p . ,!led. 146,193. Greene, M .I., Sugimoto, hl., and Benacerraf, B. (1978).J . Zittniutiol. 120, 1604. Jadassohn, J . (1895). Verh. Dt.vch. Dertticrtol. Ges. 5, 103. Janeway, C. A,, Wigzell, H . , and Binz, H. (1976). Sctrttd. I . I t t i f n U t i ~ J / .5, 993. Jerne, N. K. (1974).Atiit. I r t i t t t r r t i o l . (Poris) 125c, 373. Jones, J. M . , and Amos, H. E. (1975). Nature (Lotidon) 256, 499. Kapp, J . A., Pierce, C. W., DeLnCroix, F., and Benacerraf, B. (1976).J. l t t t i ~ t ~ t t o116, /. 305. Katz, D. H . , and Beniicerraf, B. (1975).l'ruti,spluiit. Rev. 22, 175. Katz, D. H., Davie, J. M.,Paul, W. E., and Benacerraf, B. (1971).J. E.Y)J. .\Jetf. 134, 201. Katz, S. I., Parker, D., and Turk, J. L. (1974). Nature (Lotidoti)251, 550. Kerl, H., Burg, G., and Braun-Falco, 0. (1974). Arch. Der71icitol. Forsch. 249, 207. Kirchner, H., Chusecl, T. M.,Herberman, R. B., Hoden, H. T., and Larvin, D. H. (1974). J . E x p . IlJed. 139, 1473. Klareskog, L., Tjernlunde, U. M., Forsum, U., and Peterson, P. A. (1977). Nuture (Lotidoti ) 268, 248. Kumar, V., and Bennett, M . (1977).Ntrtrrre ( L , o ~ i o t)i 265, 345. Levis, W. R., Whalen, J. J., and Powell, J. A. (1975).J. Invest. Dermutol. 64, 100. Levis, W. R., Lincoln, P. M., and Dattner, A. M. (1978).J. Zttintuttol. 121, 1496. Long, C. '4.R., and Scott, D. W. (1977). E u r . J . I n i t t t u t t d . 7, 1. Lowney, E. D. (1967).J . Inicest. Dennuto/. 48, 391. McFarlin, D. E., and Balfour, B. (1973). Znirtiuiiology 25, 995. hlacher, E., aiid Chase, M . W. (1969).J. E x / ) . Med. 129, 103. Maguire, H. C., and Ettore, \;. L. (1967).J. Zticjest. Zlenttotol. 48, 39. Miller, A. E., and Levis, W. R. (1973).J. Invest. Dermatol. 61, 261. Miller, J. F. A. P. (1978). ~ 7 ~ 1 7 T i l L i i 0Rev. /. 42, 76. Miller, J. F. A. P. (1979). Ado. Crrriccr Res. 29, 1. Miller, J. F. A. P., Vadas, M. A., Whitelaw, A., and Ganihle, J. (1975). Prof. Natl. Actrd. Sci. U . S. A. 72, 5095. Miller, S. D. (1979).J. E.qi. dfed. 150, 676. Miller, S. D., and Clarnan, H. N . (1976).J. Intttiutiol. 117, 1519. Miller, S. D., Sy, M. S., and Claman, H. N . (1977a). E u r . J . Inintuttol. 7, 165. Miller, S. D., Sy, M. S., and Claman, H. N . (1977b)./. E x p . Med. 145, 1071. Miller, S. D., Sy, M. S., and Claman, H. N. (1978a).J. Itrtttiutiol. 121, 265. Miller, S. D., Sy, M. S., and Clarnan, H. N. (197Xb).J. E x p . Med. 147, 788. Miller, S. D., Sy, M. S., and Claman, H. N. (1978c).J. Irtintunol. 121, 274. Miller, S. D., Wetzig, R. P., and Claman, H. N. (1979).J . E.Y/J.Med. 149, 758.
156
HENRY N. CLAMAN ET AL.
Miller, S. D., Conlon, P. J., Sy, M. S., Colon, S., Moorhead, J. W., Grey, H. M., and Claman, H. N. (1980)J. Zmmunol. 124, 1187. Mitsuoka, A., Baba, M., and Morikowa, S. (1976).Nature (London) 262, 77. Moller, G., ed. (1975).Transplant. Reu. 26, 1. Moorhead, J. W. (1976).J. Zmmunol. 117,802. Moorhead, J. W. (19774.J. Immunol. 119,315. Moorhead, J. W. (1977b).J. Immunol. 119, 1773. Moorhead, J. W. (1978).]. Immunol. 120, 137. Moorhead, J. W. (1979).J. E x p . Med. 150, 1432. Moorhead, J. W., and Scott, D. W. (1977).Cell. Zmmunol. 28,443. Neta, R., and Salvin, S. B. (1974).J. Zmmunol. 113, 1716. Nossal, G. J. V., and Pike, B. L. (1978).]. E x p Med. 148, 1161. Okumura, K., and Tada, T. (1980).Adu. Zmmunol. (in press). Okumura, K., Takemori, T., Tokuhisa, T., and Tada, T. (1977).J. E x p . Med. 146, 1234. Oudin, J., and Michel, M. (1963).C . R. Hebd. Seances Acad. Sci. 257,805. Oudin, J., and Michel, M. (1969).J. E x p . Med. 130, 595. Pang, T., and Blanden, R. V. (1976).J.E x p . Med. 143,469. Parrott, D. M. V., deSousa, M. A. B., Fachet, J., Wallis, V., Cenchars, E., and Davies, A. J. S. (1970).Clin. Exp. Zmmunol. 7, 387. Paul, W. E., and Benacerraf, B. (1977).Science 195, 1293. Paul, W. E., Shevach, E. M., Pickeral, S., and Rosenthal, A. S. (1977a).Cold Spring Harbor Symp. Quaiat. Biol. 41, 571. Paul, W. E., Shevach, E. M., Pickeral, S., Thomas, D. W., and Rosenthal, A. S. (1977b).J. E x p . Med. 145,618. Perry, L., Benacerraf, B., and Greene, M. I. (1978).J. Zntmunol. 121,2144. Phanuphak, P., Moorhead, J. W., and Claman, H. N. (1974a).J. Zmmunol. 112, 115. Phanuphak, P., Moorhead, J. W., and Claman, H. N. (1974b).J. Zmmunol. 112,849. Phanuphak, P., Moorhead, J. W., and Claman, H. N. (1974c).J. Zmmunol. 113,1230. Phanuphak, P., Moorhead, J. W., and Claman, H. N. (1975).J. Zmmunol. 114, 1147. Pierce, C. W., and Kapp, J. A. (1976).Contemp. Top. Zmmunobwl. 5,91. Pierres, A., Bromberg, J. S., Sy, M. S., Benacerraf, B., and Creene, M. I. (1980)J . Zmmurd. 124, 343. Pierres, M., and Cermain, R. N. (1978).J. Znamunol. 121, 1306. Polak, L. (1977).Adu. Mod. Toxicol. 4, 225. Polak, L., and Rinck, C. (1978).J. Znuest. Dermatol. 70, 98. Polak, L., and Turk, J. L. (1969).J. Znuest. Dermatol. 52, 219. Polak, L., and Turk, J. L. (1974).Nature (London)249, 654. Polak, L., Polak-Wyss, A., and Frey, J. R. (1974).Znt. Arch. Allergy Appl. Zmmunol. 46, 417. Ptak, W., and Rozycka, D. (1977).Eur. J . Zmmunol. 7,855. Ptak, W., Zembala, M., and Gershon, R. K. (1978a).J. E x p . Med. 148,424. Ptak, W., Zembala, M., Hanczakowski-Rewicka, M., and Asherson, G. L. (1978b).Eur.1. Zmmunol. 8, 645, Rajewsky, K., and Eichmann, K. (1977).Contemp. Top. Zmmunobiol. 7,69. Rich, S. S., David, C. S., and Rich, R. R. (1979).J.E x p . Med. 149, 114. Riley, J . F., and West, C. B. (1956).AMA Arch. Dermatol. 74, 471. Robinson, J. H., and Naysmith, J. D. (1976). Scand. J. Immunol. 5,299. Sabbadini, E., Neri, A., and Sehon, A. H. (1974).J. Zmmunol. Methods 5, 9. Schmitt-Verhulst, A. M., Sachs, D. H., and Shearer, G. M. (1976).]. E x p . Med. 143,211. Schultz, L. D., and Bailey, D. W. (1975).Zmmunogenetics 1, 570. Schwartz, A,, Askenase, P. W., and Gershon, R. K. (1978).J. Immunol. 121, 1573.
CONTROL OF EX~ERIMENTALCONTACT SENSITIVITY
157
Shearer, G. M., Rehn, T. G., and Gurburino, C. A. (1975).J.E x p . Med. 141, 1348. Shelley, W. B., and Juhlin, L. (1976). Nature (London) 261, 46. Sherr, D. H., Cheung, N.-K. V., Heghinian, K. M., Benacerraf, B., and Dorf, M. E. (1979). J . Zmmunol. 122, 1899. Shevach, E. M., and Rosenthal, A. S. (1973).J.E x p . Med. 138, 1213. Shirley, S. F., and Little, J. R. (1979).J.Zmmunol. 123, 2883. Silberberg, I., Baer, R. L., and Rosenthal, S. A. (1974).Arch.Dermatooener. (Stockholm) 54, 321. Silverberg-Sinakin, I., Baer, R. L., and Thorbecke, G. J. (1978). Prog. Allergy 24,268. Singhal, S. K., and Sinclair, N. R. St. C., eds. (1975). “Suppressor Cells in Immunity.” Univ. of Western Ontario Press, London, Ontario. Sprent, J. (1978a).J. E x p . Med. 147, 1142. Sprent, J. (1978b).J.E x p . Med. 147, 1159. Sulzherger, M. B. (1929).Arch. Dermatol. S y p h i l o l . 20,669. Sy, M. S. (1979). Ph. D. Thesis, Department of Microbiology and Immunology, University of Colorado, Boulder. Sy, M. S., Miller, S. D., and Claman, H. N. (19774.J. Zmmunol. 119, 240. Sy, M. S., Miller, S. D., Kowach, H. B., and Claman, H. N. (1977b).J.Zmmunol. 119, 2095. Sy, M. S., Miller, S. D., Moorhead, J. W., and Claman, H. N. (1979a).J.E x p . Med. 149, 1197. Sy, M. S., Moorhead, J . W., and Claman, H. N. (1979b).J.Zmmunol. 123,2593. Tada, T., Taniguchi, M., and David, C. S. (1976).J. E x p . Med. 144, 713. Tada, T., Taniguchi, M., and Okumura, K. (1978).In “Ir Genes and Ia Antigens” (H. 0. McDevitt, ed.). Academic Press, New York. Tagart, V. B., Thomas, W. R., and Asherson, G. L. (1978). Immunology 34, 1109. Takahashi, C., Nishikawa, S., Katsura, Y., and Izumi, T. (1977). Zmmunology 33, 589. Taniguchi, M., and Miller, J. F. A. P. (1977).J.E x p . Med. 146, 1450. Taniguchi, M., and Miller, J. F. A. P. (1978).J . Immunol. 120, 21. Taylor, R. B., and Iverson, G. M. (1971). Proc. R. SOC. London, Ser. B 176,393. Theze, J., Waltenbaugh, C., Germain, R. N., and Benacerraf, B. (1977).Eur. J. Immunol. 7, 705. Thomas, W. R., Watkins, M. C., and Asherson, G. L. (1978).Zmmunology 3!5,41. Thomas, W. R., Watkins, M. C., and Asherson, G. L. (1979).J.Zmmunol. 122,2300. Turk, J. L., and Stone, S . H. (1963). In “Cell-Bound Antibodies” (S. Amos and H. Koprowski, eds.), p. 142. Wistar Inst. Press, Philadelphia, Pennsylvania. Turkin, D., and Sercarz, E. (1978).In “The Immune System: Genetics and Regulation” (E. E. Sercarz, L. A. Herzenberg, and C. F. Fox, eds.), p. 174. Academic Press, New York. Vadas, M. A., Miller, J. F. A. P., McKenzie, I. F. C., Chism, S. E., Shen, F.-W., Boyse, E. A,, Gamble, J. R., and Whitelaw, A. M. (1976).J.E x p . Med. 144, 10. Vadas, M. A., Miller, J. F. A. P., Whitelaw, A., and Gamble, J. (1977).Zmmunogenetics 4, 137. Zembala, M., and Asherson, G. L. (1973).Nature (London) 244,227. Zembala, M., and Asherson, G. L. (1974). Eur. J. Zmmunol. 4, 799. Zemhala, M., and Asherson, G. L. (1976). Clin. E x p . Zmmunol. 23,554. Zembala, M., Asherson, G. L., Noworolski, J., and Mayhew B. (1976).Cell. Immunol. 25, 266. Zinkemagel, R. M. (1976).J.E x p . Med. 144, 776. Zinkemagel, R. M., and Doherty, P. C. (1975).J.E x p . Med. 141, 1427.
This Page Intentionally Left Blank
Analysis of Autoimmunity through Experimental Models of Thyroiditis a n d Allergic Encephalomyelitis WILLIAM 0.WEIGLE Deportment of fmmunopathology, Scrippr Clink and Research Foundation, to Jolla, Cafifomia
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Mechanism of Self-Tolerance ..................... 111. Types of Acquired Immunologic Tolerance ...............................
A. Peripheral Inhibition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Central Unresponsiveness
159 161 162 162
A. Polyclonal Activation ..........................
D. Suppressor Cells in Autoimmunity . . . . . . . . . . . . . V. Experimental Autoimmunity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Experimental Autoimmune Thyroiditis (EAT) ......................... B. Experimental Allergic Encephalomyelitis (EAE) ...................... VI. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References
184 185 224 251 253
I. Introduction
Self constituents normally d o not stimulate an immune response, but occasionally the immune system turns on its host environment in so aggressive a manner P S to cause disease. Significantly, the cellular events leading to and reg dating this destructive autoreactivity are the same as those involved in beneficial immune responses to foreign antigens. That is, all elements in the repertoire of immune defenses (antibody of various subclasses, antibody-dependent cell cytotoxicity, delayed-type hypersensitivity, etc.) also participate in autoimmunity. However, before one can understand the cellular parameters involved in autoimmune disease, one must first appreciate the conditions involved in recognition of self antigens as foreign. An essential condition for self-noiiself discrimination is the immunologic tolerance animals have to their own body constituents. When this self-tolerance falters, autoimmunity, sometimes accompanied by disease, ensues. The various mechanisms that may be responsible for the loss of tolerance to self antigens can be divided into three general categories. First, abnormalities may occur in the regulatory mechanisms that con159 Copyright 0 1980 h y Academic Preas, Inc All right\ of reproduction in any form recerved
ISBN (I-12-022430-5
160
WILLIAM 0. WEIGLE
trol the normal immune response. For example, genetic deficiences in immune regulation may permit self recognition to proceed to an autoimmune response, and then to disease. Second, a component of self that was once sequestered and nonimmunogenic may become exposed and presented in an antigenic form to the immune system. In this regard, factors generated during infection, trauma, etc., can either potentiate an immune response directly or do so indirectly by facilitating release of immunogenic levels of self antigens. Third, a normally tolerated self component may for some reason circumvent the prevailing regulatory mechanisms and activate one or more arms of a normal immune system. Such conditions may result from polyclonal activation of B lymphocytes by viral or microbial infection. Alteration of self constituents or contact with antigens with which they cross-react may also promote bypass of tolerance at the T cell level permitting activation of B cells.' Alteration of self could result from genetic error or as a consequence of infection. Therefore, the cause of autoimmune phenomena may range from a single condition to any combination of the aforementioned categories as in some complex autoimmune diseases. The relationship among self recognition, autoimmunity, and autoimmune disease is often obscure. Self recognition through activated T cell subsets may occur without autoimmunity, and autoimmunity often occurs without autoimmune disease. I n fact, autoimmunity is not such a rare event as is assumed from the infrequency of related clinical symptoms, but rather is quite often detected in individuals (especially in the aged) without overt disease. Furthermore, the onset of autoimmune disease depends on the target autoantigen's amount and location, persistent stimulation by the autoantigen, and the biologic properties (e.g., sufficient avidity) of the effector lymphocytes and their products. In this setting, the review that follows is an attempt to define and discuss the cellular and subcellular events that may be involved in several of the less complex experimental models of autoimmunity.
' Abbreviations used: ABC, antigen binding cells; ADCC, antibody-dependent cellular cytotoxicity; AHGG, aggregated human y-globulin; B cell, bone marrow-derived lymphocyte; BP, basic protein; BSA, bovine serum albumin; C3, third component of complement; CFA, complete Freund's adjuvant; CNS, central nervous system; DHGG, deaggregated human y-globulin; DNP, dinitrophenyl; DTH, delayed-type hypersensitivity; EAE, experimental allergic encephalomyelitis; EAMG, experimental autoimmune myasthenia gravis; EAT, experimental autoimmune thyroiditis; EB, Epstein-Barr (virus); HGG, human y-globulin; HSA, human serum albumin; Ig, immunoglobulin; LPS, lipopolysaccharide; MG, myasthenia gravis; MHC, major histocompatibility complex; MIF, migration inhibition factor; MIgG, murine IgG; MS, multiple sclerosis; mw, molecular weight; 0s obese strain (chickens); PAB, p-aminobenzoic acid; PFC, plaque-fornling cells; PPD, purified protein derivative; SAT, spontaneous autoimmune thyroiditis; sIg, surface immunoglobulin; T cell, thymus-derived lymphocyte; TD, thymus dependent; Tg, thyroglobulin; TI, thymus independent.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
161
II. Mechanism o f Self-Tolerance
During the past 25-30 years, studies of the cellular mechanism of acquired tolerance to foreign antigens have provided considerable insight into the cellular events involved in self-tolerance. That tolerance as well as immunity can be induced in animals was first predicted by Bumet (1959), who suggested that unresponsiveness to foreign antigens could be induced in animals if the antigens were injected during early life. This suggestion actually originated from the work of Owen (1945),who first demonstrated that contact with foreign antigenic substances during early life resulted in immunologic tolerance. He observed that mature dizygotic twin cows tolerated each other’s body tissues in that they did not reject mutual grafts. Undoubtedly, the tolerance resulted from embryonic parabiosis in which blood was exchanged between the twins. Subsequently, Billingham and coworkers (1953)found that adult mice of an inbred strain tolerated skin grafts of a second inbred strain if, as newborns, animals of the first strain were injected with replicating cells of the second strain. It has since been shown that numerous nonliving antigens can induce immunologic tolerance in a variety of animals (reviewed in Weigle, 1973; Howard and Mitchison, 1975; Dresser and Mitchison, 1968). Furthermore, experimental induction and maintenance of toIerant states have shown that specialized and distinctly different cellular events may be involved in different tolerant states. The relationship between experimentally induced tolerance to foreign antigens and naturally acquired tolerance to self has both practical and theoretical implications. In his original hypothesis, Burnet (1959) assumed that tolerance induced to foreign antigens is the same as tolerance to one’s own body constituents. For an animal to make an immune response to foreign substances, such as bacteria, viruses, tumor antigens, and yet not respond to its own body constituents, the immune mechanism must discriminate between self and foreign antigens. Thus, during prenatal andlor neonatal life, before the immune mechanisms mature, animals develop a state of immunologic unresponsiveness to their own body constituents, but this state does not interfere with their ability to respond, as adults, to foreign antigens. There is overwhelming evidence that the development of tolerance to self components is not genetically determined, but rather the result of direct contact between self components and specific antigen-reactive cells. In this regard, Triplett (1962) removed the hypophysis (buccal component of the pituitary gland) from a tree frog during early life (tadpole), allowed the gland to differentiate away from its donor, then returned it to the mature donor, whose immune system subsequently rejected the transplant. Similarly, animals make an immune response
162
WILLIAM 0. WEIGLE
to body constituents that they lack as a result of a genetic deficiency (Rosenberg and Tachibana, 1962; Cinader et al., 1964; Winchester,
1979). 111. Types of Acquired Immunologic Tolerance
Although immunologic tolerance induced b y prior exposure to antigen is defined by the inability of the host to respond to that specific antigen, the cellular and subcellular events leading to the unresponsive state may differ. We are concerned, then, with acquired immunologic tolerance, which can be classified into two categories: peripheral inhibition and central unresponsiveness (Weigle et al., 1974a).
A. PERIPHERAL INHIBITION In peripheral inhibition, cells competent in respect to immune capacity are present but their function is blocked. Lymphocytes of the tolerant host can bind the antigen in question, and the tolerant state disappears when the cells are transferred to a neutral (irradiated) host. Furthermore, the tolerant state, at times, is associated with a transient appearance of antibody. This type of unresponsiveness may not represent a true tolerant state, but suppression induced by regulatory mechanisms normally at play in controlling the immune response, such as suppressor cells, antigen blockade, or antibody (including antiidiotype) suppression.
1. Suppressor Cells A subpopulation of thymus-derived lymphocytes (T cells) that is capable of suppressing the immune response plays a major role in regulating various parameters of the immune response, once it is initiated (reviewed in Gershon, 1974; Benacerrafet al., 1975; Tada et al., 1975; Gershon, 1977).While suppressor T cells dampen the response, another population, helper T cells, enhances the response (Gershon, 1974; Scavulli and Dutton, 1975). These two populations of T cells can be separated on the basis of surface markers (Cantor and Boyse, 1977a; Vadas et al., 1976; Taniguchi and Miller, 1977; Okumura et al., 1977; Cantor et al., 1978). With antiserum to the Lyt surface antigen of mice, helper (inducer T cells have been shown to be Lyt 1+while suppressor cells are Lyt 23+ (Vadas et al., 1976). More recently, the functionally defined classes of suppressor and inducer T cells have been further subdivided with antisera to Lyt (Cantor and Boyse, 1977a,b) and Q a l (Stanton and Boyse, 1976; Stanton et al., 1978) surface determinants. Signals from both Lyt l+:Qal+cells and Lyt l+:Qal- cells are required for optimal formation of antibody by B cells (Cantor et al., 1978). The Lyt l+:Qal+ cells are also responsible for inducing feedback inhibition, in which a nonimmune Lyt 123+subset of regulatory T cells is
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITlS
163
induced to participate in generation of specific suppressor cell (Lyt 23+)activity (Cantor et al., 1978; Eardley et al., 1978). The interaction between T inducer cells and T suppressor cells is governed by genes linked to the Zg locus (Eardley et al., 1979). More recently contrasuppressioii has been suggested as an additional component of the immunoregulatory network in mice (Duram ct al., 1980). Contrasuppression apparently interferes with the ability of suppressor cells to inhibit the primary in uitro antibody response and is generated by the interaction of inducer T cells and another subset of T cells termed “acceptor” (Lyt 1+,Lyt 2 + ,and 1-J+)cells. In humans, suppressor cells have been identified and characterized as T cells by the specificity of their Fc receptor binding capacity for IgG (Moretta et a l . , 1977a,b). I n addition, Reinherz and Schlossmaii (1980) have suggested that subsets of inducer, suppressor, and feedback inhibitor T cells are components of a regulatory network within the immune response of man. Not only are suppressor cells active in the regulation of ongoing immune responses, but they are also associated with the tolerant state to a number of thymus-dependent (TD) and thymus-independent (TI) antigens (Basten, 1974; Bakeret al., 1974; Kapp et al., 1974; Debre et d . , 1975; Benjamin, 1975; Nachtigal et al., 1975; Basten et d , , 1975). This association between suppressor cells and immunologic tolerance has been convincingly demonstrated in some (Asherson and Zembala, 1974; Phanuphak et a/., 1974; Polak and Turk, 1974; Herzenberget d., 1975),but by no means all, models of tolerance (reviewed in Weigle et nl., 1975). On the other hand, a number of investigators have reported that tolerance to protein antigens can be established in the absence of detectable suppressor cells in normal adult animals (Chiller and Weigle, 1973a; Chiller et d . , 1974; Zolla and Naor, 1974; Fujiwara and Kariyone, 1978; Benjamin, 1977a; Parks et nl., 1978, 1979), neonatal mice (Benjamin, 1977b), athymic mice (Schrader, 1974; Parks et al., 1977; Etlinger and Chiller, 1977), and adult mice that have been thymectomized, irradiated, and bone niarrow-reconstituted (Nachtigal et d., 1975; Chiller et al., 1974). It is important to emphasize that the presence of antigen-specific suppressor cells in an animal immunologically tolerant to that antigen does not in itself imply a causal relationship between the suppressor cells and the establishment and maintenance of the tolerant state. Any postulate that the suppressor ceIls present in a tolerant host represent the mechanism of unresponsiveness must be fimily established experimentally. 2 . Antigen Blockude Although a central unresponsive state can readily be induced in T cells, experimentally induced tolerance in bone marrow-derived lymphocytes (B cells) may involve other mechanisms more closely related
164
WILLIAM 0. WEIGLE
to peripheral unresponsiveness. Some investigators have even suggested that antigen blockade is responsible for B cell tolerance to both foreign and self antigens (Femandez et al., 1979). A classical example of antigen blockade is the immunologic paralysis induced in adult mice by using pneumococcal polysaccharide (Howard, 1972), an antigen that persists in the host for many months because no specific depolymerases are present for its breakdown. Injecting small amounts of this antigen (0.5-5 pg) results in immunity, whereas injecting larger amounts (10-25 pg) induces an unresponsive state, despite the presence of antibody-producing cel1s:This apparent tolerant state probably results when the persisting antigen causes a “treadmill” neutralization of antibody (Howard, 1972; Dixon et al., 1955). When significantly larger amounts (more than 250 pg) of the polysaccharide are injected, tolerance is induced but no antibody-forming cells appear. However, these “tolerant” mice contain specific antigen-reactive cells, which, when transferred to a neutral host, are triggered to differentiate into cells that produce antibody. A similar mechanism is probably responsible for tolerance to inactivated lipopolysaccharide (LPS). With LPS, the unresponsive state of tolerant cells is reversed if the cells are cultured for 24 hours before transfer to irradiated recipients, (Sjoberg, 1972). It is important that with both the above antigens tolerance is a quality of B cells, since these antigens are T cell independent. Aldo-Benson and Bore1 (1974) detected cells containing surface tolerogen in mice made tolerant to DNP coupled to mouse IgG, and suggested the possibility of antigen blockade as a mechanism. Gronowicz and Coutinho (1975) reported that tolerance to the DNP hapten could be reversed by culturing spleen cells from tolerant mice with LPS, a polyclonal B cell activator. Using this same approach, Moller and co-workers (1976) reported polyclonal activation of “tolerant” B cells b y LPS in uitro, with antigen blockade suggested as a mechanism for B cell tolerance, On the other hand, Fidler (1979) used sulfonate to tolintravenous injection of 2,4,6-trinitro-l-chlorobenzene erize mice to the TNP hapten, and reported only a transient tolerant state apparently resulting from antigen blockade (and reversible by LPS polyclonal activation). However, this transient tolerance was replaced within 24 hours by irreversible tolerance. The failure to demonstrate antigen blockade as a mechanism for maintenance of tolerance to human y-globulin (HGG), using a similar approach, is documented in Section III,B,6. The tenets that tolerance results from antigen blockade of all B cells responsive to a particular antigen and is reversed by stimulation with polyclonal activators (Moller et al., 1976)were modified by Fernandez and Moller (1977, 1978). They reported that certain B cells can be irreversibly inactivated by exposure to antigen in the presence of a
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
165
polyclonal B cell activator. Cells bearing receptors for both the antigenic determinant on dextran and its polyclonal activator site were irreversibly inactivated during exposure to that TI antigen. These authors suggested that irreversible inactivation (tolerance) can be accomplished only through interactions at both the antigen receptor and the receptor for polyclonal B cell activation. Recently, Parks and Weigle (1980a) suggested that the assistance provided to B cells by helper T cells may be analogous to the polyclonal signal provided by T-independent antigens, and that T-dependent B cells possessing receptors for both antigen and T cell help should be irreversibly inactivated by tolerogen. Such inactivation would require macrophages and implies genetic restriction.
3 . Antibody Feedbuck Numerous examples in the scientific literature offer convincing evidence that specific antibody gives both positive and negative signals that may govern the immune response (reviewed in Uhr and Moller, 1968;Walker and Siskind, 1968; Weigle and Berman, 1979).A series of experiments on antibody inhibition led Bystryn et ul. (1971) to propose that production of antibody to both persisting and readily catabolized antigens is controlled via a dynamic equilibrium among circulating antibodies, antigen, and antigen-antibody complexes throughout the extracellular compartment. Thus, one should expect that both suppression (Rowley et ul., 1969) and enhancement (Morrison and Terres, 1966) depend upon the antigen-antibody ratio within the complexes. In the case of suppression, it has been reported that antibody inhibits the immune response simply by blocking the availability of antigens (Cerottini et (11.. 1969a,b; Feldmann and Diener, 1972; yet others believe that the Fc site of the antibody must react with Fc receptors on the cells involved in a particular immune response (Sinclair, 1969; Kappler et al., 1971; Lees and Sinclair, 1973; Sinclair and Chan, 1971; Sinclair et al., 1974; Kappler et a/., 1973; Hoffmann et d., 1974;Wason and Fitch, 1973).In this regard, Hoffmann and Kappler (1978) postulated two mechanisms of antibody-mediated suppression of antibody production. They suggested that one mechanism operates at a low concentration of antibody and depends on the Fc portion of the antibody molecule, and the other requires high concentrations of antibody and is independent of the Fc portion. Suppression through the latter mechanism could be mediated by intact antibody, or its F(ab), fragment, but the former one could be initiated only by intact antibody. In either case, T or B cells do not seem to be affected directly; rather antibody inhibits the immune response by interfering with the interaction between T and B cells. Additional information on the regulatory role of antibody in the im-
166
WILLIAM 0. WEIGLE
mune response has been obtained with antibody directed to surface immunoglobulin (sIg) and idiotypic determinants on Ig. When presented in a soluble form, anti-Ig elicits proliferative responses in lymphocytes from rabbits (Sell and Gell, 1965), mice (Sieckmann et al., 1978a,b; Parker, 1975;Weiner et ul., 1976; Sidman and Unanue, 1978), humans (Adenolfi et al., 1967; Daguillard et al., 1969; Oppenheim et al., 1969; Gausset et al., 1976), pigs (Maino et al., 1975), and chickens (Skamene and Ivanyi, 1969; Kirchener and Oppenheim, 1972); if presented on a solid matrix, anti-Ig causes both proliferation and differentiation (Parker et al., 1979).On the other hand, Sidman and Unanue (1975) observed inhibition of i n uitro Ig synthesis when anti-Ig was added to murine spleen cells. In adult cells, the inhibition was reversed upon removal of the anti-Ig, whereas with neonatal spleen cells the inhibition was irreversible. Since the original observation by Oudin and Michel (1969) on the existence of anti-idiotypic antibodies, identical or similar idiotypic markers have been detected on both T and B lymphocytes; both cell types are lysed by the specific anti-idiotypic antibody and complement (Binz and Wigzell, 1975; Eichmann and Rajewsky, 1975). A number of workers have used anti-idiotypic antibodies to inhibit the activation of both T and B cells (reviewed in Binz and Wigzell, 1977). More recently, anti-idiotypic antibodies have been shown to accompany normal immune responses in mice (Brown and Rodkey, 1979; Schrater et al., 1979; Goidl et al., 1979). This result, along with the postulated role for anti-idiotypic suppression in the network theory of antibody production (Jerne, 1974; Rodkey, 1974; Kohler, 1975; Pierce and Klinman, 1977),implicates anti-idiotypic antibody as an important regulatory mechanism of the immune response. Since Rodkey’s (1974) demonstration of auto-idiotypic antibody, Binz and Wigzell (1978) have induced, at least, a partial transplantation tolerance in rats by immunizing them with allogeneic idiotypes. Tolerance to a foreign antigen has also been induced with anti-idiotypic antibody in both adult and neonatal mice in uivo (Strayer et al., 1975).As with anti-Ig, the tolerant state induced with anti-idiotypic antibody is reversible in adults, but irreversible in neonates. In individuals with malignant tumors, the absence of protection from progressive tumors also seems to be a peripheral tolerant state that results from competition between cellular and humoral immunity (reviewed in Weigle, 1973).In many instances specific serum antibody actually enhances the survival of tumors. More recent studies have shown not only that cellular immunity to a tumor can coexist with the tumor, but also that its survival and growth in viuo may be attributed to a specific ability of that individual’s serum to prevent tumor destruction b y sensitized lymphocytes. That is, the host tolerating the tumor
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
167
apparently contains lymphocytes capable of killing the tumor i n vitro (Hellstrom and Hellstroni, 1969), but the killer cells are blocked by some factor present in the serum. Although convincing data have suggested that central unresponsiveness can be induced to transplantation antigens (reviewed in Streilein, 19791, in some models the mechanism of transplantation tolerance is also that of peripheral inhibition similar to that involved in tolerance to tumors (Hellstrom et al., 1971). That is, circulating antibody blocks the destructive effect of sensitized lymphocytes. Furthermore, tetraparental mice have lymphocytes that react against the cells of each parent and also have a serum factor (antibody) capable of inhibiting the lymphocytes (Phillips et al., 1971). Since these mice possess a mixture of cells containing different H-2 antigens, the lymphocytes from a single mouse react against each other i n vitro by undergoing blastogenesis. This observation cannot be extrapolated to explain self-tolerance in normal outbred animals, since the lymphocytes from these animals all have the same histocompatibility markers. This conclusion is supported by the failure of cells from F, hybrid mice to undergo spontaneous blastogenesis i t 1 vitro and the absence of serum factors in these hybrids capable of inhibiting the blastogenesis that occurs when lymphocytes of the two parents are mixed in vitro (Phillips et ul., 1971).The nature of transplantation tolerance is further complicated by the suggestion that activated specific suppressor T cells are involved in the maintenance of transplantation tolerance in some models (Hilgert, 1979).
B. CENTRAL UNRESPONSIVENESS In contrast to peripheral inhibition, central immunologic unresponsiveness is characterized b y an immune state in which the host is totally incapable of reacting specifically with the tolerated antigen (Weigle et ul., 1974a). No specific binding cells are detectable, and no antibody producing cells appear even transiently. The cellular and subcellular events involved i n this type of tolerance are probably identical to those at play in tolerance to self. Suppressor cell activity may be concomitant, but not responsible. Antigen blockade is not involved, a i d lymphocytes transferred from the tolerant donor to a neutral host remain unresponsive. Central unresponsiveness can be induced in adult animals with either nonimniunogenic fornis of the antigen or after temporary inhibition of the immune system, but is more easily and effectively induced before the immune system matures. 1 . E vperimental Model
The experinieiital models that best represent tolerance to self are induced in neonates given heterologous serum proteins and adults
168
WILLIAM 0. W I G L E
given deaggregated IgG (Weigle, 1973). Dresser (1962) was the first to demonstrate that heterologous IgG deaggregated by ultracentrifugation not only loses its ability to induce an immune response in adult mice, but also causes unresponsiveness to subsequent injections of immunogenic preparations of the same IgG. Apparently, commercial preparations of heterologous IgG owe their antigenicity to the presence of small amounts of aggregated IgG. The monomeric material remaining after deaggregation by ultracentrifugation, chemical precipitation, or biologic filtration is no longer immunogenic. Nevertheless, such cominercial preparations can be transformed into good immunogens by heating at 63°C for 25 minutes, which aggregates the IgG. Human IgG (HGG) is often used to induce central unresponsiveness, and, like many other heterologous and homologous serum proteins, after injection it rapidly equilibrates between intra- and extravascular fluid spaces and persists in the circulation until it is slowly eliminated by the normal catabolic processes of the host. In experimental animals, HGG equilibrates 50% in the intravascular spaces and 50% into the extravascular spaces (Nakamura et al., 1968) and is eliminated from mice with a half-life of approximately 6 days. Unlike particulate antigens and protein-hapten conjugates, HGG readily comes in contact with all antigen-reactive cells for a prolonged period of time. A single injection of 2.5 mg of deaggregated HGG (DHGG) (centrifuged at 150,000 g for 150 minutes) readily induces a complete and lasting unresponsive state in adult N J mice as evidenced by their failure to respond to a subsequent injection of heat-aggregated HGG (AHGG). Both the T and B lymphocytes become tolerant (Chiller et al., 1970) although the duration of tolerance differs in the two cell types (Chiller et ul., 1971), and this tolerance is maintained in cells transferred to irradiated, syngeneic hosts. Antigen-binding cells (ABC) can be detected in the spleens of mice shortly after injection of the tolerogen (DHGG) (Louis et al., 1973a). Such binding cells disappear within 12 hours after injection and reappear only when tolerance is lost in the B cells. The loss of ABC can best be explained by deletion of antigen-specific B cells. During the period of tolerance induction and maintenance, no antibody-producing cells (IgM or IgG) are detectable. In this system, tolerance is not maintained by suppressor cells or by antigen blockade.
2 . Suppressor Cells Suppressor cells are at times associated with the tolerant state to HGG, but they are not obligatory for either the induction or maintenance of tolerance (Parks and Weigle, 1979).I n mice tolerant to HGG,
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
169
suppressor activity has been observed by some workers (Basten, 1974; Benjamin, 1975; Basten et ul., 1975),but not by others (Chiller et al., 1974; Zolla and Naor, 1974; Fujiwara and Kariyone, 1978. Doyle et nl. (1976)observed suppressor cell activity in the spleens of mice injected 10 days previously with DHGG, but such activity disappeared by 40 days after injection, although complete tolerance remained. Furthermore, complete suppression required large numbers of tolerant cells. Basten et nl. (1975) also demonstrated transient suppressor activity in spleen cells of mice tolerant to HGG when large numbers of these cells were transferred with normal cells to irradiated recipients. The degree of either tolerance or suppressor cell activity depended markedly on the method of deaggregation, and these authors concluded, as did Doyle et nl. (1976), that suppression is not a requirement for maintenance of the tolerant state to HGG. Interestingly enough, although DHGG made from a variety of commercial preparations of HGG usually induces varying degrees of specific suppressor T cell activity, DHGG of a single, normal donor or from patients with myeloma induces a tolerant state without suppressor T cells. The adjuvant activity of colchicine is now known to result from its preferential inactivation of suppressor cells (Shek et al., 1978). Injection of colchicine, along with preparations of DHGG that normally induce specific suppressor T cells, inhibits the generation of suppressor activity, but such injections do not interfere with the duration of tolerance (Parks et al., 1979). Colchicine also prevents the activity of specific (HGG) suppressor T cells after they are generated.
3 . Kinetics of lnduction and Spontaneous Termination
A solid and completely tolerant state to HGG can be induced in thymus cells, bone marrow cells and peripheral T and B cells; however, the kinetics of induction and spontaneous loss may differ (Table I). The induction of tolerance in the intact mouse takes 4-5 days for completion, albeit it is virtually (75%) complete 12 hours after injection of the tolerogen (Chiller and Weigle, 1971).Induction of tolerance in either thymus cells or peripheral T cells is also rapid and parallels the kinetics of induction of tolerance observed in the intact animal (Chiller et al., 1971); peripheral B cells are only slightly slower to assume the tolerant state. Conversely, there is a latent period of 8-15 days after injection of the tolerogen before tolerance is noticeable in bone marrow cells, and the tolerant state is not complete in these cells until day 21. Of more importance to self-tolerance is the marked difference in the kinetics of the spontaneous termination of the tolerant state in peripheral T and B cells. Peripheral T cells, like intact mice, remain tolerant for 100-150 days although peripheral B cells return to com-
170
WILLIAM 0. W I G L E
TABLE I TEhfPORAL PATTERNS OF IMMUNOLOGIC UNRESPONSIVENESS TO (HGC) IN N JM I C E "
HUMANy-GLOBULIN
Days of Site
Induction
Thymus Bone niarrow Spleen: T cells B cells Whole an i nial
8-15 1 2-4 1
1
Maintenance
120-135 40-50 100- 150 50-60 130-1 50
(' Injected with 2.5 mg of deaggregated HGG on day 0.
plete competency between 50 and 60 days after injection of tolerogen. Using a different model, Rajewsky and Brenig (1974) reported similar kinetics in the spontaneous terniination of acquired tolerance in T and B cells. Thus, during a period late in the unresponsive state to HGG, T cells apparently remain tolerant but B cells become competent-a condition that permits the tolerant state to be circumvented or terniinated by bypassing either the need for, or the specificity of, T cells. Thus, the immune status of T and B cells in this model of central unresponsiveness readily lends itself to investigation of the cellular events involved in both the termination of immunologic tolerance and certain autoimmune phenomena. 4 . Dose Response
A situation in which tolerant T cells coexist with competent B cells can be established with low doses of tolerogen. The dose of DHGG required to induce tolerance in adult thymus cells is 100-1000 t'imes less than that required to induce tolerance in adult bone marrow cells (Chiller et ti/., 1971). Similarly, doses required to induce tolerance to bovine serum albumin (BSA) are considerably less in thymus cells compared to hene marrow cells (Katsuraet ul., 1972) and in peripheral T cells compared to peripheral B cells (Rajewsky and Brenig, 1974). Thus, when central unresponsiveness is induced with small doses of antigen, B cells remain competent, while T cells become tolerant. Similar dose response effects most likely apply to self antigens; antigens present in low concentrations in the body fluids would be expected to induce tolerance only in T cells, and antigens in high concentration should induce tolerance in both T and B cells. As before, when T cells are tolerant and B cells are competent, termination of the
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
171
tolerant state involves bypassing either the need for, or specificity of, T cells and the B cells activated by these self determinants would produce autoantibody, possibly resulting in autoimmune disease. *5. Circunivention of Ceiitral Uiire.Epoii.siveiiessi n T Cells
As mentioned above, tolerance either induced to foreign antigens or acquired to self components is susceptible to circumvention (or termination) by several maneuvers that activate competent B cells in the presence of tolerant T cells. The B cells may then be triggered either nonspecifically by activated T cells or directly in the absence of T cells. An obvious and effective maneuver to bypass tolerant T cells is with normal or activated T cells and possibly with thymus replacing factors. Nonnal thyinus cells are capable of circumventing the tolerance to HGG in niurine spleen cells, but such circumvention is dependent on the immune status of the B cells. That is, one overcomes the tolerant state by injecting spleen cells of tolerant mice along with normal thymus cells into irradiated recipients, provided the spleen cells are taken at a time (81 days after injection of tolerogen) when tolerance resides in the T cells, but not the B cells. Conversely, irradiated recipients are not reconstituted when normal thymus cells are injected along with spleen cells obtained at a time (17 days after injection of tolerogen) when both T and B cells maintain tolerance (Chiller and Weigle, 1973a). Similarly, Benjamin (1974) terniinated the tolerance to BSA induced in neonatal rabbits by injecting normal sibling thymocytes prior to challenge with BSA, provided that the thymocytes were injected at a time when B cells were competent. Factors isolated from activated T cells that are capable of causing differentiation of B cells (reviewed in Watson et a]., 1979) would also be expected to circumvent T cell tolerance and activate competent B cells. Injection of allogeneic cells in an appropriate temporal relation to injection of antigen dramatically enhances the immune response (Katz et al., 1971), presumably because of nonspecific activation of host T cells via a graft-versus-host reaction. Such activation of T cells by allogeneic cells terminates tolerance of rats to sheep red blood cells (McCullagh, 1970). Similarly, the tolerant state induced to HGG in ) but adult B6A F, mice can be circumvented with allogeneic ( N J cells, only if injected at a time when B cells are competent (Weigle et al., 19741)). N o effect is observed if the allogeneic cells are injected when both T and B cells are tolerant. These results suggest that tolerant T cell populations are nonspecifically activated by allogeneic cells yielding allogeneic factors that s t i n d a t e competent B cells to differentiate
172
WILLIAM 0. WEIGLE
and synthesize antibody to HGG. Although thymus replacing factors, which are capable of activating B cells in the absence of T cells, have not been used to circumvent T cell tolerance in the past, these substances seem to be likely candidates and may have both conceptual and practical importance. Of special interest is the ability of polyclonal B cell activators to cause B cell differentiation in the presence of tolerant T cells. Although most, if not all, polyclonal B cell activators probably possess this property, investigators studying circumvention of tolerance with polyclonal activators generally use LPS of gram-negative bacteria (Chiller et al., 1974). Besides being an excellent adjuvant for antibody production (Landy and Baker, 1966; Rudbach, 1971), LPS is a mitogen for B cells (Gery et ul., 1972; Anderson et at., 1972)and can substitute for the T cell helper function in animals devoid of specific T cells (reviewed in Coutinho and Moller, 1975), suggesting that LPS may bypass the need for specific T cells by otherwise TD antigens. Lipopolysaccharide also interferes with the induction of tolerance (reviewed in Louis et al., 197313)and circumvents the tolerant state when tolerant T cells are present along with competent B cells (Chiller and Weigle, 1973b). Although the mitogenic effect on B cells probably results from direct stimulation by LPS, polyclonal activation and adjuvanticity involve ancillary cells. A number of studies suggest that the ability of LPS to act as an adjuvant requires T cells as well as B cells (reviewed in McGee et al., 1979), and more recently, a requirement for macrophages has been suggested (McGee et al., 1979). The optimal polyclonal activation of B cells also requires T cells (Goodman and Weigle, 1979). If T cells are also needed by LPS to circumvent tolerance to HGG, the T cells are nonspecific because LPS activates HGG competent B cells in the presence of T cells tolerant to that antigen (Chiller and Weigle, 1973b). Again, for LPS to circumvent tolerance to HGG, competent (day 143) B cells must be present. When both the T and B cells are tolerant (25 days after receiving tolerogen), injection of LPS and antigen (AHGG) does not yield anti-HGG. In addition to LPS, other microbial agents cause polyclonal activation of B cells (reviewed in Ortiz-Ortiz et al., 1980). Furthermore, parasitic infections can cause in vivo polyclonal activation in experimental animals (Hudson et ul., 1976; Freeman and Parish, 1978; Kobayakakawa et al., 1979; Ortiz-Ortiz et al., 1980). Epstein-Barr virus is also a polyclonal B cell activator in humans (Rosen et al., 1977; Luzzato et al., 1977; Slaughteret ul., 1978). The implication of such B cell activators in autoimmunity will be considered later in the review. Termination of specific immunologic tolerance by immunization
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
173
with either altered preparations of the tolerated antigen or with antigens that cross-react with the tolerated antigen has been well documented by numerous investigators during the past 20 years (reviewed in Weigle, 1973). That the tolerance to HGG in mice with tolerant T cells and competent B cells could be terminated by altered preparations of HGG is a more recent finding (Yamashita et d,1976). Here, helper T cell activity for the hapten p-aminobenzoic acid (PAB) was induced in normal mice immunized with PAB conjugated to the self antigen niurine IgG (MIgG), but not with PAB conjugated to the foreign antigen, HGG. Thus, in mice tolerant to HGG at the T cell level, but not at the B cell level, immunization with PAB-MIgG generated PAB-specific helper T cells. Injection of such T cell primed mice with PAB-HGG resulted in activation of the competent B cells accompanied by production of anti-HGG. The circumvention of tolerance with altered or cross-reacting antigens is better documented in the experiments with neonatal rabbits tolerized to BSA. Neonatal rabbits injected shortly after birth with 500 mg of BSA remain completely tolerant to subsequent injections of BSA for at least 6 months (Weigle, 1973). However, this tolerant state readily terminates after immunization with chemically altered preparations of BSA (Weigle, 1961) or heterologous albumins that cross-react with BSA (Weigle, 1962; Benjamin and Weigle, 1970) as long as T cells are tolerant and B cells are competent. In related studies, Fujiwara and Fujiwara (1975) mixed B cells from mice tolerant to BSA with T cells primed to the sulfanil hapten, transferred the mixture to irradiated recipients and found that recipients made antibody to BSA when injected with sulfanil-BSA. Paul et nl. (1969), who studied DNP-BSA immunization of BSA-tolerant rabbits, showed that rabbits made tolerant with small amounts of BSA produced an anti-BSA response, but rabbits made tolerant with large amounts failed to respond. This result could best be explained by the failure to induce tolerance in the B cells with the small amount of BSA. Undoubtedly the same mechanism(s) enables both altered and cross-reacting antigens to terminate tolerance, but the latter system is more informative and easier to interpret. Rabbits immunized with aqueous preparations of heterologous, cross-reactive albumins 3 months after the induction of tolerance to BSA lose their tolerance and produce circulating antibodies to the heterologous albumin that also react with BSA. The antibody directed to BSA in the tolerant rabbits is quantitatively and qualitatively the same as the antibody produced in normal rabbits injected with this albumin (Benjamin and Weigle, 1970). These observations can best be explained by the presence of a
174
WILLIAM 0. WEIGLE
normal complement of precursor B cells and imply that T cells, but not B cells, are tolerant. If the spontaneous termination of tolerance induced in T and B cells of neonatal rabbits is kinetically similar to that in adult mice injected with DHGG, undoubtedly the rabbits' T cells would be tolerant at the time of immunization (90 days of age) with cross-reacting antigen, yet the B cell population would be immunocompetent. Thus, one could explain termination (or circumvention) of the tolerant state as a bypass of the need for specific T cells. In this situation the unrelated determinants on the cross-reacting albumins would activate T cells, permitting stimulation of B cells competent for both BSA and the unrelated determinants. The B cells would then produce a normal amount of antibody reactive with BSA. To further test this suggestion, rabbits tolerant to BSA were injected at 90 days of age with complexes formed between BSA and guinea pig anti-human serum albumin (HSA) (Habicht et d.,1975). This combination permits the unrelated antigen (guinea pig y-globulin) of the complex to activate T cells and allows the free BSA determinants (not covered with anti-HSA) to stimulate the BSA-competent B cells. As expected, injection of such complexes resulted in the production of antibody to BSA. The Competence of B cells in rabbits 90 days after neonatal induction of tolerance to BSA is further demonstrated by the ability of native BSA injected along with cross-reacting albumins to inhibit the termination of tolerance (Weigle, 1964). Most likely this occurs because the simultaneous injection of BSA with cross-reacting albumins reinduces a tolerant state in the competent B cells before they become stimulated. In support of this, Paul et (11. (1969) demonstrated that tennination of tolerance to BSA in rabbits with dinitrophenyl (DNP)-BSA depended on the temporal relationship between the last injection of tolerogen and the DNP-BSA challenge, suggesting that persisting tolerogen maintained tolerance in the B cells. Benjamin and Hershey (1974) also showed that tolerance induced to a fragment of BSA obtained by cleavage with cyanogen bromide could be terminated b y immunization with intact BSA. All the above results seem to stem from a bypass of specific T cell tolerance, so that competent B cells are triggered to produce circulating antibody. 6. Role of Antigen Blockade Although it is accepted that a central unresponsiveness can be induced readily in T cells, it has been suggested that the mechanism of B cell tolerance is one of antigen blockade (reviewed in Fernandez et al., 1979). This issue was addressed in the HGG-mouse model by using LPS as a probe for B cell competence. In mice whose T cells are
AUTOIMMUNITY:
d
THYROIDITIS AND ENCEPHALOMYELITIS
I
I
10
20
30
1
1
I
40 50 60 Oars Fallowing Tolerization
I
1
70
80
175
80
FIG. 1. Kinetics of the spontaneous reacquisition of responsiveness i n splenic B cells. At various tinies after tolerization, mice were challenged with 400 pg of aggregated hunian y-globulin (AHGG) and/or SO p g of lipopolysaccharide (LPS).The percentage of unresponsiveness was calculated from the indirect (IgC) plaque-fomiing cell response to HGC in tolerant and age-niatched control mice at each time point after challenge. Control mice received AHCC alone. [Reprinted from Parks and Weigle (1980b).]
tolerant, competent B cells can be stimulated to antibody production by injecting LPS with AHGG. Therefore, such injections were used to monitor the presence of competent B cells after prior injection of a tolerogeiiic dose of DHGG (Fig. 1). A/J mice injected with 2.5 mg of DHGG did not respond to combined injections of AHGG and LPS until 40 days after injection of the tolerogen, suggesting that T and B cells were both irreversibly tolerant during this interval (Parks and Weigle, 1980a). Subsequently, the gradually increased response to AHGG-LPS indicated a gradual loss of B cell tolerance. Although these data are compatible with a central unresponsiveness in B cells during the 40-day period after tolerization, sufficient tolerogen may have been present in the circulation to maintain an effective antigen blockade, even in the presence of LPS. To eliminate this possibility, spleen cells removed from the tolerant host on days 10, 30, and 60 of the study were washed extensively and transferred to irradiated hosts, which were subsequently imniunized with AHGG-LPS and tested for B cell responsiveness. Although spleen cells removed from tolerant donors on day 60 responded in the irradiated recipients, spleen cells removed on day 10 or 30 failed to respond (Table 11). Furtherniore, spleen cells from mice injected 10 days previously with tolerogen were removed, washed, and treated with pronase (Parks and Weigle, 1980b). Such treatment removed all sIg and therefore all receptorbound tolerogen. However, these cells also failed to make an antiHGG response when transferred to irradiated hosts subsequently in-
176
WILLIAM 0. WEIGLE TABLE I1 IRREVERSIBLE UNRESPONSIVENESS IN ADOFTIVELY TRANSFERRED TOLERANTSPLEENCELLS“ PFC per loRto HCG‘ Days following tolerization
Donor cellsb Tolerant Tolerant Tolerant Normal Normal Normal
Challenge of recipient‘
10
30
60
AHGG AHGG + LPS LPS AHGG AHGG + LPS LPS
251 <1 <1 501 2 111 6+6
<1
3 t 2 <1 156 5 23 303 2 80 <1
554 43 t 10 2 t 2 235 46 812 ? 255 <1
*
Reprinted from Parks and Weigle (1980b). reconstituted with either normal spleen cells or spleen cells from tolerant mice that had received 2.5 mg of deaggregated human yglobulin (DHGG) 10, 30, or 60 days earlier. Reconstituted mice received 400 p g of aggregated HGG (AHGG) and/or 50 pg of lipopolysaccharide (LPS) on the day of cell transfer. Each group represents 5-7 mice. “ T h e mean t’SE of the number of indirect plaque-forming cells (PFC) to HCG assayed 7 days after challenge. ‘I
* Irradiated (900 R) recipients were
jected with AHGG-LPS. Thus, it seems that the tolerant state induced in A/J mice is one of central unresponsiveness in both the T and B cells, not the result of either suppressor cell activity or antigen blockade.
7 . Cell Phenotype Required for B Cell Tolerance The induction of tolerance to antigenic substances clearly depends on the degree of a host’s immunocompetence in that tolerance is more readily induced when immunocompetence is suppressed-for example, before maturation of the immune system (reviewed in Weigle, 1973). However, the basic cellular requirements for tolerance induction during early and adult life have only recently been pursued. Sidman and Unanue (1975) reported that the addition of anti-Ig to cultured spleen cells from both adult and neonatal mice resulted in the cessation of Ig synthesis. However, the inhibition was reversible only in the adult cells. Similarly, Strayer et al. (1975) induced tolerance to phosphorylcholine by injecting neonatal and adult mice with antiidiotypic serum. Again, anti-Ig (idiotypic) administration led to reversible tolerance in adults and irreversible tolerance in neonates. Teale et al. (1979)also reported preferential induction of tolerance in a popu-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
177
lation of immature surface Ig negative (sIg-) B cells. Additionally, Nossal and Pike (1978) preferentially induced tolerance in an immature population of bone marrow cells. Other investigators demonstrated that splenic B cells from neonates are much more susceptible than those from adult mice in terms of tolerance induction to TD antigens (Metcalf et al., 1979; Kettman et al., 1979; Scott et ul., 1979), although neonatal and adult spleen cells responded similarly to TI antigens (Canibier et ul., 1977a). Differences in the ease of inducing tolerance in neonatal and splenic B cells have recently been associated with sIg markers on the B cells. That is, during ontogeny, the splenic B cells of mice express IgD later than IgM, and simultaneously their responsiveness to T D antigens significantly increases (Vitetta et ul., 1975). This occurrence suggested to Vitetta and Uhr (1975) that IgD serves as a triggering isotype for B cells, whereas IgM functions as a tolerizing receptor. Subsequent studies that compared tolerance induction in neonatal and adult cells suggested that the major precursor for TD responses is a p+6+ cell that appears late in ontogeny and is resistant to tolerance induction, whereas a p+ cell is the major precursor for TI responses and is highly susceptible to tolerance induction (Vitetta et al., 1977). This hypothesis gained further support when tolerance was induced with ease in adult B cells after removal of sIgD by treatment with papain; this procedure digests sIgD, but not five other classes of surface antigens (Cambier et al., 1977b). Furthermore, it was reported that the removal of IgD, but not IgM, by antibodyinduced capping increased the susceptibility to tolerance induction to TD antigens in splenic B cells from adult mice (Vitetta et d.,1977). The obvious hypothesis arising from these studies has been challenged by studies with adult splenic B cells enriched b y processing in the fluorescence-activated cell sorter (Layton et ul., 1979a,b). These studies implied that (a ) the TI response is not restricted to sIgD-. cells; (b) the response to a T D antigen is not restricted to the sIgD+subset of B cells; and ( c ) resistance to tolerance induction to T D antigens does not seem to be caused by acquisition of sIgD. Thus, the phenotype of B cells that marks either susceptibility or resistance to tolerance induction is not yet clear. Additionally, the ability to induce a central tolerance to HGG in splenic B cells by injecting adult mice with DHGG further complicates the determination of which B cell phenotypes are susceptible to tolerance induction. Szewezuk and Siskind (1977) reported that tolerance to bovine IgG was induced with equal ease in vivo or i n vitro with 17-day-old fetal liver, neonatal liver, 8-day-old spleen, adult spleen, and adult bone marrow cells of mice. This conclusion is difficult to reconcile in view of the observation showing the
178
WILLIAM 0. WEIGLE
preferential induction oftolerance in immature and neonatal B cells. It may be that tolerance is induced only in immature B cells of adults injected with nonimmunogenic tolerogens such as DHGG. The more mature B cells may escape actual tolerance induction, but are not triggered because the tolerogen is in a nonimmunogenic form. Because of a short half-life of mature, unstimulated B cells and their small clone size to the determinants on HGG, these more mature B cells may not effectively contribute to a response in mice in which tolerance has been induced in the more immature B cells. IV. Relationship between Experimentally Induced Tolerance a n d Self-Tolerance: Implications in Autoimmunity
Since the mechanisms involved in experimental models of central unresponsiveness appear to be the same as those in tolerance to self antigens, the implications of these models in mechanisms of selftolerance and autoimmunity are of both practical and theoretical interest. Although autoimmune disease may involve abnormalities in any phase of the complex regulatory system involved in the control of immune responses, events instrumental in initiating autoimmunity are probably dictated by both the manner in which self antigen is presented to the immune system and the immune status of T and B cells in regard to that self antigen. The importance of cellular events in the induction, maintenance, and termination of a central unresponsive state to foreign antigens must be understood in order to conceptualize accurately self-nonself recognition and autoimmune responses. This section of the review deals with relating cellular events involved in well defined models of central unresponsiveness with those putatively responsible for several established models of experimental autoimmune diseases. In developing a cellular model of self-tolerance, one can safely assume, first, that self-tolerance results from a central unresponsive state rather than from peripheral inhibition; second, that self-tolerance is dependent on the concentration of the self artigen in the microenvironment of potential self-reactive cells, not on the nature of the antigen; and, third, that the concentrations of self antigen required to induce tolerance in T and B lymphocytes differ markedly. Thus, the immune status of T and B cells to the self antigen(s) in question may dictate the immunologic pathway of a particular autoimmune response. A high degree of tolerance to self antigens, such as serum albumin, may be present in both the T and B cells, although with other antigens (certain classes of Ig, growth hormone and thyrogloublin) a
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
179
FIG.2. Relationship among levels of self proteins in body fluids, immune status of T cells and B cells to self proteins, and autoimmunity.
high level of tolerance may exist in the T cells while B cells are competent. For still other antigens (basic protein of myelin, cytochrome c, and idiotypic determinants), both T and B cells may be competent (Fig. 2). In the case of tolerant T cells and competent B cells, the B cells can also be triggered by procedures that bypass either the need or specificity of helper T cells. When neither cell type is tolerant both can be activated specifically when self antigen is presented in an effective manner. Competent B cells are also susceptible to activation by polyclonal B cell activators. A.
POLYCLONALACTIVATION
When self antigen is present in low concentrations, B cells with receptors with reactivities ranging from low to high avidity escape tolerance induction and are competent for this autologous antigen. Some self antigens present in high concentration may be able to maintain a tolerant state in both T and B cells having receptors of high to moderate affinity for the antigen; however, even this large amount of antigen may be unable to maintain tolerance in those B cells with antigen-reactive receptors of low affinity, since tolerance affects only B cells with the higher affinity for antigen (Theis and Siskind, 1968).On the other hand, the affinity of these B cell receptors for self antigens is often inadequate to trigger differentiation and antibody synthesis. However, polyclonal (B cell) activators could trigger such B cells to
180
WILLIAM 0. WEIGLE
produce antibody of low avidity. The ability of isolated microbial products as well as microbial, parasitic, and viral infections to cause polyclonal activation in uiuo is well documented, and often autoantibody has been detected as well. Injection of bacterial LPS into mice causes both a release of DNA into the circulation and subsequent production of circulating anti-DNA (Izui et al., 1977a). However, anti-DNA production following LPS injection can also occur without the release of large amounts of DNA (Izui et al., 1977b). Induction of autoimmunity by polyclonal activation is further suggested by reports that injection of LPS into mice results in production of rheumatoid factor (Izui et al., 1979a) and thymocytoxic autoantibodies (Izui et al., 1979b), and if LPS is injected with thyroglobulin (Tg), the result is production of anti-Tg and thyroid lesions (Esquivel et al., 1977). Production of anti-DNA is also induced in mice after injection of purified protein derivative (PPD) of tubercle bacteria (Izui et al., 1977c), another potent polyclonal B cell activator. Autoantibodies have also been reported to occur following in uitro polyclonal activation of mouse spleen cells (Femandez et aE., 1979). In addition to LPS and PPD, other bacterial products that act as polyclonal B cell activators in uitro include: protein A from Staphylococcus aureus, polymerized flagellin froni Salmonella adelaide, polysaccharides from Corynebacterium paruum, Klebsiella pneumoniae, Streptococcus diplococcus, and extracts of Nocardia opaca, N . brasiliensis, and Actinomyces uiscosus (reviewed in Ortiz-Ortiz et al., 1980). Whether any of these bacterial products also activate polyclonal responses in B cells to self antigens is unknown, although infection with certain gram-negative organisms has been associated with subsequent autoimmune disease (Larson, 1976). For example, autoantibody and disease involving the thyroid and eyes is readily induced in SMA mice injected with the polysaccharide of K . pneumoniae and the respective organ tissue (Nakashima et al., 1977).These observations and the aforementioned findings with bacterial B cell mitogens do not imply that the immunologic consequences of bacterial infection are always autoimmune disease, but rather suggest that, with appropriate temporal exposure of self antigen and bacterial products, the homeostatic balance of self-tolerance could be altered. Protozoan parasites have the capacity to induce polyclonal B cell activation in uiuo characterized by the spontaneous appearance of plaque-forming cells (PFC) and antibody to numerous antigens unrelated to the parasites (reviewed in Kobayakawa et al., 1979). An example is African trypanosomiasis, the hosts of which produce large quantities of predominantly IgM antibody (Houba et al., 1969) specific for
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
181
antigens other than the trypanosome antigens (Freeman et ul., 1970), suggesting a nonspecific polyclonal B cell activation (Greenwood, 1974). Autoantibodies of various types have also been noted in humans with this disease. In mice infected with Trypanosoma brucei, the appearance of numerous IgM-producing cells (Hudson et ul., 1976) and blastogenesis, indicated by uptake of t3H]thymidine (Corsini et al., 1977), also suggest B cell activation. More recently, autoimmune responses to DNA, red blood cells, and thymocyte antigens have been observed in mice infected with T. brucei, and the kinetics of these reponses parallel the development of polyclonal B cell activation in vivo (Kobayakawa et al., 1979). Furthermore, T. cruzi induces a polyclonal response to syngeneic erythrocytes in the infected mice when aged, but not fresh, indicator cells are tested (Ortiz-Ortiz et ul., 1980). All the incidents of polyclonal activation in uitro are explainable in terms of competent B cells possessing low-affinity receptors to self antigens. How often, or whether, such polyclonal activation occurs in uivo is unclear, but it may be in part responsible for low levels of autoreactive antibody found in the sera of normal individuals. Even if this is antibody to antigens responsible for vital biologic functions and the antibody is plentiful, it would rarely have a significant effect because of its negligible avidity and generally low levels of reactivity. The possible exception may be rheumatoid factor in patients with rheumatoid arthritis. The Epstein-Barr (EB) virus has been associated with rheumatoid arthritis in that it is instrumental in producing nuclear antigen reactive with antinuclear antibody (Alspaugh et a1., 1978). It was suggested that E B virus may be an etiological agent for this disease. As stated, EB virus is a potent polyclonal activator of human B cells in uitro (Rosen et al., 1977; Luzzato et d., 1977; Slaughter et al., 1978), resulting in the production of IgM antibody to HGG. Although the avidity of the antibody is low, once the HGG-anti-HGG (IgM) complexes form they are relatively stable and biologically active and are capable of activating the complement pathway. In all likelihood, low avidity antibody is produced to many other self antigens as the result of polyclonal activation, but an autoimmune disease does not ensue because the antibody is so low in affinity or the target antigen is sequestered. Furthermore, most polyclonal activation in uivo may be transient and disappear with elimination of the polyclonal activator, before clinically detectable tissue damage occurs. Polyclonal activation of B cells in the NZB mice has been suggested as a possible contributing factor to the spontaneous autoimmune disease in this strain. The finding that B cells appear to be excessively activated very early in life in NZB mice and their F, hybrids has
182
WILLIAM 0. WEIGLE
suggested a primary role for B cells in their autoimmune disease (Moutsoupoulas et al., 1977; Cohen and Ziff, 1977; Izui et al., 1978; Weigle and Parks, 1978). This spontaneous polyclonal hyperresponsiveness of the IgM subclass peaks early in life and then disappears with age (Weigle and Parks, 1978).Neither the nature of the polyclonal activator nor the role of T cells in contributing to such B cell activation is known at present. B. INDUCTION O F AUTOIMMUNITY BY BYPASSING T CELL TOLERANCE When the coiicentration of self antigen is sufficiently low in the body fluids, tolerance may exist only in the T cells, while B cells are competent (Fig. 2).This status o f T and B cells relative to a self antigen could lead to an autoimmune phenomenon by any manipulation that activates B cells in the presence of tolerant T cells (Weigle, 1971). A similar model has been presented independently by Allison ( 1971). Activation of B cells in this model could occur ( a ) by nonspecific thymus replacing factors resulting from nonspecific activation of T cells; ( b )by direct polyclonal activation of B cells; or ( c ) by circumventing the specificity of T cells with either cross-reacting antigens or altered self antigens. In the latter situation, self components may become altered by somatic mutations as postulated by Burnet (1977), by microbial or viral infection or b y trauma. Somatic mutation would represent persistent stimulation and the autoimmunity would progress, whereas the latter stimuli are likely to be transient. Possible examples of progressive autoimmune disease are thyroiditis and myasthenia gravis, whereas hemolytic anemia, thrombocytopenia purpura, and D-penicillamine-induced autoimmunity exemplify transient and reversible diseases. Even when a high degree of tolerance to self antigen is present in the T cell, if there is leakage in T cell tolerance, an effective level of T cell activation may still develop when an enhancing agent is present. Such may be the case when water-in-oil adjuvants containing large numbers of mycobacteria are required to elicit a given response.
c. ACTIVATION OF COMPETENTT CELLS When the concentration of self antigen is extremely low, complete tolerance is not present in either the T or B cells. In such instances, effective exposure to a self antigen may activate both T cells and B cells, resulting in a typical TD antibody response. In addition to helper T cells, cytotoxic T cells and possibly suppressor T cells can be activated. However, autoimmunity to these antigens does not usually develop because their concentration is too low. Even if sequestered
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
183
self antigens are released in high concentrations into the microenvironment of self reactive lymphocytes as a result of infection or other trauma, the response is transient and probably disappears before clinical symptoms are generated. Only when antigen is presented so as to persist in an immunogenic form, i.e., incorporated into complete Freund’s adjuvant (CFA), that a progressive autoimmune response accompanied by disease is observed experimentally. The classical example of a disease following this pathway is experimental allergic encephalomyelitis (reviewed in Paterson, 1977). D. SUPPRESSOR CELLS IN AUTOIMMUNITY Subsets of T cells seem to comprise a network of regulatory cells that control the normal immune response to a foreign antigen, once initiated. Such a regulatory network similarly monitors the various parameters of autoimmunity and may be instrumental in the clinical progression of autoimmune diseases. However, it is not clear what role if any, inducer, suppressor, and feedback inhibitory T cells play in initiating autoimmunity. Data supporting this putative regulatory network and its involvement in generating autoimmune responses have been presented by Cantor and Gershon (1979), who used strains of mice predisposed to autoimmune disease. According to their studies, the major T cell defect of N Z B mice is the apparent absence or malfunction of the Lyt 1,2,3 T cell subset responsible for feedback inhibition. A similar but more severe autoimmune disorder characterizes another strain, the MRL mice. However, their T cell lesion has been reported to lie in the inability of Lyt 1 cells to respond to feedback regulation by Lyt 1,2,3 cells. The Lyt 1 cells from MRL mice seem to be insensitive to suppressor signals generated by Lyt 1,2,3 feedback suppressor cells. Preliminary experiments suggest that the BXSB mouse, another strain with hereditary autoimmune disease, differs from NZB or MRL mice by developing Lyt 1 cells that fail to induce Lyt 1,2,3 cells to generate inhibitory activity. A role for suppressor cells in the initiation of autoimmunity is supported by studies in an inbred strain of mouse in which suppressor cells were deleted (McVay-Boudreau and Kemp, 1979). Mice whose T cells were eliminated were later reconstituted with either Lyt 1 inducer cells or T cells of all subsets. Lyt 1-reconstituted mice, but not mice reconstituted with all the Lyt subsets, developed autoantibodies against erythrocytes, thyniocytes, and thyroglobulin and accumulated immune complexes in the kidneys. However, these results must be viewed with caution, since the mice reconstituted with Lyt 1 T cells lacked other phenotypic and functional T cell subsets that might have been obligatory for the animals’ well-being. Furthermore, other workers have failed
184
WILLIAM 0. WEIGLE
to observe a loss of suppressor cell function in autoimmune prone mice (Creighton et ul., 1979). Furthermore, congenitally athymic (nude) New Zealand Black x White F, mice develop SLE-like immunopathology at the same rate as their nu/+ littermates, although the former mice lack suppressor cells (Ohsugi and Gershwin, 1980). Suppressor cell activity has also been linked to autoimmune disease in humans. Reinherz and Schlossman (1980) observed depressed SUPpressor cell activity in patients given transplants who developed acute graft-versus-host disease and multiple autoimmune disorders, and an enhanced level of suppressor cells in a patient with acquired agammaglobulinemia. They also reported a depression in feedback regulation and thereby linked inducer cells with the generation of suppressor cells in patients with juvenile rheumatoid arthritis. At present, these data do not reveal whether T cell suppressor activity is the result or the cause of the disorders and are too preliminaiy to allow an assessment of suppressor or regulatory cells in human autoimmunity. In any event, if suppressor cells act as a deterrent of autoimmune disease, they probably are effective only with self antigens of limited concentrations in the body fluid, which do not provide solid tolerance at the T cell level. Thereby, such antigens may constantly generate subclinical responses that are accompanied by suppressor cell activity. The failure to generate suppressor cells because of abnormalities in the regulatory network could lead to unchecked immune responses to these self antigens. An important question is, then, what is the immune state of suppressor cells in respect to self antigens? Can central unresponsiveness be induced in suppressor T cells as well as in helper cells? In evaluating the role of suppressor cells in the initiation of autoimmunity, it is essential to know whether such regulatory cells are susceptible or resistant to the induction of specific immunologic tolerance. If tolerance can be induced in suppressor cells, it would be of interest to compare the dose response for the induction of tolerance in both specific helper (inducer) cells and specific suppressor cells to the same antigen. It may be, as in B cells, that suppressor T cells require a higher dose of foreign (or self) antigen to become tolerant than is required to induce tolerance in helper T cells. However, at high doses there would be no need for specific suppressor cells to control overt immune responses. V. Experimental Autoimmunity
Because nearly all the many models of autoimmunity accompanied by disease described in the literature are complex and involve either unknown or multiple self antigens, it has been difficult to study their cellular andlor subcellular mechanisms. Experimental models that fa-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
185
cilitate an assessment of autoimmune disease mechanisms involve: (a ) single, well characterized antigens that are easily isolated in a relatively pure form; ( b) the availability of assays to evaluate the immune status of T and B cells; and ( c ) relevance to human disease. Some models that fit these criteria are experimental autoimmune thyroiditis (EAT), immune complex-induced glomerulonephritis, experimental allergic encephalomyelitis (EAE), myasthenia gravis, and lensinduced uveitis. This section is concerned mainly with EAT and EAE, since they represent two contrasting cellular pathways of tissue injury, and the cellular events have, for the most part, been documented. A. EXPERIMENTAL AUTOIMMUNE THYROIDITIS (EAT) In this disorder, the autoantigen responsible for both antibody production and the disease process is, in most cases, thyroglobulin (Tg): a well characterized protein that can be obtained in rather large amounts in purified form. Assays are available for detection of both T and B cell immunity within this model, and, the cellular events responsible for at least one type of EAT are well established. Furthermore, it appears that two different models of EAT can be induced, which involve contrasting cellular mechanisms. The disorder was first produced in rabbits injected with CFA containing mycobacteria and homologous thyroid extract or T g (Rose and Witebsky, 1956). More recently, the related kinetics and dose requirements for both Tg and mycobacteria have been more accurately defined in guinea pigs (McMaster et al., 1967) and rabbits (Nakamura and Weigle, 1969). Immunization with either thyroid extracts or T g incorporated into CFA produces autoantibodies and thyroiditis in dogs (Terplan et d.,1960), guinea pigs (Terplan et al., 1960; McMaster et al., 1961), goats (Rose et al., 1964), rats (Jones and Roitt, 1961), chickens (Jankovic and Mitrovic, 1963), and monkeys (Rose et al., 1966). The lesion in the thyroid gland is characterized by a marked infiltration of mononuclear cells, including lymphocytes, macrophages, and occasionally plasma cells in the perifollicular tissue as well as diffuse infiltration of neutrophils detectable at various phases of the disease. In more severe cases, the architecture of the gland is distorted by fibrosis and necrosis of epithelial cells that line the follicles. This EAT model corresponds to the human disease, Hashimoto’s thyroidits, which is also accompanied by both circulating antibodies to autologous T g and, at times, delayed-type hypersensitivity (DTH) to Tg. 1 . Antigens
Although various antigens have been identified as autoantigens in thyroid disease, Tg is the main autoantigen associated with
186
WILLIAM 0. WEIGLE
Hashimoto’s thyroiditis and is the autoantigen used to induce EAT; Tg is the major product of protein synthesis by the thyroid gland (Shulman et (/I., 1955). After iodination and coupling of a fraction of its tyrosyl residues, Tg serves as storage for inactive thyroid hormone. Tg is synthesized and iodinated in the follicular cells and stored in the follicular lumen, where it accounts for 7 5 4 0 % of the gland’s protein content; it reenters the follicular cell in the form of colloid droplets. Here it is digested by lysosomal enzymes, and free iodothyroxine diffuses from the lysosomes and filters into the extracellular space. More complete accounts of the synthesis and secretion of Tg have been written by Shulman (1971) and Van Herle et nl. (1979). 2. Structure of Thyroglobuliii Thyroglobulin is an iodinated glycoprotein with a molecular weight (mw) of approximately 660,000 (Edelhoch, 1965) and a sedimentation coefficient of 19 S (Heidelberger and Pedersen, 1935).However, 12 S half-molecules and 27 S dimers are also detectable (Shulman et ul., 1967). The 27 S dimer apparently contains the same subunits as does 19 S Tg, but is richer in thyroxine than is 19 S Tg (Salvatore et nl., 1965). Poorly iodinated 19 S T g dissociates into 12 S subunits in the presence of denaturing agents (Edelhoch and Lippoldt, 1964). There has been considerable controversy and difficulty in arriving at the quaternary structure of Tg. For example, molecular weights of mammalian Tg protomers ranging from 25,000 to 200,000 have been reported (Charlwood et al., 1970; Pierce et d , , 1965; Vecchio et d . , 1971; Lissitzky et d . , 1968; Spiro, 1973). Such difficulties are thought to result from the heterogeneity of purified T g (Haeberli et d.,1975a; Spiro, 1973) as the result of specific or nonspecific cleavage in uiuo or as the result of the method used for isolation and purification (Rolland and Lissitzky, 1976; Haeberli et d . , 1975a). Approximately 100 disulfide I)onds are present in the Tg molecule (Edelhoch and Rall, 1964). Nissley et al. (1969) reported the isolation of two subunits from partially reduced calf Tg of 330,000 and 160,000 mw, and more recently Smith and Shulman (1978) dissociated human Tg by succinylation, isolated an 8 S subunit of 165,000 mw, and suggested that it was the quaternary unit of intact Tg. Thyroglobulin from the thyroid glands of normal guinea pigs was isolated, reduced, and analyzed, yielding molecules of 295,000,210,000, and 110,000 mw (Haeberli et d., 1975b). Apparently, the subunit composition of guinea pig Tg depends upon the degree of iodination, and the 295,000 mw species is the only one present in the absence of iodination. A 33 S RNA purified from beef thyroid membrane-bound polysomes
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
187
contained all the information required for the synthesis of a complete 19 S Tg molecule (Vassart et ul., 1975). Evidence was presented that the polypeptide encoded by the 33 S RNA would be compatible with a model of a dimeric Tg of approximately 300,000 mw. According to electron microscope analysis the 19 S Tg molecule has the shape of a flexible helix (Bloth and Bergquist, 1968). This labile protein rapidly denatures at pH 4.5 and below (Heidelberger and Pedersen, 1935; Litanjua, 1961) and more slowly denatures at pH 4.8 (Heidelberger and Pedersen, 1935). The relative instability of Tg at alkaline p H in neutral salt at temperatures above 53°C and low ionic strength solvents is also well established (reviewed in Edelhoch, 1965). This instability of Tg may have some bearing on the antigenicity of autologous or homologous Tg in the microenvironments of granulomas resulting from adjuvant injections (see below). The carbohydrate portion of Tg has been subjected to detailed structural investigation by several groups of investigators and has been shown to contain galactose, mannose, fucose, glucosamine, and sialic acid (Ujejski and Glegg 1955; Wollnian and Warren, 1961; Robbins, 1963; Spiro and Spiro, 1972).The carbohydrate content varies slightly from one species to another and ranges from 8 to 10% for human, bovine, ovine, and porcine Tg (McQuillan and Trikojus, 1966). Calf Tg contains 280 sugar residues and can be divided into two distinct classes (Arima et al., 1972). The simpler carbohydrate “A” units consist only of mannose and N -acetylglucosamine; the more complex “B” units consist of mannose, N-acetylglucosamine, galactose, N-acetylneuraminic acid, and fucose. Although human and calf Tg contain a similar number of A units, their number of B units differs b y approximately twofold. Microheterogeneity characterizes the A and B carbohydrate units of human Tg, and an additional third, or “C”, unit has been found that consists of a galactosomine residue. This distinct C unit is linked to serine and threonine residues by O-glycosidic bonds, in contrast to the glycosylamine type of linkage to asparagine that attaches units A and B to the polypeptide chain. Human Tg differs from Tg of cows, sheep, and pigs in that it has a higher total carbohydrate content, which is a function of its higher inannose and glucosamine content (Spiro and Spiro, 1965).The extensive microheterogeneity of the carbohydrate units of T g obviously contributes significantly to the microheterogeneity of the intact molecule and its peptide subunits. Removal of sialic acid from Tg with neuraminidase has been reported to enhance its immunoreactivity (Salabe et al., 1976). A number of studies deal with the subunits of Tg resulting from
188
WILLIAM 0. WEIGLE
enzymic degradation (cited in Stylos and Rose, 1977; Shulman, 1971). Peptide fragments from Tg obtained by cleavage with trypsin (Metzger et al., 1962; Mates and Shulman, 1968b; Stylos and Rose, 1969; Ghayasuddin and Shulman, 1970; Salabe et al., 1973),pepsin (Stylos and Rose, 1977; Weigleet al., 1969)and papain (Stylos and Rose, 1969) are antigenically active against heterologous and/or autologous antiTg serum. Although the trypsinized fragments are fully reactive with heterologous antisera, these fragments are antigenically deficient in reactivity with autoantibodies. Both the auto- and heteroreactive determinants are conserved in the pepsin and papain fragments. Of particular interest, partial digestion of rabbit T g with papain (Anderson and Rose, 1971) or pepsin (Weigle et al., 1969) yields fragments that are antigenic without the aid of adjuvants in the rabbit. A more complete review of the structure of Tg has been published by Shulman (1971).
3. Organ and Species Specificity of Thyroglobulin Antisera prepared in rabbits to thyroid extracts or purified Tg from heterologous species are organ-specific in that they generally do not react with sera or extracts of other tissues of that particular species (Hektoen et al., 1927; Witebsky et al., 1955).On the other hand, rabbit antisera to heterologous thyroid extracts (Witebsky et al., 1956) or to heterologous Tg (Weigle, 1967) do cross-react with other heterologous thyroid extracts or Tg, respectively, as well as with rabbit Tg. The antisera are species-specific, but cross-reactive, in that absorption with the homologous (immunizing) preparations, but not with the heterologous preparations, removes all the antibody. Autoantisera produced in rabbits against homologous thyroid extracts (Rose and Witebsky, 1959; Rose et al., 1964) or purified Tg (Rose et al., 1964; Weigle, 1967) also react against heterologous preparations, but the cross-reaction is not as extensive or diversified as with antisera to heterologous Tg. Again, the autoantisera are species-specific; they can be completely absorbed with homologous, but not heterologous, thyroid extracts. It is of interest that immunization with either heterologous thyroid extracts (Terplan et al., 1960; Rose et al., 1964; Pudifin et d.,1977) or heterologous Tg (Rose et al., 1964; Romagnani et al., 1970; Jankovic et al., 1969) in CFA not only results in the production of cross-reacting autoantibodies, but also initiates thyroid lesions; however, these lesions are less severe than those induced with homologous Tg in CFA. Of particular note, anti-autologous Tg from the serum of a patient with Hashimoto’s thyroiditis reacted only with human T g or Tg from
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
189
Old World monkeys, but did not cross-react with any other mammalian Tg, in contrast to the broad cross-reactivity of experimentally induced antibody to heterologous Tg (Baur and Goodman, 1964). The crossreactivity of anti-rabbit Tg antibody in the rabbit was similarly restricted (Weigle, 1967), as was human anti-autologous Tg in reacting with tryptic digests of human Tg only (Mates and Shulman, 1968a; Stylos and Rose, 1969). These results indicate fewer recognition sites on autologous Tg than on heterologous Tg. It is not surprising that the autologous response is restricted to a limited number of antigenic determinants, possibly as the result of different degrees of selftolerance to individual determinants. Less obvious is why the immune responses to autologous Tg are restricted to determinants specific for self. Fuller comprehension of such restriction might provide additional insight into the mechanisms of self-nonself recognition.
4 . I n Vivo Behavior of Thyroglobulin Recent evidence implies that Tg is not sequestered as originally thought; furthermore, under normal conditions Tg may be present in circulating body fluids. Although Hjort and Pedersen (1962) and Assem (1964) used rather crude assays, they detected Tg in the sera of 75% of newborn infants examined. Since then, Tg has been detected in lymph nodes that drain the thyroid glands of presumably healthy rats (Daniel et al., 1967a) and monkeys (Daniel et al., 1967b). Moreover, with radioimmunoassays, Tg is being found in the sera of normal humans (Roitt and Torrigiani, 1967) at levels equivalent to levels of heterologous Tg that are required to induce and maintain the unresponsive state in rabbits (Nakamura and Weigle, 196713). Numerous other reports describe levels of Tg ranging from 2 to 180 ng/ml of serum in normal humans (reviewed in Van Herle et al., 1979). Clearance rates of endogenous Tg from sera are also being examined with radioimmunoassays, especially as a prognostic tool in evaluating the course of thyroid cancer after thyroidectomy. In fact, increased levels of Tg in the circulation of man seem to be well established as an indicator of metastases from differentiated thyroid carcinoma (reviewed in Van Herle et al., 1979). LoGerfo et al. (1978), who used radioimmunoassays to study whole sera from patients given total thyroidectomies to treat thyroid cancer, reported an average half-life of 14 hours for Tg. Feldt-Rasmussen and co-workers (1978) studied similarly treated tumor patients but found that 19 S Tg had a halflife of only 4.3 days, and Tg of a smaller molecular weight had an even shorter half-life-nly 3.7 hours. The clearance rates of Tg in rats, dogs, and monkeys (Brown and Jackson, 1956) suggest a rapid
190
WILLIAM 0. WEICLE
catabolic rate of homologous Tg, whereas rabbits (Nakamura and Weigle, 1967a) catabolize Tg more slowly. After injection of I3'Ilabeled homologous or heterologous Tg into rabbits, both proteins equilibrate between the intra- and extravascular fluid spaces, with 70% in the intravascular fluid space and 30% in the extravascular space (Nakamura et al., 1968). Both proteins are catabolized at an exponential rate, with a half-life of 2.5 days (Fig. 3 ) . As with other serum proteins, both equilibration and the half-life are independent of the route of injection. Heterologous, but not homologous, Tg undergoes immune elimination between the fifth and seventh days after injection. Shortly after elimination of heterologous Tg, antibody to it appears in the circulation; however, antibody is never induced by injection of aqueous preparations of homologous Tg (Nakamura and Weigle, 1967a). Injecting extremely small amounts of bovine T g into rabbits periodically from birth induces specific tolerance (Weigle, 1967). The amount of bovine Tg found in the blood of these rabbits just before they are tested for tolerance is the same as that of autologous Tg in the blood of normal rabbits. Since it has been demonstrated that tolerance to small amounts of antigen is a quality of T cells, but not of B cells (Chiller et al., 1971),it seems most likely that competent precursors of antibody-producing cells to the tolerated heterologous T g are present,
Days following Injection FIG.3. Elimination of paired '311-labeledbovine (-) and '"I-labeled rabbit (---) thyroglobulin (Tg) in three representative normal rabbits. The rapid elimination of bovine Tg between days 6 and 8 is the result of the production of anti-bovine Tg antibody. [Reprinted from Nakamura and Weigle (1967).]
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
191
but competent T cells have been functionally deleted. A similar inimune status to autologous T g was postulated previously (Weigle, 1971). This suggestion seems realistic in view of the ability to circumvent either the natural tolerant state to self Tg (Weigle, 1965a) or that induced to heterologous Tg (Nakamura and Weigle, 1968b) by manipulations that bypass the need for specific T cells. 5. Role of Autoantibody mid Eflector T Cells in EAT
Despite numerous different approaches to decipher the mechanisms of EAT, considerable controversy remains concerning the roles of humoral and cell-mediated immunity in this disease. Witebsky and co-workers (Witebsky et al., 1956; Rose et al., 1962a) induced EAT in rabbits and were the first to point out an association between DTH to Tg and thyroiditis. In addition, Rose et al. (1962b) could find no correlation between levels of antibody and severity of thyroid lesions in rabbits. In contrast, Lerner et ul. (1964) and McMaster et a / . (1961) correlated thyroiditis with antibody to T g in guinea pigs injected 7 weeks earlier with homologous thyroid extract, but found that lesions observed at a later time were associated with DTH. Ringertz et (11. (1971) observed that guinea pigs injected with homologous T g in CFA developed thyroiditis, DTH, and migration inhibition factor (MIF), but not autoantibody to Tg. Flax (1963) noted a similar relationship between DTH and thyroiditis in guinea pigs, whose thyroid lesions were histologically characteristic of a DTH reaction (Flax et ul., 1963). By using a different approach, Miescher et al. (1961)found that injection of guinea pigs with homologous Tg conjugated to picryl chloride dramatically reduced the level of circulating autoantibodies, but did not affect the frequency or severity of thyroid lesions. In other studies, rabbits immunized with the homologous antigen in alum (Doebbler and Rose, 1961) and guinea pigs immunized with homologous Tg in incomplete Freund’s adjuvant (McMaster et a!., 1961) produced anti-Tg antibodies, but failed to develop thyroiditis. Although Zavalita and Stastny (1967) observed a proliferative in vitro response to homologous Tg in lymphocytes of rabbits immunized with CFA containing rabbit Tg, others failed to incite proliferation that coincided with thyroiditis in rabbits (Romagnani et ul., 1970). Little additional information Concerning the cellular mechanism of EAT has developed from examining the nature of the lymphocytes infiltrating the lesion; although 75% of the lymphocytes released from lesions in guinea pig thyroids were T cells (Paget et nl., 1976), there was no evidence that these cells were specific for guinea pig Tg. The remaining 25% of cells might have been Tg-specific non-T cells that initiated the events lead-
192
WILLIAM 0. WEIGLE
ing to infiltration of nonspecific T cells. Rose et aE. (1971) reported a correlation between thyroid lesions and both Arthus and DTH reactions, irrespective of antibody concentrations in mice. In other experiments (detailed in Section V,A,6), thyroiditis was produced in rabbits or mice by injecting aqueous preparations of heterologous Tg instead of Tg in CFA. The result was a close correlation between the thyroid lesions and humoral immunity. In rabbits, the kinetics for formation of antibody-producing cells in the thyroid and of circulating autoantibodies coincided with the appearance of thyroid lesions (Clinton and Weigle, 1972). Similarly, mice injected with soluble preparations of heterologous Tg developed thyroid lesions characteristic of the Arthus reaction parallel with the appearance of circulating antibodies to murine Tg (Clagett et al., 1974). The ability to transfer thyroiditis with either lymphoid cells or antibody has been used in the past as evidence of humoral or cellular immune events in thyroiditis (Felix-Davies and Waksman, 1961). McMaster and Lerner (1967) found sera from sensitized guinea pigs to be ineffectual in transferring thyroiditis, whereas lymphocytes from the same animals were capable of adoptively transferring thyroiditis to healthy guinea pigs. Thyroiditis was similarly transferred with cells into rats (Twarog and Rose, 1970) and rabbits (Nakamura and Weigle, 1967~).Since the transferred cells could also produce antibody cytotoxic to thyroid tissue in vitro, adoptive transfer with sensitized lymphocytes did not rule out a role for antibody. Injection of high titer autoantibody from rats (Rose et al., 1973), rabbits (Rose et al., 1962b), monkeys (Roitt and Doniach, 1958), or dogs (Anigstein et al., 1957) into animals of the respective species did not produce thyroiditis. Although numerous attempts to transfer thyroiditis with antibody have failed, an equal number of attempts have been successful. Roitt and co-workers (1961) could not transfer thyroiditis to rats with heterologous antisera to rat Tg unless the rats were first stressed by injection of CFA, injection of radioactive iodine, or irradiation. Rose and Kite (1969) injected antiserum from the rhesus monkey directly into the thyroid gland of monkeys to induce thyroiditis characterized by a mononuclear cell infiltration. Such antiserum is known to contain cytotoxic antibodies to the thyroid cell surface. Sharp et al. (1967) demonstrated thyroid lesions in guinea pigs after they were injected with heterologous antibodies. Interstitial areas of the lesions were transiently infiltrated by eosinophils, which disappeared in 2 4 weeks, but not mononuclear cells. However, repeated injections of hyperimmune rabbit antiserum to guinea pig Tg protected guinea pigs from developing thyroiditis (Sharp et al., 1974a) and, additionally, de-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
193
pressed DTH. Godal and Karesen (1967a,b) described the development of thyroiditis in guinea pigs after intraperitoneal injections of either rabbit or guinea pig antisera to guinea pig Tg. After 24 hours, a massive infiltration of eosinophils was observed. In guinea pigs receiving rabbit anti-guinea pig Tg serum, a subacute stage with niononuclear cells mainly of the histiocytic type, but also lymphocytes, was seen 5 days after injection. This was not observed in the animals receiving guinea pig anti-guinea pig Tg serum. At 10-20 days after injection, the thyroiditis resolved, leaving the thyroids nearly normal. In subsequent studies, neutrophilic granulocytes were also observed (Karesen and Godal, 1969a). Fluorescent antibody (Karesen and Godal, 1969b) and electron microscopic (Karesen and Godal, 1970) studies suggested that the mechanism involved in infiltration of neutrophils and eosinophilic granulocytes was uptake of T g antigenantibody complexes. Although these studies are of general interest in regard to the mechanism of tissue damage, they give limited insight into the role of antibody in the generation of thyroiditis, since the lesions are not typical of those produced in any of the well defined EAT models. However, Nakamura and Weigle (1969), Vladutiu and Rose (19714, and Tomazic and Rose (1975) used homologous antisera to transfer thyroiditis into rabbits and rats, respectively, and the subsequent lesions were characterized by a lymphocytic infiltration in both strains. The successful transfer of thyroiditis to healthy rabbits occurred only when sera containing anti-Tg antibody were obtained from Tgimmunized (in CFA) thyroidectomized donors soon after immunization (Nakamura and Weigle, 1969). Antisera taken after the lesions had fully developed were not effective for transferring thyroiditis. Although preliminary tests failed to detect homocytotropic antibody in the donors’ sera, these tests were not sufficiently sensitive to detect trace amounts. In any event, it may be difficult to correlate circulating antibodies and lesions, even if both are present, since the nature of this antibody has not been established, and effective antibody may be rapidly removed from the circulation (Clinton and Weigle, 1972). The many different models and experimental approaches may have confused the issue concerning requirements for humoral and DTH reactivity in EAT because thyroiditis may be initiated by either effector T cells or circulating antibody, or sometimes both. In this regard, the combined transfer of thyroid-sensitized lymph node cells and hyperimmune rabbit antiserum to guinea pig T g into inbred strain 13 guinea pigs resulted in a higher incidence and greater severity of thyroiditis than that produced by transferring serum or cells alone
194
WILLIAM 0. WEIGLE
(Sharp et al., 1974b). The lesions produced by these combined passive transfers were comparable to lesions induced by active immunization. The roles of antibody-mediated and cell-mediated immunity in the induction and progression of Hashimoto’s thyroiditis are also controversial. Specifically activated T cells are not involved in the induction of EAT after a single series of immunizations with aqueous preparations of heterologous Tg; subsequently, after repeated injections over a period of 18 months, the T cells are not only activated but produce MIF to homologous Tg (Weigle and Romball, 1975). This system might exemplify a loss of self-tolerance in T celIs as a result of persisting antibody that would explain the cell-mediated immune reactions found in some humans with autoimmune thyroiditis. Soberg and Halberg (1968) demonstrated that 12 of 15 patients with Hashimoto’s thyroiditis had cell-mediated immunity to crude thyroid extracts. Moreover, Brostoff (1970) used inhibition of leukocyte migration to show that 16 of 26 such patients had cell-mediated immunity, and Calder et a l . (1972) reported that Tg in 44% of similar patients inhibited leukocyte migration. The failure to demonstrate consistent cell-mediated immunity in these patients has not been adequately explained; however, the aforementioned experiments with rabbits suggest that some of the patients’ cell-mediated immunity may have resulted from a loss of self-tolerance in the effector T cell population. It would be of interest to know how long circulating antibody had been present before testing and how the presence or absence of antibody correlates with the failure to detect cell-mediated immunity. Peripheral lymphocytes from patients with Hashimoto’s thyroiditis are cytotoxic for Tg-coated target cells (Poclleski, 1972); however, the mechanism for cytotoxicity is undetermined. Cultured human thyroid cells have been shown to be lysed in the presence of complement and cytotoxic antibody from the sera of patients with thyroiditis (Pulvertaft et ( i l . , 1959, 1961; Irvine, 1962; Chandler et nl., 1962; Forbes et al., 1962; Kite et ul., 1965; Rose et al., 1965). However, the cytotoxic antibody effective in this assay may have been directed not to Tg, but to another thyroid antigen, e.g., cell surface antigen. Calder et (11. (1973) have correlated antibody-dependent cell cytotoxicity (ADCC) with Hashimoto’s thyroiditis in humans. This ADCC was higher when concentrations of circulating antibody were small and lower when more antibody was circulating (Calder et nl., 1973). Similarly, Pudifin et nl. (1977) reported evidence for ADCC to homologous Tg in vervet monkeys immunized with human Tg. When Ringertz et al. (1971) examined guinea pigs immunized with guinea pig Tg in CFA, ADCC was also associated with thyroiditis.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
195
The evidence for and against humoral or cellular immunity as the responsible agent for Hashimoto’s thyroiditis is as incomplete as for EAT. Although the histologic similarity of thyroid lesions in patients with Hashimoto’s thyroiditis and the lesions characteristic of cellmediated immunity suggest that the latter mechanism fosters this disease, some of the round cells infiltrating the human gland in chronic thyroiditis contain antibody to Tg, implicating a humoral mechanism (Mellors et af., 1962). The appearance of plasma cells that produce antibody to T g in thyroid gIands of Tg-immunized rabbits has also been correlated with the development of such lesions in EAT (Clinton and Weigle, 1972). Both circulating antibody and cell-mediated immunity may participate, but the role of each may vary markedly from patient to patient and may be determined by different genetic and inimunoregulatory factors. 6. Models of Antibody-Mediated EAT a. EAT in the Rabbit. The EAT experimental model is excellent for testing the postulate that low concentrations of circulating autoantigens lead to the deletion of specific T cells (tolerization) even though competent B cells are present. As with foreign antigens to which the tolerant state is present in the T cell compartment alone, maneuvers that bypass either the need for or the specificity of T cells could terniinate self-tolerance, resulting in an autoimmune response possibly accompanied b y disease. Since autologous Tg is present in the body fluids in concentrations too low for maintenance of B cell tolerance, but sufficient for maintaining tolerance in T cells, it is not surprising that inimunization with aqueous preparations of either altered autologous Tg (Weigle, 1965a,b) or cross-reacting heterologous Tg (Weigle and Nakamura, 1967) results in antibodies that react with autologous Tg. Thus, “new” (nonself) determinants present in altered-self molecules or cross-reacting Tg could activate T cells, thereby triggering competent B cells to produce autoantibody to Tg. Adult rabbits immunized with aqueous preparations of rabbit T g coupled to the diazonium derivatives of arsanilic and sulfanilic acids produce autoantibodies that react with rabbit Tg and develop thyroiditis characterized by infiltration of mononuclear cells, similar to the thyroiditis induced by injecting rabbits with native rabbit T g in complete adjuvant. Similarly, rabbits injected serially with aqueous preparations of either bovine, human, or porcine Tg produce antibody reactive with rabbit Tg and develop thyroid lesions. These three heterologous Tg preparations cross-react 25-30% with guinea pig anti-rabbit Tg (Weigle and Nakamura, 1967). Injection of a mixture of
196
WILLIAM 0. WEIGLE
TABLE 111 THEPRODUCTION OF THYROIDITIS AND ANTIBODY TO RABBIT THYROGLOBULIN (TG) IN RABBITS INJECTED WITH HETEROLOGOUS TG" Tg injected
Antibody
(PP)
Fractions with lesions
Bovine Human Porcine Mixture
76.2 56.3 12.5 120.0
14/20 7110 4/11 20122
Modified from Weigle and Nakamura (1967).
the three heterologous Tg preparations results in both a higher level of autoantibody production and a greater frequency and severity of lesions (Table 111). When these injections cease, both the lesions and antibodies disappear, although an immediate injection of rabbit Tg in aqueous form transiently produces antibody and disease. However, none of the antibody is specific for rabbit Tg. Additional injections of rabbit Tg only return the rabbits to their normal unresponsive state. Thus, the rabbit Tg seems to be incapable of recruiting the T cells required to perpetuate the immune response and competent B cells are no longer triggered. Witebsky and Rose (1959)have also induced thyroiditis in rabbits by multiple injections of porcine Tg. The kinetic patterns with which cells that produce antibody to Tg appear in the spleens and thyroids of rabbits immunized with an aqueous preparation of heterologous Tg provide additional insight into the cellular mechanisms for this model of EAT. In these studies, the Jerne hemolytic assay (Jerne and Nordin, 1963) was modified to detect PFC to Tg. Rabbits were injected with two courses of bovine Tg, and the spleens and thyroid glands were then assayed for PFC during and after the injections (Fig. 4). The peak numbers of PFC to both bovine and rabbit Tg appeared in the spleens on approximately day 24 after the initial injection, although PFC to rabbit Tg were only 5% of those to bovine Tg. Plaque-forming cells reactive with both preparations were also detected in the thyroid gland, and both responses were similar in magnitude and strongest 7 days after the major peak of PFC in the spleen. Apparently the gland acted as a specific immunoabsorbent where locally exposed Tg removed migratory memory cells that reacted with rabbit Tg after stimulation in the spleen by the cross-reacting bovine Tg. Of special interest is the direct correlation between the occurrence of PFC in the thyroid gland and
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
197
60001
a
-. . - -. ..
lIi51 6 7 I 9
15
20 2 24 26 28 30 32 t t f t t Day o f Assay
36
44
FIG.4. The plaque-forming cell (PFC) response to bovine and rabbit thyroglobulin (Tg) by the spleen (a)and the thyroid gland (b). The arrows (t) indicate injections of aqueous bovine Tg at the indicated days. [Reprinted from Clinton and Weigle (1972).]
the appearance of thyroid lesions (Fig. 5). Significant lesions did not appear before PFC were detected in the gland, and the lesions were most severe shortly after the peak number of PFC developed. Serum levels of precipitating antibody to rabbit Tg also correlated with the appearance of lesions (Fig. 6). Again, antibody preceded the lesions, and as the lesions began to appear, antibody dramatically cleared from the blood considerably faster than can be accounted for by the 5.8-day half-life of IgG in the rabbit (Spiegelberg and Weigle, 1966). Apparently the antibody was removed by Tg either released into the body fluids or exposed in the gland as the result of thyroid damage. In contrast to the production of hunioral antibody reactive with rabbit Tg, cell-mediated immunity to rabbit Tg was not detected in rabbits immunized with aqueous preparations of bovine Tg. Attempts to demonstrate the release of M I F with rabbit Tg failed; however, this factor's activity was released with bovine Tg. Similarly, Romagnani et al.
1400 1300-
-Spleen -Thyroid
1200 1100.
Day o f Assay
FIG.5. The plaque-forming cell (PFC) response to rabbit thyroglobulin (Tg) of cells and thyroid glands ( L M ) of rabbits immunized with from the spleens (A-A) aqueous preparations of bovine Tg. The degree of infiltration of mononuclear cells is shown b y the filled bars. [Reprinted from Clinton and Weigle (1972).]
IS16-
I
- 14i 12 = - 0)
2 100 OD
t I
-
I
8c M
I
-
f 6a n 0 0
c
=
z
-
4-
2-
7 8 9 1523
25
21
29 31 33 D a y o f Assay
35
31
39
41 4 4
FIG.6. Level of precipitating antibody to rabbit thyroglobulin (Tg)compared with the appearance of thyroid lesions in rabbits immunized with aqueous preparations of hovine Tg. AhN = antibody N as detected by maximum precipitation of antigen in slight antigen excess. [Reprinted from Clinton and Weigle (1972).]
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
199
(1970) failed to stimulate migration inhibition activity with homologous Tg in lymphocytes from rabbits immunized with heterologous Tg in CFA, although the inhibitory activity was generated by the heterologous Tg and mild lesions developed. The role of circulating antibodies produced by activated B cells without activation of specific T cells in the EAT model is further supported by unsuccessful in uitra attempts to stimulate proliferation of T cells from rabbits immunized with bovine or rabbit Tg in CFA. T cells from experimental animals previously immunized with antigen proliferate i n uitro ([3H]thymidine uptake) in response to that antigen (Schwartz et al., 1975). In mice, the responding T cells seem to be Lyt 1+ inducer cells and are not of the Lyt 2+ phenotype ascribed to suppressor cells (Corradin et al., 1977). Rabbits were immunized with bovine Tg; their spleen cells were removed 1-24 weeks later, and the cells were assayed in vitro for their ability to proliferate in the presence of bovine and/or rabbit Tg. The presence of thyroid lesions and circulating antibodies to rabbit Tg was also determined. Spleen cells from rabbits immunized with bovine Tg proliferated in response to bovine Tg, and T cells were shown to be the responsive cell type (Romball and Weigle, 1979; Weigle and Romball, 1980). That is, the spleen cells eluted after passage over nylon wool and devoid of Ig+ cells retained the ability to mount a proliferative response to bovine Tg (Table IV). In contrast, the addition of rabbit Tg to the cultures did not cause proliferation (with one exception), despite the fact that antibody produced by these rabbits reacted with rabbit T g and thyroid lesions were present. Spleen cells from rabbits immunized with rabbit Tg in CFA also failed to proliferate in response to rabbit Tg, albeit TABLE IV ANTIGEN-INDUCEDPHOLIFERATION OF UNFRACTIONATEDOR NYLON WOOL-PURIFIED RABBITSPLEENCELLS
[3H]TdRcpm/culture" Source of spleen cells Bovine Tg-primed rabbit Unfractionated spleen Fractionated spleen Rabbit Tg-primed rabbit Unfactionated spleen Fractionated spleen
Bovine Tg
Rabbit Tg
40,333 26,362
482 0
72 0
135
0
" Values are expressed as the mean of experimental-control counts per minute of quadruplicate cultures. Thyroglobulin (Tg) was added at a final concentration of 20 pgirnl. Cultures were pulsed with 1 pCi of [3H]TdR24 hours prior to harvest on day 3.
200
WILLIAM 0. WEIGLE
these animals produced antibody to rabbit Tg and developed thyroid lesions (Table V). These results are consistent with the existence of tolerance to autologous Tg at the T cell level, which, by appropriate immunization, can be bypassed to trigger competent B cells. Progressive thyroiditis also develops in rabbits immunized with aqueous preparations of either altered autologous Tg or cross-reacting heterologous Tg. Thyroiditis induced by arsanil-sulfanil Tg (rabbit) was perpetuated by periodic injections of the altered Tg over a 6-month period (Weigle and Romball, 1975), during which amounts of autoantibody and severity of lesions increased. After 6 months, all the rabbits had inflammatory lesions involving more than 50% of each thyroid gland. An average of 0.31 mg of antibody protein per milliliter of serum was precipitable with rabbit Tg. In contrast, rabbits receiving only one series of arsanil-sulfanil Tg injections developed antibody and thyroiditis only transiently, suggesting that if progressive immune diseases like Hashimoto’s thyroiditis are the result of an immune response to altered self antigens, these antigens must persist for a fairly lengthy period. Transient trauma or infection might be expected to result only in episodic disease perpetuated briefly by the altered self antigens, whose return to normalcy would soon end the autoimmune state. Such may be the case in acute transient forms of acquired hemolytic anemia and idiopathic thrombocytopenia purpura. It is still questionable whether cellular or humoral immunity is the key participant in EAT induced by immunization with CFA containing Tg or thyroid extract in most species; nevertheless, humoral immunity seems to be largely responsible for at least initiating these lesions in rabbits. EAT usually follows immunization with homologous thyroid extracts or Tg incorporated in CFA. With the early experimental procedures, rabbits required a month or two to develop distinct lesions in their thyroid glands, and DTH reactions took still longer (Rose et al., 1962a). However, when optimal conditions were established, the lesions of EAT maximized within 14 days after injection (Nakamura and Weigle, 1969). For this regimen, 1 ml of CFA containing 10 mg of mycobacteria and 10 mg of rabbit Tg were injected subcutaneously into the rabbits’ footpads. Lesions first appeared on the fifth day after injection, and by the seventh day most of the rabbits had ummistakable lesions. McMaster et al. (1697) defined optimal conditions for the induction of thyroiditis in guinea pigs and also showed that high doses of both Tg and mycobacteria were required for maximal lesions. Complete Freund’s adjuvant probably enhances the antigenicity of Tg in general, and in some species specifically enhances cell-
TABLE V
T CELLACTIVATION (PROLIFERATION), B CELL ACTIVATION (SERUMANTIBODY), AND THYROID LESIONSIN RABBITSPRIMEDTO HETEROLOGOUS OR HOMOLOGOUS THYROGLOBULIN Serum antibody toc
[3HJTdR cpm/cultureb Immunizing antigen' Bovine Tg Rabbit 1 Rabbit 2 Rabbit Tg Rabbit 3 Rabbit 4
Bovine Tg
Rabbit Tg
Bovine Tg
Rabbit Tg
Thyroid lesions
+ + + +
+ +
43,570 40,333
0 482
+ +
44 72
0 0
NDd ND
+ +
Five milligrams of thyroglobulin (Tg) incorporated in complete Freund's adjuvant were injected intraperitoneally. Splenic lymphocytes cultured at a final concentration of 2 x lo6cells per milliliter, with Tg added at a final concentration of 20 pg/nd. Cultures were pulsed with 1 pCi of [3H]TdR 24 hours prior to harvest. Values represent experimental-control counts per minute of quadruplicate sets of cultures. Precipitating antibody. ND, not done.
202
WILLIAM 0. WEIGLE
mediated hypersensitivity. The mycobacteria in CFA may also nonspecifically stimulate T cells that subsequently act as effector cells or help in triggering competent B cells. As an example of such nonspecific stimulation, guinea pigs that are not responsive to DNPpoly-L-lysine can be converted to responders by incorporating large amounts of mycobacteria in the injections (Green et al., 1969).In addition, there has been speculation as to whether components of the mycobacterium, as essential ingredient in eliciting thyroiditis, can alter the antigenicity of homologous and autologous Tg (Weigle et al., 1969). If so, the alteration apparently occurs in vivo from a local reaction to the mycobacteria shortly after injection. Significant amounts of both BSA and rabbit Tg are released from deposits of CFA within the first 2448 hours after injection; however, the BSA, but not the Tg, appears and persists in the circulation. Apparently the Tg in the developing granuloma is significantly altered so that it either does not enter the circulation or is rapidly eliminated. Enzymic studies suggest that alteration of Tg in CFA, injected in vivo, is the result of partial degradation by neutrophils, whose lysosomes may release proteolytic enzymes (cathepsins) that are responsible for such alteration (Weigle et al., 1969).Since cathepsins are active only at a relatively high concentration of hydrogen ions, the p H at the site of degradation would have to be low. Rous (1925) suggested that the p H within leukocyte granules is 3.0 or less after phagocytosis of particles. Several groups of researchers subsequently showed that production of lactate increases under the anerobic conditions of phagocytosis (Sbarra et al., 1960; Strauss and Stetson, 1960; Cohn and Morse, 1960). Furthermore, McCarty et al. (1966) demonstrated that the p H level of joint fluid containing neutrophils progressively decreases with time synchronously with the changes in the number of neutrophils. A similar local drop in p H level may well occur in granulomas formed by CFA alone. Complete Freund’s adjuvant injected into rabbit spleens forms deposits around which large numbers of infiltrating neutrophils gather, whereas infiltration of neutrophils in spleens injected with incomplete Freund’s adjuvant (without mycobacteria) is insignificant (Weigle et al., 1969).Likewise, rabbits injected with homologous T g incorporated in CFA develop both thyroid lesions and autoantibodies, although neither is produced in rabbits injected with Tg in incomplete Freund’s adjuvant. Furthermore, significantly larger amounts of cathepsins D and E have been isolated from splenic extracts of rabbits injected with CFA than from unimmunized rabbits or rabbits injected with the mycobacteria-free adjuvant. Therefore, at low p H (2.5),rabbit Tg, but not BSA, seems to
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
203
be denatured and thus degraded by cathepsins. In agreement, rabbits injected with aqueous preparations of homologous Tg partially degraded by either pepsin (Weigle et ul., 1969) or papain (Anderson and Rose, 1971) produce autoantibodies and develop EAT. b. EAT in the Mouse. By immunization with aqueous preparations of heterologous Tg (Nakamura and Weigle, 1968a),EAT, characterized by infiltration of the thyroid with inflammatory cells, can be induced in mice, as in rabbits. Mice that receive several series of such injections containing a mixture of three heterologous Tgs produce circulating antibody to both heterologous and autologous Tg and develop thyroid lesions. These lesions in mice, or in rabbits, are initiated by circulating antibody rather than by eEector T cells (Clagett and Weigle, 1974; Clagett et al., 1974). The status of murine immunity to autologous T g is one of tolerant T cells and competent B cells, as supported by the fact that B cells, but not T cells, react with syngeneic Tg. In autoradiography studies using 1251-labeledsyngeneic Tg, B cells from A/J mice specifically bound the Tg, but the T cells lacked such antigen-binding capability (Clagett and Weigle, 1974). It was previously established by others that specifically competent B cells expressed receptors for the antigen, which accounted for this binding (Ada et ul., 1970; Humphrey and Keller, 1970). Such Tg binding cells have also been detected in peripheral blood of normal rats (Ada and Cooper, 1971) and humans (Bankhurst et d . , 1973). In tests of healthy humans, 9 of 11 subjects had peripheral blood lymphocytes that bound 1z51-labeledhuman Tg in uitro (Bankhurst et ul., 1973). These antigen binding cells (ABC) were B cells in that the capacity to bind Tg was removed by passing the cells over nylon wool-a procedure that eliminates B cells. This reactivity was specific for Tg because no lymphocytes that bound HSA were detected even when the cells were treated with protease to remove previously bound HSA, which might be blocking the receptors. Later, when the treated cells were permitted to regenerate their Ig receptors, they were still unable to bind HSA. Roberts et al. (1973) also found ABC for human T g in the peripheral blood of all 23 normal individuals tested, but a considerably larger number of such cells were found in the peripheral blood of patients with Hashinioto’s thyroiditis. Others have also detected human Tg-binding cells in the peripheral blood of both healthy individuals (Urbanick et d.,1973; Wick et ul., 1977; Salabe et ul., 1978) and patients with Hashinioto’s thyroiditis (Perudet-Badoux and Frei, 1969; Wick et ul., 1977). Immune responses to homologous thyroid extracts in CFA and aqueous preparations of heterologous Tg are thymus-dependent in
204
WILLIAM 0. WEIGLE
mice. Neither nude mice (Wick et al., 1978) nor mice that have been thymectomized, lethally irradiated, and bone marrow reconstituted (Clagett and Weigle, 1974) respond to Tg immunization. However, thymectomized, irradiated A/J mice, reconstituted with normal syngeneic spleen cells and immunized with an aqueous preparation of heterologous Tg, produce antibody to both heterologous and murine Tg (Clagett and Weigle, 1974). However, if reconstituted only with bone marrow cells prior to challenge with heterologous Tg, recipients produce antibody to neither homologous nor heterologous Tg (Table VI). When the mice are reconstituted with both bone marrow and thymus cells, they respond to injections of heterologous Tg by developing thyroid lesions and producing autoantibodies. Similar observations have been reported by Allison (1974). These experiments permitted investigation of the cellular requirements for autoimmune thyroiditis. Deleting specific B cells, but not specific T cells, before reconstituting thymectomized, lethally irradiated mice interferes with the induction of EAT. It is possible to inhibit both autoantibody production and development of lesions by preincubating syngeneic Tg heavily labeled with lZ5Iwith bone marrow cells, but not with thymus cells. Others have established that incubation of heavily labeled (T) antigen with lymphocytes eliminates specific immunocompetent cells because of local irradiation (Humphrey and Keller, 1970; Ada et al., 1970). The prevention of both EAT and production of autoantibody to syngeneic Tg by the above approaches demonstrates that the B cells or a B cell product (antibody) is etiologically involved in this murine model of thyroiditis. In mice, as in rabbits, immunization with heterologous Tg apparently bypasses the specificity of the T cells, and T cells activated by determinants specific for heterologous Tg supply a second signal needed for differentiation of competent B cells that have reacted with self-related determinants of the heterologous Tg. That heavily radiolabeled Tg eliminates B cells alone is also compatible with the finding of antigen-binding B cells, but not T cells, for syngeneic Tg. A primary role for antibody as the initiator of EAT in the above model is further implicated by a histologic examination ofthe developing thyroid lesion (Clagett et al., 1974). In A/J mice injected with aqueous preparations of heterologous Tg, the temporal appearance and quantity of serum autoantibody correlated directly with the intense formation of immune complexes within the thyroid interstitium. These immune complexes contained complement-fixing antibodies belonging exclusively to the IgG class. Immediately thereafter, the glands were invaded briefly and extensively by neutrophils, which
TABLE VI ANTIGEN SUICIDE O F NORhlAL BONE-MARROW CELLS BY '"I-LABELED SYNGENEIC THYROGLOBULIN
(TG)IN RECIPIENTSOF NORMALTHYMUS AND BONE-MARROW CELLS" Donor cells
T Cells 100 x lo6 normal thymocytes 100 x lo6 normal thymocytes 100 x 106thymocytes and mouse [1251]Tg
B Cells
25 x l@hone marrow and mouse [1251]Tg 25 x lo6 normal bone marrow 25 x lo6 normal bone marrow
Indirect PFCb per spleento bovine Tg
3,395 28,037 20,693
Incidence of thyroid lesions'
12/27 1414 18/20
Reprinted in part from Clagett and Weigle (1974). PFC, plaque-forming cells. p Denominator equals total number of animals examined and numerator equals number of animals exhibiting various degrees of inflammation at time of maximal lesions. a
206
WILLIAM 0. WEIGLE
were soon replaced by a chronic infiltration of monocytic cells. By using autoradiography and fluorescence microscopy, it was demonstrated that Tg was at least one of the antigens deposited in the interstitial immune complexes. These complexes were granular to lumpy in appearance and formed at the basal area of the follicular cells, in intimate association with the follicular basement membrane and plasma membrane. The in situ formation of immune complexes in this model resembles that occurring in an Arthus reaction and is a mechanism of immune complex injury different from that caused by tissue deposition of circulating immune complexes as occurs in serum sickness. c. Spontaneous Autoimmune Thyroiditis (SAT). The high frequency of SAT in certain species and strains of animals offers a unique opportunity to examine this disease’s kinetics of development, histologic patterns, and cellular events in relationship to human disease. Spontaneous autoimmune thyroiditis accompanied by autoantibody occurs in chickens (Cole et al., 1968), rats (Glover and Reuber, 1967), dogs (Lombardi, 1962), and monkeys (Levy et al., 1972) and is characterized by infiltration of lymphocytes, plasma cells, and macrophages. As with Hashimoto’s thyroiditis, hereditary transmission of SAT has been clearly demonstrated in chickens, rats, and dogs (cited in Bigazzi and Rose, 1975).Although the causative mechanisms have not been ascertained, SAT is included in this section because circulating antibody apparently plays a prominent role. A more complete review on SAT has been presented by Bigazzi and Rose (1975). This disease was first observed in fewer than 1% of the Cornell strain of female White Leghorn chickens (cited in Rose et al., 1976). By selective breeding of individual chickens with symptoms of hypothyroidism, the incidence of the disease has been increased to 80-90% of males as well as females (Cole et al., 1970). Because of their marked obesity, these chickens are called the “obese strain” (0s).The autoimmune character of their disease is manifested clearly by the production of both IgM and IgG autoantibodies to Tg (Witebsky et nl., 1969). The thyroids become infiltrated with lymphoid cells as early as the first week after hatching, and at 3 weeks of age 90% of these fowl have severe thyroiditis characterized by intense infiltration of lyniphocytes with large pyroninophilic and mature plasma cells. Histologically, the numerous germinal centers (reviewed in Wick et cil., 1974) duplicate those found in lymphoid organs. This physiologic picture more closely resembles that in patients with Hashimoto’s thyroiditis than that produced by EAT produced in conventional chickens (Irvine and Muri, 1963; Wick and Burger, 1971)and mammals. The increased uptake of 1 3 1 1 by embryonic thyroids (Sundick and Wick, 1974) and
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
207
abnormal thyroid functions in both 0s and the Cornell strain chickens (Sundick and Wick, 1976; Sundick et al., 1979) suggests that an inherent abnormality of the thyroid gland might be an important factor in the development of their EAT. The available data implicate a B cell dependency for SAT in the 0s chicken. Neonatal thymectomy of 0s chicks not only fails to suppress SAT, but actually results in more severe thyroid lesions (Wick et nl., 1970b). Although this phenomenon was interpreted as a lack of T cell dependency, it might also result from deletion of suppressor cells. However, bursectomy results in inhibition of SAT in 0s chickens. Interestingly, Cole et a l . (1968) showed that hormonal bursectomy, accomplished by injecting androgens into 0s embryos before B lymphocytes migrate into the periphery, prevents spontaneous hypothyroidism. Surgical bursectomy of 7-week-old 0s chickens also significantly lowers the frequency and severity of thyroiditis compared to untreated or sham-operated chickens (Wick et al., 1970a). When the bursa is removed by surgery in ouo, before cellular migration into the periphery is complete, suppression of thyroiditis is even greater. In sharp contrast, Jankovic and Isvaneski (1963) and Blaw et (11. (1967) found that neonatal thymectomy impairs the development of EAT in conventional chickens, but neonatal bursectomy does not significantly decrease their thyroiditis. Although most investigators report that thymectomy results in increased SAT, the levels of autoantibody decrease after this procedure. Similarly, reconstitution of neonatally bursectomized, irradiated 0s chickens with autologous bursa cells completely reinstates severe SAT, but does not totally restore production of anti-Tg antibodies (Nilsson and Rose, 1972). On the other hand, anti-Tg antiserum from 0s chickens repeatedly injected into the genetically related, but mildly afflicted B4B4substrain results in a significant increase in lymphoid infiltration of the thyroid gland (Jaroszewski et cil., 1978).Since the same serum does not cause thyroiditis when injected into normal Cornell strain chickens, it was concluded that although antibody plays a pathogenic role in SAT of 0s chickens, other factors are also required. In this regard, germinal centers of the infiltrated thyroids, but not the spleens of OS chickens, contain cells that produce antibody to Tg, detected b y using the indirect fluorescent antibody procedure. This point is of interest in view of similar observations in rabbits with EAT elicited by immunization with aqueous preparations of heterologous Tg (Clinton and Weigle, 1972). These rabbits have antibody-producing cells specific for rabbit Tg located in their spleens and thyroid glands associated with the appearance of thyroid lesions.
208
WILLIAM 0. WEIGLE
The failure to detect antibody-producing cells in the spleens of 0s chickens may result from the insensitivity of this indirect fluorescence technique or from migration to and preferential localization of antiTg-reacting cells in the thyroid gland. It was suggested that during the disease process the thyroid glands of 0s chickens are converted into a lymphoid organ capable of responding to a variety of foreign antigens. Not all the germinal centers contain cells producing antibody to Tg. In support of this, after the chickens are immunized with BSA, antibody-producing cells that are specific for BSA appear in the germinal centers of the thyroid glands (Schauenstein and Wick, 1974). The genetic factors in SAT of 0s chickens will be discussed in Sections V,A,7 and 8. A spontaneous disease resembling chronic lymphocytic thyroiditis in humans has been observed in Buffalo rats (Glover and Reuber, 1967; Hajdu and Rona, 1969). The thyroid infiltrate is composed of lymphocytes, plasma cells and macrophages, and the animals develop distinct lymphoid follicles, germinal centers, and perivascular cuffing by lymphocytes. In most cases, many follicles are replaced b y mononuclear cells. At times the lumen is reduced and contains mononuclear cells. Until they are 12 weeks old, SAT is not observed in Buffalo rats; after this the incidence is 14%, increasing to 30% by 14-16 weeks of age and 48% at 30 weeks (Silverman and Rose, 1971). The frequency of disease is higher in females than in males and is markedly increased in both sexes after ingestion or injection of methylcholanthrene (Reuber and Glover, 1969). Like chickens with SAT, Buffalo rats whose thymuses are removed surgically during neonatal life undergo a significant increase in the incidence of thyroiditis and decrease in time of onset (Silverman and Rose, 1974a). Penhale et al. (1973) also observed a marked increase in SAT in randomly bred Wistar rats subjected to thymectomy (60%) or to sublethal whole body irradiation (22%). The histology of the thyroid gland resembled that seen in patients with Hashimoto’s thyroiditis. Although both these increases in the incidence of thyroiditis could readily be explained as a disruption in the normal control of immune reactivity of competent T cells by regulatory T cells, circulating antibody apparently plays a major role in the spontaneous disease, at least in the Buffalo rat (Noble et al., 1976). Antibody to Tg measured by two different procedures closely correlated with thyroid infiltration. In addition, direct immunofluorescence of the diseased thyroid revealed the presence of Ig. On the other hand, these rats had no DTH reactions to thyroid antigens according to skin tests and assays for migration inhibition factor. Thus, the presence of circulating autoantibody and absence
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
209
of DTH in these Buffalo rats strongly support a major role of antibody in the pathogenesis of their SAT. Additionally, SAT has been documented in several colonies of laboratory purebred beagles in which the frequency of disease was further increased by selective inbreeding (reviewed in Bigazzi and Rose, 1975). A high incidence (16%)of canine SAT in young adult beagles was first reported by Tucker (1962)and seemed similar to Hashimoto’s thyroiditis. Another colony of purebred beagles at the Argonne National Laboratory had a 20% incidence of lymphocytic thyroiditis after 1 year of age (Fritz et ul., 1970b). These workers found 63 cases of thyroiditis in a total of 401 dogs, with approximately equal frequencies in males and females. Although antibodies to thyroid antigens and positive skin responses to thyroid extracts were associated with a high incidence of thyroiditis, no appreciable attempt was made to implicate either humoral or cell-mediated immunity. Musser and Graham (1968) reported thyroiditis in 70 of 541 males (13%)and 48 of 440 females (11%) among purebred beagles. Histopathologic patterns in the thyroids of dogs with SAT consisted of either nodular or diffuse infiltration of lymphocytes and macrophages. Quite often germinal centers replaced lymphoid follicles in the thyroid, but the remaining follicles often contained vacuolated colloid infiltrated with both lymphocytes and macrophages (Gosselin et ul., 1978). In dogs with severe lymphocytic thyroiditis, follicular cells were sometimes transformed into large eosinophilic Hurthle cells (Mawdesley-Thomas and Small, 1968). Abnormalities of the thyroid follicular basement membrane consisting of degeneration and focal disruption have been described (Mawdesley-Thomas and Jolly, 1967). Also identified were autoantibodies to Tg, to a second colloid antigen, and to the microsomal antigen (Mizejewski et uZ., 1971), as detected in Hashimoto’s disease of humans. The focal lymphocytic thyroiditis observed in laboratory beagles usually is not associated with alterations in thyroid function, although in dogs with severe thyroiditis, the biologic half-life of 1311 is significantly shorter than in dogs whose disease involves mild lesions or none at all (Fritz et d., 1970a). The maximum uptake of 1311 b y thyroid tissue is also significantly lower and occurs earlier in dogs with more severe disease. On the other hand, clinical symptoms of hypothyroidism, such as low levels of protein-bound iodine, and decreased I3lI uptake, are often observed in pet dogs with SAT (Capen et ul., 1975). The Cnllithrix jucchus species of marmoset monkeys develop SAT with a rather high frequency (Levy et ul., 1972). Of 32 such monkeys at one laboratory, 60% of the females and 28% of the males had chronic
210
WILLIAM 0. WEIGLE
thyroiditis. However, of 169 C. jacchus caught in the wild, only 12% of the females and 9% of the males had SAT. In both groups, the histology of the lesions ranged from scattered focal collections of mononuclear cells to almost complete replacement of follicular structure by iiiflaniniatory cells, which were mostly lymphocytes with occasional plasma cells. The atrophied follicles contained little or no colloid, but infiltrates of mononuclear cells were often observed. Two C. urgentatu male monkeys born in captivity also had chronic thyroiditis with lesions much like those of Hashimoto’s disease in humans. In these monkeys, SAT has not been evaluated for associated humoral or cellmediated immunity.
7. Ininiunogenetics of EAT During recent years, great interest has centered around the major histocompatibility complex (MHC) and the importance of various MHC genes in cytotoxic (reviewed in Zinkernagel and Doherty, 1979) and humoral immune responses (McDevitt and Sela, 1965; Benacerraf and McDevitt, 1972; Katz and Benacerraf, 1975). Furthermore, the MHC is repeatedly linked with susceptibility to autoimmune phenoniena in both experimental animals and man (Bodmer and Bodmer, 1978; Warner, 1977). The first suggestion of genetic predisposition to autoimmune thyroiditis came from Hall and co-workers (1960; Hall, 1962; Hall and Stanbury, 1967), who described familial clustering of various forms of thyroid disease. Roitt and Doniach (1967) then observed thyroid autoantibodies in 42% of female relatives and 30% of male relatives of patients with either Hashimoto’s thyroiditis or priniary niyxedema. Whether this genetic influence is a fundamental hereditary aberration in thyroid structure and/or function or some underlying genetic abnormality in the immunologic apparatus is not known; however, because of the difficulties in carrying out genetic studies in man, attention has turned to the genetic aspects of SAT and EAT in laboratory animals. The available evidence strongly suggests that SAT is genetically controlled in chickens, rats, and dogs. Cole (1966)established that SAT in 0s chickens is inherited as a polygenic trait with some degree of dominance when he used inbreeding to obtain a high incidence of thyroiditis. Subsequent experimentation linked alleles of the MHC with this susceptibility to SAT (Bacon et al., 1974) and identified the B locus ofthe chicken MHC as the site that codes for both a blood group antigen and a tissue alloantigen. Bacon et al. (1973)described the B locus genotypes in 0s and Coniell chickens. After these initial studies, it was shown that ofthe 12 known blood-group systems in 0s and
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
21 1
Cornell chickens, only the B locus has a major influence on SAT. However, it was suggested that other genes act as possible modifying factors in particular substrains (Rose et n l . , 1978). Recently, Wick et a l . (1979) and Rose et nl. (1978) presented evidence for multigenic control of susceptibility to SAT in chickens involving at least three loci, i.e., the B locus (MHC), a non-B locus, and a thyroid locus. Apparently irregularities in genetic control of the immune response (MHC, B locus), in suppressor cells (MHC, non-B locus), and in thyroid function (thyroid locus) all contribute to the SAT observed in OS chickens. Furthennore, when partially inbred substrains of these chickens were studied for SAT, the influence of the MHC was niarked in one strain, less pronounced in a second, and hnrely detectuble and transient in a third (Bacon and Rose, 1979). Obviously, genes other than those involved with the B haplotype of OS chickens are involved in the control of SAT, a subject of recent, inore extensive reviews b y several groups (Rose et nl., 1976, 1978; Bacon et al., 1978). Although hereditary traits associated with SAT are not always as obvious as in 0s chickens, the restriction to certain strains and family lines of rats and dogs suggests that their SAT is also under some measure of genetic control. For example, Buffalo rats (Noble et al.. 1976) have a high incidence of SAT in comparison to outbred rats or other inbred strains, compatible with a genetic predisposition for the disease. Penhale et (11. (1975a) have also shown a strain variability in the susceptibility of rats to EAT. That thyroiditis observed in a random sampling of pet dogs is restricted to certain pedigrees also suggests a familial tendency, as exemplified in purebred beagles among which thyroiditis is present only in certain family lines (Fritz et nl., 1970b). The incidence of SAT increases among dogs in these lines as inbreeding increases, again supporting the hypothesis that SAT is under genetic control. The high frequency of SAT in one species of marmoset monkeys, but rarely in another species, also indicates some genetic influence (Levy et nl., 1972). Genetic restraints on EAT in several different laboratory animals have also been reported. McMaster et (11. (1965)showed that Hartley guinea pigs are more susceptible to induction of EAT than strain 13 guinea pigs. Similarly, Braley-Mullen et (11. (1975) reported that strain 2 guinea pigs are more susceptible to EAT than the strain 13 variety. A contrasting genetic relationship between these strains was previously reported for EAE, where strain 13 was more susceptible than strain 2 (Gasser et d., 1973).In rats, EAT has also been reported to be under genetic influence (Rose, 1975). Although these studies in guinea pigs and rats implicate a genetic factor, more extensive experiments with
212
WILLIAM 0. WEIGLE
selected strains of inbred mice have firmly established that at least one model of EAT is regulated by the MHC. Mice of the same H-2 type, when immunized with homologous thyroid extract in CFA, exhibited the same pattern of thyroiditis and autoantibody production, regardless of the strain's genetic background. This result suggests a close relationship between the MHC and EAT in which the responsiveness to Tg is transmitted genetically as a dominant trait so that F1hybrids of good and poor responders are good responders (Rose et al., 1978). Good responders ( H - 2 k , H-28, and H - J Qhaplotypes) produce high levels of antibodies to mouse Tg and develop severe thyroiditis, whereas poor responders ( H - 2b, H - 2 f , and H-2" haplotypes) produce only moderate levels of these antibodies followed by mild inflammation of the thyroid (Vladutiu and Rose, 1971b). In H-2" mice, antibody production and cellular infiltration are intermediate, between good and poor responders. Apparently, the H-2 linked restriction governing susceptibility to EAT is directed to the T cell. Results from adoptive transfer experiments show that one can elicit severe thyroiditis in mice that are first thymectomized and lethally irradiated only if the cells used for reconstitution contain T cells from good responders (Vladutiu and Rose, 1975). Whether the B cells are from good or poor responder strains is much less important. Of special interest along these lines is that injection of aqueous preparations of mouse Tg with LPS, a B cell mitogen and adjuvant for antibody production, results in EAT in responder, but not nonresponder, mice (Esquivel et al., 1977).This result implies that specific helper T cells, first activated nonspecifically by LPS, are indirectly responsible for the production of EAT. Similar studies by Maron and Cohen (1979) suggest that the susceptibility to induction of EAT in inbred strains of mice results from an apparent point mutation occurring at the H-2k locus of the resistant H-2 haplotypes. Tomazic et al. (1974) have theorized that genetic control is associated with the centromeric (left) side of the H-2 gene complex, probably the K and/or Z-A subregion. According to a more recent report (Konget al., 1979), genes at the D end of H-2 may have a modifying effect on EAT in mice of both good and poor responder strains. The extent of genetic regulation appears to be dependent upon the participation of genes both at the K end (Zr) and at the D end. Possibly, the D region regulates suppressor cells and thereby prevents the interaction of T cells expressing the Zr gene for Tg with precursors of the effector cells (Kong et al., 1979). Since there is no quantitative correlation between the levels of circulating antibody and thyroiditis induced in mice with Tg in CFA, presumably antibody is not important but cannot yet be eliminated
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
213
entirely as a mediator of lesions. Clearly T cells regulated by the MHC complex are required in this model of EAT, although defining the actual mechanisms involved is complicated by the fact that the disease can be transferred with antibody in responder mice, but not in nonresponder mice (Tomazic and Rose, 1975). Furthermore, the niacrophage disappearance reaction, a measure of cell-mediated immunity, is similar in responder and nonresponder mice as reported by Tomazic and Rose (1977). In contrast, Christadoss et al. (1978) reported that lymphocytes from responder strains immunized with mouse T g in CFA proliferated in the presence of mouse Tg, whereas lymphocytes from nonresponder strains did not. However, the nature of these proliferating cells and their role in antibody production, immune regulation, or cytotoxicity is unknown. The genetic restriction ofEAT in mice immunized with mouse Tg in CFA is understandable, since the number of antigenic determinants is in all likelihood limited. However, it has been shown that the same strains that are responders and nonresponders to mouse T g are also responders and nonresponders to heterologous Tg (Tomazic and Rose, 1976; Christadoss et al., 1978). Heterologous Tgs are potent antigens and apparently activate a variety of T cell specificities, making a strict genetic restriction ofH-2 type at the T cell level less likely. The use of homologous Tg in the above studies must be considered in evaluating the genetic restriction in thyroiditis, since isoantigens may also be present on Tg (Nakamura and Weigle, 196813; Tomazic and Rose, 1976). Along this line, Tg from high responder mice (H-.2k) incited a greater response in both high and low responder mice than T g from low responder mice (H-2") (Tomazic and Rose, 1976).
8 . lnzniuiioregulation of EAT Factors involved in the control of normal immune responses to foreign antigens can function quite similarly in regulating SAT and EAT. In addition to the controlling effects governed by the interaction between T and B cells, other mechanisms such as regulation by suppressor T cells have also been associated with these models of thyroiditis. That such mechanisms are at play in thyroiditis was first suggested by procedures that preferentially deplete T cells. Neonatal thymectomy of 0s chickens (Wick et al., 1970b) and rats (Silverman and Rose, 1974a) increases the incidence of SAT; however, treatment with antithymocyte serum yields no such increase (Wick et al., 1971). This discrepancy was clarified by studies showing that anti-thymocyte serum deleted circulating T cells, but did not affect T cells within the thymus (Denman and Frenkel, 1968). Thus, apparently the thymus is
2 14
WILLIAM 0. WEIGLE
the major source of suppressor cells affecting this disease. The role of thymic suppressor cells in regulation of SAT in 0s chickens was further evidenced by the appearance of SAT in the B4B4 line of 0s chickens after thymectomy. This line does not normally develop SAT (Bacon et al., 1974). The finding that neonatal thymectomy prolongs graft survival in normal Cornell chickens, but not in 0s chickens, suggests that the effector (helper) to suppressor cell ratio may be higher in neonatal 0s chickens than in conventional chickens (Jakobisiak et al., 1976). Neonatal thymectomy also increases the incidence of SAT in Buffalo rats. The increased risk of SAT in Buffalo rats treated with methylcholanthrene may also b e associated with preferential deletion of a subpopulation of thymic T cells, since methylcholanthrene is known to diminish thymus weight and interfere with the DTH response to bacillus Calmette-Guerin (BCG), a T cell function (Silverman and Rose, 197413). A role for suppressor cell activity in regulating SAT of randomly bred Wistar rats has been suggested by Penhale and co-workers (1973, 197513, 1976). Thyroid lesions and autoantibody to Tg spontaneously developed in 60% of these rats after thymectomy and whole body irradiation. Another group given whole body irradiation alone also developed thyroiditis, but at a much lower incidence (10%). Histologically, the thyroids resembled those of humans with Hashimoto’s thyroiditis. Although rats differ somewhat by strain in susceptibility to thyroiditis after thymectomy and irradiation (Penhale et al., 1975b), most strains become lymphopenic, including reductions in levels of peripheral T cells and T cell functions. The thyroiditis developing after thymectomy and irradiation is abrogated by reconstitution with syngeneic lymph node cells. It may be that the radiosensitive cells are T ancillary cells required for expression of T suppressor cell, rather than T suppressor cells themselves (Penhalee t al., 1976). In any event, these combined results further support the hypothesis that a population of short-lived peripheral T cells is involved in suppressing both SAT and autoantibody production. The role of such suppressor activity in the initiation of thyroiditis is speculative. Although current theories concerning the role of suppressor cells in self-nonself recognition (Cantor and Gershon, 1979)would essentially predict that autoimmune thyroiditis arises only when specific suppressor cell activity wanes, the existing data d o not justify such a conclusion. Specific suppressor activity may be generated only after the immune response to Tg (or other thyroid antigens) is underway and may function more as a fail-safe mechanism than as a primary deterrent.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
215
Pretreatment of host animals with aqueous preparations of either thyroid extract or purified Tg suppresses both SAT and EAT. Jankovic and Flax (1963) reported delaying the onset of EAT in guinea pigs by injecting thyroid extract in saline both before and after sensitization with thyroid extract in CFA, although there was no reduction in the incidence or severity of the lesions. Silverman and Rose (19744 noted that multiple injections of thyroid extract without adjuvant suppressed the development of SAT in Buffalo rats. However, injecting aqueous preparations of thyroid extract had no effect on previously established SAT; in fact, the titers of circulating antibody to Tg actually increased. Similarly, the development of SAT in the PVG/c rat as the result of thyniectomy and irradiation was prevented by injections of aqueous preparations of thyroid extracts (Whitmore and Irvine, 1977). These authors’ tentative interpretation was that suppressor cells were responsible, although no attempt was made to demonstrate such a population. Both EAT and antibodies to Tg, which develop in rabbits immunized with aqueous preparations of heterologous Tg or altered preparations of rabbit Tg, are abrogated by simultaneous injections of homologous Tg (Weigle, 1964). Previous data indicated that EAT produced in this model resulted from a bypass of the specificity of T cells, thus triggering competent B cells to produce anti-Tg in the presence of tolerant T cells. In this situation, the ability of aqueous rabbit Tg to interfere with the development of autoantibody and thyroiditis may have resulted from the induction of tolerance to rabbit T g in the B cell population. It is unlikely that the generation of specific suppressor T cells was responsible, since doses larger than those used for immunization were required for significant suppression. Injection of aqueous preparations of rabbit Tg also reversed the autoimmune state (autoantibody and thyroid lesions) of rabbits previously immunized with heterologous Tg, but to acquire this result injections had to be administered simultaneously with cyclophosphamide (Nakamura and Weigle, 1970). Similarly, rabbits tolerant to BSA lost their unresponsiveness when immunized with a cross-reacting albumin such as HSA (see Section IV,B), but this loss was prevented if BSA was injected simultaneously (Weigle, 1961). Braley-Mullen et 01. (1978) observed that pretreatment of inbred strains of guinea pigs with guinea pig Tg in incomplete Freund’s adjuvant reduced both the incidence and severity of EAT in animals subsequently immunized with guinea pig Tg in CFA. It was of interest that, although the amount of circulating antibody was reduced in many of the pretreated animals, DTH to guinea pig Tg was not affected. The suppressive effect was transferable with lymphoid cells, but not with
216
WILLIAM 0. WEIGLE
serum, suggesting that suppressor cells mediated the diminished responses. Possibly, after injection of Tg in incomplete Freund's adjuvant, the mode of protection involves suppressor cells, but the protection resulting from injection of aqueous preparations of Tg is caused by either antigen blockade or induction of central unresponsiveness in the T and/or B cells. Similarly, in EAE the mechanism of protection may differ according to the carrier used with the pretreatment antigen (see below). 9. Secondary Tissue Znjury
Secondary kidney disease has been described both in laboratory animals with EAT and in patients with chronic thyroiditis. In addition to the thyroiditis induced in rabbits by injecting aqueous preparations of either altered T g or heterologous Tg, tissue damage can also result from complexes formed between Tg released from the damaged gland and circulating antibody synthesized by the animal. Such secondary lesions were first observed with acute EAT in rabbits (Weigle and High, 1967). Both antibody to autologous Tg and the thyroid lesions induced with aqueous preparations of either heterologous T g or rabbit Tg coupled to the diazonium derivatives of arsanilic and sulfanilic acids were transient and disappeared from test rabbits within 2 or 3 months. These rabbits were then given an iodine-free diet and injected with 3-6 mCi of Na'"1. At this point, their thyroids incorporated sufficient 1311 (18%)to cause disruption of the follicles, loss of colloid, and degeneration and necrosis of the epithelial cells. Autologous [1311]Tg was released into the circulation, where it equilibrated between the intra- and extravascular fluid spaces and persisted with a half-life of approximately 2.3 days. When their immune systems were restimulated with autologous [13'I]Tg released from the damaged thyroids, these rabbits produced circulating antibody (IgG) to the autologous Tg that complexed with the [13'I]Tg in the circulation (Fig. 7 ) . A number of the rabbits developed proteinuria and at autopsy showed glomerular changes in the kidneys, accompanied by localization of rabbit IgG (antibody) and Tg along the glomerular basement niembrane. These events leading to kidney injury seem to resemble those previously established in serum sickness (Dixon et al., 1958). In both the present model and serum sickness, circulating complexes form between newly synthesized antibody and circulating antigen. As the complexes are eliminated from the circulation, both antigen and antibody deposit along the glomerular basement membrane, resulting in lesions. Secondary glomerular lesions are also observed in progressive EAT.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
400.0001
10
b . . .
',.. .
x'
b
.2
,/ 1.'
I
2 17
0' tr' 25' 15' . I
Y I.I
1
2
3
4
5
6 1 8 9 Days Altar Injection of Na1l3'
10
11
FIG.7. Changes in the serum and urine of a rabbit imniunized with ;Irsaiiil-sultaliilTg (rabbit) and sulxequently injected with Nal'"1. G O , 1'311-ldieled thyroglohulin [I3'I]Tg; A . - . A, antibody; 0---0,[L3111Tg complexed to y-globulin. [Reprinted from Weigle and High (1967).]
Rabbits injected with aqueous preparations of altered rabbit Tg over a 6-month period develop progressive thyroiditis (Weigle and Nakamura, 1969). In addition, chronic glomerular lesioiis evolve, apparently resulting from the deposition of Tg-anti-Tg complexes formed between circulating antibody and injected, altered Tg or native, autologous Tg released from the damaged glands, or both. Using fluorescence microscopy, one can see IgG, the third component of complement (C3) and Tg localized along the gloinerular basement membrane. Although it is well established that secondary tissue damage resulting from circulating antigen-antibody complexes is an important component of autoimmune disease in humans, little was known until recently concerning the presence of such complexes or their biologic significance in thyroiditis of humans. In an isolated case, R. S. Schwartz (personal communication, 1967) observed a patient with thyroiditis whose complete renal failure apparently resulted from complexes containing autologous Tg released from the thyroid and circulating anti-Tg antibody deposited along the glomerular basement membrane. D. Koffler (personal communication, 1968) also noted glomerular injury and deposition of Tg and Ig in the kidneys of a patient with chronic thyroiditis. More recently, O'Regan et (11. (1976) described a patient with Hashimoto's thyroiditis accompanied by epimembranous nephropathy. After the diagnosis of thyroiditis, the patient developed massive proteinuria with high levels of both Tg antibodies and microsomal antibodies in the circulation. In addition to the histologic changes in the glomeruli, there were glonierular de-
2 18
WILLIAM 0. WEIGLE
posits of IgG, C3, C4, Tg, and microsomal antigens along the basement membrane. Jordan et al. (1978) reported that an individual with chronic thyroiditis accompanied by a mixed membranous and mesangial proliferative glomerulonephritis had subepithelial intramembranous and subendothelial renal deposits. Immunofluorescence revealed granular deposits of IgG, IgM, and C 3 along the glomerular basement membrane. Granular staining of basement membranes and mesangia was also detected by using anti-Tg and antimicrosomal antibody for an indirect fluorescence assay. Two additional patients with autoimmune thyroid disease have since developed glomerulonephritis accompanied b y granular glomerular and mesangial deposits of IgG, IgM, C3, Tg, and microsomal antigen (Jordan et al., 1979). Anticomplementary activity present in sera of 59% of patients with Hashimoto’s disease suggested the presence of associated antigenantibody complexes (Calder et al., 1974). More recently, antigenantibody complexes detected in the sera of such patients by several different approaches have been reported (Calder et al., 1975; Barkas et al., 1976; Al-Khateeb and Irvine, 1978; Mariotti et al., 1979). A more systematic monitoring of patients with chronic thyroiditis may show a much higher incidence of at least transient glomerulonephritis than previously anticipated. Similar phenomena have been consistently observed in persons with other diseases. Kunkel et al. (1961) observed the presence of y-globulin complexes in the sera of patients with rheumatoid arthritis. Tan et al. (1966)and Krishman and Kaplan (1956)were first to implicate complexes between DNA and anti-DNA as a causative agent for renal injury in patients with systemic lupus erythematosus. Since these initial observations, similar tissue damage secondary to the primary disease has been observed with a variety of autoimmune and nonautoimmune diseases (reviewed in Haakenstad and Mannik, 1977; Theofilopoulos and Dixon, 1980). 10. Related Autoimmuite Phenomena 11. Acetylcholine Receptors (AChR). Experimental autoimmune myasthenia gravis (EAMG) is a model of the human disease myasthenia gravis (MG) and is characterized by muscular weakness and fatigue. The symptoms of impaired neuromuscular transmission result from an autoimmune response to AChR (reviewed in Lindstrom, 1979). Dogs spontaneously develop a disease that clearly resembles MG (Lennon, 1978). In rabbits (Patrick and Lindstrom, 1973; Patrick et al., 1973; Heilbron and Mattson, 1974; Aharonov et al., 1975; Green et al., 1975; Sanders et al., 1976; Penn et al., 1976; Berti et al., 1976;
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
219
Eldefrawi, 1978), guinea pigs (Lennon et a / . , 197s; Tarrab-Hazdai, 1975a), rats (Lennon et ul., 1975; Lindstrom et a/., 1976a; Green et a l . , 197S),monkeys (Tarrab-Hazdai et d . , 1975b), goats (Lindstrom, 1976; Lindstrom et al., 1978a),frogs (Nastuk et al., 1979),and mice (Fuchs et al., 1976; Bermaii and Patrick, 1980), EAMG has been induced by injection of CFA containing AChR isolated from the electric fish Torpedo califonaicu. These AChR are intrinsic membrane glycoproteins, are hydrophobic, and have an approximate molecular weight of 330,000 (Eldefrawi and Eldefrawi, 1977). They are comprised of four kinds of peptide chains with apparent respective molecular weights of 38,000, 50,000, 57,000, and 64,000 (Raftery et d . , 1976; Karliii et al., 1976). These chains are referred to as a, p, y , and 6 chains, respectively. The a chain constitutes at least a part of the acetylcholine binding site. The functions of the remaining chains are unknown, although some of them may constitute the ioiiophore portion of the molecule. The AChR of Torpedo marmoratu, on the other hand, contains two chains with molecular weights of 41,000 and 43,000 (Sobel et ul., 1977). The 41,000 mw species is believed to be the acetylcholine binding site and the 43,000 niw species, the ionophore. Some controversy concerns the structure of AChRs isolated from muscle; this structure has been described as being composed of identical polypeptide chains of 41,000 mw (Merlie et al., 1979; Dolly and Barnard, 1977), but also as being composed of several different sized polypeptide chains like those of AChR from T. californicu (Boulter and Patrick, 1977; Froehner et d . , 1977). The concentration of receptors in tissue can be quantitated readily by the ability to bind small protein toxins present in venom of cobras and kraits (Lee, 1972; Engel et nl., 1977). Conformationally intact AChRs are good imniunogens, whereas denaturation by sodium dodecyl sulfate, heat, or urea greatly reduces the ininiunogenicity of AChR. In Lewis rats, EAMG has been induced by immunization with CFA containing AChR purified from organs of electric fish (Lindstrom et al., 1976a; Lennon et ul., 1955), from muscle of normal Lewis rats (Lindstrom et al., 1976b), or from muscle of fetal calves (Lindstrom, 1979). Immunization with any one of the four peptide chains isolated from T. califomica produces EAMG in these rats (Lindstrom et ul., 1978b). Both EAMG and MG are mediated by antibody rather than by sensitized cells, as first determined when MG symptoms abated with thoracic duct drainage but worsened when patients were reinfused with the cell-free lymph (Bergstrom et nl., 1973). Changes in clinical symptoms of MG also correlate well with antibody concentration (Pinching et al., 1976; Dau et al., 1979; Denys et al., 1979; Lefvert et
220
WILLIAM 0. WEIGLE
al., 1978). Furthermore, the occasional occurrence of a transient form of MG in neonates from mothers with MG (Simpson, 1960; Keesey et al., 1977; Masters et al., 1977; Nakao et al., 1977) strongly indicates that this disease is mediated by autoantibodies to AChR. The ability to transfer EAMG by injecting sera from rats with chronic EAMG into normal rats firmly established anti-AChR antibody as the major factor (Lennon et nl., 1978; Engel et al., 1979); EAMG has also been transferred to mice (Toyka et al., 1975) with serum anti-AChR antibody derived from patients with MG. A strong association between the loss of AChR receptors in tissue and anti-AChR antibody in both EAMG and MG has been well documented (reviewed in Lindstrom, 1979). In either EAMG or MG, neuromuscular transmission apparently is not inhibited by the direct antagonistic effect of antibodies specific for the acetylcholine binding site of the AChR, but it is inhibited by loss of AChRs via complement-dependent lysis and antigenic modulation (Lennon et al., 1978). Development of acute EAMG during the early phase depends critically upon the third component of complement (C3) and is prevented by depleting rats of C3 with cobra venom factor. Although the acute and chronic forms of EAMG are inducible by immunization with AChR in CFA, to induce the acute disease one must inoculate CFA along with pertussis vaccine (Lennon et al., 1975). The chronic phase is associated with high levels of circulating antibodybound AChRs and more closely mimics MG than does the acute disease. The histologic patterns of acute and chronic EAMG have been reviewed recently by Lindstrom (1979). Lewis rats immunized with AChR isolated from the electric eel first produce antibody reactive with this heterologous AChR; negligible antibody reactive with rat AChR is produced (Lindstrom et al., 1976a). Gradually, the animal develops a clone of cells producing antibody to the rat receptor, ultimately causing chronic EAMG. These events are reminiscent of those in rabbits that have been immunized with aqueous preparations of heterologous Tg and may involve competent B cells and tolerant T cells (Weigle, 1971). As in thyroiditis, EAMG may result from bypassing either the requirements or specificity of T cells for syngeneic AChR. T cells are a requirement for the induction of EAMG in rats immunized with heterologous AChR and for both the primary and secondary antibody responses to heterologous AChR (Lennon et ul., 1976). Although T cell reactivity to unaltered, syngeneic AChR has not been demonstrated, EAMG can be produced in rats immunized with heterologous as well as syngeneic AChR when incorporated into CFA (Lindstrom et al., 1976b).Only limited data are available, however, suggesting that T cells are activated in response to
AUTOIMMUNITY: THYROIDITIS A N D ENCEPHALOMYELITIS
221
syngeneic AChR; the appearance of such T cells may result from expansion of a faulty T cell tolerance by CFA. To fully understand the cellular events required for either EAMG or MG, it will be important to evaluate the immune status of helper T cells in respect to syngeneic AChR both before and after induction of EAMG. There is no available information on the relationship between suppressor T cells and EAMG, although preliminary data indicate that patients with MG are deficient in mitogen-induced suppressor T cells (Mischak et al., 1979). As with most autoimmune diseases, the spontaneous occurrence of MG and induction of EAMG seem to be genetically influenced. Myasthenia gravis is associated with the MHC in that patients with MG have abnormally high frequencies of HLA-B8 and D R W 3 specificities compared to normal subjects (Feltkampet al., 1974;Fritzeet al., 1973). In addition, MG patients with either of these histocompatibility antigens have higher titers of anti-AChR antibody (Naeim et al., 1978) than patients without these antigens. Furthermore, there is a strain variation in the susceptibility of mice to EAMG after immunization with T. californica AChR in CFA. Fuchs et nl. (1976)reported that mice with the H-2q or H-2* haplotypes are resistant to EAMG, but produce titers of hemagglutinating antibody comparable to susceptible strains of other H - 2 types. Berman and Patrick (1980)also observed a strain variation in susceptibility to EAMG, but differ from Fuchs et al. (1976)in finding that mice with H - 2 q and H - 2 8 haplotypes are not resistant to EAMG. In fact the SJL strain (H-2$haplotype) was highly susceptible. The finding that a high and a low susceptible strain possessed the same H-2 allele (H-2k)suggested that the H-2 haplotype alone does not determine susceptibility to EAMG. Studies with congenic and recombinant inbred strains of mice reveal that T cell proliferative responses to AChR are controlled by an H-2 linked Ir gene, mapping in the I-A subregion of the mouse MHC (Christadoss et al., 1979). The relationship between in uitro responses to AChR and EAMG, however, is only speculative. Moreover, the response measured was to a heterologous AChR that showed little cross-reactivity with mammalian AChR; no correlation was evident between antibody response and disease. Graves’ disease is another disorder involving autoimmune responses to receptor molecules. In these patients, autoantibody is directed to the thyroid-stimulating hormone receptor (Smith, 1976; Adams and Kennedy, 1971;McLachlan et al., 1976).Moreover, autoantibodies to insulin receptors are present in certain patients with insulin-resistant diabetes (Jarret et al., 1976;Flier et al., 1979). h. Liver Protein F. Protein F is found in the livers of all mammalian
222
WILLIAM 0.WEIGLE
species (Fravi and Lindenmann, 1968; Iverson and Lindenmann, 1972). Mice are polymorphic for protein F in that they have either of two allelic types, I or 11, which are serologically identifiable. Protein F has a molecular weight of 40,000 and migrates upon electrophoresis as a @-globulin(Lane and Silver, 1976), although the type I protein migrates more cathodally than does the type I1 protein (Anders et d . , 1978). Mice do not respond to their own allelic type-only to the opposite type, but the antibody produced reacts with both the immunizing type and the host type. Thus, it seems that mice have tolerant T cells to their own protein F, but contain protein F-competent B cells (Fravi and Lindenmann, 1968). That mice lack T cell recognition for syngeneic protein F is shown by their inability to develop DTH when immunized with either syngeneic or allogeneic protein F (Anders et al., 1978). Iverson and Lindenmann (1972) suggested that protein F consists of two distinct antigenic moieties: a carrier region that recognizes the alloantigenic part of the molecule and an antigenic determinant common to both type I and type I1 proteins. Thus, induction of autoantibodies to F antigen probably involves collaboration between helper T cells recognizing the allogeneic region of the molecule, acting as a “carrier” determinant, and B cells with specificity for the syngeneic determinant common to both types of protein F. This suggestion is supported by preliminary observations that small amounts of protein F are present in the serum. In further support, by injecting enough alloantigenic protein F to balance the in vivo concentration of the syngeneic protein, one induces a tolerant state in T cells and inhibits the antibody response to the syngeneic determinants on the alloantigenic protein (Winchester, 1979). It has recently been reported that protein F is not specific for liver, but is also present, at somewhat lower concentrations, in kidney, spleen, heart, and lung tissue (Anders et nl., 1978).These observations emphasize the importance of the nature and logistics of autoantigens, not just their presence, in determining whether or not disease accompanies autoimmune reactivity, since no disease state has been associated with autoimmunity to protein F. The protein F model also lends itself to an evaluation of acquired immunologic tolerance to self. Protein F is expressed codominantly, and F, hybrids between type I and type I1 mice are tolerant to both protein F types. However, in a bone marrow chimera made by injecting bone marrow from type I1 into irradiated type I mice, the bone marrow cells mature in the absence of type I1 protein and can respond to immunization with the type I1 protein (Winchester, 1979). Responsiveness is inhibited by injection of type I1 protein F in soluble form,
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
223
supporting the contention that tolerance to self antigens is acquired. Thus, the responsive state to protein F obviously is under genetic influence. There seems to be an absolute requirement for an H - 2 k allele at the K or I-A region of H-2 to allow a response to protein F (Silver and Lane, 1977). In F, hybrids, this requirement stands, but is not sufficient to guarantee a response; a non-H-2-linked gene is also required (Silver and Lane, 1975). Furthermore, there is a non-H-2linked gene affected by antigen dose that governs low responder-high responder phenotypes (Silver and Lane, 1979). c. WFetoprotein. Several investigators have reported that injection of animals with either heterologous or chemically altered afetoprotein results in the production of autoantibodies reactive with a-fetoprotein of the immunized host (Ruoslahti and Wigzell, 1975; Ruoslahti et nl., 1975; Ruoslahti and Seppala, 1971a; Pihko et nl., 1973). This protein of approximately 70,000 mw is produced abundantly in the fetal liver, yolk sac, and certain tumors (Abelev, 1974; Ruoslahti et d.,1974).The high concentrations (several milligrams per milliliter) also present in the sera of normal fetuses drop markedly (nanograms per milliliter) by adulthood (Ruoslahti and Seppala, 1971b, 1972). However, the a-fetoprotein in fetal rabbits and humans are serologically indistinguishable from those in adults (Ruoslahti et al., 1975). Antibodies produced in rabbits, horses, or rats against human a-fetoprotein cross-react, respectively, with rabbit, horse, or rat a-fetoprotein (Nishi et al., 1973). Injection of either rabbits (Ruoslahti et al., 1975) or monkeys (Ruoslahti and Wigzell, 1975) with homologous a-fetoprotein in CFA does not result in antibody production. Yet, rabbit a-fetoprotein in CFA injected into monkeys results in antibody to rabbit a-fetoprotein that is cross-reactive with both human and monkey a-fetoprotein (Ruoslahti et nl,, 1975). Likewise, the immunization of rabbits with rabbit a-fetoprotein modified b y coupling with a hapten yields antibody to both the modified protein and native rabbit a-fetoprotein (Ruoslahti and Wigzell, 1975). Like the Tg model, the a-fetoprotein model seems to provide tolerant helper T cells in the presence of competent B cells. The tolerant state in T cells is bypassed b y the cross-reacting a-fetoproteins, and B cells are triggered to make antibody to this a-fetoprotein, which then cross-reacts with native, homologous a-fetoprotein. Limited attempts to show tumor resistance in adult animals treated with anti-a-fetoprotein (Mizejewski and Allen, 1974) or immunized with CFA containing homologous a-fetoprotein (Goussev and Yasova, 1974) have been unsuccessful; however, definite degenerative changes have been demonstrated in the liver of fetuses from rats immunized with homologous a-
224
WILLIAM 0. W'EIGLE
fetoprotein (Nishi et ml., 1973). Information on either the genetic control or the T cell regulation of this response is not yet available.
B. EXPERIMENTAL ALLERGIC ENCEPHALOMYELITIS (EAE) Encephalomyelitis resulting from immunization was first observed in humans vaccinated with rabies virus extracted from infected heterologous tissue of the central nervous system (CNS) (reviewed in Kolb, 1950). Rivers and associates (1933; Rivers and Schwentker, 1935) later showed that most experimental animals repeatedly injected with heterologous brain extracts in emulsions over a prolonged period developed inflammatory lesions accompanied by demyelination. Twelve more years passed before Freund et ul. (1947), Kabat et nl. (1947), Morgan (1947), and Morrison (1947) independently reported that a single injection of CNS tissue emulsified in CFA regularly induced severe disseminated encephalomyelitis in guinea pigs, monkeys, and rabbits within a relatively short period of time. Now EAE is a widely studied animal model for determining cellular and subcellular events in the pathogenesis of demyelinating diseases and is, beyond any doubt, immunologically based. It can be best induced by immunization with CFA containing CNS tissue, myelin basic protein (BP) and peptides either isolated from CNS tissue or synthetically prepared. The disease is characterized by deposition of fibrin in CNS, acute vasculitis, associated with demyelination of nerve fibers and paralysis (Paterson, 1977, 1978).In contrast to some models of EAT in which antibody seems to initiate the disease, cellular immunity is apparently responsible for the lesions and clinical symptoms of EAE. Although there are differences between EAE and multiple sclerosis (MS), these diseases are sufficiently similar so that EAE is regarded as an experimental counterpart of MS in humans (Paterson, 1979a). The production of cytotoxic antibodies by animals with EAE and patients with MS is taken as evidence that immunologic events play an important role in the development of both diseases. These complementdependent cytotoxic antibodies are toxic for glial cells and cause myelin breakdown in cultures of myelinating brain cells (reviewed by Bornstein, 1973). In MS patients, IgG seenis to be deposited selectively within the plaques of demyelination (Tourtellotte and Parker, 1966, 1967; Tourtellotte, 1972), and there is some evidence that the IgG extracted from these patients' lesions causes demyelination of brain cell cultures (Kim et ul., 1970). However, no circulating antibodies have been detected in MS patients. Moreover, only a few animals with EAE seem to have IgG deposits associated with their
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
225
CNS lesions (Oldstone and Dixon, 1968). A direct correlation between DTH, as reflected by cutaneous reactivity to CNS tissue antigens, and development of EAE has been well documented in experimental animals (reviewed by Alvordet al., 1975). Yet, in the few relevant studies, none of the patients with MS had cutaneous reactivity to CNS tissue (Lisak et al., 1968), although the peripheral blood cells of some generated MIF in response to CNS tissue antigens (Bartfeld and Atoynatan, 1970; Bartfeld et al., 1972; Rocklin et al., 1971; Sheremata et al., 1976), a response that is usually considered to be an in uitro correlate of DTH. However, purified B cell preparations have also been shown to secrete MIF (Cohen and Yoshida, 1977). The morphologic changes found in MS are similar to some of those in EAE, particularly in the chronic recurrent variant of the experimental disease (Prineas et al., 1969; McFarlin et al., 1974; Raine et al., 1974; Snyder et al., 1975; Wisniewski and Keith, 1977), which is characterized by lesions of different ages, including large demyelinated plaques like those seen in MS patients. The two diseases are dissimilar with respect to the long-term effects on oligodendrocytes and remyelination. With MS, oligodendrocytes are lost in the center of plaques and during remission do not regenerate except at the margin. In contrast, during the recovery from EAE, oligodendrocytes proliferate and remyelination occurs throughout the areas of earlier demyelination (Lambert, 1967, 1978).These findings are of special interest in view of a recent report that sera from MS patients contain antibodies to oligodendrocytes (Abramsky et al., 1977a). Mackay and Camegie (1977) have also theorized that patients with MS may develop autoimmune reactivity to serotonin receptors on oligodendrocytes, eventually blocking the receptors’ interactions with serotonin. Prolonged blockage could lead to death of the oligodendrocyte and demyelination resulting from failure to replace metabolically active protein in myelin. A role for serotonin receptor blockage in EAE is suggested by the work of Weinstock et al. (1977), who described guinea pigs with EAE in which the neuronal receptors for serotonin in the ileum showed evidence of blockage. A relationship between serotonin and EAE has been reported by other investigators as well: the levels of serotonin are increased in the brains of rabbits with EAE (Cazzulo et al., 1969); the encephalitogenic determinant for BP is structurally similar to the serotonin receptor (Lennon and Carnegie, 1971; Carnegie, 1972), and serotonin and other hallucinogenic indoles block the reaction between encephalitogenic protein and lymphocytes sensitized to the protein (Carnegie et aZ., 1972; Carnegie, 1973).
226
WILLIAM 0. WEIGLE
1. Models of EAE Acute EAE is a transient, monophasic disease inducible by a single injection of heterologous, homologous, or syngeneic CNS tissue, BP, or either natural or synthetic polypeptides of BP incorporated in CFA (reviewed in Paterson, 1978). Clinical symptoms appear in 2-3 weeks and include loss of tonus of the tail and ataxic gait resulting from weakness of the hind legs that can progress to total hind leg paralysis, urinary incompetence, and usually death. The main pathologic feature is acute vasculitis accompanied by perivascular myelin damage confined to the nervous system (Paterson, 1966,1976). Myelin is present in the form of sheaths tightly packed in multilayers around axons (Lambed, 1978). In the CNS tissue these sheaths are formed by compaction and fusion of plasma membranes from cytoplasmic extensions of oligodendrocytes. A single oligodendrocyte is connected to many myelin segments by long cytoplasmic projections (Bunge, 1968; Hirano and Dembitzer, 1967). In addition to demyelination of the axons, an early pathologic change is the deposition of fibrin inside and outside vessel walls. The infiltrating mononuclear cells include lymphocytes, histiocytes, and some plasma cells (Paterson, 1978). In this model, guinea pigs and rats that survive the injection usually recover completely. A hyperacute form of EAE is produced when aqueous preparations of CNS tissue are injected with a suspension of Bordetella pertussis organisms (Levine and Wenk, 1965; Levine et al., 1966) or when B . pertussis vaccine is injected along with CNS tissue (Lee and Olitsky, 1955; Levine and Wenk, 1967) or BP peptides (Westall and Lennon, 1977) in CFA. Paralytic signs appear in some of these animals within 7-8 days. Lesions in the brain and spinal cord are characterized by extensive deposits of fibrin, vessel wall damage, and accumulations of polymorphonuclear cells (Levine and Wenk, 1967). The role of B . pertussis in promoting EAE is unknown, although pertussis vaccine causes T and B cells to migrate from lymphoid tissue into the circulation as well as marked lymphocytosis (Morse, 1965; Taub et al., 1972) and enhanced release of mediators of immediate hypersensitivity (Reed et al., 1972). Although pertussis vaccine induces IgE antibody in rats (Tada and Okumura, 1971),this IgE apparently plays no part in the pathogenesis of their EAE (Moore et al., 1974a,b). Additionally, injections of CNS tissue along with the histamine sensitizing factor pertussigen, a protein isolated from B . pertussis, cause acute EAE in rats that is accompanied b y lymphocytosis and a marked increase in vascular permeability in the CNS (Bergman et al., 1978). Consequently, the role of pertussis vaccine in EAE may be that of enhancing
AUTOIMMUNITY: THYROIDITIS A N D ENCEPHALOMYELITIS
227
vascular permeability (Munoz and Bergman, 1977) and thus allowing competent lymphocytes to accumulate in CNS tissue (Bergman et al., 1978). No relationship between this experimental model and the occasional occurrence of encephalopathies following pertussis immunization in man is known. Chronic forms of EAE have also been described in which there are periods of remission and exacerbation. This type of EAE occasionally occurs in immunized monkeys, cats, and dogs (Paterson, 1959a) and particularly in rabbits that survive acute episodes of EAE (Prineas et all., 1969; Raine et al., 1971).A chronic relapsing form of EAE, which is highly reproducible, has been induced by immunizing immature (juvenile) strain 13 guinea pigs with a single injection of CFA containing isologous spinal cord (Raine et al., 1974; Snyderet al., 1975). After a l-month delay, the onset of EAE becomes obvious and is followed by periods of remission and exacerbation. The resultant lesions contain areas of both acute demyelination and chronic demyelinationremyelination. Demyelination is accompanied by gliosis and fibrosis of blood vessels. Chronic EAE is considered by many to bear a strong resemblance to MS (Wolf et al., 1947; Ferraro and Cazzullo, 1948; Adams, 1959; Stone and Lemer, 1965). 2. Antigens
EAE-inciting antigens include a number of proteins, such as acidic proteins, glycoproteins, proteolipids, and BP, all of which are specific for CNS tissue, have been isolated, and are at least partially characterized (reviewed in Rauch and Einstein, 1974; Kies, 1975; Hashim, 1978; Paterson, 1978). Basic protein is readily extracted from delipidated CNS tissue at an acidic p H and purified by various combinations of column chromatography (Kies, 1975; Rauch and Einstein, 1974). The BP constitutes approximately 33% of the total myelin protein in spinal cord tissue and about 1% of the total dry weight of CNS tissue. It is b y far the most extensively characterized autoantigen associated with a major autoimmune disease. Considerable research has documented that the BP in myelinated nerve fibers accounts in large part for the encephalitogenic activity of whole CNS tissue when incorporated in CFA and injected into experimental animals. It is also the CNS tissue protein responsible for the development of postvaccination encephalomyelitis (Paterson, 1978). a . Physicochemical Properties. Basic protein is cationic because of its high content of basic amino acids (lysine, histidine, and arginine), which represent 25% of the total amino acids (Rauch and Einstein, 1974). The amino terminus on bovine and human BP has been iden-
228
WILLIAM 0. W I G L E
tified as N-acetyl-L-alanine, and the carboxyl terminus as arginine (Hashim and Eylar, 1969a; Eylar, 1970; Chao and Einstein, 1970a).As determined by a variety of procedures, the molecular weight ranges between 18,000 and 20,000 (reviewed in Rauch and Einstein, 1974; Kies, 1975). However, according to amino acid analysis, BP from human and bovine myelin has a molecular weight of 18,200 (Carnegie et al., 1967; Eylar and Thompson, 1969) and its isoelectric point is approximately 10.5 (Carnegie, 1971). One of its most important features is its open configuration, as implied by its resistance to heat, low pH, urea, and other denaturing agents (Eylar and Thompson, 1969). The unfolded configuration of this protein is further evidenced by optical rotatory measurements indicating few or no alpha helices (Eylar and Thompson, 1969; Palmer and Dawson, 1969; Chao and Einstein, 1970a). Thus, the tertiary or secondary structures of BP are less important than the primary structure in defining the determinants responsible for its biologic activities. Immunizing with peptides and testing for antibody reactivity to the intact BP molecules have demonstrated the presence of buried antigenic determinants, suggesting that sufficient secondary structure exists to cause internalization of a portion of the molecule (Whitaker et al., 1977). Among the unusual features of BP amino acid sequences is a Pro-Arg-Thr-Pro-Pro-Pro sequence, which might allow a sharp bend in the molecule so that it folds back on itself (Smyth and Utsumi, 1967) and shields some determinants. Of all mammalian BP studied thus far, each contains a total of 170 amino acid residues (Eylar, 1970; Carnegie, 1971),with the exception of rat BP, which contains in addition to the 170 residue molecule a smaller molecular species that lacks the 40 amino acid sequence at the carboxyl terminus (Martenson et al., 1972a). In guinea pigs, the single tryptophan residue at position 116 of the BP molecule is particularly important to its encephalitogenic properties, The amino acid sequences of both bovine (Eylar et al., 1971) and human (Carnegie, 1971) BP appear in Table VII. b. Peptides with Encephalitogenic Activity. The specific structural requirements of BP necessary to induce EAE have been clarified by numerous investigators during the past 10 years using both artificially synthesized peptides and peptides obtained by enzymically degraded BP (reviewed in Rauch and Einstein, 1974; Kies, 1975; Hashim, 1978). It is clear from these studies that encephalitogenic peptides have unique amino acid sequences. On the other hand, peptides with unique sequences are encephalitogenic in some species, but not in others, and the regions in the molecule that yield encephalitogenic peptides with BP from one species may yield only nonence-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
229
TABLE VII THE AMINO ACID SEQUENCE O F THE BASICPROTEIN"," (-) His-Gly N-Ac-Ala-Ser-Ala-Gln-Lys-Arg-Pro-Ser-Gln-Ar~-Ser-Lys-Tyr-Leu-Ala-
10
Thr
Ser-Ala-Ser-Thr-Met-Asp-His-Ala-Arg-His-Gly-Phe-Leu-Pro-Ar~-His20
30
Ile
GlY
Arg-Asp-Thr-Gly-Ile-Leu-Asp-Ser-Len-G1y-Ar~-Phe-Phe-Gly-Ser-Asp40
Ser
Ar~-Gly-Aln-Pro-Lys-Arg-Gly-Ser-Gly-Lys-Asp-Gly-His-His-Ala-Ala-Arg50
60
Ala
Thr
Ser ( - )
Thr-Thr-His-Tyr-Cly-Ser-Leii-Pro-Glii-Lys-Ala-Gln-Gly-His-Arg-Pro-Gln 70
80
Asp-Glu-Asll-Pro-Val-Val-His-Phe-Phe-Lys-Asn-Ile-Val-Thr-Pro-Arg-Thr90
Me
Pro-Pro-Pro-Ser-Gln-Gly-Lys-Gly-Arg-Gly-Leu-Ser-Le~i-Ser-Arg-Phe-Ser100
110
.4rg
Trp-Cly-Ala-Glu-Gly-Gln-Lys-Pro-Gly-Phe-Gly-Tyr-Gly-Gly-Ar~-Al~-Ser 120
130
Phe
Val
Asp-Tyr-Lys-Ser-Ala-His-Lys-Cly-Len-Lys-Gly-His-Asp-Ala-Gln-Gly-Thr140
Leu-Ser-I,ys-Ile-Phe-Lys-Leu-Cly-~ly-Arg-Asp-Ser-Arg-Ser-Gly-Ser-Pro 150
160
R.Iet-.4la-Arg-Arg-COOH 170 Reprinted from Hashim (1978). '!The ;unino acid sequence of the hovine myelin basic protein is shown. Deletions and sulxtitutions in the huiman myelin basic protein are superimposed. "
230
WILLIAM 0. WEIGLE
phalitogenic peptides in BP isolated from another species. Furthermore, there is more than one encephalitogenic area within the intact BP molecule (Hashim, 1978). Peptides obtained by pepsin digestion of bovine BP and injected into guinea pigs provided the first data suggesting that the encephalitogenic properties of BP reside in determinants with unique amino acid sequences. Later, peptides derived from trypsin digestion of bovine BP established that the encephalitogenic sequence was Phe-Ser-Trp-Gly-Ala-Glu-Gly-Gln-Lys (Eylar, 1970) and represented residues 114-122 in the parent molecule (Table VIII). Human BP contains a similar nine amino acid peptide with arginine replacing the C-terminal Lys residue, and this peptide is encephalitogenic in guinea pigs (Camegie, 1971). Both bovine and human BP peptides have now been synthesized and shown to be encephalitogenic in guinea pigs (Eylar et al., 1970; Westall et al., 1971; Hashim et al., 1973).Of special interest is the requirement for the tryptophan residue for activity. If this residue is either blocked or deleted, the encephalitogenic properties are lost. When the C-tryptophyl bond in the intact BP is cleaved with N-bromosuccinimide (Eylar and Hashim, 1969) or blocked with 2-hydroxy-5-nitrobenzyl bromide, the protein becomes inactive in guinea pigs (Chao and Einstein, 1970b; Swanborg, 1970; Eylar et ul., 1972). Furthermore, deletion or substitution of the tryptophan residue in synthetic peptides abolishes their activity in guinea pigs. Although blockage of the tryptophan residues renders intact BP nonencephalitogenic in guinea pigs, it has no effect on its activity in other species. Furthermore, the peptide comprising residues 114-122 has little or no activity in rabbits, rats, and monkeys (Swanborg, 1970; Eylar et al., 1972a). Studies with various synthetic peptides indicate that five amino acids of the nonapeptide are required for EAE induction in guinea pigs (Westall et al., 1971; Westall, 1972; Westall and Thompson, 1975). The required amino acids are ( a ) glutamine; ( b )a positively charged amino acid adjacent to glutamine; (c ) tryptophan located five residues from the glutamine toward the amino terminal TABLE VIII AMINOACIDSEQUENCE OF THE ENCEPHALITOGENIC TRYITOPHANREGION BASICPROTEIN Source Bovine Human
Amino acid sequence
H-Phe-Ser-Trp-Gly-Ala-Glu-Gly-Cln-Lys-OH H-Phe-Ser-Trp-Gly-Ah-Glu-GIy-Gln-Arg-OH
OF
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
231
end with the ( d ) glycine-C-terminally adjacent to it and ( e ) a hydrophobic region seven residues from the glutamine. It was also suggested that the alanine adjacent to the glycine may be important for disease induction. A uniquely encephalitogenic amino acid sequence is evident also in rats in which guinea pig BP is a highly effective inducer of EAE, whereas bovine BP is considerably less active (Martenson et al., 1975a). An amino acid sequence located in a peptide comprising residues 43-88 embodies the encephalitogenic properties of guinea pig BP (McFarlin et ul., 1973; Martenson et ul., 1972b, 1975a,b),with the major activity contained in the peptide composed of residues 68-88 (Westall et al., 1976). Martenson et (11. (1977) reported that a peptide with residues 72-84 is considerably encephalitogenic, although a peptide with residues 75-88 is inactive. This latter result is of particular interest, since Hashim et al. (1978) recently reported that a peptide with residues 75-84 of guinea pig BP is highly active. The means b y which an additional four terminal residues suppress encephalitogenic activity is not clear at present. Nevertheless, the minimum amino acid sequence for induction of EAE in Lewis rats is defined by the eight amino acid residues (H-Ser-Gln-Arg-Ser-Gln-Asp-Glu-Asn-OH). Interestingly, a slightly active peptide from bovine BP showed considerably greater encephalitogenic activity when a Gly-His sequence was cheniically deleted-a sequence that is naturally deleted in the peptide 75-84 isolated from guinea pig BP (Hashim et d . , 1978). The activity of a C-terminal peptide (residues 89-169) of bovine BP in Lewis rats (Martenson et al., 1975b) was shown to reside entirely within residues 89-115 (Martenson et al., 1977). Another protein obtained during the isolation of BP from the spinal cord of cows (Kibler et nl., 1964) and several other species (Shapiro et nl., 1971; Marks et al., 1974) was encephalitogenic in rabbits. This protein was 4140 mw compared to 18,200 mw for intact BP from each species and apparently had been degraded from the parent protein b y endogenous acid proteases (Kibler et ul., 1969).This smaller fragment contained 46 amino acids corresponding to residues 44-89. Tryptic and chymotryptic digestions of this fragment yielded a 10 amino acid peptide (H-Thr-Thr-His-Tyr-Gly-Ser-Leu-Pro-Gln-Lys-OH) corresponding to residues 65-74 that seemed to possess the total encephalitogenic activity. Recently, Westall and Thompson (1978) reported a 9 amino acid peptide (Phe-Lys-Leu-Gly-Gly-Arg-Asp-Ser-Arg) that corresponds to residues 154-162 of rabbit BP, which is another endogenous encephal itogen. Establishing the EAE-producing peptide(s) in mice has been
232
WILLIAM 0. WEIGLE
somewhat difficult, since EAE is not readily induced in most strains of mice. Olitsky and Yager (1949)induced EAE in outbred Swiss mice by repeated intramuscular and subcutaneous injections of mouse brain in CFA. Levine and Sowenski (1973) induced EAE in 3 of 14 inbred strains and 3 of 4 outbred strains with a single injection of mouse spinal cord in CFA with pertussis vaccine. Subsequently, Bernard and Carnegie (1975) were able to induce EAE in SJWJ mice with CFA containing mouse spinal cord, whole myelin, myelin BP, and rat small BP. Like the rat, the SJL/J mouse responds to determinants other than the tryptophan region that induces EAE in guinea pigs. A 14 amino acid peptide (residues 154-167) isolated from bovine BP (Karkhanis et ul., 1975) appears to carry the main encephalitogenic activity for monkeys and to differ from the main encephalitogenic regions f9r guinea pigs and rats. On the other hand, this peptide has striking similiarities to the 154-162 residue peptide of rabbit BP which is encephalitogenic in rabbits. In fact, the first nine residues of the former peptide are identical with the nine-residue (154-162) peptide encephalitogenic for rabbits. A peptide (residues 44-89) isolated from human BP, which contains the residues 65-74, known to be encephalitogenic in rabbits, (Shapiro et d . , 1971), and a peptide with residues 44-89 isolated from monkey (Kibler et ul., 1972)or bovine BP (Eylar et ul., 1972a) all have some encephalitogenic activity in monkeys. This latter peptide is in the same location as are the major residues encephalitogenic for the rat. Clearly the requirements for EAE induction vary from species to species. The inducing molecules may require specific peptides located at particular positions in the sequence. Nevertheless, one can observe similar sequences among some of these peptides. Westall and Thompson (1978)compared the amino acid sequences of four peptides containing 8-14 amino acids and having different patterns of encephalitogenicity depending upon the species. He concluded that their similar amino acid sequences seem to fulfill the requirement that encephalitogenic molecules, in general, have like structures (Westall and Thompson, 1978). c . Peptides with Immunogenic Actiuity. The areas on BP that react in DTH responses are considerably more numerous than those having encephalitogenic properties (Rauch and Einstein, 1974; Hashim, 1978). As evidenced by their ability to either block migration inhibition, generated by intact BP, or induce skin reactivity, DTH reactive peptides seem to b e distributed throughout the BP molecule, perhaps with some deletion of reactivity toward the middle of the bovine BP molecule. At least two antigenic determinants for induction of DTH in guinea pigs are located within the encephalitogenic peptides 44-89
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
233
(Bergstrand, 1973; Bergstraiid and Killen, 1973b). Regions 1 4 3 also seem to contain two determinants capable of eliciting DTH reactivity (Bergstrand and Kallen, 1973a). Data from several laboratories imply that one deteniiiriant is located in the sequence of residues 115-133 (Lennon et [il., 1970; Sptiler et nl., 1972; Lanioureux et al., 1972; Bergstrand, 1973);however, Bergstrand (1973)proposed that there are two determinants for DTH in the sequence 112-133. Other determinants for DTH apparently occupy residues 134-150 and 154-170 (Bergstrand, 1972a; Bergstrand and Killen, 1972). One determinant for DTH in guinea pigs clearly resides in the tryptophan fragment (Bergstrand, 19724, and guinea pigs injected with synthetic peptides that correspond to sequences 112-123 or 112-130 develop DTH to whole BP and EAE (Lennon e t al., 1970; Lennon and Cariiegie, 1974). Obviously a number of determinants on BP show DTH reactivity, but are not involved in eiicephalitogeiiesis. That the blockage of the single tryptophan residue in BP greatly impairs its encephalitogenic activity without altering its reactivity in i n vivo and i n v i t ro assays for DTH in guinea pigs (Eylar, 1970; Chao and Einstein, 19701); Spitler et a!., 1972) has serious implications for the role of DTH in EAE. Spitler et (11. (1972) also observed a synthetic peptide identical to the nonapeptide eiicephalitogenic for the guinea pig, but containing additional residues Ser and Arg on the N terminus that had the capacity to induce EAE in guinea pigs without eliciting DTH. Antigens that elicit BP antibody activity can be dissociated from antigens that induce EAE, For example, trypsiiiization or pepsin digestion prevents BP from reacting with anti-BP antibodies without affecting its encephalitogenic activity in guinea pigs. Conversely, digestion of BP with a-chymotrypsin does not alter the reactivity with anti-BP antibody, but does destroy the eiicephalitogenic activity (Hashini and Eylar, 196913). These results suggest that the antibody combining sites are located in regions other than the encephalitogenic tyrptophan region destroyed by the enzyme. Furthermore, studies in guinea pigs implicated three regions of the BP niolecule that were involved in eliciting antibody production, none of which are part of the 112-122 tryptophan fragment (Driscoll et nl., 1974a). McFarlin et al. (1975a) also observed that Lewis rats produced anti-BP antibody reactive with a site on the BP molecule other than the encephalitogenic site. d . Other Eiicephalitogenic Compotients ofChr.5 Tissue. The greater encephalitogenic activity per unit weight of intact CNS tissue than BP preparations suggests that either (a ) BP in intact CNS tissue, i.e., when associated with other components of CNS tissue, is more im-
234
WILLIAM 0. WEIGLE
munogenic; or ( h ) components of CNS tissue other than BP may contribute to the overall encephalitogenic activity. If the second explanation is valid, then other components more encephalitogenic than BP would have to contribute to the overall activity of CNS tissue. Although there is no evidence that CNS components other than BP have greater encephalitogenic activity, some data suggest that such components are, at least partially, responsible for EAE. Hoffman e t al. (1973) reported that equal numbers of guinea pigs injected with preparations of whole spinal cord, BP, or the encephalitogenic H-Phe-Ser-Trp-GlyAla-Glu-Gly-Gln-Lys-OH sequence developed clinical signs of EAE. Yet, perivascular demyelination was common in animals injected with spinal cord, less frequent with BP injections, and rare with the peptide injections. That this did not result from quantitative differences in the amount of effective antigens in each preparation emphasizes the possibility that other factors in CNS tissue are instrumental in EAE induction. This idea is supported by the fact that experimental animals injected with whole CNS tissue in adjuvant develop complementdependent cytotoxic antibodies that cause demyelination and glial cell changes, whereas such antibodies are not produced in animals injected with BP in adjuvant (Seil et n/., 1973). Conversely, rabbits immunized with cerebroside develop demyelinating and anti-glial cell antibody (Dubois-Dalcq et nl., 1970; Fry et al., 1974). Along this line, cerebroside levels in CNS tissue of guinea pigs change after injection of whole CNS tissue in adjuvant or a mixture of BP and cerebroside in adjuvant; injection of BP in adjuvant alone does not elicit such change (Maggio and Cuniar, 1975). Cerebrosides may also be involved in the induction of experimental allergic neuritis. Repeated immunization of rabbits with galactocerebroside (present in both central and peripheral nervous system tissue) results in the development of experimental allergic neuritis, which may relate more to circulating antibody than to DTH (Saida et nl., 1979). On the other hand, Lebar e t 01. (1976) reported that the specificity of demyelinating sera from guinea pigs iinniunized with CNS tissue in CFA was directed to neither BP nor cerebroside. Thus, although BP of myelin is obviously the major encephalitogenic antigen in CNS tissue, other antigens do play a yet undetermined role in EAE.
3 . Autoantibody nnd Cell-Mediated Immunity in EAE
In EAE it is well established that cell-mediated immunity plays the major role, although antibody may also participate in the progression of the disease. As in EAT, antigen-reactive lymphocytes to BP are present in both experimental animals and man. That is, lymph nodes
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
235
and peripheral blood from human subjects without neurologic disease, as well as lymph nodes from normal guinea pigs, contain lymphocytes that can bind '"I-labeled human BP (Yung et al., 1973), demonstrating that lymphocytes reactive with the self antigen BP apparently do exist in normal animals, as do lymphocytes capable of binding syngeneic Tg. In contrast to binding of Tg by B cells but not by T cells, both T and B cells apparently bind syngeneic BP (Table IX). Ortiz-Ortiz and Weigle (1976) evaluated 12sI-labeledsyngeneic BP bound by enriched populations of T and B cells of normal Lewis rats and concluded that both cell types bind this self antigen. These results are not surprising considering that thymocytes taken from normal Lewis rats can be sensitized in vitro with syngeneic brain antigen so that their injection into syngeneic recipients results in lesions suggestive of EAE (Orgad and Cohn, 1974). The ability to transfer EAE with cells but not serum strongly iniplicated cell-mediated immunity as the mediator of this autoimmune disease. Lipton and Freund (1953)first transferred EAE to normal rats by parabiosis with immunized rats. More recently, EAE was passively transferred to inbred rats by using lymph node cells (Paterson, 1960; Koprowski et d.,1960; Paterson and Didakow, 1961; Koprowski, 1962; Levine and Wenk, 1965, 1966, 1967; Paterson and Weiss, 1965; Paterson et al., 1975), spleen cells (Levine and Wenk, 1967), or blood leukocytes (Wenk et al., 1967)from donors previously immunized with CNS tissue antigens. In other studies, Rauch and Griffin (1969) sensitized donors with purified guinea pig encephalitogen and used their lymphocytes to transfer EAE, including paralysis and CNS lesions, to histocompatible recipients. Whitacre and Paterson (1977) extended these observations by showing that supernatants from lymph node cells of Lewis rats immunized with guinea pig spinal cord in CFA transferred EAE to syngeneic recipients. However, EAE was not TABLE IX ANTIGENBINDING CELLSIN NORMAL T AND B CELLSTO THYROGLOBULIN (TG)AND BASICPROTEIN (BP) OF MYELN
Species
Cell type
Antigen
Mice
T cell B cell T cell B cell
Tg TR BP BP
Rat
Antigen binding cells -
+ +
+
236
WILLIAM 0. VVEICLE
transferred with such supernatants when brain antigen was added, consistent with earlier studies suggesting that exposure of sensitized lymph node cells to BP may inhibit their ability to transfer EAE (Flad et al., 1968; Paterson, 1968). Contrasting data have now clearly demonstrated that in uitro incubation with BP under appropriate conditions greatly enhances the capacity of guinea pig peritoneal cells to transfer EAE (Driscoll et al., 1979), and as few as lo7cells gave reproducible results. Furthermore, adoptive transfer of EAE with splenic lymphocytes, but not lymph node lymphocytes, was potentiated by incubating the cells in uitro with concanavalin A (Con A) (Panitch and McFarlin, 1977). Thus, during the course of EAE, a population of T cells with immunologic memory for BP is generated and persists in the spleen. Incubation with Con A may activate these cells, resulting in enhanced T cell activity. The ability of T cells sensitized with BP to proliferate when exposed to Con A but not to BP prior to transfer was interpreted as signifying that T cell proliferation is not necessary for passive transfer of EAE (Richert et al., 1979). However, assays of cells mixed with BP were really too brief to eliminate a possible role for proliferation. In fact, others have reported proliferation of lymphoid cells from BPsensitized animals during subsequent in uitro stimulation with BP (Levinson et al., 1977; Panitch and McFarlin, 1977). Along these lines, lymphocytes from Brown Norway rats, compared to those from Lewis rats, do not proliferate as extensively in response to BP (Sheffield et al., 1977). These rats are resistant to EAE induced by immunization with BP (Levine and Sowinski, 1975; Kornblum, 1968; Gasser et al., 1973; Williams and Moore, 1973) and lack detectable T cells that bind BP, although their B cells bind BP (Ortiz-Ortiz and Weigle, 1976). Lymphocyte transformation, that is, increased proliferative responses by cells from sensitized animals in the presence of BP (Dau and Peterson, 1969; Shioiri-Nakano et al., 1971; Bergstrand, 1972b), has been used to implicate cell-mediated immunity in EAE. However, the heterogeneity of the T lymphocytes-helper T cells, cytotoxic T cells, suppressor T cells-must be considered because all may contribute to such proliferative responses. Until functional, purified T cell preparations are available and fully characterized as to proliferative ability and role in EAE, such studies add little insight into the cellular mechanisms of EAE. In contrast to the successhl transfer of EAE with lymphoid cells, transfer with serum is possible only under restricted conditions (Chase, 1959).Transfer experiments in which normal monkeys were injected intravenously, intracisternally, or intraperitoneally with sera
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
237
from diseased animals were unsuccessful in inducing EAE (Kabat et
al., 1948; Morgan, 1947). Nor could Waksman and Morrison (1951) transfer EAE into rabbits with intravenous injections of such serum. The failure to transfer EAE with serum is in line with previous reports regarding the poor correlation between antibody titers and the presence of CNS lesions (Thomas et d.,1950; Paterson, 1977, 1978; Rauch and Einstein, 1974) and the independence of antibody and EAE initiating determinants in BP (Section V,B,2,c). In addition, suppression of EAE by injection of encephalitogenic agents in incomplete Freund’s adjuvant (see Section V,B,5) has no effect on the level of circulating antibody to BP (Ortiz-Ortiz and Weigle, 1976; Coates et al., 1974; Lisaket al., 1970; Bernardet al., 1976),but at times is associated with a decline in DTH (Shaw et al., 1965b; Webb et al., 1974; Bernard et al., 1976; McGraw and Swanborg, 1978). On the other hand, antibody is sometimes implicated in EAE. Although myelination inhibitor factors have not been isolated from experimental animals immunized with BP (Kies et al., 1973; Seil et al., 1973, 1975), there is a close relationship between demyelination and cytotoxic glial (Koprowski and Fernandes, 1962) and myelin (Bornstein, 1973) antibodies in patients with MS and in experimental animals immunized with whole CNS tissue. Furthemiore, clinical manifestations and histologic lesions that resemble those seen in animals with EAE result from injecting antibody via routes that bypass the blood-brain barrier. Jankovic et (11. (1965) first reported the induction of EAE lesions in guinea pigs that were injected intraventricularly with serum containing antibody to brain. Subsequently Simon and Simon (1975) showed that serum containing anti-brain antibody injected into the ventricle or suboccipital subarachnoid space caused rabbits to develop mild symptoms and histopathologic changes seen in EAE. Antibody has also been iniplicated in the regulation of EAE. Paterson and Harwin (1963) showed that sera taken from rats that had recovered from EAE and injected into normal rats both during sensitization and 2 weeks afterward markedly reduced the frequency of EAE. A similar inhibition of EAE in rats by antiserum was reported by Hughes (1974) and Nakao and Einstein (1965). Although cell-mediated immunity is the major immunologic niechanism in EAE, the aforementioned studies d o not rule out the possibility that antibody may also have an important role; for example, the transfer of EAE could have resulted from activation of effector T cells or from antibody produced by the transferred cells. Further, it is unclear whether the ability to resist EAE after the depletion of T cells by thoracic duct drainage (Gonatas and Howard, 1974) or neonatal
238
WILLIAM 0. WEIGLE
thymectomy (Arnason et al., 1962; Wick and Steiner, 1972) is based upon removal of helper T cells required in the humoral response or effector T cells directly responsible for EAE lesions. In a more direct approach to evaluating effector and helper T cells in the initiation of EAE, the EAE-inducing activity of sensitized lymphocytes was assessed after removal of T cells (Ortiz-Ortiz et al., 1976). That T cells are required was confirmed by reconstituting irradiated Lewis rats with splenic and lymph node cells of normal rats, depleted of T cells, prior to challenge of the recipients with BP in CFA. Neonatal rats were thymectomized, irradiated, and then reconstituted with syngeneic lymph node cells pretreated with anti-thymus serum and complement; subsequent immunization with CFA containing syngeneic BP failed to induce either clinical symptoms or histologic lesions characteristic of EAE. Of greater importance, lymph node cells were removed from BP-sensitized donors 9 days after injection, when T cells were no longer required to sustain antibody production, and then used to reconstitute syngeneic, irradiated recipients that developed both the clinical symptoms and histologic lesions of EAE. However, prior treatment of the transferred cells with anti-thymus serum plus complement completely circumvented all symptoms and lesions in the recipients but had no effect on antibody production (Table X). Thus, although these recipients had levels of circulating antibody TABLE X EFFECTOF T CELLS ON INDUCTION OF EXPEFUMENTAL ALLERGIC ENCEPHALOMYELITIS (EAE) IN THYMECTOMIZED, IRRADIATED RATS RECONSTITUTED WITH PRIMEDCELLS",* EAE Treatment of transferred cells
Clinically
None ATS + C ATS (abs)e + C ~
~~
7/10'' 0/10 516 ~~
Histologically
Serum antibodyc
10/10 0/10 6/6
2.3 2.3 ND'
~~~
~
Reprinted in part from Ortiz-Ortiz et al. (1976). Lewis rats were thymectomized, irradiated (900 R), and reconstituted with 250 x lo6 lymph node and 350 x lo6 spleen cells from rats sensitized 9 days earlier with BP-CFA. The transferred cells were either untreated or treated with anti-lymphocyte serum (ATC) + complement (C). Values represent micrograms of basic protein bound per milliliter tested. Fraction of animals positive. Absorbed with thymus cells. 'ND, not done.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
239
comparable to recipients that received untreated sensitized donor cells, EAE was not induced. The roles of T and B cells in the initiation of EAE were further defined by antigen “suicide” experiments in which cells that bound BP and were separable according to either T or B cell function were specifically deleted by treatment with BP heavily labeled with Iz5I (Ortiz-Ortiz and Weigle, 1976).In preliminary experiments, irradiated Lewis rats reconstituted with lymphocytes from normal syngeneic donors readily developed EAE when immunized with BP in CFA. Since both T and B cells capable of binding BP are present in this population, it is not surprising that treatment with BP heavily labeled with I z 5 I abolished the cells’ ability to transfer EAE, apparently the result of their death by irradiation. Thymectomized, irradiated rats were also readily reconstituted with a mixture of purified thymus and bone marrow cells from normal rats in that the recipients injected with BP in CFA developed clinical symptoms of EAE, typical histologic lesions in the brain, and antibody to BP. On the other hand, neither symptoms, lesions, nor antibody to BP followed when the reconstituting thymus cells were treated (“suicided”) with [1251]BPprior to the transfers. Therefore, pretreatment with the [1251]BP eliminated specific T ceIIs and apparently abrogated an essential factor for cellmediated immunity. Treatment of the thymus cells with [12sI]BPalso inhibited the formation of antibody to BP, but not to burro red blood cells, demonstrating that specific helper T cells were also deleted. Supplementation of the [12sI]BP-inactivatedthymus cells with normal thymus cells restored the capacity to induce all the parameters of EAE. More precise evidence for cell-mediated immunity in EAE was obtained in the suicide experiments with B cells. When the bone marrow cells were treated with the heavily 1251-labeledBP and injected into thymectomized, irradiated recipients along with normal thymus cells, antibody formation to a subsequent challenge with BP in CFA was inhibited. In contrast, both clinical symptoms of EAE and histologic lesions were similar to those in rats that received normal thymus cells and normal bone marrow cells (Table XI). The absence of antibody production in animals with EAE again confirms the prior failure to transfer passively EAE with sera from sensitized animals or to induce EAE in bursectomized (agammaglobulinemic) chickens (Blaw et ul., 1967). Additional insight into the diversity of cellular events that may lead to autoimmunity is disclosed by studies with Brown Norway rats. As previously stated, not only are the Brown Norway rats resistant to
240
WILLIAM 0.WEIGLE TABLE XI ANTIGEN SUICIDE OF RAT THYMUS AND BONE MARROWCELLS BY 1251-LAHELEDSYNGENEIC BASICPROTEIN (BP)" _________
~~~~
EAE Bone marrow cells" Nomid Treated with IYS!!-Bp Noimal Treated with lZSII-BP
Thymus cells" Nomial N onnal Treated with lZsII-Bp Treated with I2 5 I - Bp
Clinically
Histologically
Antibody
10110 10110
10110 10110
+
015
015
-
0/10
0/10
-
-
Reprinted in part from Ortiz-Ortiz and Weigle (1976). 250 x 106bone marrow cells. L. 200 x 10"thyniocytes.
"
induction of EAE with BP in CFA, but they lack antigen binding T cells specific for BP (Ortiz-Ortiz and Weigle, 1976) and are deficient in T cells that proliferate in vitro in response to BP (Sheffield et al., 1977). This model may resemble EAT in which the T cells are unresponsive, while the B cells are competent. Then, as in EAT, the T cells may he bypassed and the B cells stimulated directly, resulting in the production of antibody to BP (but not disease). Such may be the case with the studies of Pitts et al. (1975), who induced antibody production, but not EAE, after immunizing Brown Norway rats with BP in CFA and pertussis vaccine, an adjuvant known to stimulate B cell proliferation (Murgo and Athanassiades, 1975) probably via nonspecific T cell help. The latter results further support the presence of tolerance to BP at the T cell, but not the B cell, level in the Brown Norway rat. The ability to induce some properties of EAE in these rats with homogenates of CNS tissue in CFA further supports the contention of difl'erences in the EAE-inducing potentials of BP and whole CNS tissue, which may be explained by a greater level of specific T cell reactivity in the latter preparation. Although there is little question that cell-mediated immunity is essential for the initiation of EAE, the exact T cell requirements are undetermined. Numerous reports describe a close relationship between induction of EAE and DTH as revealed by either skin testing or generation of migration inhibition activity (reviewed in Alvord et al., 1975).Recently, McFarlin et ul. (1975a,b) suggested a close relationship between the induction of EAE with a 43 amino acid
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
241
peptide and delayed skin reactions. In another study, neonatally thymectomized Lewis rats challenged with BP never developed clinical EAE and the incidence of lesions was markedly reduced (Lennon and Byrd, 1973);nevertheless, DTH (generation of MIF) was statistically associated with the occurrence of lesions. Of all the immunologic parameters studied, EAE seenis to correlate most closely with DTH responsiveness. However, the documented experiments with BP peptides reveal a dramatic discrepancy between DTH and EAE (see Section V,B,2,c). These clear-cut exceptions are difficult to ignore and suggest that different or additional cellular events other than those involved in DTH are required for the induction of EAE. Possibly, the lesions and resulting clinical symptoms are caused by cytotoxic T cells rather than T cells responsible for DTH. In mice, these two cell types are clearly separable because T cells responsible for DTH bear the Lyt 123- phenotype and cytotoxic T cells are of the Lyt 1-23+ phenotype (Vadas et ul., 1976). That cytotoxic T cells are the effector cells is compatible with the genetic restriction the MHC imposes in both EAE and MS (see Section V,B,4) since the target cell specificity of cytotoxic cells is regulated by the MHC locus. A major role for cytotoxic T cells would require that the target cells be oligodendrocytes whose surfaces express BP antigens as well as MHC antigens. In this model, the cytotoxic T cells sensitized to BP would react with both BP and an antigen of the MHC, either via dual recognition or altered self (Zinkeniagel and Doherty, 1979), on the surfaces of oligodendrocytes, thereby interfering with the ability of these cells to lay down myelin. In both EAE and MS, the oligodendrocyte has been proposed as a major target for antibody (Berg and Kallen, 1962a,b; Bomstein, 1973) and also sensitized cells (Koprowski and Fernandes, 1962). A further suggestion is that the antigen on the oligodendrocyte may be a serotonin receptor rather than myelin (Mackey and Camegie, 1977). As mentioned previously (Section V,B,3), one can find antibody capable of interfering with myelin production b y cultured glial cells in sera from animals immunized with whole CNS tissue in CFA and from patients with MS (reviewed in Bornstein, 1973). Moreover, the presence of cytotoxic cells in Lewis rats immunized with BP in CFA was suggested by the ability of lymphocytes from such rats to lyse fibroblasts preincubated with BP (Allbritton and Loan, 1975). In addition to T cells, macrophages ha've also been implicated in the induction of the histologic lesions of EAE and apparently are directly involved in demyelination (Lanibert and Carpenter, 1965; Yonezawa et uZ., 1968; Lambert, 1969; Raine and Bomstein, 1970);the deficiency of macrophages in immature rats may be partially responsible for the
242
WILLIAM 0. WEIGLE
resistance of these rats to EAE (Fujinami and Paterson, 1977; Dal Canto et al., 1977). Although immunization with CFA containing whole guinea pig CNS tissue causes suckling Lewis rats to respond with EAE, they d o not respond in this manner to injections of CFA containing soluble rat BP (Fujinami and Paterson, 1977).These immature rats mobilize macrophages poorly at inoculation sites, and possibly are defective in processing soluble antigens, which may account for their development of EAE after sensitization with whole CNS tissue, but not with soluble BP. Their immunologic deficiency, however, must be at the effector arm of the response rather than one of antigen recognition, since injecting macrophages from normal adult rats into neonatal Lewis rats already sensitized to BP renders them susceptible to EAE (Paterson, 197913; Paterson and Day, 1979).
4 . Zmmuriogenetics of EAE Like thyroiditis and a number of other autoimmune diseases, EAE is under genetic influence. Early work by Stone (1962) has shown that Hartley strain guinea pigs are more prone to EAE than those of strain 13, and strain 2 guinea pigs are resistant to EAE. Gasser et al. (1973) first suggested that the susceptibility of various strains of rats to EAE is closely linked to the MHC. Later, Williams and Moore (19731, who used susceptible Lewis rats and resistant Brown Norway rats, concluded that an autosomal dominant gene (Zr)linked to the histocompatibility locus determines susceptibility to EAE. Working with these same two strains, McFarlin et (11. (1975a,b) provided evidence that an Zr gene not only determines susceptibility to EAE, but also controls all other immune responses to both BP and the encephalitogenic fragment (residues 45-86). Others have also shown strain variability in rats (Hughes and Stedronska, 1973; Swanborg et al., 1974) and guinea pigs (Swanborg et ul., 1974) in susceptibility to EAE. More recently, Singer et u1. (1979) suggested that an ZT-EAE gene controls the response in backcrosses between Lewis and Brown Norway rats. These authors observed that susceptibility to disease depended upon at least one MHC haplotype (AgB') of Lewis rats and suggested that the genetic control of demyelination was at the level of sensitization rather than at the effector level of lymphocytes. Similarly, Gasser et ul. (1978) concluded that susceptibility and resistance to EAE in the Lewis and Brown Norway strains are controlled by a single dominant gene closely linked to the AgB locus. On the other hand, the observation that the susceptible DA strain and resistant AVN strain are both of the AgB4 halotype forced these investigators to conclude that other genes must be involved. Thus, they suggested that the susceptibility of rats
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
243
to EAE must be considered a polygenic trait and that nonhistocompatibility genes can substantially modify the effects of the MHC. Teitelbaum et al. (1978) extended these observations of Stone and concluded that the susceptibility of guinea pigs to EAE is linked to the MHC locus and probably controlled by two genes that segregate separately. Ben-Nun and Cohen (1978) believe that the restriction lies at the level of the interaction between macrophages and responding lymphocytes and suggest, as do Singer et al. (1979),that the main role of the l r gene is in controlling the interaction of the antigen-presenting inacrophages and responding T cells. In contrast to the conclusion reached with rats and guinea pigs, the influence of genetics in the susceptibility of mice to EAE is controversial. Bernard (1976) claims that susceptibility to EAE in mice is inherited as a dominant trait and is controlled in part by genes linked to the H - g Sand H-2 haplotype. However, Levine and Sowiiiski (1974) have concluded that the MHC does not control susceptibility of mice to EAE, and Teitelbaum et al. (1978) agree. Unfortunately, the stringent requirement for pertussis vaccine in addition to CFA for inducing EAE in susceptible mice may complicate the evaluation of such genetic studies. In patients with MS, one also finds a pattern of susceptibility for the disease related to histocompatibility determinants that control the immune response (reviewed in Kuwert and Bertrams, 1977).The risk of developing MS is considerably greater among family members than in the general population. Furthermore, there is a well-documented increase in the frequency of MS among persons with HLA phenotypes A3 and B7, and the lymphocyte determinant 7a (Jersild et al., 1975; Lamoureux et al., 1976). Patients with MS also seem to have abnormal immune responses to a variety of naturally occurring antigens (Lamoureux et al., 1976).
5 . linmunoregulation of EAE In addition to the genetic controls, EAE seems to be regulated by autoantigen, circulating antibody, and regulatory T cells. Although induction of EAE is dependent on a subpopulation of T cells that recognize and are activated by encephalitogenic determinants on BP, the reactivity of these T cells may be partly attenuated by circulating fragments of autologous BP (Paterson and Day, 1979). The fragments were first detected by using a radioimmunoassay to test sera from suckling rats (Day et ul., 197813; Fujinami et al., 1978).A factor was isolated from the sera that inhibited binding of '251-labeledrat BP to syngeneic anti-BP antibodies (Day et al., 1978a). Concentrations of
244
WILLIAM 0. WEIGLE
this serum factor became as high as 21 ng/pl in Lewis rats 2-3 weeks old (Fujinami et nl., 19781, then progressively declined to extremely low but still detectable levels in 7-week-old animals. It is particularly important that diminishing levels of the serum factor correlated inversely with the age-related increasing capacity of Lewis rats to develop EAE after sensitization with syngeneic BP. It was previously documented that BP of myelin lost encephalitogenic activity after in uitro incubation in sera (Lamoureux et nl., 1970; Bernard et nl., 1970; Gerstl et al., 1972; Vandenbark et ul., 1974; Bernard and Lamoureux, 1975a,b). Pescovitz et nl. (1978) similarly found evidence that encephalitogen is extensively degraded by trypsin-like proteolytic enzymes. Since the fibrin deposits in the lesions of animals with EAE implicated the clotting system, these researchers hypothesized that the serum factor was BP fragments resulting from serum-mediated degradation of BP and that encephalitogenic activity was lost simultaneously owing to enzymic digestion associated with the coagulation pathway. Thus, BP serum factor seems to be an endogenous regulatory product of potential importance for immunologic tolerance to BP at the T cell level. Endogenous BP may be responsible for the resistance to EAE in suckling Lewis rats injected with syngeneic BP in CFA, because of the T cells’ tolerant state, which declines with time. The difficulties in inducing EAE without the use of adjuvants may suggest that BP serum factor maintains some degree of (but not complete) tolerance in the T cells even during adult life. Steinnianet al. (1977)demonstrated autoantigen-mediated regulation of lymphocyte sensitization in vitro. Their model involved presenting normal lymphocytes to BP-pulsed macrophages and transferring the activated lymphocytes to normal syngeneic guinea pigs. Lymphocytes are recruited and proliferate in response to the BP-pulsed macrophages. The addition of soluble BP to the macrophages before these transfers abrogated their ability to recruit responsive lymphocytes. In other studies the addition of soluble BP to lymphocyte preparations from sensitized donors both inhibited and enhanced the transfer of EAE to irradiated recipients (Section V,B,3). As documented above, circulating antibody is not responsible for the initiation of either the clinical symptoms or the histologic lesions of EAE. However, in addition to a possible secondary role in the progression of the disease, circulating antibody has been implicated in the regulation of EAE. Immune sera collected from rats long after recovery of EAE, and then transferred to recipient rats within 2 weeks after sensitization with CNS tissue in CFA, prevented the development of EAE (Paterson and Hanvin, 1963). Hughes (1974) confirmed this re-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
245
sult but subsequently showed that the EAE-inhibiting activity of immune serum did not reside in antibodies directed to galactocerebroside (Hughes and Leibowitz, 1975). Hashim et al. (1975) used rabbits to induce a nonlethal form of EAE characterized by remissions and exacerbations. An inverse relationship between the presence of circulating antibodies and development of EAE again suggested that antibodies are a regulatory factor in this disease. Nakao and Einstein (1965) also showed that sera from animals with EAE have some protective effects. Immune sera from rats sensitized with CNS tissue in CFA suppress the clinical symptoms of EAE in rats, but only if injected close to the time of challenge. After the onset of paralysis, the injections of sera are no longer effective. In regulating the immune response, high concentrations of antibody may function by directly blocking specific determinants, and low concentrations in the form of antigen-antibody complexes may operate through interaction with Fc receptors (Weigle and Berman, 1979) on lymphocytes and/or macrophages. As in thyroiditis and other autoimmune phenomena, regulation of EAE b y suppressor cells has gained considerable attention over the past several years. Of even greater interest is the fact that such regulation may have considerable potential for controlling MS, the suggested human counterpart of EAE. Rats made tolerant to CNS tissue during neonatal life resisted EAE when challenged with CNS antigen in CFA (Paterson, 1958, 1959b; Koprowski et al., 1963). However, it is questionable whether this unresponsive state is similar to self-tolerance, because even early studies showed that aqueous preparations of CNS tissue injected either before or after sensitization of adult animals inhibited the development of this disease (Ferraro and Cazzullo, 1949a,b; Ferraro et al., 1950). Since then, numerous investigators have found that injecting aqueous preparations of either CNS tissue (Condie et al., 1975; Paterson, 1959a; Shaw et al., 1960; Svet-Moldavsky et al., 1959), or BP (Willenborg et al., 1978), or injecting incomplete Freund’s adjuvant containing either BP (Einstein et al., 1968; Alvord et al., 1965; Cunningham and Field, 1965; Lisak et al., 1970; Coates et al., 1974; Bernard et al., 1976) or CNS tissue (Svet-Moldavsky et al., 1959; Falk et al., 1969; Shaw et al., 1962) prevents the induction of EAE. Not only is EAE inhibited by pretreatment with BP, but even after sensitization the administration of BP prevents the disease (Einstein et al., 1968; Shaw et al., 1965a; Levine et al., 1972; Eylar et al, 1972b, 1979; Driscoll et al., 1974b; von Muller et al., 1978). Passive transfer of EAE with sensitized cells can be prevented by administering BP to the recipients (Levine et al., 1968, 1970), further indicat-
246
WILLIAM 0. WEIGLE
ing that suppression does not result solely from the failure to sensitize specific lymphocytes. Protection is also afforded by treatment with encephalitogenic (Driscoll et al., 1976; Eylaret al., 1979) and nonencephalitogenic peptides (Barton et al., 1972; Spitler et al., 1972; Hashim and Schilling, 1973) and CNS protein (MacPherson et al., 1977). In addition, BP rendered nonencephalitogenic b y blocking the critical tryptophan residue is effective in inhibiting EAE (Swanborg, 1972; Einstein et d., 1972). Moreover, the small molecule of rat BP, which lacks an essential part of the encephalitogenic site for guinea pigs, confers significant protection against EAE in the guinea pig (Hashim, 1978). Injecting guinea pigs with embryonic rabbit brain preparations in CFA does not result in EAE, but rather renders such guinea pigs refractory to EAE from a second injection with adult rabbit brain and spinal cord in CFA (Svet-Moldavskaya and Svet-Moldavsky, 1958). Injection of nonencephalitogenic peptides derived from chymotryptic digestion of BP (Hashim and Schilling, 1973) also protects subjects from EAE. Swanborg (1975) reported that guinea pigs given peptides 44-89, obtained by limited pepsin digestion of BP, were significantly protected from the EAE that subsequent immunization with BP in CFA should have induced; however, neither the peptide with residues 1-20 nor the synthetic EAE peptide with residues 114-122 prevented EAE in the same situation. Since the peptide with residues 44-89 is not encephalitogenic for guinea pigs (but is for rats and rabbits), although the peptide with residues 114-122 is, Swanborg concluded that the protective site differs from the encephalitogenic site. Hashim (1978) described a synthetic peptide (H-Phe-Ser-Trp-Gln-Lys-Phe-Ser-TrpGln-Lys-Phe-Trp-Gln-Lys-Phe-Ser-Trp-Gln-Lys-OH) that is not encephalitogenic in guinea pigs and offers protection even when injections are initiated after symptoms of the disease appear. If additional studies establish the existence of separate determinants for induction of EAE and protection, it will not be surprising since different antigenic determinants previously have been demonstrated for immunity and suppression with other antigens (Sercarz et nl., 1978). Nonencephalitogenic synthetic random polypeptides are also capable of suppressing EAE in guinea pigs (Teitelbaum et al., 1971; Webb et nl., 1976). Teitelbaum et nl. (1971, 1972) reported that a basic synthetic L-copolymer of alanine, glutamic acid, lysine, and tyrOsine with a molecular weight of 23,000 inhibits the induction of EAE in guinea pigs subsequently challenged with BP in CFA. The inhibition is specific in that neither the Damino acid copolymer nor acidic
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
247
copolymers prevent EAE (Teitelbaum et aZ., 1971). The basic L-copolymer also suppresses induction of EAE in rabbits (Teitelbaum et al., 1973) and monkeys (Teitelbaum et al., 1974). Furthermore, replacing tyrosine with tryptophan results in particularly effective suppression of EAE (Webb et aZ., 1976). It is of importance that the various synthetic copolymers shown to be suppressive also cross-react with bovine BP at the cellular level (Webb et al., 1973). Of special importance is the suppression of chronic relapsing EAE, the experimental disease most closely resembling MS. Beginning 1-2 weeks after juvenile guinea pigs were injected with BP in CFA, they received 10 more injections of incomplete Freund’s adjuvant containing isologous BP, which suppressed EAE in a highly effective and enduring manner (Raine et al., 1978). Unfortunately, attempts to induce similar suppression in patients with MS have not met with the same uniform success. However, Campbell et al. (1973) reported that of 64 MS patients, those desensitized to human BP did statistically better late in the trials than did patients receiving a placebo. Abramsky et al. (1977b) studied three patients with acute disseminated encephalomyelitis and four patients with MS injected with a nonencephalitogenic synthetic copolymer previously shown to inhibit EAE. Although the patients with acute disseminated encephalomyelitis recovered completely within 3 weeks, one of two control patients also recovered. The patients with MS showed no significant changes in motor function, but two improved somewhat in vision and speech. Another clinical trial reported by Gonsette et al. (1977) showed no benefit of BP for patients with MS. However, the difficulties in analyzing clinical trials of patients (Schuniacher, 1974) with MS make it impossible to evaluate the therapeutic value of BP or related peptides conclusive1y . Inhibition and reversal of EAE by using CNS antigens reinforces the dissociation between humoral antibody and the induction of EAE. Guinea pigs (MacPherson et al., 1977; Bernard et al., 1976) and rats (Lisak et al., 1970; Swierkosz and Swanborg, 1977a) treated with BP or CNS tissue developed antibody to BP but not disease. On the other hand, protection from EAE by pretreatment with BP often inhibited DTH (Lisak et al., 1970; Webb et al., 1974; Bernard et al., 1976; Bernard, 1977; McGraw and Swanborg, 1978). McGraw and Swanborg (1978) reported that pretreatment of rats with BP inhibited EAE and delayed skin reactivity, and despite having no effect on MIF, suggested an active suppressor mechanism at play in the animal. Conversely, von Muller et (11. (1978) reported that suppression in guinea
248
WILLIAM 0.WEIGLE
pigs was accompanied by inhibition of M I F but had no effect on skin reactivity. However, in these latter studies, suppressive injections of the encephalitogenic protein were not initiated until 7 days after sensitization. Suppressor lymphocytes have been strongly implicated as a mechanism b y which pretreatment with antigens of CNS tissue suppresses EAE. Swierkosz arid Swanborg (1975)demonstrated that lymph node cells from Lewis rats pretreated with guinea pig BP in incomplete adjuvant, when transferred to normal recipients, prevented the induction of clinical symptoms of EAE in response to BP in CFA, but not histologic lesions. The suppressor cells were T cells, as evidenced by their inactivation b y treatment with anti-thymocyte serum and complement (Welch and Swanborg, 1976). This suppressor cell activity was relatively short-lived, because protection from EAE waned in 3 weeks, although some suppressor activity was still present up to 8 weeks after treatment (Swierkosz and Swanborg, 1977a). Attempts were made to rule out the possibility of antigen carry-over or blocking antibody. Suppressor activity was present in both the lymph nodes and spleens, however, in the lymph nodes suppressor function was specific in that there was no deficiency in responsiveness to Con A or phytohemagglutinin (Welch et nl., 1978). In contrast, splenic cells froni the pretreated rats had depressed responses to the mitogens, indicating nonspecific suppression. Removal of adherent cells, presumably niacrophages, from the splenic cells abrogated the ability to transfer unresponsiveness to normal recipients. Thus, two different mechanisms of suppressor activity were probably involved: formation of specific suppressor T cells in lymph nodes, and nonspecific blockade of antigen presentation b y macrophages in the spleen. Suppressor cell activity is also generated in SJL/J mice pretreated with either murine spinal cord preparations or BP in incomplete adjuvant (Bernard, 1977). Suppression was present in bone marrow and splenic cells, but not in lymph node and thymus cells, and was abrogated by treatment with anti-theta serum and complement. In a similar study, (SJL/J x BALB/c) F, hybrid mice were injected with murine spinal honiogenates, murine BP, or a synthetic copolymer in incomplete adjuvant, before attempts to induce EAE (Lando et d.,1979).All three preparations induced an unresponsive state associated with suppressor cell activity, as evidenced by adoptive transfer studies. Treatment with small doses of cyclophosphamide, which preferentially inhibits suppressor cells and augments cell-mediated hypersensitivity (Askenase et nl., 1975; Rollenghoff et al., 1977; Schwartz et ul., 1978; Gill and Liew, 1978), also inhibited the induction of protection to EAE
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
249
1979). Conversely, it was demonstrated that large doses (Lando et d., of cyclophosphamide also inhibit the induction of EAE (Paterson and Drobish, 1969) and lead to a specific tolerance to EAE (Salvin and Liauw, 1968). Suppressor T cell activity was generated during the induction of acute EAE in Lewis rats immunized with BP in CFA plus pertussis vaccine (Adda et al., 1977). In the first stages of disease, suppressor cells were present, although they were absent during recovery. In a similar acute self-limiting model of EAE, Willenborg (1979) was also unable to show suppressor cell activity in Lewis rats recovering from EAE. With increasing time after primary sensitization with spinal cord or BP in CFA, rats recover and become ever more resistant to reinduction of either clinical or histologic signs of EAE. I t $ vitro, the lymphocytes of these resistant rats proliferated in response to BP, and when transferred to lethally, irradiated rats exposed to BP in CFA, the lymphocytes induced EAE. In addition, resistant, convalescent rats reconstituted with lymphocytes from sensitized rats developed disease, indicating the absence of suppressor cell activity in these convalescent rats. Suppression of EAE by soluble preparations of BP aparently does not involve suppressor cells. Ortiz-Ortiz and Weigle (1976) were unable to demonstrate suppressor cell activity in Lewis rats rendered unresponsive b y pretreatment with large amounts of soluble syngeneic BP. Similarly, Swierkosz and Swanborg (197721) failed to detect suppressor cells in Lewis rats pretreated with high dosages of soluble guinea pig BP, although EAE was suppressed. There is little question concerning the induction of suppressor cells during the response to antigens of the CNS when incorporated in incomplete Freund’s adjuvant. However, whether this suppression is responsible for resistance to EAE is questionable. It is unlikely that suppressor T cell activity is responsible for the reversal of ongoing EAE by treatment with CNS tissue antigens in incomplete adjuvant initiated after sensitization, since suppressor cells are incapable of inhibiting sensitized cells in adoptive transfer systems (Swierkosz and Swanborg, 1977b). Furthennore, injection of soluble BP induces resistance to EAE, without generating suppressor cells. Suppressor cells are also more closely related to events occurring early in the immune response than to the later period of recovery. A major role for suppressor cells is further complicated by finding nonspecific suppression in adherent cells from spleens of pretreated animals. If the immune response to BP in incomplete adjuvant involves cullular events similar to those involved in the immune response to foreign antigens, the
250
WILLIAM 0. WEIGLE
participation of regulatory cells would be expected; however, their association with disease could be coincidental. If suppressor cells are involved in resistance to EAE, they are probably only one of several participatory cellular and/or subcellular events. This does not mean that suppressor cells may not influence the disease once it is manifested. Recently, Huddlestone and Oldstone (1979) reported an association of nonspecific suppressor cells with periods of remission in MS patients. Nonspecific suppressor cells bearing the Fc receptor for IgG were quantitated by rosette formation. Patients in remission had a significantly greater number of these suppressor cells than normal subjects, whereas during acute exacerbation the levels were lower than normal. Whether such suppressor cells govern the induction of MS or merely regulate the disease once it is initiated cannot be answered at the present time.
6. Secondary Phenomena Both the human and experimental demyelinating diseases include numerous secondary neurologic disorders in the CNS that contribute to the overall clinical symptoms of the disease. Additionally, lesions occasionally occur in the ocular regions. Fog and Bardon (1953) reported uveitis in pigs with EAE resulting from immunization b y CNS tissue in CFA. In related studies, rabbits given injections of optic nerve in CFA developed typical EAE comparable to that produced with other sources of CNS tissue (Bullington and Waksman, 1958). Optic neuritis was frequently present and often involved the optic papilla and the nerve fiber layer of the retina. However, in rabbits with experimental allergic neuritis produced by injecting peripheral nerve in CFA, no intraocular lesions occurred. Either endogenous myelin antigen or an exogenous antigen can induce DTH reactions in the vitreous humor of rabbits’ eyes resulting in a mononuclear cell invasion of adjacent myelinated retinal axons with primary demyelination (Wisniewski and Bloom, 1975; Raine, 1976). Recently, Stoner et al. (1977) described an experimental model of demyelination within retinal fibers of rabbits that developed EAE after sensitization with spinal cord. This model lends itself to studying subcellular mechanisms, since there is no barrier preventing diffusion of proteins from the vitreous into the bundle of myelinated nerve fibers of the retina. Factors were prepared from supernatants of rabbit lymph node cells that had been nonspecifically activated with Con A-Sepharose and injected into the optic vitreous of the rabbits with EAE. The result was infiltration of mononuclear cells accompanied by demyelinating lesions within the retinal fibers. Furthermore, the injection of antibody
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
251
to CNS tissue along with the supernatant factors into the vitreous in normal rabbits’ eyes also caused primary demyelination and mononuclear cell infiltration (Brosnan et al., 1977).These results could indicate that antibody dependent cell-mediated mechanisms are partially responsible for the primary demyelination in both experimental and human demyelinating diseases. The above observations on EAE are of practical interest since MS patients commonly have afflictions of the optic nerve tracts or chiasm leading to retrobulbar neuritis, usually with transient blurring of vision and loss of sight. Also, congestion of the optic disks, characterized by optic papillitis or intrabulbar optic neuritis, is occasionally seen in these patients (reviewed in Paterson,
1978). VI. Concluding Remarks
Obviously numerous factors are involved in autoimmune diseases including genetic control at both MHC and non-MHC loci, immunologic regulatory mechanisms, self-nonself recognition, and logistics andlor nature of the target autoantigens. The primary requirement is that the autoantigen is reacted against by lymphocytes of the immune system. Conceivably, the presence of cells capable of reacting against self antigens may depend upon the concentration of the self antigens in the niicroenvironnient of potentially reactive cells rather than on the nature of the antigen. Because the level of antigen, either self or foreign, required to induce and maintain an immunologically tolerant state is considerably greater for B cells than T cells, T cells would be the major barrier to autoimmunity in respect to many self antigens. Whenever concentrations of self antigens are sufficiently high so that both T and B cells are tolerant, the onset of autoimmune disease would require abnormalities in the immune status of one or both of these cell types. At lower concentrations of self antigen, when T cells are tolerant and B cells are competent, autoimmunity could result from ( a ) bypass of either the specificity of or requirements for T cells or ( b )polyclonal activation of the competent B cells. In either event, the autoimmune disease would be mediated by antibody. If the antigen is either nonexistent in the body fluids, or in extremely low concentration, neither cell type would be tolerant; then both T and B cells could b e activated under appropriate conditions and the disease might be mediated by effector T cells and/or antibody. Polyclonal activation of B cells could be an additional mechanism of autoimmunity in the latter case. Suppressor cell activity must also be given careful consideration in
252
WILLIAM 0.WEIGLE
both the initiation and progression of autoimmune disease. The role of these cells in the course of the disease is probably similar to that of suppressor cells in regulating the immune response to foreign antigens. Here inducer cells, effector cells and antibody-producing cells are controlled by a network of mechanisms that involve feedback regulation b y inducer cells and generation of suppressor cells. How suppressor T cells or intermediate regulatory cells participate in the initiation of autoimmune disease is speculative. As addressed in this review, suppressor T cells can be detected during autoimmune processes in experimental models of autoimmunity, as they are detected in normal immune responses to foreign antigens. Furthermore, manipulations that limit suppressor cell activity enhance the induction of experimental and spontaneous autoimmunity in experimental animals, just as procedures that preferentially interfere with suppressor T cells also enhance the immune response to foreign antigens (Shek et al., 1978). The findings of both enhanced and reduced suppressor cell activity in immune deficiency diseases and depressed suppressor cell activity in certain autoimmune diseases (Reinherz and Schlossman, 1980) suggest a role for suppressor T cells in these human diseases. However, whether these abnormalities in suppressor cell activity are instrumental in the initiation of the diseases or are a consequence of the disease is not clear. If the lack of suppressor cells is instrumental in the onset of autoimmunity, these cells may b e involved only with self antigens that induce a tolerant state that is “leaky” at the T cell level. Suppressor T cell activity could then be generated by antigen concentrations that are ineffective in mounting an overt immune response. In any event, the role of suppressor cells may be that of a fail-safe mechanism rather than the first line of defense against autoimmune reactivity. Experimental autoimmune thyroiditis (EAT) and experimental allergic encephalomyelitis (EAE) are two experimental models that allow evaluation of contrasting subcellular and cellular mechanisms in autoimmunity. The self antigens responsible for each of these diseases are well defined and have permitted extensive experimentation on the related mechanisms of autoimmunity over the years. By using these models, the development of lesions in these diseases has been defined in at least two modes, one of which involves humoral antibody resulting from a bypass of T cell specificity (EAT), and another in which both T and B cells are activated by a defined self antigen and specific effector T cells initiate tissue injury (EAE). Since these diseases are readily induced in animals without inborn autoimmune manifestations, the preimmunization status of their regulatory cells is unimportant in initiating the reaction. However, the role of such cells in reg-
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELZTIS
253
dating the ongoing response to self antigens is most likely a factor in the disease that follows.
ACKNOWLEDGMENTS The author would like to thank Carole G. Roniball for her comments and help i n completing this manuscript. He also thanks Ms. Andrea Rothman and Phyllis Minick for editorial assistance and Ms. Janet Kuhns for expert secretarial assistance in the preparation of the manuscript. This is publication No. 2075 from the Department of Immunopathology, Scripps Clinic and Research Foundation, La Jolla, California. The experimental work reported was supported by United States Public Health Service Grants AI-07007, AI-15761, AG-00783, American Cancer Society Grant IM-421, and Biomedical Research Support Program Grant RRO-5514.
REFERENCES Abelev, G. I. (1974). Tran.r.plant. Reu. 20, 3. Abramsky, O., Lisak, R. P., Silberberg, D. H., and Pleasure, D. E. (1977a). N. Errgl. /. Med. 297, 1207. Abramsky, O., Feilelbaum, D., and Arnon, R. (1977b).J. Neurol. Sci. 31, 433. Ada, G. L., and Cooper, M. G. (1971). Ann. N.Y. Acud. Sci. 181, 96. Ada, G. L., Byrt, P., Mandel, T., and Warner, N. (1970). Z n “Developmental Aspects of Antibody Formation and Structure” (J. Sterzl and I. Riha, eds.), Vol. 1, p. 503. Academic Press, New York. Adams, D. D., and Kennedy, T. H. (1971)./. Clin. Endocrinol. 33, 47. Adams, R. D. (1959). I n “Allergic Encephalomyelitis” (M. W. Kies and E. A. Alvord, eds.), p. 183. Thomas, Springfield, Illinois. Adda, D. H., Beraud, E., and Depieds, R. (1977). Eur. J, I n t i n u t i d . 7, 620. Adenolfi, M., Gardner, B., Gianelli, F., and McGuire, M. (1967). Experientirr 23, 271. Aharonov, A., Kalderon, N., Silman, I., and Fuchs, S. (1975).Zinniutrochenristry 12, 765. Aldo-Benson, M., and Borel, Y. (1974).J . Zrnrnunol. 112, 1973. Al-Khateeb, S. F., and Irvine, W. J. (1978).J. Clin. Lab. InimutiO/. 1, 55. Allbritton, A. R., and Loan, R. W. (1975). Cell. Znzrnunol. 19, 91. Allison, A. C. (1971). Lancet 2, 1401. Allison, A. C. (1974). Zn “Immunological Tolerance: Mechanisms and Potential Therapeutic Applications” (D. H. Katz and B. Benacerraf, eds.), p. 25. Academic Press, New York. Alspaugh, M. A,, Jensen, F. C., Rabin, H., and Tan, E. M. (1978).J. Exp. Med. 147,1018. Alvord, E. C., Jr., Shaw, C.-M., Hruby, S., and Kies, M. W. (1965).Ann. N.Y. Acud. Sci. 122,333. Alvord, E. C., Jr., Shaw, C.-M., Hruby, S., Petersen, R., and Harvey, F. H. (1975). I n “The Nervous System” (D. B. Tower, ed.), Vol. 1, p. 647. Raven Press, New York. Anders, R. F., Cooper, P. C., Coffey, M. F., and Mackay, I. R. (1978).I n “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p. 393. Elsevier, Amsterdam. Anderson, C. L., and Rose, N. R. (1971).J. Zrnmunol. 107, 1341. Anderson, J. L., Moller, G., and Sjoberg, 0. (1972). Cell. Zrnrnunol. 4, 381. Anigstein, L., Eklund, G. W., and Whitney, D. M. (1957). Tex. Rep. B i d . bled. 15,205. Arima, T., Spiro, M. J., and Spiro, R. G. (1972).J. B i d . Chern. 247, 1825. Amason, B. G., Jankovik, B. D., Waksman, B. H., and Wennersten, C. (1962).J . E x p . Med. 116, 177.
254
WILLIAM 0.WEIGLE
Asherson, G. L., and Zembala, M. (1974).Proc. R. Soc. London, Ser. B. 187,329. Askenase, P. W., Hayden, B. J., and Gershon, R. K. (1975).J.E x p . Med. 141,697. Assem, E. S. K. (1964). Lancet 1, 139. Bacon, L. D., and Rose, N. R. (1979). Proc. Natl. Acad. Sci. U.S. A. 76, 1435. Bacon, L. D., Kite, J. H., Jr., and Rose, N. R. (1973). Transphtation 16,591. Bacon, L. D., Kite, J. H., Jr., and Rose, N. R. (1974). Science 186,274. Bacon, L. D., Cole, R. K., Polley, C. R., and Rose, N. R. (1978). In “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p. 259. Elsevier, Amsterdam. Baker, P. J., Stashak, B. P. W., Amsbaugh, D. F., and Prescott, B. (1974).J.Immunol. 112,2020. Bankhurst, A. D., Torrigiani, G., and Allison, A. C. (1973). Loncet 1,226. Barkas, T., Al-Khateeb, S. F., Irvine, W. J., Davidson, N. McD., and Roscoe, P. (1976). Clin. E x p . Immunol. 25, 270. Bartfeld, H., and Aloynatan, T. (1970). Int. Arch. Allergy Appl. Immunol. 39,361. Bartfeld, H., Aloynatan, T., and Dannenfeld, H. (1972). In “Multiple Sclerosis” (F. Wolfgram, ed.), p. 333. Academic Press, New York. Barton, M. A., McPherson, T. A., Lemieux, R. U., and Bain, G. 0. (1972). Can. /. Biochem. 50,689. Basten, A. (1974). In “Immunological Tolerance: Mechanisms and Potential Therapeutic Applications” (D. H. Katz and B. Benacerraf, eds.), p. 107. Academic Press, New York. Basten, A., Miller, J. F. A. P., and Johnson, P. (1975). Transplant. Rev. 26, 130. Baur, S., and Goodman, M. (1964).J.Immunol. 93, 183. Benacerraf, B., and McDevitt, H. 0. (1972).Science 175, 273. Benacerraf, B., Kapp, J. A., Debre, P., Pierce, C. W., and DeLaCroix, F. (1975).Tranyplant. Rev. 26, 21. Benjamin, D. C. (1974).J.Zmmunol. 113, 1589. Benjamin, D. C. (1975).]. E x p . Med. 141, 635. Benjamin, D. C. (1977a).J.Immutiol. 118, 2125. Benjamin, D. C. (1977b).J.Zmmunol. 119, 311. Benjamin, D. C., and Hershey, C. W. (1974).J.Zmmunol. 113, 1593. Benjamin, D. C., and Weigle, W. 0. (1970).J.E i p . Med. 132, 66. Ben-Nun, A., and Cohen, I. R. (1978). I n “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p, 381. Elsevier, Amsterdam. Berg, O., and Kallen, B. (1962a).Acta Pnthol. Microhiol. Scatid. 54, 425. Berg, O., and Kallen, B. (196213).Lancet 1, 1051. Bergman, R. K., Munoz, J. J., and Poltes, J. L. (1978). Infect. Zmmun. 21, 627. Bergstrand, H. (19724.Eur. J. Biochem. 27, 126. Bergstrand, H. (1972b).Eur. J. Immunol. 2,266. Bergstrand, H. (1973). Immuiiochemistry 10, 61 1. Bergstrand, H., and Kkllen, B. (1972). Cell. Immunol. 3, 660. Bergstrand, H., and Kallen, B. (1973a). Immunochemistry 10, 229. Bergstrand, H., and Kallkn, B. (197313).Inamunochemistry 10,471. Bergstrom, K., Frankson, C., Matell, G., and von Reis, G . (1973).Eur. Neurol. 9, 157. Bemian, P. W., and Patrick, J. (1980).J.E x p . Med. 151, 204. Bernard, C., Chabot-Debabes, P., and Lamoureux, G . (1970).Proc. Can. Fed. Biol. Sci. 13, 164. Bernard, C . C . , and Lamoureux, G. (1975a). Int. Arch. Allergy Appl. Immunol. 48, 597. Bernard, C. C., and Lamoureux, G . (1975b). Cell. Immunol. 16, 182. Bernard, C. C. A. (1976).J. Immunogenet. 3,263.
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
255
Bernard, C. C. A. (1977). Cliir. E x p . Inimunol. 29, 100. Bernard, C. C. A., and Carnegie, P. R. (1975).J.Immunol. 114, 1537. Bernard, C. C. A., Mackay, I. R., Whittingham, S., and Brous, P. (1976). Celt. Iminunol. 22, 297. Berti, F., Clemente, F., Conti-Tronione, B., and Falco, G. (1976).Br. J. Phnrmucol. 57, 17. Bigazzi, P. E., and Rose, N. R. (1975).Prog. Allergy 19, 245. Billingham, R. E., Brent, L., and Medawar, P. B. (1953).Nature (London ) 172,603. Binz, H., and Wigzell, H. (1975).J.E x p . Med. 142, 197. Binz, H., and Wigzell, H. (1977). Contemp. Top. Imniunohiol. 7, 111. Binz, H., and Wigzell, H. (1978).J.E x p . Med. 147, 63. Blaw, M. E., Cooper, M. D., and Good, R. A. (1967). Science 158, 1198. Bloth, B., and Bergquist, R. (1968).J.E x p . Med. 128, 1129. Bodnier, W. F., and Bodmer, J. C. (1978).Br. Med. But/. 34, 309. Bomstein, M. B. (1973). Prog. Neuropnthol. 11, 69. Boulter, J., and Patrick, J. (1977). Biochemistry 16, 4900. Braley-Mullen, H . , Sharp, G. C., and Kyriakos, M. (1975).J.Imnunol. 114, 371. Braley-Mullen, H., Sharp, G. C., Kyriakos, M., Hayes, N., Dunn, C., Jepsen, P., and Sanders, R. D. (1978). Cell. Inimunol. 39, 289. Brosnan, C. F., Stoner, G. L., Bloom, B. R., and Wisniewski, H. M. (1977).J.Znirnunol. 118,2103. Brostoff, J. (1970). Proc. R . Soc. Med. 63, 905. Brown, F., and Jackson, H. (1956).Biochem. J . @2,295. Brown, J. C., and Rodkey, L. S. (1979).J. E x p . Med. 150, 67. Bullington, S. J., and Waksman, B. H. (1958).Arch. Ophthnlmol. 59,435. Bunge, R. P. (1968). Physiol. Ret;. 48, 197. Bumet, F. M. (1959). “The Clonal Selection Theory of Acquired Immunity.” Cambridge Univ. Press, London and New York. Burnet, F. M. (1977).I n “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects” (N. Talal, ed.), p. 513. Academic Press, New York. Bystryn, J.-C., Schenkein, J., and Uhr, J. W. (1971). 111 “Progress in Immunology” (B. Amos, ed.), p. 627. Academic Press, New York. Calder, E. A., McLeman, D., Barnes, E. W., and Irvine, W. J. (1972).Clin. E x p . Zmmunol. 12,429. Calder, E. A,, Penhale, W. J., McLennan, D., Barnes, E. W., and Irvine, W. J. (1973). C / k E s p . Iinntunol. 14, 153. Calder, E. A., Penhale, W. J., Barnes, E. W., and Irvine, W. J. (1974). Br. Med.1. 2,30. Calder, E. A., Penhale, W. J., and Irvine, W. J. (1975). Proc. R . S O C . Med. 68, 250. Cambier, J. C., Vitetta, E. S., Uhr, J. W., and Kettman, J. R., Jr. (1977a).]. E x p . Med. 145, 778. Cambier, J. C., Vitetta, E. S,, Kettman, J. R., Wetzel, G. M., and Uhr, J. W. (1977b).J . E x p . Med. 146, 107. Campbell, B., Vogel, P. J . , Fisher, E., and Lorenz, R. (1973).Arch. Neurol. (Chicago)29, 10. Cantor, H., and Boyse, E. A. (1977a).Imnunol. Reu. 33, 105. Cantor, H., and Boyse, E. A. (1977h). Contemp. Top. Inimunohiol. 7,47. Cantor, H., and Gershon, R. K. (1979). Fed. Proc., Fed. Am. Soc. E x p . B i d . 38,2058. Cantor, H., Hugenberger, J., McVay-Boudreau, L., Eardley, D. D., Kemp, J., Shen, F. W., and Gershon, R. K. (1978).J.E x p . Med. 148,871. Capen, C. C., Belshaw, B. E., and Martin, S. L. (1975). In “Textbook of Veterinary Internal Medicine. Diseases of the Dog and Cat” (S. Ettenger, ed.), Vol. 2, p. 1351. Saunders, Philadelphia, Pennsylvania.
256
WILLIAM 0.WEIGLE
Camegie, P. R. (1971).Nature (London) 229,25. Carnegie, P. R. (1972).Z n “Multiple Sclerosis, Progress in Research” (E. J. Field, T. M. Bell, and P. R. Carnegie, eds.), p. 63. North-Holland Pub., Amsterdam. Camegie, P. R. (1973). “Neuroscience Research Study Program.” Pergamon, Oxford. Camegie, P. R., Bencina, B., and Lamoureux, G. (1967). Riochem. J. 105, 559. Camegie, P. R., Smythies, J . R., Caspary, E. A., and Field, E. J. (1972).Nature (London) 240, 56 1. Cazzulo, C. L., Brazelli, C. A., and Giardano, P. L. (1969). Znt. Arch. Allergy Appl. Zmmunol. 36, 234. Cerottini, J.-C., McConahey, P. J., and Dixon, F. J. (1969a).J. Zmmunol. 102, 1008. Cerottini, J.-C., McConahey, P. J., and Dixon, F. J. (1969b).J.Immunol. 103,268. Chandler, R. W., Kyle, M. A., Hung, W., and Blizzard, R. M. (1962).Pediatrics 29,961. Chao, L. P., and Einstein, E. R. (1970a).J.Neurochem. 17, 1121. Chao, L. P., and Einstein, E. R. (1970b).J.Biol. Chem. 245, 6397. Charlwood, P. A., Pitts-Rivers, R., and Schwarz, H. L. (1970).Biochem. J. 116, 769. Chase, M. W. (1959).I n “Allergic Encephalomyelitis” (M. W. Kies and E. C. Alvord, Jr., eds.), p. 348. Thomas, Springfield, Illinois. Chiller, J. M., and Weigle, W. O., (1971).J.Zmmunol. 106, 1647. Chiller, J. M., and Weigle, W. 0. (1973a).J.Immunol. 110, 1051. Chiller, J. M., and Weigle, W. 0. (1973b).J.E x p . Med. 137, 740. Chiller, J . M., Habicht, G. S., and Weigle, W. 0. (1970).Proc. Natl. Acad. Sci. U . S . A. 65, 551. Chiller, J. M., Habicht, G. S., and Weigle, W. 0. (1971).Science 171,813. Chiller, J. M., Louis, J. A,, Skidmore, B. J., and Weigle, W. 0.(1974).In “Immunological Tolerance: Mechanisms and Potential Therapeutic Applications” (D. H. Katz and B. Benacerraf, eds.), p. 373. Academic Press, New York. Christadoss, P., Kong, Y. M., Elrehewy, M., Rose, N. R., and David, C. S. (1978). I n “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and E. L. Cooper, eds.), p. 445. Elsevier, Amsterdam. Christadoss, P., Lennon, V. A., and David, C. (1979).J. Zmmunol. 123, 2540. Cinader, B., Dubiski, S., and Wardlaw, A. C. (1964).J. E x p . Med. 120,897. Clagett, J. A., and Weigle, W. 0. (1974).J.E x p . Med. 1:39, 643. Clagett, J. A., Wilson, C. B., and Weigle, W. 0. (1974).J.E x p . Med. 140, 1439. Clinton, B. A., and Weigle, W. 0. (1972).J.E x p . Med. 136, 1605. Coates, A., Mackay, I. R., and Crawford, M. (1974). Cell. Zmmunol. 12, 370. Cohen, P. L., and Ziff, M . (1977).J.Zmmunol. 119, 1534. Cohen, S., and Yoshida, T. (1977).J. Immunol. 119, 719. Cohn, Z. A., and Morse, S. I. (1960).J.E x p . Med. 111, 667. Cole, R. K. (1966). Genetics, 53, 1021. Cole, R. K., Kite, J. H., Jr., and Witebsky, E. (1968). Science 160, 1357. Cole, R. K., Kite, J. H., Jr., Wick, G . , and Witebsky, E. (1970). Poult. Sci. 49, 839. Condie, R. M., Kelly, J. T., Campbell, B., and Good, R. A. (1957).Fed. Proc., Fed. Am. Soc. E x p . Biol. 16, 24. Corradin, G., Etlinger, H. M., and Chiller, J. M. (1977).]. Immunol. 119, 1048. Corsini, A. C., Clayton, C., Askonas, B. A., and Ogilvie, B. M. (1977). Clin. E x p . Zmw1unol. 29, 122. Coutinho, A., and Moller, G. (1975). Ado. Zmntunol. 21, 113. Creighton, W. D., Katz, D. H., and Dixon, F. J. (1979).J.Zmmunol. 123,2627. Cunningham, V. R., and Field, E. J. (1965).Ann. N.Y. Acud. Sci. 122, 346. Daguillard, F., Heiner, D. C., Richter, M., and Rose, B. (1969). Clin. E x p . Immunol. 4, 203. Dal Canto, M. C., Fujinami, R. S., and Paterson, P. Y. (1977). Lab. Invest. 37, 395.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
257
Daniel, P. M., Pratt, 0. E., Roitt, I. M., and Torrigiani, G. (1967a). Q . J . E x p . Physiol. Cogtr. Med. Sci. 52, 184. Daniel, P. M., Pratt, 0. E., Roitt, I. M., and Torrigiani, G. (1967b). Irrtnturtofogy 12,489. Dau, P. C., and Peterson, R. D. A. (1969).I t i t . Arch. AUergy Appl. Irrttnurd. 35, 353. Dau, P. C., Lindstrom, J. M., Cassel, C. K., and Clark, E. C. (1979). I n “Plasmapheresis and Immunobiology of Myasthenia Gravis” (P. C. Dau, ed.), p. 229. Houghton, Boston, Massachusetts. Day, E. D., Varitek, V. A,, Fujinami, R. S., and Paterson, P. Y. (1978a). IitinruirocheiiiistrlJ 15, 1. Day, E. D., Varitek, V. A,, and Paterson, P. Y. (1978b). Ir,ii,rurtochemistry 15, 4.37. Debre, P., Kapp, J. A,, and Benacerraf, B. (1975).J. E x p . hled. 142, 1436. Denman, A. M., and Freukel, E. P. (1968).Imriiutiology 14, 107. Denys, E. H., Dau, P. C., and Lindstrom, J. M. (1979).I n “Plasmapheresis and the Immunobiology of Myasthenia Gravis” (P. C. Dau, ed.), p. 248. Houghton, Boston, Massachusetts . Dixon, F. J., Maurer, P. H., and Weigle, W. 0. (1955).J . Intntutrol. 74, 188. Dixon, F. J., Vazquez, J . J., Weigle, W. O., and Cochrane, C. G. (1958).AAIA Arch. Pathol. 65, 18. Doebbler, T. K., and Rose, N. R. (1961). Fed. Proc., Fed. Am. Soc. E x l ~ B. i d . 20, 39. Dolly, J. O., and Bamard, E. A. (1977). Biochtnistry 16, 5054. Doyle, M. V., Parks, D. E., and Weigle, W. 0. (1976).J. Zmmutiol. 116, 1640. Dresser, D. W. (1962).Zmniuttok~gl~ 5, 378. Dresser, D. W., arid Mitchison, N. A. (1968). A d c . Z t n m u i d . 8, 129. Driscoll, B. F., Kramer, A. J., and Kies, M. W. (19744. Science 184, 73. Driscoll, B. F., Kies, M. W., and Alvord, E. C. (1974b).J. Intmutud. 112, 392. Driscoll, B. F., Kies, M. W., and Alvord, E. C., Jr. (1976).J. Irnrnuno/. 117, 110. Driscoll, B. F., Kies, M. W., and Alvord, E. C. (1979). Science 203, 547. Dubois-Dalcq, M., Niedieck, B., and Buyse, M. (1970). P o t l d . Eur. 5 , 331. Dunim,. S. K., Eardley, D. D., Green, D. R., Ywniauchi, K., Murphy, D. B., Cantor, H., and Gershon, R. K. (1980). Fed. Proc., Fed. Anr. Soc. E x ? ) . B i o l . 39, 353. Eardley, D. D., Hugenherger, J., McVay-Boudreau, L., Shen, F. W., Gershon, R. K., and Cantor, H. (1978).J. Ewp. Med. 147, 1106. Eardley, D. D., Shen, F. W., Cantor, H., and Gershon, R. K. (1979).J. E x / ) . .%fed.150,44. Edelhoch, H. (1965). Recent Prog. Horm. Res. 21, 1. Edelhoch, H., and Lippoldt, R. F. (1964). Biochitn. B i o l ~ h y sActcl . 79, 64. Edelhoch, H., and Rall, J. E. (1964). I n “The Thyroid Gland” (R. Pitt-Rivers and W. R. Trotter, eds.), p. 113. Buttenvorth, London. Eichmann, K., and Rajewsky, K. (1975). Eur. J. Imnrurrol. 5, 661. Einstein, E. R., Csejtey, J., Davis, W. J., and Rauch, H. C . (1968). Im,,rtrttochet?tistrrJ 5, 567. Einstein, E. R., Chao, L.-P., and Csejtey, J. (1972). Imitruitochetriistr~/9, 1013. Eldefrawi, M . E. (1978). Fed. Proc., Fed. Am. Soc. Ex?). Biol. 37, 2823. Eldefrawi, M. E., and Eldefrawi, A. T. (1977). Rccept. Recognition, Ser. A 4, 197. Engel, A. G., Lindstrom, J., Lambert, E. H., and Lennon, V. A. (1977). Neurology 27, 307. Engel, A. G., Sakakibara, H. Sahashi, K., Lindstrom, J., Lambert, E., and Lennon, V. (1979). Neurology 29, 179. Esquivel, P. S., Rose, N. R., and Kong, Y. M. (1977).j. E x p . Med. 145, 1250. Etlinger, H. M., and Chiller, J. M. (1977). Cell. Znrmuttol. 33, 297. Eylar, E. H. (1970). Proc. N d . Acnd. Sci. U.S.A. 67, 1425. Eylar, E. H., and Hashim, G. H. (1969).Arch. Biochetn. Biophys. 131, 215. Eylar, E. H., arid Thompson, M. (1969).Arch. Biocfieirt. Biophys. 129, 468.
258
WILLIAM 0.WEIGLE
Eylar, E . H., Caccam, J., Jackson, J., Westall, F., and Robinson, A. P. (1970). Science 168, 1220. Eylar, E. H., Brostoff, S . W., Hashim, G. A., Caccam, J., and Bumett, P. (1971).J.Biol. Chem. 246,5770. Eylar, E . H., Brostoff, S., Jackson, J., and Carter, H. (19724. Proc. Natl. Acad. Sci. U S A . 69, 617. Eylar, E . H., Jackson, J., Rothenberg, B., and Brostoff, S . W. (1972b). Nature (London) 236, 74. Eylar, E. H., Jackson, J. J., and Kniskem, P. J. (1979).Neurochem. Res. 4,249. Falk, G. A., Kies, M. W., and Alvord, E. C. (1969).J.Zmmunol. 103, 1248. Feldmann, M., and Diener, E. (1972).J.Zmmunol. 108, 93. Feldt-Rasmussen, U., Petersen, P. H., Nielsen, H., Date, J., and Madsen, C. M. (1978). Clin. Endocrinol. (Oxford)9, 205. Felix-Davies, D., and Waksman, B. H. (1961).Arthritis Rheum. 4,416. Feltkamp, T., Van Den Bergt-Loonen, P., Njenhuis, L., Engelfriet, C., Van Rossum, A., Van Loghem, J., and Oosterhuis, H. (1974). Br. Med. J. 1, 131. Femandez, C., and Moller, G. (1977).J.Exp. Med. 146, 308. Femandez, C., and Moller, G . (1978). Scand. J . Immunol. 7, 137. Fernandez, C., Hammarstrom, L., Moller, G., Primi, D., and Smith, C. J. E. (1979). Znmunol. Reo. 43, 3. Ferraro, A., and Cazzullo, C. L. (1948).J.Neuropathol. E x p . Neurol. 7, 235. Ferraro, A., and Cazzullo, C. L. (1949a).J.Neuropathol. E x p . Neurol. 8,61. Ferraro, A., and Cazzullo, C. L. (1949b).J.Neuroputhof. Exp. Neurol. 8,226. Ferraro, A., Roizin, L., and Cassullo, C. L. (195O).J.Neuropathol. E x p . Neurol. 9, 18. Fidler, J. M . (1979).J.Exp. Med. 150,491. Flad, H . D., Paterson, P. Y., and Miescher, P. A. (1968).J. Immunol. 100,647. Flax, M. H. (1963).Lab Invest. 12, 199. Flax, M. H., Jankovik, B. D., and Sell, S. (1963).Lab. Znoest. 12, 119. Flier, J., Kahn, C., and Roth, J. (1979).N . Engl. J . Med. 300, 413. Fog, T., and Bardon, M. (1953). Nord. Med. 49, 851. Forbes, I. J., Roitt, I. M., Doniach, D., and Solomon, I. L., (1962)J. Clin. Znuest. 41,996. Fravi, G., and Lindenmann, J. (1968).Nature (London) 218, 141. Freeman, J., Smithers, S . R., Targett, G. A. T., and Walker, P. J. (1970).]. Infect. Dis. 121, 401. Freeman, R. R., and Parish, C. R. (1978). Clin. E x p . Zmmunol. 32,41. Freund, J., Stem, E. R., and Pisani, T. M. (1947).J.Zmmunol. 57, 179. Fritze, D., Herrmann, C., Jr., Smith, G. S., and Walford, R. L. (1973). Luncet 2, 211. Fritz, T. E., Norris, W. P., and Kretz, N. D. (1970a).Proc. SOC. E x p . Biol. Med. 134,450. Fritz, T. E., Zeman, R. C., and Zelle, M. R. (1970b).E x p . Mol. Pathol. 12, 14. Froehner, S. C., Reiness, C. G., and Hall, Z. (1977).J . Biol. Chem. 252, 8589. Fry, J. M., Weissbarth, S., and Lehrer, G. M. (1974). Science 183, 540. Fuchs, S., Nevo, D., and Tarrab-Hazdai, R. (1976).Nature (London) 263,329. Fujinami, R. S., and Paterson, P. Y. (1977).J.Zmmunol. 119, 1634. Fujinami, R. S., Paterson, P. Y., Day, E. D., and Varitek, V. A. (1978).J.E x p . Med. 148, 1716. Fujiwara, M., and Fujiwara, S. (1975) Immunology 29, 1171. Fujiwara, M., and Kariyone, A. (1978). Immunology 34, 51. Gasser, D. L., Newlin, C. M., Palm, J., and Gonatas, N. K. (1973). Science 181, 872. Gasser, D. L., Boloyannis, S. J., and Gonatas, N. K. (1978). In “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p. 355. Elsevier, Amsterdam.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
259
Gausset, P., Delespesse, G., Hubert, C., Kennes, B., and Covaerts, A. (1976l.J.Zmmunol. 116, 446. Gershon, R. K. (1974).Contemp. Top. Irnmunohiol. 3, 1. Gershon, R. K. (1977).In “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects” (N. Talal, ed.), p. 171. Academic Press, New York. Gerstl, B., Uyeda, C. T., Hunt, H., and Eng, L. F. (1972).Immunology 23, 395. Gery, I., Kruger, J., and Spiesel, I. Z. (1972).J.Zmmunol. 108, 1088. Ghayasuddin, M., and Shulman, S. (1970). Zmmunology 19,225. Gill, H. K., and Liew, F. Y. (1978). Eur. 1.fmmunol. 8, 172. Glover, E. L., and Reuber, M. D. (1967). Endocrinology 80,361. Godal, T., and Karesen, R. (1967a).Actn Pathol. Microbiol. Scand. 69, 332. Godal, T., and Karesen, R. (1967b). Acta Pathol. Microhiol. Scand. 69, 343. Goidl, E. A., Schrater, A. F., Siskind, G. W., and Thorbecke, G. J. (1979).J . E x p . Med. 150, 154. Gonatas, N. K., and Howard, J. C. (1974).Science 186, 839. Gonsette, R. E., Delmotte, P., and Delmonty, L. (1977).J.Neurol. 216,27. Goodman, M. G., and Weigle, W. 0. (1979).J . Zmmunol. 122,2548. Gosselin, S., Capen, G. C., and Martin, S. L. (1978).Am. 1.Pathol. 90,285. Goussev, A., and Yasova, A. (1975). Cited in Ruoslahti and Wigzell (1975). Green, D. P. L., Miledi, R., and Vincent, A. (1975).Proc. R. Soc. London, Ser. B 189,57. Green, I., Benacerraf, B., and Stone, S. H. (1969).J . Zmmunol. 103, 403. Greenwood, R. M. (1974). Lancet 1,435. Gronowicz, E., and Coutinho, A. (1975).Eur. J. Immunol. 5, 413. Haakenstad, A. O., and Mannik, M. (1977).In “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects” (H. Talal, ed.), p. 277. Academic Press, New York. Habicht, G. S., Chiller, J. M., and Weigle, W. 0. (1975).J.E x p . Med. 142, 312. Haeberli, A., Belstad, J., Edelhoch, H., and Rall, J. E . (1975a).J.B i d . Chem. 250,7294. Haeberli, A,, Salvatore, G., Edelhoch, H., and Rall, J.E. (1975b).J . B i d . Chem. 250, 7836. Hajdu, A., and Rona, G. (1969).Experientiu 25, 1325. Hall, R. (1962). N. Engl. J . Med. 266, 1204. Hall, R., and Stanbury, J. B. (1967). C f i n .E x p . Imniunof. 2, 719. Hall, R., Owen, S. G., and Smart, G. A. (1960). Luncet 2, 187. Hashim, G. A. (1978).Zmmunol. Rev. 39,60. Hashim, G. A., and Eylar, E. H. (1969a).Biochenr. Biophys. Res. Cominun. 34, 770. Hashim, G. A., and Eylar, E. H. (1969b).Arch. Biochem. Biophys. 129, 635. Hashim, G. A., and Schilling, F. J. (1973).Arch. Biochem. Biophys. 156, 287. Hashim, G. A., Hwang, F., and Schilling, F. J. (1973).Arch. Biochem. Biophys. 156, 298. Hashim, G. A., Sharpe, R. D., Carvalho, E., and Stevens, L. E. (1975).Proc. SOC. EXTJ. Biol. Bled. 149, 646. Hashim, G. A., Carvalho, E. F., and Sharpe, R. D. (1978).J.Immunol. 121, 665. Heidelberger, M., and Pedersen, K. 0. (1935).J.Cen. Physiol. 19, 95. Heilbron, E., and Mattson, C. (1974).J. Neurochem. 22, 315. Hektoen, L., Fox, H., and Schulhof, K. (1927).J.Infect. Dis. 40,641. Hellstrom, I., and Hellstrom, K. E. (1969).Znt. J . Cancer 4, 587. Hellstrom, I., Hellstrom, K. E., and Allison, A. C. (1971).Nature (London) 230, 49. Herzenberg, L. A., Okumura, K., and Metzler, G. M. (1975).Trunsplunt. Reo. 27, 57. Hilgert, I. (1979).Zmmunol. Reu. 46, 27. Hirano, A,, and Dembitzer, H. M. (1967).J.Cell B i d . 34, 558. Hjort, T., and Pedersen, G. T. (1962).Lancet 2,259.
260
WILLIAM 0.WEIGLE
Hoffman, P. M., Gaston, D. D., and Spitler, L. E. (1973). Clin. Zmmunol. Zmmunopathol. 1,364. Hoffmann, M. K., and Kappler, J. W. (1978). Nature (London) 272,64. Hoffmann,M. K., Kappler, J. W., Hirst, J. A., and Oettgen, H. F. (1974). E u r . J . Zmmunol. 4,282. Houba, V., Brown, K. N., and Allison, A. C. (1969).Clin.E x p . Zmmunol. 4, 113. Howard, J. G. (1972). Transplant. Reo. 8, 50. Howard, J. G., and Mitchison, N. A. (1975).Prog. Allergy 18,43. Huddlestone, J. R., and Oldstone, M. B. A. (1979).J.Zmmunol. 123, 1615. Hudson, K. M., Byner, C., Freeman, J., and Terry, R. J. (1976).Nature (London) 264,256. Hughes, R. A. C. (1974). J. Zmmunol. 26, 703. Hughes, R. A. C., and Leibowitz, S. (1975).Zmmunology 28, 213. Hughes, R. A. C., and Stedronska, J. (1973). Zmmunology 24,879. Humphrey, J. H., and Keller, H. A. (1970). In “Developmental Aspects of Antibody Formation and Structure” (J. Sterzl and I. Riha, eds.), Vol. 1, p. 485. Academic Press, New York. Irvine, W. J. (1962). Br. Med. J. 1, 1444. Irvine, W. J., and Muri, A. R. (1963).Q. J . E x p . Physiol. Cogn. Med. Sci. 48, 13. Iverson, G. M., and Lindenmann, J. (1972). E u r . j . Zmmunol. 2, 195. Izui, S., Lambert, P.-H., Fournie, G . J., Turler, H., and Miescher, P. A. (1977a).J. E x p . Med. 145, 1115. h i , S., Zaldivar, N. M., Scher, I., and Lambert, P.-H. (1977b).J. Zmmunol. 119, 2151. Izui, S., Kobayakawa, T., Zryd, M.-J., Louis, J., and Lambert, P.-H. (1977c).J. Zmmunol. 119,2157. Izui, S., McConahey, P. J., and Dixon, F. J. (1978).J. Zmmunol. 121,2213. h i , S., Eisenberg, R. A., and Dixon, F. J. (1979a).J. Zmmunol. 122,2096. h i , S., Kobayakawa, T., Louis, J., and Lambert, P.-H. (197913).E u r . J. Zmmunol. 9,339. Jakobisiak, M., Sundick, R. S., Bacon, L. D., and Rose, N. R. (1976). Proc. Natl. Acud. Sci. U.S.A. 73, 2877. Jankovic, B. D., and Flax, M. H. (1963).J. Zmmunol. 91, 1978. Jankovi6, B. D., and Isvaneski, M. (1963). Znt. Arch. Allergy Appl. Zmmunol. 23, 188. Jankovic, B. D., and Mitrovib, K. (1963). Nature (London) 200, 186. Jankovic, B. D., Draikoci, M., and Janjib, M. (1965).Nature (London) 207,428. Jankovic, B. D., Mitrovic, K., Popeskovic, L., and MiloSevik, D. (1969). Clin.E x p . Zmmunol. 5,29. Jaroszewski, J.. Sundick, R. S., and Rose, N. R. (1978). Clin. Zmmunol. Zmmunopathol. 10, 95. Jarret, D., Roth, J., Kahn, R., and Flier, J. (1976). Proc. Natl. Acad. Sci. U . S. A. 73, 4115. Jeme, N. K. (1974).Ann. Zmmunol. (Paris) 125c, 373. Jeme, N. K., and Nordin, A. A. (1963). Science 140,405. Jersild, C., Dupont, B., Fog, T., Platz, P. J., and Svejgaard, A. (1975).Transplant. Reo. 22, 148. Jones, H. E. H., and Roitt, I. M. (1961).Br. J. E x p . Pathol. 42,546. Jordan, S. C., Johnston, W. H., and Bergstein, J. M. (1978). Arch. Pathol. 102,530. Jordan, S . C., Buckingham, B. A., Kogut, M. D., and Fine, R. N. (1979).Am. Soc. Nephrol. Abst., p. 142a. Kabat, E. A., Wolf, A., and Bezer, A. E. (1947).J. E x p . Med. 85, 117. Kabat, E. A., Wolf, A., and Bezer, A. E. (1948). 1. E x p . Med. 88,417. Kapp, J. A., Pierce, C. W., Schlossman, S., and Benacerraf, B. (1974).J. E x p . Med. 140, 648.
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
26 1
Kappler, J. W., H o h a n n , M., and Dutton, R. W. (1971).J.E x p . Med. 134, 577. Kappler, J. W., van der Hoven, A., Dharmarajan, U., and Hoffmann, M. (1973).J . Immunol. 111, 1228. Karesen, R., and Godal, T. (1969a).Zmmunology 17, 847. Karesen, R., and Godal, T. (1969b). Immunology 17, 863. Karesen, R., and Godal, T. (1970).Immunology 19, 301. Karkhanis, Y. D., Carlo, J. C., Brostoff, S. W., and Eylar, E. H. (1975).J.Biol. Chem. 250, 1718. Karlin, A., Weill, C. L., McNamee, M. G., and Valderrama, R. (1976). Cold Spring Harbor Symp. Quant. Biol. 60,203. Katsura, Y., Kawaguchi, S., and Muramatsu, S. (1972).Zmmunalogy 23, 537. Katz, D. H., and Benacerraf, B. (1975).Transplant. Reu. 22, 175. Katz, D. H., Paul, W. E., Goidl, E. A., and Benacerraf, B. (1971).J.E x p . Med. 133, 169. Keesey, J., Lindstrom, J., Cokely, H., and Hemnann, C. (1977).N . Eng2.J. Med. 296,55. Kettman, J . R., Cambier, J. C., Uhr, J. W., Ligler, F., and Vitetta, E. L. (1979).Zmmunol. Rev. 43, 69. Kibler, R. F., Fox, R. H., and Shapiro, R. (1964). Nature (London) 204, 1273. Kibler, R. F., Shapiro, R., McKneally, S., Jenkins, J., Selden, P., and Chou, F. (1969). Science 164, 577. Kibler, R. F., Re, P. K., McKneally, S., and Shapiro, R. (1972).J.Biol. Chem. 247, 969. Kies, M. W. (1975). I n “The Nervous System” (D. B. Tower, ed.), Vol. 1, p. 637. Raven Press, New York. Kies, M. W., Driscoll, B. F., and Seil, F. J. (1973).Science 179, 689. Kim, S. U., Murray, M. R., Tourtellotte, W. W., and Parker, J. A. (197O).J.Neuropathol. E x p . Neurol. 29, 420. Kirchener, H., and Oppenheim, J. J. (1972).Cell. Znmunol. 3, 695. Kite, J. H., Jr., Rose, N. R., Kano, K., and Witebsky, E. (1965).Ann. N.Y. Acad. Sci. 124, 626. Kobayakawa, T., Louis, J., Izui, S., and Lambert, P.-H. (1979).J.Immunol. 122, 296. Kohler, H. (1975).Trunsplant Rec. 27, 24. Kolb, L. C. (1950).Medicine (Baltimore)29, 99. Kong, Y. M., David, C. S., Giraldo, A. A., Elrehewy, M., and Rose, N. R. (1979).J. Immunol. 123, 15. Koprowski, H. (1962).Am. J. Dis. Child. 103, 273. Koprowski, H., and Femandes, M. V. (1962).J. E x p . Med. 116,467. Koprowski, H., Paraf, A., Billingham, R. E., and Jervis, G . A. (1960).C.R. Hebd. Leonces Acad. Sci. Ser., D 250,2950. Koprowski, H., Paraf, A,, and Jervis, G. A. (1963).Colloq. Znt. C. N . R. S. 116, 339. Komblum, J. (1968).J.Zmmunol. 101, 702. Krishman, G., and Kaplan, M. H. (1956).Fed. Proc., Fed. A m . Soc. Exp. B i d . 25, 309. Kunkel, H. G., Muller-Eberhard, H. J., Fudenberg, H. H., and Tomasi, T. B. (1961).J. Clin. Inue.yt. 40, 117. Kuwert, E. K., and Bertrams, H. J. (1977). I n “Slow Virus Infections of the Central Nervous System” (V. Meulen and M. Katz, eds.), p. 1861. Springer-Verlag, Berlin and New York. Lanibert, P. W. (1967).Acta Neuropathol. 9, 99. Lanibert, P. W. (1969).Lab. Znoest. 20, 127. Lambert, P. W. (1978).A m . J. Pathol. 91, 176. I a n b e r t , P. W., and Carpenter, S. (1965).J.Neuropathol. Exp. Neurol. 24, 11. Lamoureux, G., Carnegie, P. R., McPherson, T. A., Makay, I. R., and Bernard, C. (1970). Ann. Inst. Pasteur, Paris 118, 562.
262
WILLIAM 0. WEIGLE
Lamoureux, G., Thibealt, G., Richer, G., and Bernard, C. (1972). Union Med. Can. 101, 674. Lamoureux, G., Giard, N., Jolicoeur, R., Toughlian, V., and Desosiers, M. (1976).Br. Med.1. 1, 183. Lando, Z., Teitelbaum, D., and Amon, R. (1979).J.Zmmunol. 123,2156. Landy, M., and Baker, P. J. (1966).J . Zmrnunol. 97, 163. Lane, D. P., and Silver, D. M. (1976).Eur. J. Zmmunol. 6,480. Larson, J. H. (1976). In “Infection and Immunity in the Rheumatic Diseases” (D. C. Dumonde, ed.), p. 133. Blackwell, London. Layton, J. E., Pike, B. L., Battye, F. L., and Nossal, G. J. V. (1979a).J.Zmmunol. 123,702. Layton, J. E., Teale, J. M., and Nossal, G. J. V. (1979b).]. Zrnmunol. 123, 709. Lebar, R., Boutry, J.-M., Vincent, C., Robineaux, R., and Voisen, G. A. (1976).J. Zmmunol. 116, 1439. Lee, C. Y. (1972). Annu. Reo. Pharmacol. 12,265. Lee, J. M., and Olitsky, P. K. (1955). Proc. SOC. E x p . Biol. Med. 89,263. Lees, R. K., and Sinclair, N. R. St. C. (1973). Zmmunology 24,735. Lefvert, A. K., Bergstrom, K., Matell, G., Osterman, P. O., and Pirskanen, R. (1978).J. Neurol., Neurosurg. Psychiatry 41,394. Lennon, V. A. (1978). In “Genetic Control of Autoimmune Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p. 295. Elsevier, Amsterdam. Lennon, V. A., and Byrd. W. J. (1973).Eur. J. Zmmunol. 3,243. Lennon, V. A., and Camegie, P. R. (1971). Lancet 1,630. Lennon, V. A., and Camegie, P. R. (1974). Eur. J. Zmmunol. 4,60. Lennon, V. A., Wilks, A. V., and Carnegie, P. R. (1970).J . Zmmunol. 105, 1223. Lennon, V. A., Lindstrom, J., and Seybold, M. E. (1975).J.E x p . Med. 141, 1365. Lennon, V. A,, Lindstrom, J. M., and Seybold, M. E. (1976).Ann. N.Y. Acad. Sci. 274, 283. Lennon, V. A., Seybold, M. E., Lindstrom, J., Cochrane, C., and Ulevitch, R. (1978).J. E x p . Med. 147,973. Lerner, E. M., McMaster, P. R. B., and Exum, E. D. (1964).J.E x p . Med. 119,327. Levine, S . , and Sowinski, R. (1973).J.Zmmunol. 110, 139. Levine, S., and Sowinski, R. (1974). Immunogenetics 1,352. Levine, S., and Sowinski, R. (1975).J.Zmmunol. 114, 597. Levine, S., and Wenk, E. J. (1965).Am. /. Pathol. 47,61. Levine, S., and Wenk, E. J. (1966). Proc. Soc. E x p . Biol. Med. 121, 301. Levine, S., and Wenk, E. J. (1967).Am. /. Pathol. 50,465. Levine, S . , Wenk, E. J., Devlin, H. B., Pieroni, R. E., and Levine, L. (1966).J. Zmmutiol. 97,363. Levine, S., Hoenig, E. M., and Kies, M. W. (1968). Science 16, 1155. Levine, S., Hoenig, E. M., and Kies, M. W. (1970). Clin. E x p . Zmmunol. 6, 503. Levine, S., Sowinski, R., and Kies, M. W. (1972). Proc. Soc. E x p . B i d . Med. 139, 506. Levinson, A. I., Lisak, R. P., Zweiman, B., and Wilkerson, L. D. (1977). Zmmunology 33, 621. Levy, B. M., Hampton, S., Dreizen, S., and Hampton, J. K., Jr. (1972).J.Comp. Pathol. 82, 99. Lindstrom, J. (1976).J.Supramol. Struct. 4, 389. Lindstrom, J. (1979).Ado. Zmmunol. 27, 1. Lindstrom, J., Lennon, V. A., Seybold, M. E., and Whittingham, S. (1976a).Ann. N.Y. Acad. Sci. 274,254. Lindstrom, J., Einarson, B., Lennon, V. A., and Seybold, M. E. (1976b).J.Exp. Med. 144, 726.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
263
Lindstrom, J., Campbell, M., and Nave, B. (1978a).Muscle Neme 1, 140. Lindstrom, J., Einarson, B., and Merlie, J. P. (1978h).Proc. Natl. Acad. Sci. U . S . A. 75, 769. Lipton, M.M., and F r e u d , J . (1953).J.lmmunol. 71, 380. Lisak, R. P., Heinze, R. G., Falk, G. A., and Kies, M. W. (1968).Neurology 18, 122. Lisak, R. P., Falk, G. A., Heinze, R. C . , Kies, M. W., and Alvord, E. C., Jr. (1970).J . lmtnutaol. 104, 1435. Lissitzky, S., Rolland, M., Raynaud, J., Savary, J., and Lassy, S. (1968).Eur.1. Biocheni. 4, 464. Litanjua, A. D. (1961).Nature (London) 191, 356. LoCerfo, P., Colacchio, T., Colacchio, D., and Feind, C. (1978). Cancer 42, 164. Lomhardi, M. H. (1962).Am. /. Vet. Res. 33,412. Louis, J., Chiller, J. bl., and Weigle, W. 0. (1973a)./. E x p . hled. 137, 461. Louis, I., Chiller, J. M., and Weigle, W. 0. (1973b).J.E x p . N e d . 138, 1481. Luzzato, A. L., Hengartner, H., and Schreier, M. H. (1977).Nature (Londo~i) 269,419. McCarty, D. J., Jr., Phelps P., and Pyenson, J. (1966).1.E x p . Med. 124, 99. McCullagh, P. J. (1970)./. E x p . Med. 132, 916. McDevitt, H. O., and Sela, M. (1965).J.E x p . Med. 122,517. hfcFarlin, D. E., Blank, S. E., Kibler, R. F., McKneally, S., and Shapiro, R. (1973). Science, 179, 478. McFarlin, D. E., Blank, S. E., and Kibler, R. F. (1974)./. ltnmunol. 113, 712. McFarlin, D. E., Hsu, S. C.-L., Slenianda, S. B., Chou, F. C.-H., and Kibler, R. F. (1975a)./. E x p . Med. 141, 72. McFarlin, D. E., Hsu, S. C.-L., Slenianda, S. B., Chou, F. C.-H., and Kibler, R. F. (1975h)./. lmmunol. 118, 1456. blccee, J. R., Farrar, J . J., Michalik, S. M., Mergenhagen, E., and Rosenstreich, D. L. (1979).J . E x p . Bled. 149, 793. McCraw, T. P., and Swanborg, R. H. (1978). Eor. J . Immutiol. 8, 955. Mackay, I. R., and Carnegie, P. R. (1977). I n “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects” (N. Talal, ed.), p. 597. Academic Press, New York. McLachlan, S., Smith, B. R., Peterson, V., Davies, T., and Hall, R. (1976).Nature (Lotidon ) 270, 447. McMaster, P. R. B., and Lerner, E. M. (1967)./.lmmuiiol. 99,208. McMaster, P. R. B., Lerner, E. M., and Exum, E. 13. (1961)./.E x p . Med. 113,611. A4cMaster, P. R. B., Lemer, E. M., and MueIler, P. (1965). Scietice 147, 157. McMaster, P. R. B., Lemer, E. M., Kyriakos, M., and Mueller, P. S. (1967)../.lmmutwl. 99, 201. hfacPherson, C. F. C., Armstrong, H., and Tan, 0. (1977).immunology 33, 161. h4cQuillan, M. T., and Trikojuz, V. M. (1966).171 “Glycoproteins” (A. Gottschalk, ed.), p. 516. Elsevier, Amsterdam. McVay-Boudreau, L., and Kemp, J . (1979).Cited in Cantor and Gershon (1979). Maggio, B., and Cumar, F. A. (1975). Nature (London ) 253, 364. Maino, V. C., Hayman, M. J., and Cmmpton, M. J. (1975).Biochem. /. 146, 247. Mariotti, S., DeCroat, L. J . , Scarborough, D., and Medlof, M . E. (1979)./.Clin. Endocr i n d . Metul,. 49, 679. h4arks, N., Benuck, M., and Hashini, G. (1974).Riochem. B w p h y s . Res. Cornmun. 56,68. Maron, R., and Cohen, I. R. (1979).Nature (Lotidon) 279, 715. Martenson, R. E., Deibler, G. E., Kies, M. W., Levine, S., and Alvord, E. C. (1972a).]. Inimunol. 109, 262. Martenson, R. E., Deibler, G. E., Kies, M. W., McKneally, S. S., Shapiro, R., and Kibler, R. F. (1972b). Biochirn. Biopliys. Acta 263, 193.
264
WILLIAM 0. WEIGLE
Martenson, R. E., Deibler, G. E., Kramer, A. J., and Levine, S. (1975a).J.Neurochem. 24, 173. Martenson, R. E., Levine, S., and Sowinski, R. (1975b).J.Zmmunol. 114, 592. Martenson, R. E., Nomura, K., Levine, S., and Sowinski, R. (1977).J.Immunol. 118, 1280. Masters, C., Dawkins, R., Zilko, P., Simpson, J., Leedman, R., and Lindstrom, J. (1977). Am. J. Med. 63,689. Mates, G. P., and Shulman, S. (1968a).Zmmunology 14, 89. Mates, G. P., and Shulman, S. (1968b). lmmunochemistnj 4,475. Mawdesley-Thomas, L. E., and Jolly, D. W. (1967).Vet. Rec. 80, 553. Mawdesley-Thomas, L. E., and Small, J. (1968).Anim. Pract. 9, 539. Mellors, R. C., Brzosko, W. J., and Sonkin, L. S. (1962).Am. J. Puthol. 41,425. Merlie, J., Heinemann, S., and Lindstrom, J. (1979).J . Biol. C h m . 254, 6320. MetcaIf, E. S., Schrater, A. F., and Klinman, N. R. (1979). Immunol. Rev. 43, 143. Metzger, H., Sharp, C. C., and Edelhoch, H. (1962). Biochemistry 1,205. Miescher, P., Gorstein, F., Benacerraf, B., and Cell, P. C. H. (1961).Proc. Soc. E x p . Biol. Med. 107, 12. Mischak, P., Dau, P. C., Gonzalez, R., and Spitler, L. (1979).In “Plasmapheresis and the Immunobiology of Myasthenia Gravis” (P. C. Dau, ed.), p. 72. Houghton, Boston, Massachusetts. Mizejewski, G., and Allen, R. P. (1974).Nuture (London ) 250, 50. Mizejewski, G. J., Baron, J., and Poissant, G. (1971).J.Zmmunol. 107, 1152. Moller, G., Gronowicz, E., Person, U., Coutinho, A., Moller, E., Hammarstrom, L., and Smith, C. I. E. (1976).J . E x p . Med. 143, 1429. Moore, M. J., Behan, P. O., Kies, M. W., and Matthews, J. M. (1974a). Res. Commun. Chem. Pathol. Pharmacol. 9, 119. Moore, M . J., Matthews, J. M., Matthews, T. K., Behan, P. O., and Kies, M. W. (197413). Res. Conimun. Chem. Puthol. Pharmacol. 9, 133. Moretta, L., Webb, S. R., Grossi, C. E., Lydyard, P. M., and Cooper, M. D. (1977a).J . E x p . Med. 146, 184. Moretta, L., Mengari, M. D., Webb, S. R., Pearl, E. R., Lydgard, P. M., Grossi, C. E., Lawton, A. R., and Cooper, M. D. (1977b).Eur. J. Zmmunol. 7, 696. Morgan, I. (1947).J. E x p . Med. 85, 131. Morrison, L. R. (1947). Arch. Neurol. Psychiutrg 58,391. Morrison, S. L., and Terres, G . (1966).J. Zmmunol. 96, 901. Morse, S. I. (1965).J.E x p . Med. 121,49. Moutsoupoulas, H. M., Boehm-Truitt, M., Kassan, S. I., and Chused, T. M. (1977).J. Zmrnunol. 119, 1639. Munoz, J. J., and Bergman, R. K. (1977).Immunol. Ser. 4, 148. Murgo, A. J., and Athanassiades, T. J. (1975).J.Zmmunol. 115,928. Musser, E., and Graham, W. R. (1968).Lab. Anim. Care 18,58. Nachtigal, D., Zan-Bar, I., and Feldman, M. (1975). Transplant. Reo. 26, 87. Naeim, F., Keesey, J., Henman, C., Lindstrom, J., Zeller, E., and Walford, R. (1978). Cited in Lindstrom (1979). Nakamura, R. M., and Weigle, W. 0. (1967a). J . Zmmunol. 98,653. Nakamura, R. M., and Weigle, W. O., (196%).J. Immunol. 99,357. Nakamura, R. M., and Weigle W. 0. ( 1 9 6 7 ~ )Znt. . Arch. Allergy Appl. lmniunol. 32,506. Nakamura, R. M., and Weigle, W. 0. (1968a).Proc. Soc. E x p . B i d . Med. 129, 412. Nakamura, R. M., and Weigle, W. 0. (1968b).J.Zmmunol. 101,876. Nakamura, R. M., and Weigle, W. 0. (1969).J.E x p . Med. 130, 163.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
265
Nakaniura, R. M., and Weigle, W. 0. (1970). C / i i i .E s p . Irnrtiunol, 7, 541. Kakaniura, R. M., Spiegelberg, H. L., Lee, S., and Weigle, W. 0. (1968).J. Irtiniroio/. 100, 376. Nakao, A,, and Einstein, E. R. (1965).Ann. N.Y. Acrid. Sci. 122, 171. Kakao, K., Nishitani, H., Suzuki, M., Ohta, M., and Hayashi, K. (1977).N . E n g l . J . Med. 297, 169. Nakashima, I., Yokochi, T., Kato, N., and Asai, J. (1977). ;Micro/7io/. I n i r t i i m o / . 27, 279. Nastuk, W., NCemi, W., Alexander, J., Chong, H., and Nastuk, M. (1979). Am. J . Phy.uio/. 236, 653. Nilsson, L. A., and Rose, N. R. (1972). Irnr?iuno/ogy 22, 13. Nishi, S., Watahe, H., and Hirai, H. (1973). Turnor Re.9. 8, 17. Nissley, P., Cittanova, N., and Edelhoch, H. (1969). Biochemistry 8, 443. Noble, B., Yoshida, T., Rose, N. R., and Bigazzi, P. E. (1976).J. Ztnniuito/. 117, 1447. Nossal, G. J. V., and Pike, B. L. (1978).J. E x p . Med. 148, 1161. Ohsiigi, Y., and Gershwin, M. E. (1980). Fed. PrrJc., Fed. Am. SOC.E.rp. Biol. 39, 1130. Okumura, K., Takemori, T., Tokuhisa, T., and Tada, T. (1977).J. E x p . Med. 146, 1234. Oldstone, M. B. A., and Dixon, F. J. (1968). Am. J . P a t h / . 52, 251. Olitsky, P. K., and Yager, R. H. (1949).J. Esp. Med. 90, 213. Oppenheim, J . J., Rogentine, G. W., and Terry, W. D. (1969). Ininiuno/ogy 16, 123. O’Regan, S., Fong, J . S. C., Kaplan, B. S., DeChadarevian, J.-P., Lapointe, N., and Druniniond, K. N. (1976).Clin. Ittimuno/. Zniitizirio),ritho/. 6, 341. Orgad, S., and Cohen, I. R. (1974). Science 183, 1083. Ortiz-Ortiz, L., and VFgle, W. 0. (1976).J . Ex),. Med. 144, 604. Ortiz-Ortiz, L., Nakamura, R. M., and Weigle, W. 0. (1976).J. Irnrtiunol. 117, 576. Ortiz-Ortiz, L., Parks, D. E., Rodriguez, M., and Weigle, W. 0. (1980).J. Iinmuiio/. 124, 121. Oudin, J., and Michel, M .(1969).J. E x p . ,$.led. 130, 595. Owen, R. D. (1945). Science 102, 400. Paget, S. A., McMaster, P. R. B., and van Boxel, J . A. (1976).J . I r t i r n i u i o / . 117, 2267. Palmer, F. B., and Dawson, R. M . C. (1969). Riochern. J. 111, 637. Panitch, H. S., and McFarlin, D. E. (1977).J. Zmniunol. 119, 1134. Parker, D. C. (1975). Nature (London ) 258, 361. Parker, D. C., Fothergill, J. J., and Wadsworth, D. C. (1979).J. Iirimunol. 123, 931. Parks, D. E., and Weigle, W. 0. (1979). Z m n i u d . Rec. 43, 217. Parks, D. E., and Weigle, W. 0. (1980a). In “T and B Lymphocytes: Recognition and Function” (F. L. Bach, B. Bonavida, E. S. Vitetta, and C. F. Fox, eds.), p. 229. Academic Press, New York. Parks, D. E., and Weigle, W. 0. (1980b).J. Imniunol. 124, 1230. Parks, D. E., Doyle, M. V., and Weigle, W. 0. (1977).J. Ininiuno/.119, 1923. Parks, D. E., Doyle, M. V., and Weigle, W. 0. (1978).J. E x p . hled. 148, 625. Parks, D. E., Shaller, D. A,, and Weigle, W. 0. (1979).J. E x p . M e d . 149, 1168. Patterson, P. Y. (1958). Anti. N.Y. Acad. Sci. 73, 811. Paterson, P. Y. (19594.I n “Cellular and Humoral Aspects ofthe Hypersensitive States” (H. S . Lawrence, ed.), p. 469. Harper, New York. Paterson, P. Y. (1959b). I n “Allergic Encephalomyelitis” (M. W. Kies and E. C. Alvord, Jr., eds.), p. 444. Thomas, Springfield, Illinois. Paterson, P. Y. (1960).J. E x p . Med. 111, 119. Paterson, P. Y. (1966). Adu. Zmrnunol. 5, 131. Paterson, P. Y. (1968).I n “International Convocation on Immunology” (N. R. Rose and F. Milgrom, eds.), p. 260. Karger, Basel.
266
WILLIAM 0. WEIGLE
Paterson, P. Y. (1976). In “Textbook of Immunopathology” (P. A. Miescher and H. J. Miiller-Eberhard. eds.), 2nd ed., p. 179. Grune & Stratton, New York. Paterson, P. Y. (1977). In “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects (N. Talal, ed.), p. 643. Academic Press, New York. Paterson, P. Y. (1978). In “Immunological Diseases” (M. Samter, D. W. Talmage, B. Rose, K. F. Austen, and J. H. Vaughan, eds.), p. 1400. Little, Brown, Boston, Massachusetts. Paterson, P. Y. (1979a). Reu. Infect. Dis. 1,468. Paterson, P. Y. (197913).In “The Menarini Series on Immunopathology” (P. Miescher, L. Bolis, S. Gorini, T. A. Lambo, G. J. V. Nossal, and G. Tonigiani, eds.), p. 165. Schwabe, Basel. Paterson, P. Y., and Day, E. D. (1979). H o s p . Pract. 14,49. Paterson, P. Y., and Didakow, N. C. (1961). Proc. Soc. Exp. Biol. Med. 108, 768. Paterson, P. Y., and Drobish, D. G. (1969). Science 165, 191. Paterson, P. Y., and Harwin, S. M. (1963).J . Exp. Med. 117, 755. Paterson, P. Y., and Weiss, H. S. (1965).Proc. Soc. Exp. B i d . Med. 119, 267. Paterson, P. Y., Richardson, W. P., and Drobish, D. G. (1975). Cell. Zmmunol. 16, 48. Patrick, J., and Lindstrom, J. (1973). Science 180, 871. Patrick, J., Lindstrom, J., Culp, B., and McMillian, J. (1973). Proc. Natl. Acad. Sci. U . S. A. 70, 3334. Paul, W. E., Thorbecke, G. J . , Siskind, G. W., and Benacerraf, B. (1969).Zmmunology 1 7 , s . Penhale, W. J., Farmer, A., McKenna, R. P., and Irvine, W. J . (1973).Clin. Exp. Zmmunol. 15,225. Penhale, W. J., Farmer, A., Urbanick, S. J., and Irvine, W. J. (1975a). Clin. Exp. Zmmunol. 19, 179. Penhale, W. J., Farmer, A., and Irvine, W. J. (1975b). Clin. Exp. Zmmunol. 21, 362. Penhale, W. J., Irvine, W. J., and Farmer, A. (1976). Clin. Exp. Zmmunol. 25, 6. Penn, A., Chang, H., Lovelace, R., Niemi, W., and Meranda, A. (1976).Ann. N.Y. Acad. Sci. 274, 354. Perudet-Badoux, A., and Frei, P. C. (1969). Clin. Exp. Zmmunol. 5, 117. Phanuphak, P., Moorhead, J. W., and CIaman, H. N. (1974). J . Zmmunol. 113, 1230. Phillips, S. M., Martin, W. J., Shaw, A. R., and Wegmann, T. G. (1971).Nature (London ) 234, 146. Pierce, J. G., Rawitch, A. R., Brown, D. M., and Stanley, P. G. (1965). Biochim. Biophys. Acta 111, 247. Pierce, S., and Klinman, N. R. (1977).J . Exp. Med. 146, 509. Pihko, H., Lindgren, J., and Ruoslahti, E. (1973). Zmmunochemistry 10,381. Pinching, A. J., Peters, D. K., and Davis, J. N. (1976). Lancet 2, 1373. Pitts, O., Varitek, V. A., and Day, E. D. (1975).J. Zmmunol. 115, 1114. Podleski, W. K. (1972). Clin. Exp. Zmmunol. 11, 543. Polak, L., and Turk, J. L. (1974). Nature (London) 249,654. Pescovitz, M. D., Paterson, P. Y., Kelly, J., and Lorand, L. (1978). Cell. Zmmunol. 39, 355. Prineas, J., Raine, C. S., and Wisniewski, H. (1969). Lab. Znuest. 21, 472. Pudifin, D. J., Duursma, J., and Brain, P. (1977). Clin. Exp. Zmmunol. 29, 256. Pulvertaft, R. J . V., Davies, J. R., Weiss, L., and Wilkinson, J. H. (1959).Pathol. Bacterwl. 77, 119. Pulvertaft, R. J. V., Doniach, D., Roitt, I. M., and Hudson, R. V. (1961).Br.J. E x p . Pathol. 42,496. Raftery, M. A., Vandlen, R. L., Reed, K. L., and Lee, T. (1976). Cold Spring Harbor Symp. Quant. Bid. 60, 193.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
267
Raine, C. S. (1976). Brain Res. 102, 355. Raine, C. S., and Bomstein, M. B. (1970).]. Neuropathol. E x p . Neurol. 29, 177. Raine, C. S., Wisniewski, H., Dowling, P. C., and Cook, S. D. (1971).Lab. Znuest. 25,28. Raine, C. S., Snyder, D. H., Valsamis, M. P., and Stone, S. H. (1974).Lab. Znuest. 3,369. Raine, C. S., Traugott, U., and Stone, S. H. (1978).Science 201,445. Rajewsky, K., and Brenig, C. (1974). Eur. J . Zmrnunol. 4, 120. Rauch, H. C., and Einstein, E. R. (1974). Reu. Neurosci. 1,263. Rauch, H. C., and Griffin, J. (1969). Znt. Arch. Allergy Appl. Zrnrnunol. 36, 387. Reed, C. E., Benner, M., Lockey, S. D., Enta, T., Makino, S., and Carr, R. H. (1972).J. Allergy Clin. Zmmunol. 49, 174. Reinherz, E. L., and Schlossman, S. F. (1980).Z t i “Regulatory T Cells” (B. Pemis and H . J. Vogel, eds.), Biorned. Sci. Symp., p. 85. Academic Press, New York. Reuber, M . D., and Glover, E. L. (1969).Endocririology 84, 443. Richert, J. R., Driscoll, B. F., Kies, M. W., and Alvord, E. C. (1979).]. Zrnrnunol. 122,494. Ringertz, B., Wasserman, J., Packalen, T., and Perlmann, P. (1971).Znt. Arch. Allergy Appl. Zmmunol. 40, 917. Rivers, T. M., and Schwentker, F. F. (1935).J.E s p . Med. 61, 689. Rivers, T. M., Spnint, D. H., and Berry, G. P. (1933).J.E x p . Med. 58, 39. Robhins, J. (1963).J . B i d . Chem. 238, 182. Roberts, I., Whittingham, S., and Mackay, I. R. (1973). Loncet 2, 936. Rocklin, R. E., Sheremata, W. A., Feldnian, R. G., Kies, M. W., and David, J. R. (1971).N . E n g ! . 1. M e d . 284, 803. Rodkey, L. S. (1974).J.E x p . Med. 139, 712. Roitt, I . M., and Doniach, D. (1958).Lamet 2, 1027. Roitt, I . &I., and Doniach, D. (1967).Clin. Exp. Zmmunol. 2, 727. Roitt, I . M., and Torrigiani, G. (1967). Endocrinology 81, 421. Roitt, I . M., Jones, H. E. H., and Doniach, D. (1961).In “Mechanism o f c e l l and Tissue Damage Produced by Immune Reactions” (P. Grabar and P. Miescher, eds.), p. 174. Schwabe, Basel. Rolland, M., and Lissitzky, S. (1976).Biochim. Biophy.~. Act“ 427, 696. R6llenghoff, M. S., Starzinski-Powitz, A., Pfizenmaier, K., and Wagner, H. (1977)J. E x p . Me(/. 145, 455. Romagnani, S., Ricci, M., Passaleva, A., and Bilotti, G. (1970).Znarnunology 19, 599. Roinball, C. G., and Weigle, W. 0. (1979).J.Supmmol. Struct., S u p p l . 3, 272. Rose, N. R. (1975). Cell. Zmmunol. 18, 360. Rose, N . R., and Kite, J. H. (1969). In “International Convocation on Immunology” ( N . R. Rose and F. Milgroni, eds.), p. 247. Karger, Basel. Rose, N. R., and Witebsky, E. (1956).J. Zmmuno!. 76, 417. Rose, N . R., and Witebsky, E. (1959).J . Zmmunol. 83, 34. Rose, N . R., Kite, J . H., and Doebbler, T. K. (1962a).I n “Mechanisms of Cell and Tissue Damage Produced by Immune Reactions” (P. Grabar and P. Miescher, eds.), p. 161. Schwabe, Basel. Rose, N . R., Shulman, L., and Witebsky, E. (1962h).J. Zmmunol. 88, 229. Rose, hi. R., Kite, J. H., Jr., Doehhler, T. K., and Witebsky, E. (1964). I n “Injury, Inflammation and Immunity” (L. Thomas, J. W. Uhr, and L. Grant, eds.), p. 1.35.Williams & Wilkins, Baltimore, Maryland. Rose, N. R., Kite, J . H., Jr., Doebbler, T. K., Spier, R., Skelton, F. R., and Witebsky, E. (1965).Ann. N . Y . Acud. Sci. 124,201. Rose, N. R., Skelton, F. R., Kite, J. H., Jr., and Witehsky, E. (1966).Clin. E x p . Znmtrnol. 1, 171. Rose, N . R., Twarog, F. J., and Crowle, A. J. (1971).]. Zmrnunol. 106, 698.
268
WILLIAM 0. WEIGLE
Rose, N. R., Molotchnikoff, M.-F., and Twarog, F. J. (1973). Zmmunology 24, 859. Rose, N. R., Bacon, L. D., and Sundick, R. S. (1976). Transplant. Rev. 31,264. Rose, N. R., Bacon, L. D., Sundick, R. S., and Briles, W. E . (1978).In “Comparative and Developmental Aspects of Immunity and Disease” (M. E. Gershwin and E. L. Cooper, eds.), p. 143. Plenum, New York. Rosen, A., Gergely, P., Jandel, M., Klein, G., and Britton, S. (1977). Nature (London) 267, 52. Rosenberg, L. T., and Tachibana, D. K. (1962).]. Zmmunol. 89, 861. Rous, P. (1925).J . E x p . Med. 41, 399. Rowley, D. A., Fitch, F. W., Axelrad, M. A., and Pierce, C. W. (1969). Zmmunology 16, 549. Rudbach, J. A. (1971).J.Zmmunol. 106, 993. Ruoslahti, E., and Seppala, M. (1971a). Znt. 1. Cancer 7,218. Ruoslahti, E., and Seppila, M. (1971b). Znt. J. Cancer 8, 374. Ruoslahti, E., and Seppala, M. (1972). Nature (London) 235, 161. Ruoslahti, E., and Wigzell, H. (1975).Nature (London) 255, 716. Ruoslahti, E., Pihko, H., and Seppala, M. (1974). Transplant. Reo. 20,38. Ruoslahti, E., Pihko, H., Becker, M., and Makela, 0. (1975).Eur. J. Zmmunol. 5, 7. Saida, T., Saida, K., Dorfman, S. H., Silberberg, D. H., Sumner, A. J., Manning, M. C., Lisak, R. P., and Brown, M. J. (1979).Science 204, 1103. Salabe, G. B., Davoli, C., and Fontana, S. (1973).Znt. Arch. Allergy Appl. Zmmunol. 44, 181. Salabe, G. B., Salabe, H., Accini, L., and Dominici, R. (1978).Clin. E x p . Zmmunol. 32, 159. Salabe, H., Dominici, R., and Salabe, G. 8. (1976). Clin. E x p . Zmmunol. 25, 234. Salvatore, G., Vecchio, G., Salvatore, M., Cahnman, H. J., and Robbins, J. (1965).J. B i d . Chem. 240,2935. Salvin, S. B., and Liauw, H. L. (1968).]. Zmmunol. 101, 33. Sanders, D. B., Schleefer, L., Eldefrawi, M., Norcross, N., and Cobb, E. (1976). Ann. N.Y. Acad. Sci. 274, 319. Sbarra, A. J., Bardawil, W. A., Shirley, W., and Gilfillan, R. F. (1960).J.E x p . Med. 111, 667. Scavulli, J., and Dutton, R. W. (1975).J.E x p . Med. 141, 524. Schauenstein, K., and Wick, G. (1974). Clin.E x p . Zmmunol. 17, 637. Schrader, J. W. (1974).]. E x p . Med. 139, 1303. Schrater, A. F., Goidl, E. A., Thorbecke, G. J., and Siskind, G. W. (1979).J.E x p . Med. 150, 138. Schumacher, G. A. (1974). Neurology 24, 1010. Schwartz, A., Askenase, P. W., and Gershon, R. K. (1978).J. Zmmunol. 121, 1573. Schwartz, R. H., Jackson, L., and Paul, W. E. (1975).J.Zmmunol. 115, 1330. Scott, D. W., Venkataraman, M., and Jandinski, J. J. (1979). Zmmunol. Reo. 43,241. Seil, F. J. Rauch, H. C., Einstein, E. R., and Hamilton, A. B. (1973).J.Zmmunol. 111,96. Seil, F. J., Smith, M. E., Leiman, A. L., and Kelley, J. M. (1975). Science 187, 951. Sell, S., and Cell, P. G. H. (1965).J . Exp. Med. 122, 423. Sercarz, E. E., Yowell, R. L., Turkin, D., Miller, A., Araneo, B. A., and Adorini, L. (1978). Zmmunol. Reo. 39, 108. Shapiro, R., McKneally, S., Chou, F. C.-H., and Kibler, R. F. (1971).J.Biol. Chem. 246, 4630. Sharp, G . C., Wortis, H. H., and Dunmore, B. (1967). Zmmunology 12,39. Sharp, G. C., Woltis, H. H., LaRoque, R. L., Mullin, H., and Kyriakos, M. (1974a).J. Zmmunol. 112, 468.
AUTOIMMUNITY:
THYROIDITIS AND ENCEPHALOMYELITIS
269
Sharp, G. C., Mullen, H., and Kyriakos, M. (1974b).J. Immunol. 112, 478. Shaw, C.-M., Wilson, M. D., Fahberg, J., Kies, N . W., and Alvord, E. C. (196O).J. E x p . Med. 111, 171. Shaw, C.-M., Alvord, E. C., Jr., Fahlberg, W. J., and Kies, M. W. (1!362).J.Immunol. 89, 54. Shaw, C.-M., Alvord, E. C., Jr., Kaku, J.. and Kies, M. W. (1965a).Ann. N.Y. Acad. Sci. 122, 318. Shaw, C.-M., Alvord, E. C., Jr., Kaku, J., and Kies, M. W. (1965b).Ann. N.1’. Acad. Sci. 122,333. Sheffield, W. D., Gonatas, N. K., and Kim, S. U. (1977).J. Zmrnunol. 119, 1762. Shek, P. N., Waltenbaugh, C., and Coons, A. H. (1978).]. E x p . Med. 147, 1228. Sheremata, W., Cosgrove, J. B. R., and Eylar, E. H. (1976).J. Neurol. Sci. 27, 413. Shioiri-Nakano, K., Nagai, Y., Otani, L., and Tsumita, T. (1971).Jpn.]. E x p . Med. 41, 159. Shulman, S. (1971). Ado. Zmrnunol. 14, 85. Shulman, S., Rose, N. R., and Witebsky, E. (1955).J. Immunol. 75, 291. Shulman, S., Mates, G., and Bronson, P. (1967). Biochim. Biophys. Actm 147, 208. Sidman, C. L., and Unanue, E. R. (1975). Nature (London) 257, 149. Sidman, C . L., and Unanue, E. R. (1978). Proc. Natl. Acad. Sci. U. S. A . 75, 2401. Sieckmann, D. G . , Asofsky, R., Mosier, D. E., Zitron, I . M., and Paul, W. E. (1978a).J. Ex)).Med. 147, 814. Sieckmann, D. G., Scher, I., Asofsky, R., Mosier, D. E., and Paul, W. E. (1978b).J. E x p . Med. 148, 1628. Silver, D. M., and Lane, D. P. (1975).J. Exp. Med. 142, 1455. Silver, D. M., and Lane, D. P. (1977). Zrnmunogenetic.s 4, 295. Silver, D. M., and Lane, D. P. (1979). Zmmutiogetietic,~8, 65. Silverman, D. A., and Rose, N. R. (1971). Proc. Soc. E.rp. B i o l . Med. 138, 579. SiIvemian, D. A,, and Rose, N. R. (1974a). Science 184, 162. Silverman, D. A., and Rose, N. R. (1974b).J. Natl. Cancer Znst. 53, 1721. Silverman, D. A., and Rose, N. R. (1974~). Lancet 1, 1257. Simon, J., and Simon, 0. (1975). Exp. Nezrrol. 47, 523. Simpson, J. (1960).Scott. Med. J. 5, 419. Sinclair, N. R. St. C. (1969).J. Exp. Med. 129, 1183. Sinclair, N. R. St. C., and Chan, P. L. (1971). Z n “Morphological and Functional Aspects of Immunity” (K. Lindahl-Kiessling, G. Alm, and M. G. Hanna, eds.), p. 609. Plenum, New York. Sinclair, N. R. St. C., Lees, R. K., Ahrahams, S., Chan, P. L., Fagan, G., and Stiller, C . R. (1974).J . Zmniunol. 113, 1493. Singer, D. E., Moore, M. J., and Williams, R. M. (1979). Transplmt. Proc. 11, 1598. Sjoberg, 0. (1972).J . E x p . Med. 135, 850. Skamene, E., and Ivanyi, J. (1969). Nuture (London) 221, 681. Slaughter, L., Carson, D. A. Jensen, F. C . , Holbrook, T. L., and Vaughan, J. H. (1978).J. Exp. Med. 148, 1429. Smith, B. R. (1976). Ininiunol. Commun. 5, 345. Smith, D. J., and Shulman, S. (1978). Clin. Exp. lninircnol. 31, 341. Smyth, D., and Utsunii, S. (1967). Nuture (London) ZIG, 332. Snyder, D. H., Valsaniis, M. P., Stone, S. H., and Raine, C. S. (1975).J. Neuropathol. Exp. Nrurol. 34, 209. Sobel, A., Weber, M., and Changeux, J. P. (1977). Eur. J. Biockem. 80, 215. Soberg, M., and Halberg, P. (1968).Acln Med. S c u d . 18:3, 101. Spiegelberg, H. L., and Weigle, W. 0. (1966).J . Imnlunol. 98, 1034. Spiro, M. J. (1973).J. B i d . Chem. 248, 4446.
270
WILLIAM 0. WEIGLE
Spiro, R. G., and Spiro, M. J. (1965).J.Biol. Chem. 240, 997. Spitler, L. E., von Muller, C. M., Fudenberg, H. H., and Eylar, E. H. (1972)J E x p . Med. 136, 156. Stanton, T. H., and Boyse, E. A. (1976). lmmunogenetics 3, 525. Stanton, T. H., Calkins, C. E., Jandinski, J., Schender, D. J., Stutman, O., Cantor, H., and Boyse, E. A. (1978).J.Erp. Med. 148,963. Steinman, L., Cohen, I. R., Teitelbaum, D., and Amon, R. (1977).Nature (London) 265, 1973. Stone, S. H. (1962). Znt. Arch. Allergy Appl. lmmunol. 20, 193. Stone, S. H., and Lemer, E. M. (1965).Ann. N.Y. Acad. Sci. 122,227. Stoner, G . L., Brosnan, C. F., Wisniewski, H. M., and Bloom, B. R. (1977).J.lmmunol. 118,2094. Strauss, B. S., and Stetson, C. A. (196O).J.Exp. Med. 112,653. Strayer, D. S., Lee, W., Rowley, D. A., and Kohler, H. (1975).J.lmmunol. 114,728. Streilein, J. W. (1979). lmmunol. Reu. 46, 123. Stylos, W. A., and Rose, N. R. (1969). Clin. Exp. lmmunol. 5, 285. Stylos, W. A., and Rose, N. R. (1977). Clin. E x p . lmmunol. 27,245. Sundick, R. S., and Wick, G. (1974). Clin. E x p . Zmmunol. 18, 127. Sundick, R. S., and Wick, G. (1976).J.lmmunol. 116, 1319. Sundick, R. S., Bagchi, N., Livezey, M. D., Brown, T. R., and Mack, R. E. (1979). Endocrinology 105,493. Svet-Moldavskaya, I. A., and Svet-Moldavsky, G. J. (1958).Nature (London) 181, 1536. Svet-Moldavsky, G . J., and Svet-Moldavskaya, I. A., and Raffkina, L. (1959). Nature (London) 184, 1552. Swanborg, R. H. (1970).J.Zmmunol. 105,865. Swanborg, R. H. (1972).J.Zmmunol. 109, 540. Swanborg, R. H. (1975).J.lmmunol. 114, 191. Swanborg, R. H., Swierkosz, J. E., and Saieg, R. G. (19741.1. Immunol. 112,594. Swierkosz, J. E., and Swanborg, R. H. (1975).J.lmmunol. 115,631. Swierkosz, J. E., and Swanborg, R. H. (1977a).J.lmmunol. 119, 1501. Swierkosz, J. E., and Swanborg, R. H. (1977b). lmmunol. Commun. 6,499. Szewezuk, M. R., and Siskind, G . W. (1977).J.E x p . Med. 145, 1590. Tada, T., and Okumura, K. (1971).J.Zmmunol. 107, 1137. Tada, T., Taniguchi, M., and Takemori, T. (1975). Transplant. Reu. 26, 106. Tan, E. M., Shur, P. H., and Kunkel, H. G. (1966).J.Clin.lnuest. 44, 1104. Taniguchi, M., and Miller, J. F. A. P. (1977).J.E x p . Med. 146, 1450. Tarrab-Hazdai, R., Aharonov, A,, Abramsky, O., Yaar, I., and Fuchs, S. (1975a).J. E x p . Med. 142, 785. Tarrah-Hazdai, R., Aharonov, A., Silman, I., Fuchs, S., and Abramsky, 0. (1975b).Nature (London) 256, 128. Taub, R. N., Rosett, W., Adler, A,, and Morse, S. I. (1972).J.E x p . Med. 136, 158. Teale, J. M., Layton, J. E., and Nossal, G. J. V. (1979).J.E x p . Med. 150, 205. Teitelbaum, D., Meshorer, A., Hershfield, T., Amon, R., and Sela, M. (1971). Eur. 1. lmmuiiol. 1, 242. Teitelbaum, D., Webb, C., Amon, R., and Sela, M. (1972). Nature (London) 246, 564. Teitelbaum, D., Webb, C., Meshorer, A., Amon, R., and Sela, M. (1973). Eur. J. lmmunot. 3,273. Teitelbaum, D., Webb, C., Bree, M., Meshorer, A., Amon, R., and Sela, M. (1974). Clin. lmmunol. Immunopathol. 3, 256. Teitelbaum, D., Lando, L., and Amon, R. (1978). In “Genetic Control of Autoimmune
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
271
Disease” (N. R. Rose, P. Bigazzi, and N. L. Warner, eds.), p. 365. Elsevier, Amsterclam. Terplan, K. L., Witebsky, E., Rose, N. R., Paine, J. R., and Egan, R. W. (1960).Am. J . Pathol. 36, 213. Theis, G. A., and Siskind, G. W. (1968).J . Immunol. 100, 138. Theofilopoulos, A. N., and Dixon, F. J. (1979). Ado. Immunol. 28,89. Thomas, L., Paterson, P. Y., and Smith, B. (1950).J . E x p . Med. 92, 133. Tomazic, V., and Rose, N. R. (1975). Cliti. Immunol. Immunopathol. 4, 511. Tomazic, V., and Rose, N. R. (1976). Zmmutiolog!! 30, 63. Tomazic, V., and Rose, N. R. (1977). E u r . J. Ztnmunol. 7, 40. Tomazic, V., Rose, N. R., and Shreffler, D. C. (1974).J. Zmmunol. 112, 1457. Tourtellotte, W. W. (1972). I n “Multiple Sclerosis Ultrastructure: Immunology, Virology, and Ultrastructure” (F. Wolfgram, G. Ellison, J. Stevens, and J. M. Andrews, eds.), p. 285. Academic Press, New York. Tourtellotte, W. W., and Parker, J. A. (1966). Science 154, 1044. Tourtellotte, W. W., and Parker, J. A. (1967). Nature (London)214, 683. Toyka, K. V., Drachman, D. B., Pestrone, D. B., and Kao, I. (1975). Science 190, 397. Triplett, E. L. (1962).J. Imrnunol. 89, 505. Tucker, W. E. (1962). Am. J . Clitt. Pathol. 38, 70. Twarog, F. J., and Rose, N. R. (1970).J. Immunol. 104, 1475. Uhr, J. W., and Moller, G . (1968).Adz;. Zmmunol. 8, 81. Ujejski, L., and Glegg, R. E. (1955).Cnn. J . Biochem. Physiol. 33, 199. Urbanick, S. J., Penhale, W. J., and Irvine, W. J. (1973). Clin. E x p . Zmmunol. 15, 345. Vadas, M . M., Miller, J . F. A. P., McKenzie, I. F. C., Chism, S. F., Shen, F. -W., Boyse, E. A., Gamble, J. R., and Whitelaw, A. M. (1976). J. E x p . Med. 144, 10. Vandenbark, A. A., Gibbins, B. L., and Hinrichs, D. J. (1974). Cell. Zmmunol. 11, 1. Van Herle, A. J., Vassart, G., and Dumont, J . E. (1979). N. Etigl. J. Med. 301, 307. Vassart, G., Refetoff, S., Brocas, H., Densart, C., and Dumont, J. E. (1975). Proc. N n t l . Acad. Sci. U . S. A. 72, 3839. Vecchio, G., Carlomagno, M. S., and Consiglio, E . (1971).J . Biol. Chem. 246, 6476. Vitetta, E. S., and Uhr, J. W. (1975). Science 189, 964. Vitetta, E. S., Melcher, U., McWilliams, M., Phillips-Quagliata, J., Lamm, M., and Uhr, J. W. (1975).J. E x p . Med. 141,206. Vitetta, E. S., Cambier, J. C., Ligler, F. S., Kettman, J. R., and Uhr, J. W. (1977).J . E x p . Med. 146, 1804. Vladutiu, A. 0..and Rose, N. R. (1971a).J. Immunol. 106, 1139. Vladutiu, A. O., and Rose, N. R. (1971h). Science 174, 1137. Vladutiu, A. O., and Rose, N. R. (1975). Cell. Immunol. 17, 106. von Muller, C. S., Spitler, L. E., and LeCocq, J. (1978). E u r . J. Immunol. 8, 771. Waksman, B. H., and Morrison, L. R. (1951).J.Inirnunol. 66,421. Walker, J. G., and Siskind, G. W. (1968).Immunology 14, 21. Warner, N. L. (1977). In “Autoimmunity: Genetic, Immunologic, Virologic and Clinical Aspects” (N. Talal, ed.), p. 33. Academic Press, New York. Wason, W. M., and Fitch, F. W. (1973).J. Irnmunol. 110, 1427. Watson, J., Aarden, L. A., Shaw, J., and Paetkau, V. (1979).J. Immunol. 122, 1633. Webb, C., Teitelbaum, D., Amon, R., and Sala, M. (1973). Eur. J . Irnmunol. 3, 279. Webb, C., Teitelbaum, D., Herz, A., Amon, R., and Sela, M. (1976). Inimunochernistrg 13, 333. Weigle, W. 0. (1961).J. E x p . Med. 114, 111. Weigle, W. 0. (1962).J. E x p . Med. 116, 913.
272
WILLIAM 0. WEIGLE
Weigle, W. 0. (1964). lmniunology 7, 239. Weigle, W. 0. (1965a).J. E x p . Med. 121,289. Weigle, W. 0. (1965b).J . E x p . Med. 122, 1049. Weigle, W. 0. (1967). “Natural and Acquired Immunologic Unresponsiveness.” World Publ. Co., Cleveland, Ohio. Weigle, W. 0. (1971).Clin. E x p . Immunol. 9, 435. Weigle, W. 0. (1973).Ado. Immunol. 16, 61. Weigle, W. O., and Berman, M. A. (1979).In “Cells of Immunoglobulin Synthesis” (B. Pernis and J. Vogel, eds.), p. 223. Academic Press, New York. Weigle, W. O., and High, G. J. (1967).J. lmmunol. 98, 1105. Weigle, W. O., and Nakamura, R. M. (1967).J . Immunol. 99,223. Weigle, W. O., and Nakamura, R. M. (1969). Cfin. E x p . fmmunol. 4,645. Weigle, W. O., and Parks, D. E. (1978). Fed. Proc., Fed. Am. Soc. E x p . Biol. 37, 1253. Weigle, W. O., and Romball, C. G. (1975). Cliti. E x p . Imrnunol. 21, 351. Weigle, W. O., and Romball, C. G. (1980).In “Symposium on Autoimmune Aspects of Endocrine Disorders” (in press). Weigle, W. O., High, G. J., and Nakamura, R. M. (1969).J. E x p . Med. 130,243. Weigle, W. O., Chiller, J. M., and Louis, J. A. (1974a).Prog. Imrnunol., Proc. l n t . Congr., 2nd, 1974 Vol. 3, p. 187. Weigle, W. O., Louis, J. A., Habicht, G. S., and Chiller, J. M. (1974b).Ado. Biosci. 12,93. Weigle, W. O., Sieckmann, D. G., Doyle, M. V., and Chiller, J. M. (1975). T r a n s p l a n t . Reu. 26, 186. Weiner, H. L., Moorhead, J. W., and Claman, H. (1976).J . Immunol. 116, 1656. Weinstock, M., Shoham-Moshonor, S., Teitelbaum, D., and Armon, R. (1977).Bruin Res. 125, 192. Welch, A.M., and Swanborg, R. H. (1976).Eur. J . Immunul. 6, 910. Welch, A. M., Swierkosz, J. E., and Swanborg, R. H. (1978).]. Intmunol. 121, 1701. Wenk, E. J., Levine, S., and Warren, B. (1967).Nature (London) 214,803. Westall, F. C. (1972).I n “Multiple Sclerosis: Progress in Research” (E. J. Field, T. M. Bell, and P. R. Camegie, eds.), p. 72. North-Holland Publ., Amsterdam. Westall, F. C., and Lennon, V. A. (1977).Nature (London) 269, 425. Westall, F. C., and Thompson, M. (1975). lmmunol. Cotnniun. 4, 187. Westall, F. C., and Thompson, M. (1978).Imniunochemistry 15, 189. Westall, F. C., Robinson, A. B., Caccam, J., Jackson, J . J., and Eylar, E. H. (1971).Nature (London ) 229, 22. Westall, F. C., Thompson, M., and Lennon, V. A. (1976).Fed. Proc., Fed. Ant. Soc. E x p . Biol. 35, 436. Whitacre, C. C., and Paterson, P. Y. (1977).J. E x p . Med. 145, 1405. Whitaker, J. N., Chou, C. -H. J., Chou, F. C. -H., and Kibler, R. F. (1977).J. E x p . Merl. 146, 317. Whitmore, D. B., and Irvine, W. J. (1977). Clin. E x p . Imtnunol. 29,4704. Wick, G., and Burger, H. (1971).Z. Immunitaetsforch. E x p . K l i n . Immunol. 142, 54. Wick, G., and Steiner, R. (1972).J. Iinmunol. 109, 1031. Wick, G., Kite, J. H., Jr., and Witebsky, E. (1970a).J. Immunol. 104,45. Wick, G., Kite, J. H., Jr., and Witehsky, E. (1970b).J. Immunol. 104,54. Wick, G., Kite, J . H., and Cole, R. K. (1971).Znt. Arch. AlZergy Appl. Immunol. 40,603. Wick, G., Sundick, R. S., and Albini, B. (1974).Clin.Iinmunol. lmmunopathol. 3, 272. Wick, G., Richter, E., Zambelis, N., Ludwig, H., and Schemthaner, G. (1977).Lancet 2, 1180. Wick, G., Scharz, S., and M d l e r , P. U. (1978).2. Itrimunitnets For.sch. 154, 162.
AUTOIMMUNITY: THYROIDITIS AND ENCEPHALOMYELITIS
273
Wick, G., Gundolf, R., and Hala, K. (1979).J.Ztwnirrtogetiet. 6, 177. Willenlmrg, D. 0. (1979).J . 11nrnut1o/.123, 1145. Willenlm-g, D. O., St;iten, E. A,, and Witting, G. F. (1978). E x p . Neirrof. 612, 527. Williams, R. M., and Moore, M.J. (1973).J.E.sp. Xfed. 1:38, 775. Winchester, G . (1979j.j. S u p r m w f . S t n t c t . , S r d ] i p / . 3, 274. Wisniewski, H . hl., and Bloom, B. R. (1975j.J.Neirrol. Sci. 24, 257. Wisniewski, H. M., and Keith, A. B. (1977).Attit. Neicrol. 1, 144. Witellsky, C., and Rose, N. R. (1956).J . Z t m i u t i o / . 76, 408. Witellsky, E., and Rose, N. R. (1959).J.Z i i i i n i t ~ w l .83, 41. Witellsky, E., Rose, N. R., and Shulman, S. (1955).J.Znitnr~tiol.75, 269. Witehsky, E., Rose, N. R., and Shulman, S. (1956).Cancer Res. 16, 831. Witel>sky, E., Kite, J. H., Jr., Wick, G., and Cole, R. K. (1969).J.Zrtxtizuttol. 103, 708. Wolf, A., Kabat, E. A,, and Bezer, A. E. (1947j.J.Nertropothol. E x p . Neurol. 6, 333. Wollin;cn, S. H., kind Warren, L. (1961).Riochi?)t. H i o p h ! / . ~Acto . 47, 251. Y;imashita, A., Takoini, T., Hamaoka, T., and Kitagawa, M. (1976).Cell. Z m n t t r d . 25, 15. Yoneznwn, T., Ishihara, Y., and Matsuyama, H. (1968).J.Neuropnthol. E s p . Neurof. 27, 45.3. Yiung, L. L., Dieiier, E., McPherson, T. A , , Barton, M. A., and Hyde, H. A . (1973).J . Z t n n i t r t i o / . 110, 1383. Ziivalita, A., and Stastny, P. (1967).Clin. E x p . Itnnitcrtol. 2, 543. Zinkeningel, R. M., and Doherty, P. C. (1979).Ada. Znttnrctiol. 27, 51. Zolla, S . , and Naor, D. G. (1974).J.E.sp. .lied. 140, 1421.
This Page Intentionally Left Blank
ADVANCES IN IMMUNOLOGY, VOL. 30
The Virology a n d lmmunobiology o f Lymphocytic C horiomen ing itis Virus Infection M. J. BUCHMEIER,
R. M. WELSH, F. J. DUTKO, AND M. B. A. OLDSTONE
Deportment of Immunopothology, Scripps Clinic
11.
111.
IV. V.
ond Reseoreh Foundation, l o Jollo, Californio
B. LCMV: Clinical Features of Human Illness and Safety Precautions.. . . . C. LCMV: Historical Features and Experimental Model Systems Used for Immunobiologic and Immunopathologic Research ..................... Virus and Host Cell Interactions . . . . . . . . . A. Virion Structure and Organization.. . . . B. Virus Replicative Cycle . . . . . . . . . . . . . . C. Nonassociation of Viral Polypeptides and H-2 Antigens ................ D. Virus Mutants and Variants ..................... .... E. Relationship between LCMV and Other Arenaviruses ................. F. Interactions of LCMV with Differentiated Cells ................ LCMV-Induced Acute Immune Response Disease . ................ A. Pathology of the Acute Infection ..................................... B. Generalized Host Responses and the Production of Interferon . . . . . . . . . C. Natural Killer Cell Activation and Function in Acute Infection.. . . . . . . . D. Macrophage Activation and Function in Acute Infection E. Cytotoxic T Cel! Activation and Function in Acute Disease F. B Cell Activation and Function in Acute Disease, ..................... LCMV-Induced Chronic Immune Response Disease ...................... A. B Cell Activation and Function in Persistent Disease . . . . . . . . . . . B. T Cell Activation and Function in Persi .......... Epilogue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
277 279
291 292 298 299 302 302 303 308
311 313 313 321 325 326
I . Prologue
Lymphocytic choriomeningitis virus (LCMV) and the infection it causes have been and are important subjects of biomedical investigation. First, this virus sporadically causes human illness. In the United States the related disease has been observed in those handling pets, usually hamsters, or cultured cells persistently infected with the virus. This disease has been especially troublesome among researchers, especially nonvirologists in cancer and immunobiologic investigations, as reviewed recently (Oldstone and Peters, 1978). Second, the study of LCMV and the disease it causes in its natural murine host has provided the initial findings to open new fields in viral immunobiology, viral immunopathology, and cell-cell recognition. It is these observa275 Copyright 0 19XO by Academic Press. Inc. A l l rights of reproduction in any form resewed. ISBN 0.12-022430-5
276
M. J. BUCHMEIER ET AL.
tions, spawned from the LCMV infection model, that form the basis of this review. Immunologically mediated injury is familiar to most readers of this series. Less familiar is an appreciation of how viruses assemble and code antigens expressed at the cell surface and how nonimmune factors regulate viral antigenic expression, events that also form a major part of this review.
A. INTRODUCTION Lymphocytic chorionieningitus virus is a member of the arenavirus group, which obtains its name from arenosus-Latin €or sandy-on the basis of the characteristic fine granularities seen inside the virion in ultrathin electron microscopic sections. Arenaviruses contain segmented noninfectious ribonucleic acid, are covered by an outer lipoprotein membrane, and bud from the plasma membranes of cells. Several of these viruses contain group-specific antigen detectable by immunofluorescence and complement fixation tests (Rowe et al., 1970a). This antigen is now known to be related to common determinants of the nucleoprotein (Buchmeier and Oldstone, 1978a).The arenaviruses are susceptible to heat and desiccation. In the electron microscope (Fig. l ) , these pleomorphic virus particles are round, oval, or irregularly shaped and range from 50 to 300 nm with a mean diameter of 110 to 130 nm. The particles consist of a dense, well defined unit membrane envelope having closely spaced projections. Virions of LCMV
FIG. 1. Electron micrographs of lymphocytic choriomeningitis virus (LCMV). (A) Negative stain preparation of purified virus; (B) thin section of a virion released from infected L-929 cells; (C) thin section from the cerebellum of a rat acutely infected with LCMV. Note the presence of surface projections in (A) and encapsidated ribosomes in (B) and (C). (Electron micrographs by courtesy of Dr. Peter Lampert.)
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
277
and several other arenaviruses contain two glycopeptides and one nonglycopeptide, the nucleoprotein. The interior of the virion contains variabIe numbers of electron dense granules 20-30 nm in diameter, which are host cell ribosomes. Both acute and persistent infections occur naturally and can be induced experimentally with viruses in the arenavirus group. The relationship of LCMV to other members of the arenavirus group and the biochemical and nucleic acid definitions of this virus are provided in Section I1 of this review.
€3. LCMV: CLINICAL FEATURES OF HUMANILLNESS AND SAFETYPRECAUTIONS
The virus was named by Armstrong and Lillie (1934)on the basis of the pathologic picture produced in monkeys and mice after inoculation of material taken from a patient who died of St. Louis encephalitis. In the same year Traub (1935) described a similar virus that produced an equivalent clinical picture when inoculated into adult outbred Swiss mice. He also noted that the virus could be carried by apparently healthy mice throughout their lives. These mice usually shed virus into the urine (Traub, 1936),which is thought to be the source of transmission to humans. Scott and Rivers (1936) first documented the role of LCMV in human disease with the study of two patients presenting with acute aseptic meningitis. Viruses isolated from the cerebrospinal fluid of each patient were shown to be serologically identical to the two strains of virus obtained by Armstrong and Lillie (1934)and by Traub (1935). Human disease from LCMV has been reported in the Americas, Europe, and Asia (Armstrong, 1940-1941; Farmer and Janeway, 1942; Lehmann-Grube, 1971; World Health Organization, 1975). The virus is usually transmitted via field mice or other rodents and occurs most frequently in persons between the ages of 15 and 40 years of age, although infections have been described both in fetuses and in individuals over 70. Epidemiologic surveys of random populations in the United States in the 1930s indicated that 11% of sera contained neutralizing antibodies against LCM virus (Armstrong, 1940-41). The clinical syndromes produced by LCMV infection in man can b e separated into four categories: ( a ) inapparent or subclinical infection; ( b ) nonmeningial influenza-like illness; (c ) aseptic meningitis; and (d ) meningoencephalitis. Of these, the vast majority of cases fall into categories (u ) and (h). The reasons why previously unexposed individuals vary in the degree and severity of resultant tissue injury and clinical manifestations following LCMV infection are not clear. Of importance is the realization that passage of LCMV through hamster and perhaps dog tissues may enhance its virulence for hu-
278
M. J. BUCHMEIER ET AL.
mans. Safety precautions recommended by the Center for Disease Control and the NIH Advisory Board on Recombinant DNA have characterized mouse passaged LCMV, Armstrong strain, as an agent that can be handled in P2 facilities. Those handling fresh field isolates of LCMV or LCMV passaged through hamster tissues should use P3 facilities.' The apparent increased virulence of LCMV on passage through hamsters and the fact that LCMV infection can be transmitted from hamsters to humans (Lewis et al., 1965; Baum et al., 1966; Armstrong et al., 1969; Akermann et al., 1972; Vanze et al., 1975; Oldstone and Peters, 1978) suggest important concerns not only for investigators working with LCMV but for experimentalists who use hamsters or hamster tissues in their studies. This is because hamsters or cultured lines derived from hamster tissues may be persistently infected and carry infectious LCMV. Such animals show no illness and cell lines show no cytopathology. Hence, persons working with these reagents are frequently unaware of the potential danger. Of the 181 cases of LCMV infection reported to and documented by the Center for Disease Control in the early to mid 1970s, all originated from two sources: LCMV was transmitted to laboratory personnel when hamsters were implanted with tumor cells not known to be infected with LCMV and to families having pet hamsters later found to be persistently infected with LCMV. For example, in 1973 Hinman (Hinman et a!., 1975) reported 48 cases among personnel in the radiotherapy and vivarium departments of the University of Rochester School of Medicine. All these persons had contact with Syrian hamsters into which tumors persistently infected with LCMV were injected. These tumors were obtained from an outside research institute that distributed tumor cell lines to numerous laboratories throughout the country. I n another instance, at least 9 cases of LCMV infection were recorded by the Center for Disease Control in families having pet hamsters. Cases were scattered from upstate New York to Reno, Nevada, but investigations indicated that hamsters were obtained from a common supplier in Florida. Testing of the supplier's hamster breeding stocks indicated that one of every 13 hamsters assayed carried LCMV. Subsequently the infected tumor cell lines and hamsters were removed from the market. Clearly, since LCMV can cause persistent infections in tissue culture lines and experimental or pet animals, those handling hamster tissues or exposed to any of these situations are at high risk. Hamster
' Personal communications: Dr. J. Richardson, Director, Office of Biosafety, CDC; and Dr. S. Barban, Office of Recombinant DNA Activities, NIH.
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV INFECTION
279
fhcilities should be segregated from areas where LCMV is passaged in niice. Further, LCMV passaged in hamsters should be used with caution and only in appropriate containment facilities. C. LCMV: HISTOWCALFEATURES AND EXPERIMENTAL MODEL SYSTEMSUSED FOR IMMUNOBIOLOGIC AND I MMUNOPATHO LO GIC RESEARCH Observations with LCMV have yielded several of the fundaniental concepts in modern biology. Pioneering work by Rowe (1954) utilizing various assays that depleted immune responses showed that death associated with acute LCMV infection in the mouse resulted from the ininiune response. This work opened the field of viral ininiunopathology. Many investigators using a variety of techniques including neonatal thymectomy, irradiation, or treatment of adult niice with antilymphoid drugs or antithyniocyte sera have confirmed and extended Rowe’s observations (Rowe et d., 1963; East et al., 1964; H a s and Stewart, 1956; Hirsch et al., 1967; Hotchin and Weigand, 1961; Lundstedt and Volkert, 1967; Oldstone and Dixon, 1971a). The action of immune lymphocytes in mediating virus-induced imniunopathologic disease associated with LCMV infection was suggested by i i i vitra experiments performed independently by Lundstedt (1969) and Oldstone (Oldstone et al., 1969; Oldstone and Dixon, 1971b). Later investigations by Cole et (11. (1973) and Marker and Volkert (1973) showed that in uitro killing by immune lymphocytes was restricted to a population of cytotoxic cells bearing the theta marker. This first observation of cytotoxic T cell killing in a virus infection was subsequently confirmed by several groups for LCMV infection and extended to several other virus infections (reviewed b y Zinkernagel and Doherty, 1979). Similar results occur i i i uivo. Gilden et (11. (1972a,b) demonstrated that immunosuppressed mice infected with LCMV survived an ordinarily lethal dose of virus. Upon reconstitution of such niice with syngeneic T cells specifically primed to LCMV, acute LCM disease ensued. The experimental model for demonstrating acute tissue injury iri viva with LCMV infection is shown in Fig. 2. Infection with LCMV also provided the unique observations first recorded by Zinkernagel and Doherty ( 1974a) that cell-cell recognition between specifically primed cytotoxic T lymphocytes and virusinfected targets required recognition both of specific viral antigens on the cell surface and also the major histocompatibility ( H - 2 )complex at the D or K region. This H-2 restriction phenomenon, first described with LCMV infection, has been extended to other viral infections as
M. J. BUCHMEIER E T AL.
280 Acute Infection
Perivascular Inliltration
I. C. Inoculation
0
Adult Mouse
Dead
Balance Between
* lmmunopathologv Peripheral Inoculation
and lmmunitv
n
k
Oead
FIG. 2. Experimental model system for acute tissue injury with lymphocytic choriomeningitis virus (LCMV). (A) An adult mouse infected with LCMV by the intracerebral route usually dies in 6-9 days owing to acute meningial and ventricular inflammation. The effector cell of this response is likely to be the cytotoxic T lymphocyte. (B) If the infection is by a peripheral route the animal may either survive or die. In survivors a balance is struck favoring immunity over immunopathology. In those animals that succumb, the reverse occurs and inflammatory foci are found in multiple organs. Spleens of mice infected intraperitoneally are a rich source of cytotoxic T cells.
well as other cell-cell recognition systems (reviewed by Zinkernagel and Doherty, 1979) and is known to function both in vitro and in vivo (Zinkemagel and Doherty, 1973; Mims and Blanden, 1972; Zinkernagel and Welsh, 1976). Still other major advances in biomedical concepts were to come from the study of LCMV infection. Eric Traub (1936)noted the natural occurrence of persistent LCMV infection in mice whose tissues retained infectious virus throughout the animals’ lives. On the basis of Traub‘s early observations concerning natural infections of mice with LCMV as well as of Owen’s (1945) examination of chimeric cattle, Bumet and Fenner postulated (1949) that clonal elimination of immunocompetent cells was the basis for immunologic tolerance to viruses and self antigens, respectively. Hotchin and Cinits (1958) described a model whereby newborn mice, inoculated with LCMV within the first 24 hours of life, survived an ordinarily lethal dose of virus and showed a clinical and biologic picture similar to the naturally occurring disease described by Traub. The use of this model led several investigators (reviewed by Lehmann-Grube, 1971; Oldstone, 1979) to study immune responses, immune tolerance, and viruses. As described in
VIROLOGY AND IMMUNOBIOLOGY OF LCMV INFECTION
281
Section IV of this review, the evidence now at hand clearly indicates that there is no B cell tolerance in mice infected in utero or at birth with LCMV. Thus, using the LCMV model, Oldstone and Dixon in the late 1960s provided evidence that persistently infected mice did, in fact, mount humoral immune responses against their infecting virus. Antibody was present but previously undetected owing to its binding to virus and formation of virus-antibody immune complexes (Oldstone and Dixon, 1969, 1971a; Buchmeier and Oldstone, 1978).Several investigators used this information to show that in most, if not all, virus infections a host immune response is elicited resulting in the formation of immune complexes. The biomedical importance of the initial observation of virus-antibody immune complexes in LCMV infection lay in defining the events of virus-induced immune complex disease, and these events are now known to occur in a large variety of experimental models and natural infections of many animal species (reviewed by Oldstone, 1975).Figure 3 illustrates the model system used for studying persistent LCMV infection in uiuo. It has become clear that persistent LCMV infection in specialized Persistent Infection
A.
Immune lymphocytes
*/-4
Newborn Mouse
Virus Carriei
LCMV + lmmunosuppresslon
&
Immune Lymphocytes
\\ /
/ Adult Mouse
Dead
Virus Carriei Antiviral Antibody
FIG.3. Experimental model system for persistent lymphocytic choriomeningitis virus (LCMV) infection. (A) Neonatal infection of a mouse by any route results in a lifelong virus carrier state with persistence of infectious virus and formation of antiviral antibody. Such mice develop immune complex disease, the severity of which is strain dependent. Transfer of lymphocytes primed to react against LCMV to these persistently infected mice does not initiate the acute LCMV disease. (B) Adult infection coupled with immunosuppression (X-ray, thymectomy, antithymocyte serum, Cytoxan, etc.) results in a virus carrier state similar to that observed in the neonatally infected mouse. However, in contrast, transfer of immune lymphocytes results in acute LCMV disease and death of the mouse due to acute inflammatory disease. Antibody transfer does not.
282
M. J. BUCHMEIER ET AL.
cells in vitro can alter the cell’s physiology without changing the survival rate or growth characteristics. The observation that persistent LCMV infection of neuroblastoma cells impairs the cell’s ability to make enzymes that synthesize or degrade acetylcholine, the major neurotransmitter, while not affecting vital functions of the cell (Oldstone et af., 1977), offers a new mechanism whereby viruses can cause disease via cell dysfunction without cell destruction. Again, the initial observation of LCMV-induced loss of luxury functions of cells has been extended to other viruses and cell culture systems (Koschel and Halbach, 1979; Halbach and Koschel, 1979). In this review we are concerned with bringing together data that clearly indicate that tissue injury and disease in both acute and persistent LCMV infection result from the host immune response against LCM virions or infected cells. Data supporting this concept have developed in several laboratories and resulted in an avalanche of information that can now be sorted and assembled critically. This should provide a picture of the current knowledge of LCMV infection specifically and viral immunopathologic investigation in general. Most of this information from the last decade has been built on the carefully and critically devised foundations described by Rowe (1954) for acute, and by Traub (1936) and Hotchin (1962) for persistent, LCMV infection. The data in this field, prior to early 1970 have been reviewed thoroughly in monograph forni by Lehmann-Grube (1971); therefore, the following review will predominantly focus on the events occurring since then. We will not cover the biomedical importance of LCMV and other arenaviruses as human health problems, and we refer those interested in this area to recent reviews on that subject (World Health Organization, 1975; Oldstone and Peters, 1978).
II. Virus a n d Host Cell Interactions
A. VIRIONSTRUCTURE AND ORGANIZATION 1 . Morphology
Several morphologic studies of the arenaviruses have been published (Murphy et al., 1969, 1970; Murphy and Whitfield, 1975; Dalton et al., 1968; Mannweiler and Lehmann-Gmbe, 1973), and, on the basis of these reports, the following general structural features have emerged. Arenaviruses as a group are round to pleomorphic in shape and range in size from approximately 50 to 300 nm. Negatively stained
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
283
preparations show a virion envelope (membrane) containing closely spaced surface projections (Murphy et al., 1970), which consist of the viral glycoproteins (Gard et al., 1977; Vezza et al., 1977; Buchmeier et ul., 1978). Thin-section electron microscopic examination of the interior of the virions typically reveals a membrane-bound viroplasm containing 20 nm electron dense granules embedded in an amorphous matrix. These granules, which give the virion a characteristic sandy appearance, have been shown to consist of host-cell derived ribosomes (Carter et al., 1973; Pederson, 1973; Pederson and Konigshofer, 1976; Farber and Rawls, 1975). No distinct nucleocapsid structure has been visualized within the virion by electron microscopy. Tissues from animals experimentally infected with LCMV (Fig. 1) (Walker et n l . , 1977) contain ribosome-bearing virions similar to those found in cells similarly infected in cultures. Lehmann-Grube and co-workers (Mannweiler and LehniannGrube, 1973) have expressed doubt that the infectious unit in LCMV infection is the widely observed ribosome-containing particle. These workers reported finding a smaller 50-65 nm particle with a dense core structure in cells infected by LCMV. Similar structures have been reported in tissues from patients with Argentine hemorrhagic fever (Maiztegui et al., 1975); however, it must be emphasized that these findings have not been reproduced in other laboratories, and that a number of investigators have clearly found the larger arenavirus particles associated with purified virus fractions containing high levels of infectivity (Murphy et al., 1969, 1970; Carter et al., 1973; Gard et izl., 1977; Vezza et ul., 1977).
2. Virio t i Subcompor he nt .s For the biochemist studying the chemical structure and basic rnechanisms of replication of the arenaviruses and for the immunologist studying the host response to the arenaviruses, it is important to understand how the components of the virus are assembled and what topological relationship exists among these structural components, both in the virion and on the surfaces of virus-infected cells. The virion envelope contains the structural glycopeptides in spikes projecting from the membrane (Murphy et nl., 1969, 1970; Gard et al., 1977; Vezza et a/., 1977). When virions are solubilized with nonionic detergents such as Nonidet P-40 or Triton X-100, the surface glycopeptides are stripped from the virion, yielding a dense ribonucleoprotein structure consisting of virion RNA and the nucleocapsid protein (Ramos et al., 1972; Gard et al., 1977; Vezza et al., 1977; Buchmeier et al., 1978). Unlike most other enveloped RNA and DNA viruses, the ribonucleo-
284
M. J. BUCHMEIER E T AL.
protein of arenaviruses appears amorphous upon isolation. Recent careful electron microscopy of Tacaribe virus (Palmer et al., 1977) and Pichinde virus (Vezza et al., 1978) has revealed a circular filamentous nucleocapsid containing globular subunits arranged in a “string of beads” fashion. If such an arrangement of filamentous circular RNA with attached protein subunits is proved to be universal among arenaviruses, this finding may yield important insight into the mode of replication and packaging of the virion. Ribosomes derived from the host cell are enclosed within the arenavirus virion. The number of ribosomes observed within virions appears to vary (Murphy et al., 1969, 1970), and no specific association of the ribosomes with other structural components has been described (Farber and Rawls, 1975). Whether the envelopment of ribosomes at the time of budding is merely fortuitous or is predetermined by the maturation mechanism of the virus is unknown. Evidence is mounting from several laboratories (Leung and Rawls, 1977; Vezza et al., 1978) that the ribosomes are not obligatory for infectivity.
$3. Specific Polypeptides and Antigenicity Individuals studying the immunologic properties of LCMV and the varied disease states associated with LCMV infection must consider the virus as a complex replicating mosaic of antigens. Viral antigens are found in many forms during the infection: as components of free virions, as virus antigens expressed on surfaces of infected cells, and as products liberated by degradation of virions or by immunologically mediated destruction of infected cells. Because the expression of viral antigens is complex, a detailed knowledge of their nature is essential for examining host responses to infection. A considerable portion of our effort has recently been directed toward understanding the molecular basis of disease (Welsh and Oldstone, 1977; Welsh et al., 1977; Buchmeier and Oldstone, 1978a,b, 1979; Buchmeier et al., 1978; Welsh and Buchmeier, 1979; Oldstone et al., 1980). The Arnistrong strain (CA 1371) of LCMV is composed of three major virus-specific polypeptides. These are the nucleocapsid protein (NP), which is a nonglycosylated 63,000 molecular weight polypeptide, and two glycopeptides (GP-1 and GP-2) with molecular weights of 44,000and 35,000, respectively (Buchmeier et a1 ., 1978; Buchmeier and Oldstone, 1979) (Fig. 4). Surface localization of the two glycopeptides was demonstrated by digesting purified virions with high concentrations of the proteolytic enzyme bromelain, then repurifying the digested virions to a buoyant density characteristic of intact virus. When analyzed by sodium dodecyl sulfate-polyacrylamide gel elec-
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV INFECTION
285
Slice Number
FIG. 4. Structural polypeptides of purified lyniphocytic choriomeningitis virus radiolabeled with [3H]glucosamine (O----O)and [35S]niethionine( L O ) . A preparation of purified virus was disrupted with sodium dodecyl sulfiate and P-mercaptoethanol then electrophoresed on a 12.5% gel. Radioactivity in 1mm slices was detennined. The molecular weights of the viral polypeptides are: nucleocapsid protein (NP),63,000; and two glycopeptides-GP-1, 44,000, and GP-2, 35,000. (Reproduced from Buchmeier et a / . , 1978.)
trophoresis (SDS-PAGE), the digested virions had lost most of their GP-1 and GP-2 glycoproteins. In contrast, the NP was unaffected b y proteolysis. Internal localization of NP was further supported by experiments in which virions were disrupted with Nonidet-40, and the ribonucleoprotein complex consisting of NP and ribonucleic acid was isolated on density gradients. Similar results have been obtained with a number of other arenaviruses (reviewed in Rawls and Leung, 1980), the chief differences being in the molecular weights of the various proteins. During the course of acute LCMV infection and convalescence in the mouse or guinea pig, antibodies to all the major viral polypeptides appear (Buchmeier and Oldstone, 1978b; Buchmeier et al., 1978; Oldstone et al., 1980). We have observed an additional virus-specific glycopeptide in LCMV-infected cells (Buchmeier et ul., 1978; Buchmeier and Oldstone, 1979). This glycopeptide, termed GP-C, has an apparent molecular weight of 74,000-75,000 and is found in the infected cells but never in purified virions. Pulse chase labeling studies demonstrated that [3H]glucosamine or [35S]methionine labels incorporated into GP-C during a short pulse could be chased into GP-1 and GP-2 in extracellular virus. Further, the tryptic and chymotryptic peptide maps of GP-C contained all the peptides of GP-1 and GP-2 (Fig. 5 ) .
286
M. J. BUCHMEIER E T AL.
FIG.5 . Precursor-product relationship between the lymphocyte choriomeningitis virion glycopeptides GP-1 and GP-2 and the cell-associated glycopeptide GP-C. The [3JS]niethionine chymotryptic peptides of GP-1 (A) and GP-2 (B) when mixed together and coniapped (C) reproduce the chymotryptic map of the precursor glycopeptide, GP-C (D). The same results are observed using trypsin digests. (Reproduced from Buchmeier and Oldstone, 1979.)
Taken together, these findings indicate that GP-C is a proteolytic cleavage precursor of GP-1 and GP-2. A similar precursor glycopeptide has been described in Tacaribe virus infection (Salehet al., 1979). Classical studies of LCMV (Smadel and Wall, 1940; Smadel et al., 1939, 1940; Lehmann-Grube e l al., 1975; Gschwender et al., 1976; Rutter and Gschwender, 1973) have identified two major categories of virion antigens. These are the complement-fixing and virusneutralizing antigens. We can now assign polypeptide specificities to
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV INFECTION
287
these antigens based upon recent work. The complement-fixing antigen was long thought to be a nonvirion soluble antigen (review by Lehmann-Grube, 1971; Lehmann-Grubeetal., 1975;Rawls and Leung, 1980). This antigen was produced in infected tissues in vivo and in cell cultures and could be dissociated from virus infectivity by high speed centrifugation. Under conditions in which virus infectivity was pelleted, the complement-fixing antigen activity remained in the supernatant (Smadel et al., 1940; Bro-Jorgensen, 1971). Studies of LCMV (Gschwender et al., 1976) and of Pichinde virus (Buchmeier et al., 1977) demonstrated that the complement-fixing antigens of these viruses are antigenically identical to an internal structural component of the virion. This component is released by solubilization of virions and, in the solubilized form, is resistant to heat denaturation and to proteolysis (Buchmeier et al., 1977). These findings correlate well with those reported by Bro-Jorgensen (1971) for antigen extracted from infected cells. Antiserum prepared against purified cell-associated complement-fixing antigen of Pichinde virus specifically immunoprecipitated the viral NP (Buchmeier and Oldstone, 1978a).In addition, antiserum raised by immunization with purified NP of LCMV has a high titer of antibody to complement-fixing antigen, does not neutralize virus, and reacts only with antigens expressed in the cytoplasm of infected cells (Buchmeier and Oldstone, 1978b). The means by which excess complement-fixing antigen (i.e., nucleocapsid) is released from the infected cell to become “soluble” antigen is unknown; however, its well documented resistance proteolysis (Bro-Jorgensen, 1971; Buchmeier et al., 1977) may contribute to persistence of this soluble antigen . Neutralizing antibody to LCMV is almost certainly directed against one or both of the virus surface glycopeptides. Sera containing high neutralizing antibody titers are efficient in immunoprecipitating GP-1 and GP-2; conversely, sera prepared against specifically purified NP (Buchmeier and Oldstone, 1978b; Buchmeier et al., 1978) do not neutralize virus, nor do such sera immunoprecipitate GP-1, GP-2, or their precursor, GP-C. Whether these surface antigens are identical to those recognized by cytotoxic T lymphocytes is a subject of ongoing studies. Evidence from studies in which cytotoxic T cell-mediated lysis of virus-infected cells was blocked by antisera containing anti GP-1, GP-2 antibodies, but not by monospecific anti NP, suggests that the surface glycopeptides are important in T cell cytolysis (Welsh and Oldstone, 1977; Buchmeier et al., 1978). Monoclonal hybridoma antibodies generated to all of the viral polypeptides can be used as molecular probes to examine this and other questions of specific polypeptide participation in LCMV-induced disease.
288
M. J. BUCHMEIER ET AL.
4 . Virul RNA It has long been thought that LCMV contains RNA, since inhibitors that are effective against DNA viruses do not influence the growth of the virus. Halogenated deoxyuridines (BUdR, FUdR, and IUdR), as well as arabinosyl cytosine, have no effect on LCMV. In contrast, LCMV synthesis is inhibited by 6-azauridine, an inhibitor of RNA synthesis (Pfau, 1974; Hotchin, 1971; Lehmann-Grube, 1971). Direct evidence came from PAGE analysis of viral RNA labeled with [3H]uridine and isolated from highly purified LCMV. Two virus-specific single-stranded RNA segments were found: a large RNA segment ( L ) with a sedimentation coefficient of 31 S and molecular weight of 2.1 x lo6,and a small RNA segment (S) with a sedimentation coefficient of23 S and molecular weight of 1.1 x lo6(Pederson, 1971).In our hands (Dutko et ul., unpublished observations), 32P-labeled LCMV RNA denatured with glyoxal (McMaster and Carmichael, 1977) can be resolved into 4 RNA species (L-1, L-2, S-1, and S-2) with molecular weights of 3.15 x lo8,2.95 x lo6, 1.35 x lo6, and 1.25 x lo6, respectively (Fig. 6). The relationship of these RNAs to each other, as well as the gene products encoded by each of the RNAs, are not known, but are active areas of investigation. Viral ribonucleoproteins (RNP) were isolated by disruption of purified LCMV with Nonidet P-40 followed by sucrose gradient centrifugation. Two size classes of viral RNP were observed: a small RNP that sedimented at 83 S and contained only the S RNA, and a larger heterogeneous RNP that sedimented between 123 S and 148 S and contained both L and S RNA (Pederson and Konigshafer, 1976). The genomes of arenaviruses are of negative stranded polarity (i.e., complementary to messenger RNA). Pichinde virus contains a virionassociated RNA-dependent RNA polymerase (Carter et uZ., 1974; Leung et uZ., 1979). Furthermore, Pichinde viral RNA lacks two characteristics of a positive stranded genome, namely polyadenylic acid at the 3' termini and methylated 5' termini. Pichinde viral RNA hybridizes to polysomal messenger RNA from infected cells (Leung et ul., 1977). The LCMV also contains an RNA-dependent RNA polymerase, but it is present at a lower specific activity than that of Pichinde virus (W. -C. Leung and M. J. Buchmeier, unpublished data). B. VIRUSREPLICATIVE CYCLE To understand host responses to LCMV infection, it is crucial to bear in mind that viral replication and antigen expression are not static processes, but may undergo a number of marked changes as the infec-
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV
INFECTION
289
FIc:. 6 . Lyniphocytic choriomeningitis virus (LCMV) RNA. 32P-labeled LCMV was prepared by infecting BHK cells at a niultiplicity of infection of 0.1, and 1 day after infection culturing in medium containing one-tenth the nonnal amount of phosphate and 3'P. Virus was purified from the slipernatant, and viral RNA was isolated by extraction with pheno1:chloroform ( 1 : l), followed by precipitation with ethanol. The RNA, either not denatured (A) or denatured (B) with glyoxal in the presence of dimethylsulfoxide, was analyzed by electrophoresis in R 1.5% agarose gel. Indicated by arrows are the positions of 28 S and 18 S rihosonial RNAs, and Seniliki Forest virus 42 S RNA (F.J. Dutko and S. I. T. Kennedy, unpublished results).
tion progresses from the early acute stage through long term persistence (Cole et d.,1973; Welsh and Oldstone, 1977). Figure 7 illustrates the expression of viral antigens, production of infectious virus, and susceptibility to immune attack mechanisins for both L-929 and
M. J. BUCHMEIER ET AL.
290
=I
E
1
Acute
3 5 7 >720 Days after Virus Infection
Persistent
FIG.7. Correlation of expression of viral antigens and infectious virus production with susceptibility to immune attack. L-929 cells infected with lymphocytic choriomeningitis virus produce maximum infectious virus (plaque-forming units, PFU) and express high density surface antigens on 2 to 3 days after infection. This correlates with peak susceptibility to lysis by antiviral antibody + complement (C)and by cytotoxic T cells. Surface antigen and infectious virus decrease rapidly as DI virus is synthesized. Susceptibility to T cell lysis remains moderately high in persistently infected cultures that no longer produce infectious virus but continue to express viral antigens in their cytoplasms, (Data were obtained from Welsh and Oldstone, 1977.)
N115 neuroblastoma cells after infection with LCMV. It is readily apparent from such studies that infected cells express maximal concentrations of surface antigens 2-4 days after infection under these conditions and simultaneously achieve their highest sensitivity to both humoral and cell-mediated immune attack mechanisms. After this initial period of high susceptibility, both virus synthesis and susceptibility to immune lysis decrease markedly. Notable is the observation that persistently infected cells still retain susceptibility to cytotoxic T cells (about 50% of maximum) even though sensitivity to lysis by antiLCMV antibody and complement has waned. This may indicate that antigen density required on the membrane for T cell killing is quantitatively less than for antibody and complement lysis, or, alternatively, that each mechanism recognizes different cell surface determinants. The former mechanism appears more likely, because of Fab’2 guinea pig antibody to LCMV can block cytotoxic T cell killing (Welsh and Oldstone, 1977).
VIROLOGY AND FMMUNOBIOLOGY OF
LCMV INFECTION
291
Synthesis of the viral NP is apparently dissociated from replication of viral glycopeptides. The NP appears in the cell 4-6 hours after infection (Buchmeier et al., 1978) and remains in long-term persistently infected cells that no longer express surface antigen (glycopeptides) or produce detectable infectious virus (Welsh and Oldstone, 1977; Welsh and Buchmeier, 1979). Similar findings of virus persistence in the absence of infectivity have been described in vivo in Tamiami virus infection of the cotton rat (Sigmadon hispidus) (Murphy et al., 1976). The molecular mechanism for dissociation of virus production from antigen synthesis is unknown, but recent work from this (Welsh and Buchmeier, 1979) and other laboratories (Dutko and Pfau, 1978) suggests an involvement of defective interfering virus. The specificity of the shutoff is apparently exquisite. Cells persistently infected by LCMV are totally or partially resistant to challenge infection with LCMV and other arenaviruses, yet fully susceptible to heterologous unrelated virus (Welsh and Buchmeier, 1979; Lehmann-Grube et al., 1969; Staneck et al., 1972). In addition to controI of surface antigen expression exerted by virus replication, additional host-dependent mechanisms may function i n d u o . The best described of these is the phenomenon of antigenic modulation (Joseph and Oldstone, 1975; Fujinami and Oldstone, 1979). Mice that are LCMV carriers are known to have circulating antiviral antibody (Oldstone and Dixon, 1969, 1970; Benson and Hotchin, 1969; Buchmeier and Oldstone, 1978b; Oldstone et al., 1980). This antibody, which has demonstrable specificity for all the viral proteins, could potentially function by suppressing surface expression of LCMV antigens on infected cells.
c. NONASSOCIATIONOF VIRAL POLYPEPTIDES AND H-2 ANTIGENS The major histocompatibility gene complex (MHC) codes for a series of antigens that can be easily and quantitatively followed during virus infection. The amount of H-2D and K present on the surface of L-929 (H-2k)or N-115 (H-2") cells does not change during acute or persistent LCMV infection from the amount detected on uninfected cells (Welsh and Oldstone, 1977). Further, according to inimunoelectron microscopy and double fluorochronie immunofluorescence staining, the H-2 antigens and LCMV glycopeptides found on the surfaces of infected cells do not cocap (Fig. 8). Hence these two antigens are not associated as a complex within the sensitivity of the assay methods used.
292
M. J. BUCHMEIER ET AL.
FIG.8. Lymphocytic chorionieningitis virus (LCMV) surface antigen is not associated with H-Jk on the cell surface. L-929 cells infected 48 hours earlier were allowed to react with fluoresceinated antibody to LCMV under capping conditions, then kept at 4°C in sodium azide. Rhodamine-conjugated antibody to H-Zk was added to stain H-2 determinants on the cell surface. In this cell, the LCMV antigen has capped (left) while the H-2 remains distributed over the entire cell surface. Reversing the order of reaction and capping H-Zk first gave similar results.
D. VIRUSMUTANTSAND VARIANTS
1. Virus Assuy Traditionally, LCMV was considered to b e a noncytopathic virus, and this putative inability to cause cytopathic effects (CPE) seemed to be an important characteristic of its ability to produce persistent infections. For reasons that are not readily apparent, early results of LCMV plaque assays were not reproduced in other laboratories (reviewed by Pfau et al., 1973; Dutko and Pfau, 1978). Subsequently, though, LCMV has been plaque assayed on monolayer or suspension cultures of cells from many species, including mouse, hamster, monkey, human, cow, and vole (Pfau et ul., 1973). Most laboratories now use either BHK 21/13 S agarose suspensions (Sedwick and Wiktor, 1967; Pulkkinen and Pfau, 1970; Hotchin et a/., 1971), L-929 cell monolayers (Welsh and Pfau, 1972; Popescu and Lehmann-Grube, 1976), or Vero cell monolayers (Pfau et ul., 1973; Buchrneier et al., 1978; Doyle and Oldstone, 1978).
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
293
2 . Homologous lnterference and Production of Defective lnterfering Virus Early work had shown that cultured lyniph node cells from mice infected with a viscerotropic strain of LCMV resisted productive infection b y a neurotropic LCMV strain (Traub and Kesting, 1963). Further, peritoneal cells from mice persistently infected with LMCV produced only low levels of virus and did not produce higher levels upon challenge (Mims and Subrahmanyan, 1966). Although factors such as interferon production, viral recombination due to mixed infection, and macrophage activation may have played a role, these studies suggested the presence of a viral interference phenomenon. More conclusive evidence that homologous interference occurs in LCMV-infected cultures was documented by Lehmann-Grube and co-workers (Lehmann-Grube, 1967; Lehmann-Grube et al., 1969). Cultures of L-929 cells infected with LCMV (Armstrong) were repeatedly passaged in vitro. By day 80 after infection, production of infectious virus assayed by mouse inoculation had stopped. However, transfer of culture fluids to uninfected cells resulted in the infection of a small percentage of cells, suggesting that attenuated virus was being made. Nearly all cells in these persistently infected cultures contained viral protein in the cytoplasm and resisted superinfection with either Armstrong or WE strains of LCMV but were susceptible to heterologous viruses. Three additional points indicated that virus products were being made. First, it was possible to “cure” the cultures of the infection b y cultivation in the presence of antibody to LCMV. Second, the supernatants from these persistently infected cultures released a product that immunized mice against LCMV challenge. Third, the product was neutralized with LCMV antiserum. Lehmann-Grube et al. (1969) suggested that the responsible agent was a defective virus, although a silent infection resulting from the attenuated virus could not be ruled out. Using then newly developed assays to measure LCMV plaque formation and infective centers, Welsh and Pfau (1972) found an interfering component that inhibited LCMV synthesis infective center formation and cytopathology. This interfering component was generated rapidly during the acute LCMV infection. Interference was specific for LCMV and to a lesser degree for closely related arenaviruses. Heterologous virus synthesis was not inhibited. Welsh and Pfau (1972) also demonstrated that the interfering component appeared shortly before maximal infectious virus synthesis and peaked later. Dilution of the inoculuni before infection reduced the interference and resulted in
294
M. J. BUCHMEIER ET AL.
100-fold more infectious virus. On the basis of several physicochemical characteristics, they suggested that the interfering component was a defective interfering (DI) virus. A number of studies indicate that LCMV produces DI virus (reviewed by Huang, 1973), which may play a role in persistent infections. Defective interfering virus was first separated from infectious LCMV in the culture fluid of BHK 21/13 S cells persistently infected with LCMV (Welsh et aZ., 1972; Staneck et al., 1972). These culture fluids did not contain infectious LCMV, but released DI virus (Welsh et al., 1972). It is now clear that LCMV D I ( a ) sediments to a lower density in sucrose or renografin than infectious LCMV (Welsh and Buchmeier, 1979); ( b ) contains infectious LCMV structural proteins (Welsh and Buchmeier, 1979); (c) does not reproduce in culture unless infectious LCMV is also present (Welsh et al., 1972; Oldstone et al., 1977; Welsh and Oldstone, 1977); ( d ) interferes specifically with LCMV and to a lesser extent with related arenaviruses but not heterologous viruses (Welsh et al., 1972; Welsh and Buchmeier, 1979); and ( e )has a target size for ultraviolet inactivation significantly less than infectious LCMV, suggesting a reduced genome size (Welsh et al., 1972; Popescu et al., 1976). Although the precise nucleic acid deletion in DI LCMV is unknown (Welsh et al., 1975), the DI virus of a related arenavirus, Pichinde virus, lacks the viral S RNA (Dutko et al., 1976). Several quantitative assays exist for measuring DI LCMV, including infective center inhibition (Welsh et al., 1972), effect on clone formation (Dutko and Pfau, 1978), and a “negative plaque” assay (Popescu et al., 1976). The latter two deserve mention because of their uniqueness to the LCMV system. Dutko and Pfau (1978) tested several “low interference” cell lines for susceptibility to LCMV infection and found extensive CPE and inhibition of clone formation in MDCK cells. Treatment of cells with D I virus before infectious virus challenge protected the cells and enabled them to form clones. Popescu et al. (1976) developed a sensitive assay for D I virus. Using LCMV passaged in mouse brain, they obtained a stock that was highly cytotoxic to monolayers of L-929 cells. If cells were pretreated with culture fluids containing D I virus, areas of resistant cells developed after infectious LCMV challenge. These “negative plaques” resulting from coinfection with D I and standard virus were quantitated as an assay for DI virus. The use of the various assays for DI LCMV has revealed interesting biological properties. Statistical data indicate that preinfection with one DI virus can protect a cell from infectious LCMV challenge. Interference with infective center formation, production of plaque-forming
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
295
units (PFU), fluorescent antigen production, and cytolysis is usually an all-or-none phenomenon, that is, pretreatment of a cell with DI virus may completely inhibit the production of any detectable viral product or effects (Welsh et al., 1972; Oldstone et al., 1977; Welsh and Oldstone, 1977; Popescu et al., 1976; Dutko and Pfau, 1978; Welsh and Buchmeier, 1979). The primary effect of DI LCMV in a persistent infection may be to inhibit the cytolytic potential of the infectious LCMV. Dutko and Pfau (1978) examined the interaction of infectious and DI LCMV in low interference cell lines (Perrault and Holland, 1972) that restrict the generation and interference of DI virus. In this “DI virus-free” system, infectious LCMV killed MDCK cells as assayed by colony formation. But DI LCMV (isolated from persistently infected cells) could prevent infectious LCMV cell killing and establish a persistent infection in vitro. 3 . Production of Other Mutants Other persistent virus infections are characterized by the generation of aberrant virus forms occurring by mutation, selection, or recombination (Holland et al., 1979; Youngner, 1977; Elder et al., 1977). It is possible, therefore that, in addition to the production of DI virus, other variants may be involved during LCMV persistent infections. We looked for temperature-sensitive mutants in BHK or L-929 cell cultures infected with LCMV Armstrong strain, but found none (Welsh and Buchmeier, 1979). However, other investigators have described persistently infected L-929 ceIls that appeared to produce an agent in the supernatant that transferred LCMV-specific fluorescent antigen to less than 1 in 10,000 susceptible cells (Lehmann-Grube et al., 1969). In such cultures, Popescu and Lehmann-Grube (1976) later detected variants that formed turbid plaques. Marked variations in plaque morphology among strains of LCMV were first reported by Pulkkinen and Pfau (1970), and correlations between plaque morphology and pathogenicity in uitro and in uiuo have been made by Hotchin et al. (1971). These investigators isolated variants of LCMV (WE strain) that formed clear and turbid plaques on BHK cells. The clear variant lysed BHK cells, but the turbid variant caused little CPE. However, the turbid variant interfered with the CPE of the clear variant. In light of current knowledge, a possible explanation for these findings is that the turbid variant generated more DI virus than the clear variant. One other LCMV variant has been implicated as necessary for maintenance of a persistent infection in uitro (Jacobson et al., 1979). A turbid-plaque variant (designated SP) was detected in culture fluids
296
M. J. BUCHMEIER ET AL.
from LCMV persistently infected L-929 cells. The SP variant, unlike infectious LCMV, was resistant to interference by DI virus. The biological importance of this observation is that the SP variant was required to establish a D1 infectious LCMV-initiated persistent infection in MDCK cells. Thus, the function of the SP variant, which is DI virus resistent, may be to maintain the infectious genome in persistent infection. It will be important to determine whether DI resistant variants are also generated i n vivo.
+
4 . Regulation of Virul Antigen Synthesis b y Mutants Cells (BHK, L-929, N-115) persistently infected by LCMV express considerably lower levels of viral antigens on the cell surface than acutely infected cells (Welsh and Oldstone, 1977; Welsh and Buchmeier, 1979). As a result, these persistently infected cells are more resistant to LCMV-specific immune attack mechanisms (antibody to LCMV plus complement; LCMV-specific cytotoxic T cells), while remaining comparably sensitive to effector mechanisms not specific for LCMV (antibody to H-2 plus complement; activated natural killer cells). Quantitative studies by Welsh and Buchmeier (1979) showed that the production of DI LCMV in persistently infected cultures was 10- to 100 fold lower than infectious virus production during acute infections. Most persistently infected cells expressed the LCMV necleocapsid antigen in the cytoplasm, but the synthesis rate of this protein was too low to be detected by the radiolabeling methods used. These persistently infected cells produced relatively low levels of detectable DI virus. In such cells no additional synthesis of LCMV-specific protein was detectable upon superinfection with infectious virus. Therefore, such a mechanism may make it possible for cells to escape immunosurveillance.
5 . Possible Role of DI Virus in Vivo Assuming that the negative plaque assay (Popescu et al., 1976) is a true assay for DI LCMV, persistent LCMV infection in the mouse is the only naturally occurring persistent infection in which DI virus has been quantitated. “Interference focus-forming units” were found in tissues of adults and newborn mice during both acute and persistent infections (Popescu and Lehmann-Grube, 1977).Persistently infected mice differed from persistently infected cells in uitro in that significant levels of infectious virus could be detected. This i n vivo generation of D I virus strongly suggests the existence of, and a mechanism for, viral interference in uivo. Based on infective center inhibition, Jacobson and Pfau (1980) have detected a similar interfering entity in
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
297
the kidneys of 1-week-old mice infected at birth with LCMV. Both infectious arid D I LCMV are synthesized in newborn mice at levels 100- to 1000-fold higher than in older chronic carriers or acutely infected adults (Popescu and Lehmann-Grube, 1977). Peritoneal cells from persistently infected mice contain LCMV antigen, but produce only low levels of LCMV PFU and infective centers and resist superinfection with LCMV, all characteristics of in vitro persistent infections (Mims and Subrahmanyan, 1966; Welsh and Oldstone, 1977). Addition of DI virus iu vitro to primary peritoneal cells from uninfected mice renders these cells resistant to LCMV infection (Welsh and Oldstone, 1977). Injection of a very high concentration of DI LCMV with low levels of infectious LCMV into the brains of rats (Welsh et al., 1977) or mice (Welshet ul., 1975) aborts the LCMV infection and protects the animals. Thus, DI LCMV is produced i i i vivo and can interfere in uivo. Much still needs to be resolved concerning its role in vivo during natural infections. Significant diversity of LCMV plaque morphology can be generated in vivo, even when infections are initiated with cloned viruses (Hotchin et a ] . , 1971; Popescu and Lehmann-Grube, 1976). Generally speaking, virus isolates from the brain are clear-plaque formers, whereas isolates from the liver or spleen are turbid-plaque formers (Hotchin et nl., 1971; Popescu and Lehmann-Grube, 1976; R. M. Welsh, unpublished observation). The reason for this reproducible strain variation is unknown. Hotchin et al. (1971) reported that a clear-plaque variant lysed cultured cells in uitro. I n vivo this isolate induced convulsions and early (7 day) death in intracerebrally inoculated adults, and it uniformly killed newborn mice. In contrast, the turbid-plaque variant did not cause extensive CPE in culture but interferred with CPE by the clear-plaque variant. The turbid variant caused persistent infections in newborn mice and no convulsions and only late (day 11) deaths in intracerebrally inoculated adult mice. Follow-up studies were not done on these variants, and the possible role of D I virus generztion was not evaluated. Popescu and Lehmann-Grube (1976) studied the generation of plaque variants i n v i m , but could find no correlation between plaque morphology and disease. It is likely that multiple variants are generated during the LCMV infection and that a spectrum of pathobiological potentials may be represented. Jacobson and Pfau (1980) have reported the isolation of a clear plaque-forming LCMV variant that appears to be highly resistant to interference b y DI LCMV. Characterization of this variant and its capacity to cause disease may be very useful in dissecting the complex interrelationship of plaque variation and D I virus genesis during the acute and persistent infections.
298
M. J. BUCHMEIER ET AL.
E. RELATIONSHIP BETWEEN LCMV AND OTHER ARENAVIRUSES Initially, impetus for the formation of the arenavirus taxon came from observations of similar natural histories for LCM and Machupo viruses (Johnson et al., 1965; Webb, 1965). Murphy et al. (1969, 1970) noted similar morphological features between Machupo and Tacaribe viruses and LCMV, and on this basis proposed the arenavirus group. Demonstration of serologic cross-reactivity between LCMV and members of the Tacaribe complex of viruses by Rowe and co-workers (1970a) provided the final link necessary to define the arenaviruses, and the grouping was formally made in 1970 (Rowe et al., 1970b). The viruses are indistinguishable morphologically and share varying degrees of group-specific antigen measurable by complement fixation or immunofluorescence. They do not cross-react in neutralization assays. In broad terms, the arenaviruses consist of two major subgroups. One group contains LCMV, Lassa, and, more recently, Mozambique viruses. The second includes an interesting group of eight viruses referred to as the Tacaribe complex. The Tacaribe complex consists of Amapari, Junin, Latino, Machupo, Parana, Pichinde, Tacaribe, and Tamiami viruses, all of which were isolated in South America or regions around the Carribbean (reviewed in Johnson et at., 1973; Rawls and Leung, 1980). A serologic relationship between LCMV and the Tacaribe complex of viruses was first noted by Rowe et al. (1970a), and the similarity to Lassa virus later became clear (Buckley and Casals, 1970). The full extent and the patterns of serologic cross-reactions among the arenaviruses are complex subjects that have been reviewed elsewhere (Casals, 1976) and hence will not be pursued further here. Limited studies of the molecular basis for serologic cross-reactivity among viruses of the Tacaribe complex have been camed out in our laboratory (Buchmeier and Oldstone, 1978a).Antisera directed against Amapari, Tacaribe, and Junin viruses specifically immunoprecipitated only the NP of Pichinde virus, whereas antiserum directed against Pichinde virus immunoprecipitated the two virion glycoproteins in addition to the NP (Fig. 9). A similar experiment with LCMV and Pichinde virus (Buchmeier and Oldstone, 1978b) showed no apparent cross-reactivity between these viruses when guinea pig antisera were used. In other studies (M. J. Buchmeier, unpublished data) with immunofluorescence, although guinea pig antibody was not crossreactive, hamster antiserum to Pichinde virus cross-reacted with LCMV-infected cells. An important consideration in human arenavirus disease is whether a relatively benign virus could be utilized as an attenuated vaccine to protect against the more dangerous human pathogens. Weissenbacher et al. (1976) demonstrated that guinea pigs previously infected with
VIROLOGY AND IMMUNOBIOLOGY OF
600
Anti
LCMV INFECTION
299
Pichinde
Gel Slice Number
FIG.9. Immune precipitation of Pichinde virus proteins by antisera against various Tacaribe complex arenaviruses. Purified Pichinde virus labeled with [35S]methionine (-) and [3Hlglucosarnine (---) was disrupted with Nonidet-40 and RNase, then immune precipitated with the antisera indicated. The immune precipitates were washed and analyzed on sodium dodecyl sulfate gels. Homologous antiserum precipitated all three major viral polypeptides, whereas the heterologous antisera precipitated only the viral nucleocapsid protein. (Reproduced from Buchmeier and Oldstone, 1978a.)
Tacaribe virus were resistent to lethal challenge with Junin virus. Kiley and co-workers (1979) reported protection of monkeys from lethal challenge with Lassa fever virus by prior immunization with the closely related Mozambique virus. The latter results are particularly exciting, since they represent a potential means of prophylaxis against this dangerous agent. We are currently raising and defining the specificities of monoclonal antibodies against LCMV, Pichinde virus, and Tacaribe virus. Obviously, one useful application of this technology will be to define relationships among arenaviruses at the level of single antigenic determinants. It is hoped that this approach will yield substantive information about evolutionary relationships among these viruses.
F. INTERACTIONS OF LCMV
WITH
DIFFERENTIATED CELLS
1 . Infection of Cells of the Immune System fri uitro studies of the growth of arenaviruses in tissue culture (Pfau et al., 1973; Rawls et d.,1976) have shown that these viruses grow optimally in replicating cells. This may explain the i n uivo observation
300
M. J. BUCHMEIER ET AL.
that LCMV apparently replicates best in tissues with high turnover rates. One of the tissue sites that has been shown repeatedly to support virus growth is the reticuloendothelial system (reviewed in Lehmann-Grube, 1971). Traub (1964) found that virus growth in lymph nodes peaked on the fifth day after infection. Mims and coworkers (Mims, 1966; Mims and Subrahmanyan, 1966; Mims and Wainwright, 1968) demonstrated high concentrations of viral antigen in spleens and livers of infected mice. In vitro, peritoneal macrophages supported virus growth (Mims and Subramahnyan, 1966). Wilsnack and Rowe (1964) also found high concentrations of viral antigen by immunofluorescence in the reticulum cells of the spleens and livers of acutely infected mice, but were not able to demonstrate viral antigen in lymphocytes. The in uivo consequences of replication in specialized cells of the reticuloendothelial system may compromise their function. Mims and Wainwright (1968), originally observed temporary suppression of the anti-sheep red blood cell plaque-forming cell response in adult mice acutely infected with LCMV. This observation was confirmed and extended by Bro-Jorgensen and Volkert (1974, reviewed by Bro-Jorgensen, 1978).These workers defined a period of transient immunosuppression after infection that correlated with a general suppression of hemopoiesis. Studies of immune responses by virus carrier mice to a number of unrelated soluble protein and particulate antigens as well as of their ability to reject allografts have in general not shown consistent deficits (Oldstone et a/., 1973; Mims and Wainwright, 1968; Bro-Jorgensen and Volkert, 1974; Holterman and Majde, 1971). Although mice responded normally to most antigens, there were a few notable exceptions. Decreased responses to heterologous mammalian and avian IgG were observed in the carrier mouse (Oldstone et al., 1973). Overall, these studies emphasize that a variety of antigens over a wide dose range must be used when evaluating a host’s immune response to nonviral antigens. Two reports (Doyle and Oldstone, 1978; Popescu et al., 1979) have demonstrated that lymphoid cells of the carrier mouse contain LCMV in an infectious form. Using infectious center assays, both groups recovered virus from a low percentage of mononuclear cells from spleens, thymuses, lymph nodes, and peripheral blood of persistently infected mice (Fig. 10).The activity was found in T cells and B cells as well as in macrophages. Recovery of virus required cocultivation of viable cells with susceptible monolayer cells. Pretreatment of the cells with antiviral antibody and complement abrogated the infectivity. Using adoptive transfers of unique, surface-marked lymphocytes from persistently infected mice into uninfected recipients, M . V.
VIROLOGY AND IMMUNOBIOLOGY OF LCMV INFECTION
301
. . 1
Mononuclear Cells Harvested from: Peripheral Spleen Thymus Bone Lymph Blood Marrow Node
FIG.10. Mononuclear cells scoring as infectious centers and harvested from lymphoid tissues of adult BALBlWEHI inice which had been infected at birth with LCMV. Ficoll-Hypaque purified mononuclear cells from a variety of lymphoid tissues were added to 60 x 15 niin petri dishes contiiining confluent Vero target cell monolayers. Plates were overlaid with 0.5% agarose in culture medium and incubated for 7 days. The plaques were then developed by overlaying with 0.5%agarose in culture medium containing 0.01% neutral red. Results are expressed a s the number of infectious centers per lo6 viable nioiionuclear cells. Each point represents data from a single experiment. Similar results were seen with SWRiJ mice. (Reproduced from Doyle and Oldstone, 1078.)
Doyle, L. V. Bloom, and M. B. A. Oldstone (unpublished data) showed that these infected lymphocytes circulated in the new host for several weeks. Thus lymphocytes can be a reservoir for infectious virus and can carry infectious virus to multiple body sites. Doyle and Oldstone (1978)also demonstrated the transient appearance of virus in lymphocytes of acutely infected mice. The infectivity peaked on the fourth day after infection and disappeared by the eighth day. It is tempting to speculate that the presence of virus in lymphocytes is somehow linked to the immunosuppressive effects of virus upon the host; however, at present insufficient data are available to warrant such a conclusion.
2. lrtfection of Neuroblasts and Neurolis Persistence of LCMV antigen in neurons has been described in viuo (reviewed by Nathanson et al., 1975) in mice infected at birth. The question of whether continued virus replication compromises the
302
M. J. BUCHMEIER ET AL.
function of the infected neurons has been approached at two levels. Our group (Oldstone et d.,1977) measured the enzymes acetylcholinesterase and choline acetyltransferase in N-115 neuroblastoma cells persistently infected with LCMV. Levels of these two luxury enzymes were significantly decreased in infected cells carried for over 3 years in culture. Such cells had normal growth rates and cloning efficiencies and had normal RNA, DNA, and total protein synthesis. Interestingly, levels of these enzymes were also altered in brain extracts of mice persistently infected with LCMV when compared to ageand sex-matched controls. Although the behavior of LCMV carrier mice has historically been regarded as “normal,” quantitative studies by Hotchin and Seegal (1977) revealed alterations in behavior and performance in open field and several learning tests, Whether these findings are related to the persistence of virus in neurons per se or to the other effects of persistent virus infection in general is not known. Ill. LCMV-Induced Acute Immune Response Disease
A. PATHOLOGY OF THE ACUTE INFECTION The pathology of LCMV infection varies greatly with the age of the mouse and the route of infection. Mice infected congenitally or within 24 hours after birth usually develop a lifelong persistent infection associated with infectious virus production in the blood and all major organs. These LCMV “carriers” may live a normal life-span without clinical symptoms, but commonly develop a disease consisting of glomerulonephritis, focal hepatic necrosis, arteritis, and lymphoid infiltration of many body tissues (Oldstone and Dixon, 1969, 1970). This syndrome is associated with a continuous antibody response to viral antigens. In contrast, immunocompetent adult mice infected with LCMV either succumb to infection or survive with permanent immunity. Although there is some variance with the virus strain and dose, adult mice infected intracerebrally usually develop an acute leptomeningitis and die of the infection. A small percentage may survive and become immune, but mice injected by other routes frequently survive and have asymptomatic infection. It should be noted, however, that viscerotropic strains with the capacity to kill intraperitoneally infected mice do exist (reviewed by Lehmann-Grube, 1971). The pathology of the acute disease was first described 40 years ago and reexamined by a number of investigators (Findlay and Stern, 1936; Lillie and Armstrong, 1945; Rowe, 1954; Hotchin, 1962;
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV INFECTION
303
Wilsnack and Rowe, 1964; Walker et al., 1975). Intracerebrally inoculated animals begin to show signs of disease by the fifth day postinfection. Symptoms include ruffled fur, a hunched posture, and blepharitis. On succeeding days the symptoms intensify, with weight loss, lethargy, irritability, and lead to convulsions. Death usually occurs 6-8 days after infection. As described by Hotchin (1962) the usual postmortem posture is “the convulsion position with rear limbs extended, forelimbs flexed, neck extended, and thoracic spine flexed.” The pathology of the acute LCMV infection is associated with a markedly enlarged spleen due to splenic hyperplasia, swollen lymph nodes, lymphocyte infiltration in many tissues, and perivascular cuffing (Fig. 11). Choroiditis, ventriculitis, pericarditis, cardiac valvulitis, and pneumonitis are also common features.The pathologic lesion in intracerebrally infected mice that distinguishes them from mice infected by other routes is infiltration of lymphocytes and macrophages into the central nervous system, resulting in marked lymphocytic meningoencephalitis. B. GENERALIZED HOST RESPONSES AND THE PRODUCTION OF INTERFERON Figure 12 is a composite diagram depicting the course of the first 2 weeks of LCMV infection, using data from our laboratory as well as data of others (Hotchin, 1971; Lehmann-Grube, 1971; Bro-Jorgensen, 1978). For most strains of immunocompetent mice studied, intraperitoneal injection of lo2 to lo5 PFU results in the following responses: viremia peaks at about the fourth to fifth day postinfection and then rapidly declines. Usually little if any infectivity can be detected in the blood beyond the eighth day after infection. Interferon is induced as early as the first day after infection, reaches maximum titers around the third day, then continuously declines to base levels by the eighth day. Concurrent with interferon synthesis is the generation of activated natural killer (NK) cells, which may be directly activated by interferon (Welsh, 1978). Also, beginning to occur at the peak of interferon synthesis is a marked deficiency in heniopoietic function, which does not return until the eighth day postinfection (not shown). Beginning on the fifth and peaking on the seventh to ninth days after infection is the appearance of LCMV-specific cytotoxic T cells. Synthesis of anti-LCMV antibody is detectable by the fourth day after iiifection. On the fifth day, the generation of highly activated macrophages is evident. These responses are shown in Fig. 12, and those after intracerebral inoculation are depicted in Table I.
FIG.11. Inflammatory infiltration of the choroid plexus (A, B) and meninges (C, D) of a RALBlw mouse infected intracerebrally 6 days earlier with 1000 PFU of lymphocytic choriomeningitis virus (LCMV). At this time the infiltrating cells consist of lymphocytes, macrophages, and polymorphonuclear leukocytes. In uitro, specific T cell-mediated immune attack on virus-infected target cells can be demonstrated. (E) LCMV-primed lymphocytes attacking a virus-infected L-929 cell. (Scanning electron micrograph by courtesy of Dr. Peter Lampert.)
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV INFECTION
305
Infectious Virus
Interferon
NK Cell Activation
>
Antiviral Antibody
Cytotoxic T Cell Activation
A
t
Activation
2
4
6
8
10
12
14
Days After Infection
FIG.12. A composite diagram of host responses to acute lymphocytic choriomeningitis virus (LCMV) infection following intraperitoneal inoculation derived from data 01)tained i n this and other laboratories. See text for explanation. CF, coniplement-fixiii~; NK, natural killer.
1 . Growth of Viru.s and Ziiterferun Productioli
After infection of adult mice, LCMV grows in many tissues. Some of the organs surveyed include blood, liver, spleen, lyinph nodes, thymus, lung, kidney, and brain (Rowe, 1954; Rivers and Scott, 1936; Traub, 1936; reviewed by Lehmann-Grube, 1971; Hotchin, 1971). TABLE I I hfMUNOLOGIC PARAMETERS IN ACUTE INTRACEREBRAL LYMPHOCVTIC CHORIOMENINGITIS VIRUSINFECTION" Day after iiioculation
"
Infectious vinis titer senmi (LDd0.03 nil)
Circulating V-Ig complexes
1.2 3.0 4.2
0
0 0
5.6 5.8
+ +
4.7
t
1000 LD,, inoculated into 6-week-old SWWJ mice
Cytotoxic T cells
0 0 0 0
+ ++
Tissue injury 0 0 0
0
+ ++
306
M. J. BUCHMEIER E T AL.
Gilden et al. (1972b) postulated that replication of virus in particular target cells of the brain may determine whether or not an acute choriomeningitis prevails. The virus has also been shown to replicate in macrophages (Mims and Subrahmanyan, 1966; Welsh et al., 1976; Schwartz et al., 1978) and lymphocytes, including both T and B cell subclasses (Doyle and Oldstone, 1978; Popescu et al., 1979) as well as in epithelial cells and fibroblasts. Doyle and Oldstone (1978) showed that a small population of peripheral blood T lymphocytes become infected and harbor virus during acute LCMV infection (Fig. 13).Although it is not known whether the infected T cells carry markers for cytotoxic T cells and participate in killing LCMV-infected targets, it is clear that both events occur during acute infection. It is now generally agreed that type I interferon is induced during the acute infection, peaking in titer on the second to fourth day and then declining gradually (Padnos et al., 1971; Riviere and Bandu, 1977; Bro-Jorgensen and Knudtson, 1977; Merigan et al., 1977). The question of interferon generation during LCMV infection is, however, surprisingly complex. In early studies, Wagner and Snyder (1962)
loo0l
'1 . -E -
1
0
. .
4 . :2 "3 '4 "5
LCMV 1
6
m 1112
Days Postinfaction
FIG. 13. Kinetics of appearance of infectious centers in )lymphoid cells of acutely infected BALB/w mice. Peripheral blood (A)or splenic (0)mononuclear cells were collected at the times indicated after intraperitoneal infection with 1500 PFU of lymphocytic choriomeningitis virus (LCMV) and analyzed for LCMV-specific infectious center formation. Each point represents data from a single experiment. Similar results are seen with SWWJ mice. (Reproduced from Doyle and Oldstone, 1978.)
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
307
could not detect any interferon in mice acutely infected with LCMV, and Veltri and Kirk (1971) isolated an inhibitor from the brains of infected mice that blocked the synthesis of some viruses (e.g., COXsackie, vesicular stomatitis virus, influenza), but not others (vaccinia, reovirus). This inhibitor was similar to interferon in that it required an induction period and was of low molecular weight, but its properties contrasted with interferon in that it was acid labile, was not species specific, and could be eliminated by a change of medium upon addition of infectious virus. This inhibitor has not been further characterized. The role of interferon in the control of virus infection is unclear. Reports differ concerning the sensitivity of LCMV to interferon. Gresser et (11. (1978) have shown that suckling mice treated with antibody to interferon synthesize 100-fold more virus than untreated mice ( 104.5 vs 106.6).Mice that are carriers of LCMV synthesize only low levels of interferon, and then only early in the infection (reviewed in Lehmann-Grube, 1971). The virus does not induce interferon from infected cells i n vitro (Welsh and Pfau, 1972). Why LCMV induces interferon better during the acute infection than during the persistent infection in uiuo and fails to induce interferon in tjitro is unclear. Interferon can be made by activated macrophages and expanding clones of B and T cells (reviewed in Stewart, 1979). However nude mice, which mount poor immune responses to LCMV infection, synthesize high levels of interferon (Merigan et ul., 1977; Welsh, 1978). Further, interferon production in nude or normal mice correlates with infectious virus synthesis, suggesting that interferon may be made by the infected cell rather than released by a primed immunocoinpetent cell.
2 . General Effect of LCMV
O I L Heniopoiesis
Work from two groups (Bro-Jorgensen and Knudtzon, 1977; BroJorgensen and Volkert, 1972; Silberman et al., 1978) has shown dramatic alterations in the heniopoietic activity of mice acutely infected with LCMV. Normally, sublethal irradiation of mice (500 rads) depletes many lymphoid cells whereas pluripotential heniopoietic bone marrow cells, which are relatively resistant to this dose of irradiation, repopulate the spleen and colonize there. When mice acutely infected with LCMV are irradiated with 500 rads, the repopulation of the spleen by hemopoieticcells is markedly inhibited (Bro-Jorgenseii, 1978; Silbernian et ( i f . , 1978).When normal inice are irradiated with higher doses (850-1000 rads), there is no hemopoietic cell repopulation of the spleen unless bone marrow cells are transferred from nornial mice.
308
M. J. BUCHMEIER ET AL.
Immune mice, or mice persistently infected with LCMV, have no apparent hemopoietic dysfunction and can be used as recipients for bone marrow or spleen cells from acutely infected mice. Cells from mice at different stages of the acute infection were transferred into irradiated immune or persistently infected mice (that did not undergo a cle nova acute LCMV infection). Using these techniques, Bro-Jorgensen and Knudtzon (1977) showed that pluripotential cells, quantitated by colony-forming units (CFU),decreased in number to about 10-30% of control values by 3 days after infection. Bone marrow CFU remained depressed for over 3 weeks, but spleen CFU rose to 4 times normal Ievels by day 10 postinfection, and then gradually subsided. The hemopoietic disorder accompanying LCMV infection may be associated with the induction of interferon and/or natural killer (NK) cell activity. Hemopoietic dysfunction is inversely proportional to interferon levels, and in several other systems interferon has been shown to inhibit immune responses (Chester et al., 1973; DeMaeyer, 1976). The hemopoietic dysfunction is not totally dependent on viral replication since carrier mice have normal hemopoietic tissues. However, NK cells may regulate hemopoiesis. It has been shown that sublethally irradiated mice and nude mice reject allogeneic bone marrow allografts, and a genetic system called the hemopoietic histocompatibility (Hh) system has been mapped. The effector cell of the Hh system has several properties common to NK cells (Kiesslinget al., 1977), including ( a ) maturation independent of the thymus; ( b )maturation at the third to fourth week of age; (c) dependence on the bone marrow, as judged by strontium-89 sensitivity; ( d ) resistance to X-irradiation; and (e) in vivo sensitivity to the antimacrophage agents silica and carrageenin. LCMV can induce interferon which makes NK cells cytotoxic against certain syngeneic primary tissues (Welshet al., 1979). By these means NK cells may directly lyse hemopoietic cells.
c. NATURAL KILLER CELL ACTIVATIONAND FUNCTION IN ACUTE INFECTION Correlating closely with the generation of interferon is NK cell activation (Welsh, 1978) (Fig. 12). Natural killer cells are nonadherent, nonphagocytic lymphocytes that lack surface immunoglobulin and express low concentrations of theta antigen. These cells have the capacity to lyse certain types of target cells (e.g., T cell lymphomas, thymocytes) without prior immunization (Kiessling et al., 1975; Herberman et al., 1975, 1978; Hansson et at., 1980). However, NK cells activated by LCMV infection increase their range of susceptible targets (Welsh and Zinkemagel, 1977; Welsh et al., 1979) and increase their cytotoxic
VIROLOGY AND IMMUNOBIOLOGY O F
LCMV
INFECTION
309
activity toward normally susceptible cells. This reflects two different manifestations of activation: ( a )an increase in the cytotoxic activity of previously active cells, such that they can then lyse previously resistant targets; and ( b ) an increase in the number of cytotoxic effector cells (Welsh and Kiessling, 1980a). This increase in the number of effectors does not necessarily implicate a clonal expansion, since NK cells are induced in X-irradiated mice (Welsh, 1978) and can be activated by short-term (1 hour) interferon exposure in vitro (Djeu et al., 1979). The NK cell activated during the LCMV infection differs in several ways from the nonactivated “endogenous” NK cell (Welsh and Kiessling, 1980a). The activated NK cell displays increased adherence to nylon wool, increased quantities or avidities of Fc receptors, and by the sixth day postinfection, increased concentrations of theta antigen and enlarged cell size. Activated NK cells lyse continuous cell lines as well as syngeneic primary fetal fibroblasts and syngeneic thymocytes (Welsh et ul., 1979; Hanssoii et al., 1980). Natural killer cells are clearly activated during LCMV infection; however, to date there is no evidence that activated NK cells play a major role in resistance to acute LCMV infection. C57BL/6 mutant beige mice, which have markedly reduced NK cell activity, do not synthesize LCMV in their spleens to any greater extent than normal beige/+ littermates, suggesting that N K cells do not play a significant role in lysing virus-infected cells early in the infection (Welsh and Kiessling, 1980b).
D. MACROPHACEACTIVATIONAND FUNCTION IN ACUTE INFECTION Blanden and Minis (1973) found that liver, spleen, and peritoneal macrophages were activated by 8 days after LCMV infection as indicated by increased bacteriocidal effect on Listeria monocytogewes. We have found an increase in the size and number of peritoneal macrophages of intraperitoneally infected mice beginning at the fifth day after infection and peaking at the tenth day, approximately 2-3 days afterthepeakcytotoxicTcel1 response (R. M. Welshand W. M. Doe, unpublished data) (Fig. 12). There is about a 20-fold increase in macrophage number, and the cells are large and heavily vacuolated. Twenty to forty percent of the macrophages contain LCMV antigen up to the sixth day after infection, but after that time antigen is rarely detected, perhaps because cytotoxic T cells eliminate infected niacrophages. No increase in size or number of peritoneal macrophages is detectable in athyniic nude mice, suggesting that macrophage activa-
310
M. J. BUCHMEIER ET AL.
tion may be a T cell-related function as it is in other model systems. McFarland (1974), working with Sindbis and yellow fever viruses, found that macrophages play an essential role in controlling virus titers. Jacobs and Cole (1976) reported that spleen cells from mice acutely infected with LCMV 1-8 days earlier had depressed proliferative responses to mitogens (lipopolysaccharide, phytohemagglutinin, concanavalin A) as determined by uptake of [3H]TdR. Addition of normal macrophages or P-mercaptoethanol to the cultures eliminated the defect. These authors speculated that a macrophage dysfunction was responsible for the depressed response. The macrophage dysfunction observed by Jacobs and Cole appeared to correlate with interferon levels in viva, also suggesting that interferon may perhaps be responsible for the immunodepression toward certain antigens seen during acute LCMV infection (Silberman et al., 1978). E. CYTOTOXIC T CELLACTIVATIONAND FUNCTION IN ACUTE DISEASE Virus persistence occurs when the host is unable to respond appropriately to viral antigens, either because of insufficient maturation or experimental depletion of the immune system. A number of experimental models have been utilized to dissect various aspects of this host-virus interaction. The most straightforward of the parameters tested is the dependence upon age of inoculation. Traub (1936) established that mice infected in utero carried virus for life. Hotchin and Cinits ( 1958) examined mortality among mice inoculated intracerebrally during the first few days of life and showed that a high percentage of mice infected during the first 24 hours of life survived and became chronic virus carriers. Mice infected later in life died of the typical acute inflammatory disease. Rowe (1954,1956)was the first investigator to attempt experimental dissection of the factors responsible for resistance to LCMV infection. He established that sublethal irradiation before infection protected mice from lethal infection. These irradiated and virus-infected mice were found to be persistently infected, suggesting that death resulted from immune attack on virus-infected tissues rather than from the CPE of the virus. Hotchin (reviewed in Hotchin, 1962) confirmed these studies. Rowe and co-workers (1963) presented the first evidence that LCMV-induced acute pathology was thymus dependent. Neonatally thymectomized mice infected 3-4 weeks later with LCMV survived infection and had high concentrations of infectious virus in their tis-
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
311
sues. These mice were spared the inflammatory disease seen in controls, but most made antiviral antibody. Later studies by Hirsch and co-workers (1968)using antithymocyte serum also confirmed the role of the thymus in the development of disease. Mice treated with antithymocyte serum for the first week of life were spared when subsequently infected with LCMV. Again, such mice made antiviral antibody but apparently could not mount a significant cell-mediated immune response against the virus. Subsequent studies using a variety of immunosuppressive regimens (reviewed in Cole and Nathanson, 1974; Nathanson et al., 1975) have confirmed these observations. Elegant studies by Cole and co-workers (1971, 1972) defined the cell population responsible for acute LCM disease. Virus-carrier adult mice formed experimentally by using cyclophosphamide-induced immunosuppression (Cole et a/., 1971; Gilden et al., 1972b) were then adoptively transferred with spleen cells from mice acutely infected with LCMV. The carrier mice succumbed to acute LCM disease. If the spleen cells were treated with anti-theta serum and complement their ability to adoptively transfer disease was abrogated. Transfer of inimune serum containing LCMV antibody did not cause acute disease. Lundstedt (1969)and Oldstone (Oldstone et nl., 1969)demonstrated that splenic lymphocytes with the ability to kill specifically LCMVinfected target cells were generated during acute infection. Such cells did not kill uninfected targets or targets infected with different viruses. Cole and co-workers (1973) definitively characterized this cytotoxic cell as a theta antigen bearing lymphocyte first appearing 4 days after infection and reaching peak activity approximately 10 days after infection. Maximum target cell susceptibility to lysis coincided with the expression of high density virus-specific antigen on the plasma membrane. Doherty and Zinkernagel (1974) provided another link in the story with the isolation of cytotoxic T cells from infiltrates in cerebral spinal fluids of acutely infected mice. These investigators (Zinkernagel and Doherty, 1974a) later made the important observation that killing of virus-infected targets required that the target cell and the cytotoxic T cell bear H-2K or H-2D antigens in common. This phenomenon of H - 2 restriction has since been explored and confirmed by many other groups (reviewed in Zinkernagel and Doherty, 1979) and will not be discussed further here.
F. B CELL ACTIVATION AND FUNCTION IN ACUTE DISEASE Free complement-fixing antibody to LCMV can be demonstrated as early as the sixth day postinfection and b y 2-3 weeks reaches titers of
312
M. J. BUCHMEIER ET AL.
64 to 128, which usually remain unaltered for several months. Free neutralizing antibody appears later (review in Lehmann-Grube, 1971). The reason for this is unclear, because neutralizing antibody can be readily made in guinea pigs (Webster and Kirk, 1974; Buchmeier and Oldstone, 1978b). Antibody in the form of infectious virus-antibody complexes can be detected as early as 4 days after infection by immunoprecipitation with antibody to IgG (Table I) and by the Raji cell assay (Theofilopoulos et d.,1974). We have quantitated the development of specific anti LCMV antibody in adult athymic "nude" (~zulnu) mice and their thymus sufficient ( n u / + )heterozygous littermates (M. J. Buchmeier, M. V. Doyle, and M. B. A. Oldstone, unpublished data, 1979). When such mice were infected intraperitoneally with 1000 PFU of the Armstrong strain of virus, antibody production measurable by immune precipitation of [35S]methionine-labeledLCMV was observed both in nulnu and n u /+ mice. Table I1 illustrates data extracted from one experiment. From the data we can see that precipitating antibody was present as early as 7 days after infection in both nude and heterozygous mice. This early antibody observed in both strains was most likely of the IgM class, since its activity could be abrogated by reduction and alkylation and by immunoprecipitation with heavy chain specific (anti G) antiserum. By 21 days after infection, heterozygous mice were producing IgG antibody, while only IgM antibody production was evident in the nude mice. The production of IgM antibody be the nude mouse suggests that LCMV is a thymus-independent antigen. TABLE I1 ANTIBODYRESPONSE OF ATHYMIC( n u h u ) AND THYMUSSUFFICIENT ( n U / + ) MICE TO LYMPHOCYTIC CHORIOMENINGITIS VIRUS (LCMV) INFECTION Percentage of viral CPM precipitated* Day after infection"
0 7 21 0 7 21 "
Mouse strain
BALB/WEHI BALBIWEHI BALBIWEHI BALB/WEHI BALB/WEHI BALBlWEHI
(nu/+) (nu/+) (nu/+) (nulnu) (nulnzr) (nuhu)
Unreduced
ReducedC
6.9 ? 1.0 23.9 2 2.1 40.9 ? 1.1 5.7 2 1.75 10.3 ? 0.42 22.1 ? 1.5
ND" 2.3 2 1.9 38.3 2 3.7
ND ND 6.9 ? 1.2
Intraperitoneal infection with 1000 PFU of Amistrong strain LCMV.
' Virus labeled with ["Slmethionine and disrupted.
Reduction with 0-mercaptoethanol followed by alkylation with iodoacetamide. ND, not determined.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
313
Concomitant with the production of antibody in heterozygous and nude mice is the appearance of C l q binding material in the plasma. In a study of large numbers of nude and heterozygous mice of the BALBNVEHI and CBANVEHI strains, the quantity of C l q binding material was elevated 7 days after infection and remained elevated in most mice for over 3 weeks after infection (M. B. A. Oldstone and M. J. Buchmeier, unpublished data, 1979). Although levels of C l q binding material roughly paralleled the early rise in antibody production in these mice, correlation of individual mice on the basis of antibody concentration and subclass with their C l q levels must still be made. The role that antibody plays in the immunopathology of acute infection is uncertain. Oldstone and Dixon (1971b) demonstrated that antibody to LCMV and complement could lyse LCMV-infected cells and that i n uivo administration of cobra venom factor to deplete C 3 significantly reduced the mortality of intracerebrally infected mice. The LDsUof a stock of LCMV was lo4to lo5for controls, but only to lo3 for treated mice. The authors proposed that this implied a role for antibody in the acute infection. However, more recent work showing the ability of C3b to activate macrophages and the chemotactic properties of complenient byproducts raises the possibility that the effect of complement in this system is antibody-independent. Further, Welsh and Oldstone (1977) have shown that LCMV-infected cells may be lysed by complement in the absence of antibody (see Fig. 7). No appreciable differences in mortality have been noted between a large number of C5 sufficient and C5 deficient mice so that the protective effect of cobra venom factor may occur before C5 at the C 3 stage. The most convincing evidence against a major role for antibody in the acute central nervous system disease has been provided by Johnson and co-workers (1978). These investigators treated mice from birth with a goat antimouse M chain antibody. These mice were rendered B cell deficient. When infected intracerebrally with LCMV, these B cell (antibody) deficient adult mice developed acute LCMV disease clinically and histopathologically indistinguishable from untreated controls. Substantial additional evidence indicates that T cells are the major effector cells of the acute immune response disease (see Section 111,E). IV. LCMV-Induced Chronic Immune Response Disease
A. B CELL ACTIVATION AND FUNCTION IN PERSISTENT DISEASE Mice infected i n utero or at birth with LCMV harbor virus for life (Traub, 1936). This infection, described by Traub as a perfect
3 14
M. J. BUCHMEIER ET AL.
parasitism, was not associated with any overt disease symptoms. In addition, many attempts to measure antibody in the serum of these virus carrier mice by complement fixation techniques were negative (reviewed by Lehmann-Grube, 1971). On the basis of these early observations, the virus carrier mouse was assumed to be immunologically tolerant to the virus. Oldstone and Dixon (1969, 1970, 1971a) showed that immunologic tolerance did not occur, since mice persistently infected with LCMV made specific antiviral immunoglobulin. The antibody in the serum bound virus antigens to form virus-antibody immune complexes and deposited in the renal glomeruli, arteries, and choroid plexus (reviewed in Oldstone, 1975) (Fig. 14). Antibody deposited in the glomeruli could be dissociated from the complexed antigen by acid
FIG. 14.Immune complex disease induced by lymphocytic choriomeningitis virus. Deposition of IgC (A) and viral antigen (B) in a glomerulus and deposition of IgG in the renal arterial wall (C) of a persistently infected SWWJ mouse. Electron micrograph (D) of an affected glomerulus shows electron dense deposits (D,D,D)in subepithelial location; EP designates an epithelial cell; and gbm, glomerular basement membrane. (E) Deposition of immune complexes in the choroid plexus. Host IgC is found deposited along the choroid plexus ofthe lateral ventricle ofa persistently infected SWWJ mouse.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
315
elution then isolated, and the specificity was demonstrated by adsorption with virus-infected cell cultures (Oldstone and Dixon, 1971a). This observation spurred a reevaluation of “immune tolerance” associated with other chronic virus infections and led several investigators to conclude that a B cell response with associated virus-induced immune complex fomiation was the usual occurrence (reviewed in Oldstone, 1975). Deposition of antiviral antibody-viral antigen-complement complexes leads to progressive development of glomerulonephritis and arteritis in virus carrier mice (Oldstone and Dixon, 1969, 1970). Disease was found to be more severe in certain strains of mice (e.g., SWWJ) than in others (e.g., C3H). When disease severity was conipared with host antiviral antibody levels, a positive correlation was found between anti-LCMV IgG and disease (Oldstone and Dixon, 1970). Additional support for the antiviral-antibody induced etiology of the immune complex glomerulonephritis came from two additional studies. First, mice infected at birth were foster nursed on mothers immune to LCMV (Oldstone and Dixon, 1972). The foster-nursed, newborn infected mice had more rapid and severe onset of immune complex glomerulonephritis and arteritis than conventionally reared carrier mice and substantially shorter lives. Maternal antibody was found complexed to LCMV antigens in the glomeruli. Second, antiviral antibody transferred to persistently infected mice or the parabiosis of an immune syngeiieic mouse to a persistently infected mouse resulted in enhancement and severe manifestation of the chronic LCMV disease (Oldstone and Dixon, 1970). Recently, we deterniined whether B cell responses were directed to all the LCMV polypeptides. We studied the specificity and concentration of antibody produced by several strains of virus carrier mice (Buchmeier and Oldstone, 1978b; Oldstone et al., 1980). Initially antibody was eluted from the renal glomeruli of several strains of LCMV carrier mice as well as from the pooled glomeruli of AKWJ mice (Oldstone et al., 1974) spontaneously infected with murine leukemia virus. The concentrations of IgG obtained from the LCMVinfected mice ranged from 5.7 pg per kidney for BALB/WEHI ( t i u / + ) mice to 17.4 Fg per kidney for SWWJ mice. The specificity of the recovered antiviral antibody was assessed by means of a radioimmune precipitation assay (Buchmeier and Oldstone, 197%) in which purified virus was labeled in the total proteins with [35S]methioi~ine and in the glycoproteins with [3H]glucosamine (Buchnieier et al., 1978) then allowed to react with the IgG eluted from the glomeruli. The resultant complexes of antiviral antibody and antigen were pre-
316
M. J. BUCHMEIER ET AL.
cipitated with antibody to mouse IgG then analyzed both for the total counts per minute of each label precipitated and for specific viral polypeptides by SDS-PAGE. Specificity of the reactions was assured by two controls: first, by demonstrating the failure of LCMV antibody to precipitate Pichinde virus, an arenavirus not closely related serologically to LCMV; and second, by showing that IgG recovered from the kidneys of AKWJ mice (lacking antibodies to LCMV but having antibodies to murine leukemia virus) did not precipitate LCMV proteins. The results of the immune precipitation are shown in Table 111, and results of the polypeptide analysis of the immune precipitates on SDS-PAGE in Fig. 15. One can see from these data that the eluted IgG had specificity for all three of the LCMV structural polypeptides. Some quantitative differences were observed in the precipitation of the viral glycopeptides among strains of mice. The significance of this observation is not known. The specificities observed illustrate that there is no selective exclusion of a given virion antigen that could account for failure to eliminate virus. Further studies (Oldstone et al., 1980) in which the concentration of antiviral antibody in the sera of virus-carrier mice was assessed directly indicate that the quantity of this antibody was equivalent to that found in mice or guinea pigs hyperimmune to LCMV. Using the radioimmune precipitation assay described above, we have examined the sera of a large number of virus-carrier mice. Antibody was demonstrable in virtually TABLE 111 IMMUNE PRECIPITATION OF LYMPHOCYTIC CHORIOMENINGITIS VIRUS (LCMV) AND PICHINDE VIRUS (PV) PROTEINS BY ~MMUNOGLOBULIN(IG) ELUTEDFROM LCMV CARRIER AND CONTROLMOUSEKIDNEYS
Ig eluted from persistently infected mice (strain)
BALBIWEHI nu/+ (LCMV infected) CBA/WEHI nu/+ (LCMV infected) SWWJ (LCMV infected) AWJ (MuLV infected)
Percentage of input CPM precipitated" Reaction (C Ld
Viral antigen
4.2
LCMV PV LCMV PV LCMV PV LCMV PV
2.8 6.5 3.2
[9H]Glucosamine [35S]Methionine
16.1 4.1 30.5 4.1 14.2 3.0 4.8 5.3
31.0 4.9 34.2 6.4 48.3 4.9 4.9 7.2
Mean input counts per minute: LCMV, 3H 17,000; "S 9600; PV, 3H 17,000; 35S8300.
VIROLOGY AND IMMUNOBIOLOGY OF LCMV INFECTION
C I A nu/*
317
Ilcniafacttd AKRII
Fraction Number
FIG.15. Immune precipitation of the polypeptides of lymphocytic choriomeningitis virus by immunoglobulin eluted from the kidneys of virus carrier (LCM V.C.) mice of three strains and control mice. Viral antigen was disrupted by incubation with Nonidet P-40 and RNase, then allowed to react with the following quantities of eluted immunoglobulin: SWWJ, 6.5 pg; CBA n u / + , 2.8 p g ; BALB nu/+ 4.2 pg; and AKWJ 3.2 pg. The resultant immune precipitates were collected by precipitation with antibody to mouse immunoglobulin, washed, and analyzed by SDS-PAGE on 12.5% gels. (Reproduced from Buchmeier and Oldstone, 1978b.)
all these mice, representing several strains and ages from 1 to 6 months. Hence, virus-carrier mice make large amounts of antibody to all the viral polypeptides. Although IgM antibody production does not require T cell function (Section III,F), other experiments (Oldstone et al., 1980; M. V. Doyle and M. J. Buchmeier, unpublished observations) suggest that the continued production of IgG antibody in virus carrier mice is thymus dependent. Athymic nude (BALBNVEHI)mice and their thymus sufficient littermates were infected at birth with LCMV then tested at 4 months of age for evidence of persistent viremia and for circulating antiviral antibody. Figure 16 shows the results of the antibody assays done on six individual nude mice and heterozygous littermates as well
318
M. J. BUCHMEIER E T AL.
1
2.5
5 pI Serum Added
10
2.5
5
10
pl SeNm Added
FIG. 16. Antibodies to lymphocytic chorionieningitis virus (LCMV) in the sera of 12-week-old mice infected at birth with LCMV. Sera obtained from six heterozygous nudes ( n u / + ) and from six homozygous (nulnu) athymic nudes with BALBlw backgrounds were allowed to react with a constant virus concentration of 1400 ng of viral protein labeled with [35S]methionine.The resultant complexes were precipitated with excess rabbit antibody to mouse immunoglobulin. (Reproduced from Oldstone et al., 1980.)
as six control mice immunized against LCMV. The data clearly indicate that precipitating antibody was present in all the heterozygous mice but in none of the nude mice. Further, levels of precipitating antibody in the virus-carrier heterozygotes were similar to those in the immunized controls. It is important to point out that the sera of such carrier mice, while clearly containing precipitating antibody, have been tested by complement-fixation methods and by that criterion were negative. The possibility that this reflects the selective production of a noncomplement-fixing antibody subclass in the virus carrier mouse is under study. Polypeptide specificity of the circulating antibody was demonstrated to be the same as that of antiviral antibody recovered from the kidneys of virus carrier mice. Neutralizing antibody has been sought in virus-carrier mice by a number of groups in the past, generally with negative results (reviewed by Lehmann-Grube, 1971). Recently, we examined sera from individual virus-carrier mice of several strains. Definitive but low levels of virus neutralizing activity were found only in two of six BALBlc carrier mice. Popescu (1977) has reported finding virusneutralizing activity in the sera of LCMV carrier NMRI mice. Using indirect immunofluorescence, we found in sera from virus-carrier mouse antibody that bound specifically to the surface of virus-infected cells, but did not mediate complement-dependent lysis of such cells (M. V. Doyle; M. J. Buchmeier, and M. B. A. Oldstone, unpublished observations). The possibility exists either that this surface binding
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
319
antibody acts in viuo as “blocking antibody” preventing immune recognition or attack, or that the surface reactive antibody induces modulation of viral antigen off the surfaces of infected cells. Clearly, antiviral antibody is present in virus-carrier mice and is responsible for the development of immune complex disease. The nature of the antibody in terms of IgG subclass and its role in regulating expression of viral antigens in viuo are subjects for future studies. The pathologic consequence of persistent LCMV infection is chronic disease characterized by glomerulonephritis, arteritis, focal hepatic necrosis, and widespread lymphocyte infiltrations (Oldstone and Dixon, 1969, 1970; Accinni et al., 1978) (Fig. 17). Carriers of LCMV contain high concentratioiis of infectious virus, viral antigen, and antiviral antibody. Immune complexes, which result from the union of these reactants, deposit in tissues and lead to the chronic lesions observed. Particularly interesting are the widespread foci of infiltrating cells (Oldstone and Dixon, 1969, 1970; Accinni et al., 1978).Histopathologic studies of such interstitial foci have shown that they contain abundant plasma cells. The use of immunofluorescence to examine several tissues, expecially reiial tubule cells, demonstrated the continued presence of viral antigens and suggested that the response was local immunoproliferation in the tissues (Accinni et d., 1978).Experimental evidence for this point came from experiments in which the passive transfer of anti-LCMV antibody into persistently infected mice led to sequential infiltration by polyniorphonuclear leukocytes, plasma cells, and lymphocytes (Oldstone and Dixon, 1972). Immune complexes consisting of antiviral antibody and infectious virus are present in the sera of virus carrier mice (Oldstone and Dixon, 1969). Considering the size of the virus particle, averaging approximately 110 nni (Murphy et d.,1969, 1970), it is unlikely that this is the complex that deposits and initiates tissue injury. In an effort to define smaller complexes in the serum, we have analyzed such plasma by analytical ultracentrifugation (Oldstone, 1975). Sera from LCMVcarrier mice contained heterogeneous complexes ranging from 9 S to 17 S. In contrast, sera from mice persistently infected with lactate dehydrogenase virus, which have high concentrations of antiviral antibody and infectious virus but do not develop severe immune complex disease (Notkins et a/., 1966; Porter and Porter, 1971; Oldstone and Dixon, 1971a), had only a homogeneous 11 S species of immune complexes. Chronic LCMV-induced immune complex disease has great potential value as a model system in which to study the mechanisms and
320
M . J. RUCHMEIER ET AL.
FIG.17. Some histopathologic features of chronic lymphocytic choriomeningitis virus (LCMV) infection. (A) Glomerular injury: capillary occlusion by deposition of amorphous periodic acid-Schiff-positive material in the kidney of a 6-month-old SWWJ-carrier mouse. (B) Focal necrosis and round cell infiltration in the liver of a 10-month-old LCMV-carrier B10D2 old mouse. (C) Interstitial round cell infiltrations in the lung of a 10-week-old LCMV-carrier SWWJ mouse. (D) An area of necrosis and inflammatory cell infiltration in the myocardium of a 6-week-old virus-carrier SWWJ mouse. (E) Arteritis in a renal artery of a 14-day-old virus-carrier SWWJ mouse. Necrosis and infiltration with inflammatory cells are seen. Persistence of viral antigen in neurons of the hypothalamus (F) may be responsible for subtle neurologic dysfunctions in the viruscarrier mouse. Viral antigen was demonstrated in the brain of a 6-month-old SWWJ mouse using a fluorescein-conjugated antibody to LCMV.
VIROLOGY AND IMMUNOBIOLOCY O F
LCMV INFECTION
321
kinetics of immune complex formation and deposition as well as the precise molecular composition of such complexes. We have recently studied the development of circulating immune complexes after neonatal infection with LCMV (Table IV) (Oldstone et al., 1980).According to a Clq binding assay, immune complexes appeared in the sera of neonatally infected SWWJ mice by 4 weeks of age and reached peak levels by 8-16 weeks. BALB/WEHI mice, which develop less severe glomerulonephritis, did not develop high titers of C l q binding complexes. Like IgG recovered from the kidneys of virus-carrier mice (Buchnieier and Oldstone, 1978b), antibody found in the sera of both SWWJ and BALBNEHI mice precipitated all the LCMV polypeptides. However, virus carrier SWWJ mice contained approximately fourfold more antibody to LCMV than did comparable BALB/WEHI mice. Further studies are needed to analyze critically the role of antibody concentration, subclass, and affinity (Steward et al., 1975) in the development of immune complex disease during chronic virus infections. B. T CELL ACTIVATION AND FUNCTION IN PERSISTENT DISEASE Lymphocytic choriomeningitis virus is continuously present in mice it persistently infects, yet persistently infected mice are not tolerant at the B cell level or the T helper cell level, as deduced from the active B cell response. Cytotoxic T cells generated in acute LCMV infection are associated with the clearing of virus i n vivo (Zinkernagel and Welsh, 1976). Therefore the evaluation of cytotoxic T cell activity in persistently infected mice is of value. However, the results and interpretation of experimental findings are still not clear, and thus are controversial. Most groups of investigators looking for evidence ofcytotoxic T cells in virus carrier mice have either not found them at all (Cole et nl., 1973; Cihak and Lehmann-Grube, 1978) or found them only inconsistently and at low levels of activity (Oldstone and Dixon, 1971b). This observation has been interpreted by some (Cihak and LehmannGrube, 1978; Cole, 1977;Thomsen et u l . , 1979)as evidence that clones of cytotoxic T cells capable of responding to LCMV antigens are eliminated. The idea that immunologic tolerance to LCMV exists in the persistent infection is not a new one. Persistent LCMV infection was one of the models upon which Burnet and Fenner (1949) based the clonal selection theory of antibody fomiation. More recent interpretations of this idea as it relates to LCMV infection (Cole, 1977; Cihak and Lehmann-Grube, 1978) suggest that a critical time exists in the maturation of T lymphocyte function in the neonate. If LCMV is
TABLE IV c l q BINDINGMATERIALIN THE SERA OF MICE PERSISTENTLY INFECTED WITH LYMPHOCYTIC CHORIOMENINGITTS VIRUS Percent Clq bound (age of mouse: weeks) Group
Strain
Sex
4
8
12
16
20
24
Virus carrier Virus carrier Virus carrier Virus carrier Uninfected Uninfected Uninfected Uninfected
SWWJ SWWJ BALBNEHI BALBWHI S W J SWWJ BALBWHI BALBWHI
F M F
12.3 f 2.0a 6.7 c 1.3 NDb ND 2.2 2 0.1 2.0 f 0.3 ND ND
63.0 c 8.9 52.6 c 5.0 13.0 t- 1.8 8.6 f 1.2 3.6 c 0.2 3.1 c 0.4 3.6 & 0.3 5.0 c 1.3
58.8 f 6.5 40.7 f 5.3 8.2 f 1.4 8.3 2 1.2 3.9 c 0.7 3.1 f 0.6 8.6 f 1.1 6.4 2 0.5
54.9 ? 5.1 32.5 f 8.4 9.0 f 3.2 5.6 2 0.7 4.2 2 0.7 3.5 f 0.4 6.9 4 1.7 4.2 t- 0.3
51.3 5 11 21.9 f 8.0 10.5 t- 3.8 4.8 c 0.3 ND ND 5.6 2 0.8 4.4 c 0.1
43.3 k 7.0 ND 9.2 c 5.7 4.2 2 0.7 ND ND 3.1 2 0.3 3.3 f 0.2
M
F M F M
Mean f standard error. Minimum of 15 mice in each group. ND. not determined.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV INFECTION
323
introduced during this time span (i.e., <24 hours after birth), T cell clones destined to become cytotoxic to LCMV, by virtue of their ability to recognize H-2 plus viral antigen, are aborted. If virus is introduced after this critical period, these clones are capable of responding and the characteristic acute disease ensues. The effect on this cytotoxic T cell response is known to be LCMV specific, because neonatally infected LCMV carrier mice challenged with other arenaviruses can mount normal cytotoxic T cell responses to them (M. J. Buchmeier, unpublished data). Based on the data now available, several problems exist with this interpretation. We now know that persistently infected mice make IgG antibody to all the known viral antigens in virions (Buchmeier and Oldstone, 1978b; Oldstone et al., 1980)and on the surfaces of infected cells (M. V. Doyle and M. J. Buchmeier, unpublished data) and that the synthesis of this antibody is dependent upon T cell help. Neonatally infected, virus-carrier nude mice have little if any detectable LCMV specific precipitating antibody (Oldstone et a1., 1980). Further, adult nude (athymic) mice acutely infected with LCMV make only a transient IgM antibody response to virus (M. J. Buchmeier and M. V. Doyle, unpublished data), whereas their thymus sufficient littermates produce both IgM and IgG anti LCMV antibodies. To accommodate these observations in the clonal elimination theory of LCMV persistence, one would have to propose a mechanism by which clones of T cells destined to become cytotoxic cells would be aborted while clones of B cells and clones of T cells programmed to become helper cells for antibody forniation would be spared. The inability to demonstrate cytotoxic T cells consistently may result from factors other than clonal elimination of these cells. Given the precise kinetics with which the cytotoxic T cell response develops after acute infection, one might predict that similarly restricted kinetics occur in persistently infected mice. Hence, kinetic studies should be done in the virus carrier before the presence of a cytotoxic T cell response can be ruled out. Toward this end we have made attempts to generate a secondary cytotoxic T cell response against LCMV-infected targets. We have utilized the in uitro cytotoxic cell generation system described b y Dunlop and co-workers (Dunlop and Blanden, 1976; Dunlop et a]., 1976). Using as sensitizing cells H-2 compatible fibroblasts either infected with LCMV or uninfected, and as responder cells memory spleen cells from mice immunized 30-60 days earlier with LCMV, we were able to generate an in vitro secondary response similar in kinetics and magnitude to that described by Dunlop. But with memory cells taken from LCMV-carrier mice, we could not demon-
324
M. J. BUCHMEIER E T AL.
strate such a response. Secondary stimulation of cytotoxic T cells in vivo has yielded interesting information regarding the specificity of unresponsiveness among LCMV-carrier mice (Fig. 18). Normal mice respond to infection with either LCMV or Pichinde virus by generating a cytotoxic T cell response. When LCMV-carrier mice were secondarily infected with each virus no secondary cytotoxic response was measurable to LCMV; however, the response generated against Pichinde virus was essentially identical to that observed in normal mice. This indicates that the block in expression of anti-LCMV cytotoxicity is antigen specific. Alternative mechanisms to explain the lack of cytotoxic T cell activity in virus-carrier mice have been suggested. Zinkernagel and Doherty (1974b) theorized that a state of active suppression of virusspecific cytotoxicity exists in carrier mice. By adoptively transferring allogeneic spleen cells into virus-carrier recipients, LCMV specific T
'1 2
1
.-
5
~
L,,,,I-; P'.
20 40 5 20 Lymphocyts/Targst Ratio
40
FIG. 18. Killing of lymphocytic choriomeningitis virus (LCMV) and Pichinde virus (PV) infected target cells by cytotoxic T lymphocytes. Panels A and B: normal C3H/St mice inoculated with 1000 PFU of LCMV (A) or PV (B) and cytotoxic T cell activity tested 8 days later on LCMV-infected ( L O , A-A) or PV infected (O---O, A - - - - A )target cells. Virus specificity of the cytotoxic T cell response is seen. Panels C and D: LCMV carrier mice inoculated with 1000 PFU of LCMV (C) or PV (D) and assayed as above. N o LCMV specific response is seen in (C); however, the killing of PV-infected targets by cytotoxic T cells from LCMV-carrier mice immunized against PV (D) is essentially identical to that seen with normal mice (B). These responses could be abrogated by anti-theta antiserum and complement pretreatment of the effector cells (M. J. Buchnieier, R. Ott, and M. B. A. Oldstone, unpublished observations).
VIROLOGY AND IMMUNOBIOLOGY OF LCMV INFECTION
325
cell cytotoxicity was generated. The kinetics of this induction showed a peak response 9 days after the cell transfer. The interpretation offered was that allogeneic transfer effectively broke the state of active suppression in carrier mice in a manner similar to that observed in rats (McCullagh, 1972). Also suggestive of a peripheral mechanism for immune unresponsiveness of carrier T cells are experiments by G. A. Cole (personal communication) and by M. V. Doyle et al. (unpublished results) in which concanavalin A treatment of spleen cells harvested from mice persistently infected with LCMV seemed to generate LCMV-specific cytotoxic T cells. These experiments are preliminary and must be better defined and confirmed. It is of interest that Dunlop and Blanden (1977) showed that cytotoxic T cells specific for LCMV and generated during acute infection were suppressed by a variety of approaches, all of which had in common the addition of either LCMV or LCMV-infected cells. Obviously more data are needed in this area to account for virus persistence and the absence of detectable T cell-mediated cytotoxicity. V. Epilogue
The study of LCMV and the infection it causes has led to several novel findings in experimental biology and is expected to lead to more in the coming years. This is predictable because LCMV is among the best models available for the study of acute and persistent infections and related disease in a natural host. In addition to the elegant in vivo models, there are quantitative assays for measuring both viral and immune responses in uitro. Further, there is a large group of investigators in many countries actively pursuing different aspects of this work. With the rapid advance of technology in biochemistry, molecular biology, and viral genetics it is likely that in the next decade the molecular definition of the pathology of LCMV infection will appear. Indeed, several laboratories are currently looking at the genes and gene products that control infection, immune responses, and disease. Other interest is focused on the nature of the cytotoxic T cell receptor, the structural recognition units observed by cytotoxic T cells, factors that regulate them, strain susceptibility, and tissue or organ tropism. In addition, chemical characterization of circulating immune complexes, comparison of circulating and deposited complexes, determinations of why complexes are deposited from the circulation and what might be the mechanism of in vivo viral persistence and in vitro regulation of cell function are being actively pursued. Yet in spite of the advances
326
M. J. BUCHMEIER E T AL.
made in this field and the number of publications, the initial events recorded by Eric Traub in the beginning are still not entirely clear. That is, how does a virus persist throughout the life-span of an animal that is infected in utero or at birth with LCMV?
ACKNOWLEDGMENTS This is Publication No. 2031 from the Department of Immunopathology, Scripps Clinic and Research Foundation, La Jolla, California 92037. The authors thank the many colleagues both at this institution and elsewhere who have contributed to work in this area. We especially wish to acknowledge the ongoing work on nucleic acids with Ian Kennedy, Department of Biology, University of California at San Diego, and with electron microscopy by Peter Lampert, Department of Pathology, University of California at San Diego. We acknowledge the technical aid of Antoinette Tishon, Hanna Lewicki, Ruth Ott, Tania Popov, and Linda Hallenbeck and manuscript preparation by Laura Taxel. This work has been supported by U.S. Public Health National Institute grants A1 09484, A1 16102, A1 12438, and NS 12428. Michael Buchmeier is an established investigator of the American Heart Association, and Raymond Welsh possesses an RCDA A1 00253.
REFERENCES Accinni, L., Archetti, I., Branca, K., Hsu, C., and Andres, G . (1978).Clin. Immunol. lmmunopathol. 11,395. Akermann, R., Stille, W., Blumenthal, W., Helm, E. B., Keller, K., and Baldus, 0. (1972). Dsch. Med. Wochenschr. 97, 1725. Armstrong, C., and Lillie, R. D. (1934).Public Health Rep. 49, 1019. Armstrong, C. R. (1940-1941). Harvey Lect. 39. Armstrong, D., Fortner, J. G . , Rowe, W. P., and Parker, J. C. (1969).]. Am. Med. Assoc. 209, 265. Baum, S. G., Lewis, A. M., Rowe, W. P., and Huebner, R. J. (1966).N . Engl. J . Med. 274, 934. Benson, L., and Hotchin, J. (1969).Nature (London) 222, 1045. Blanden, R. V., and Mims, C. A. (1973).Aust. J . E x p . B w l . Med. Sci. 51, 393. Bro-Jorgensen, K. (1971).Acta Pathol. Mlcrobiol. Scund., Sect. B 79, 466. Bro-Jorgensen, K. (1978).Ado. Virus Res. 22,327. Bro-Jorgensen, K., and Knudtzon, S. (1977).Blood 49,47. Bro-Jorgensen, K., and Volkert, M. (1972).Acta Pathol. Microbwl. Scand., Sect. B 845. Bro-Jorgensen, K., and Volkert, M. (1974).Infect. Immun. 9, 605. Buchmeier, M. J., and Oldstone, M. B. A. (1978a).In “Negative Strand Viruses and the Host Cell” (B. W. J. Mahy and R. D. Barry, eds.), p. 91. Academic Press, New York. Buchmeier, M. J., and Oldstone, M. B. A. (1978b).J.Immunol. 120, 1297. Buchmeier, M. J., and Oldstone, M. B. A. (1979).Virology 99, 111. Buchmeier, M. J., Gee, S. R., and Rawls, W. E. (1977).J.Virol. 22, 175. Buchmeier, M. J., Elder, J. H., and Oldstone, M. B. A. (1978).Virology 89, 133. Buckley, S. M., and Casals, J. (1970).Am. J . Trop. Med. Hyg. 19, 680. Burnet, F. M., and Fenner, F. (1949). “The Production ofAntibodies.” Macmillan, New York.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV
INFECTION
327
Carter, M. F., Murphy, F. A., Brunschweig, J. P., Noonan, C., and Rawls, W. E. (1973).J. Virol. 12, 33. Carter, M. F., Biswal, N., and Rawls, W. E. (1974).J . Virol. 13, 577. Casals, J. (1976). I n “Viral Infections of Humans-Epidemiology and Control” (A. S. Evans, ed.), Chapter 5, p. 103. Plenum, New York. Chester, J. J., Paucker, K., and Merigan, T. C. (1973).Noture (London) 246, 92. Cihak, J., and Lehmann-Grube, F. (1978). Imniuiiology 34,265. Cole, G. A. (1977).In “Microbiology” (D. Schlessinger, ed.), p. 494. Am. Soc. Microbwl. Washington, D.C. Cole, G. A., and Nathanson, N. (1974).Prog. Metl. Virol. 18, 94. Cole, G. A., Gilden, D. J., Monjan, A. A., and Nathanson, N. (1971).Fed. Proc., Fed. A m . Soc. E x p . B i d . 30, 1831. Cole, G. A., Nathanson, N., and Prendergast, R. A. (1972).Noture (London) 238, 335. Cole, G. A., Prendergast, P. A., and Henney, C. S. (1973). In “Lymphocytic Choriomeningitis Virus and Other Arenaviruses” (F. Lehmann-Grube, ed.), p. 60. Springer-Verlag, Berlin and New York. Dalton, A. J., Rowe, W. P., Smith, G. H., Wilsnack, R. E., and Pugh, W. E. (1968).J.Virol. 2, 1465. IleMaeyer, E. (1976).]. Irifect. Dis. 133, A36. Djeu, J. Y., Heinbaugh, J. A., Holden, H. T., and Herberman, R. B. (1979).J.Imrnuriol. 122. 175. Doherty’, P. C., and Zinkemagel, R. M. (1974).Trcirtsplant. Rev. 19, 89. Doyle, M. V., and Oldstone, M. B. A. (1978).J.Zmmunol. 121, 1262. Dunlop, M. B. C., and Blanden, R. V. (1976). Ininzunology 31, 171. IlunIop, M. B. C., and Blanden, R. V. (1977).J.E x p . Med. 145, 1131. Dunlop, M. B. C., Doherty, P. C., Zinkemagel, R. M., and Blanden, R. V. (1976). Imniuriology 31, 181. Dutko, F. J., and Pfau, C. J. (1978).J . Geti. Virol. 38, 195. Dutko, F. J., Wright, E. A., and Pfau, C. J. (1976).J.Gen. Virol. 31, 417. East, J. D., Parrott, M. V., and Seamer, J. (1964). Virology 22, 160. Elder, J. H., Jensen, F. C., Bryant, M. L., and Lemer, R. A. (1977).Nrrture (London)267, 23. Farber, F. E., and Rawls, W. E. (1975).]. Gen. Virol. 26, 21. Fanner, T. W., and Janeway, C. A. (1942). Medicine (Baltimore) 21, 1. Findlay, G. M., and Stem, R. 0. (1936).J . Pathol. Bacteriol. 43, 327. Fujinami, R. S., and Oldstone, M. B. A. (1979).Nature (London) 279, 529. Card, G. P., Vezza, A. C., Bishop, D. H. L., and Compans, R. W. (1977).Virology 83,84. Gilden, D. H., Cole, G. A., Monjan, A. A., and Nathanson, N. (1972a).J.E x p . Med. 135, 860. Gilden, D. H., Cole, G. A., and Nathanson, N. (1972b).J.E x p . Med. 135, 874. Gresser, I., Morel-Maroger, L., Verroust, P., Riviere, Y., and Guillon, J. C. (1978). Proc. N n t f . Acnd. Sci. U . S. A. 75, 3413. Gschwender, H. H., Rutter, G., and Lehmann-Grube, F. (1976). Med. Microbiol. I t n munol. 162, 119. Haas, V. H., and Stewart, S. E. (1956).Virology 2, 511. Halbach, M., and Koschel, K. (1979).J . Gen. Virol. 42, 615. Hansson, M., Kiessling R., and Welsh, R. (1980). Iri “Natural Cell-Mediated Immunity Against Tumors” (R. B. Herbeman, ed.), Academic Press, New York (in press). Herberman, R. B., Nunn, M. E., and Lavrin, D. H. (1975). Z n t . J . Caucer 16,216. Herbeman, R. B., Nunn, M. E., and Holden, H. J. (1978).]. Immutiol. 121, 304.
328
M. J. BUCHMEIER E T AL.
Hinman. A. R., Fraser, D. W., Douglas, R. G., Jr., Bowen, G. S., Kraus, A. L., Winkler, W. G., and Rhodes, W. W. (1975).Am. 1. Epidemwl. 191,103. Hirsch, M. S., Murphy, F. A., Russe, H. P., and Hicklin, M. D. (1967). Proc. Soc. Exp. Biol. Med. 125, 980. Hirsch, M. S., Murphy, F. A., and Hicklin, M. D. (1968).J. E x p . Med. 127, 757. Holland, J. J., Graban, E., Jones, C., and Sernler, B. (1979). Cell 16, 495. Holterman, 0. A., and Majde, J. A. (1971). Transplantation 11,20. Hotchin, J. (1962). Cold Spring Harbor Symp. Quant. B w l . 27,479. Hotchin, J. (1971).Monogr. Virol. 3, 1. Hotchin, J. E., and Cinits, M. (1958). Can. J . Microbiol. 4, 149. Hotchin, J., and Seegal, R. (1977). Science 196, 671. Hotchin, J., and Weigand, H. (1961).J.Zmmunol. 86,392. Hotchin, J., Kinch, W., and Benson, L. (1971).Infect. Immun. 4, 281. Huang, A. S. (1973).Annu. Reu. Microbiol. 27, 101. Jacobs, R. P., and Cole, G. A. (1976).J.Inimunol. 117, 1004. Jacobson, S., and Pfau, C. J. (1980).Nature (London) 283,311. Jacobson, S., Dutko, F. J., and Pfau, C. J. (1979).J.Gen. Virol. 44, 113. Johnson, E. D., Monjan, A. A., and Morse, H. C., 111 (1978). Cell. Zmmunol. 36, 143. Johnson, K. M., Mackenzie, R. B., Webb, P. A., and Kuns, M. L. (1965). Science 150, 1618. Johnson, K. M., Webb, K. A., and Justines, G. (1973).In “Lymphocytic Choriomeningitis Virus and Other Arenaviruses” (F. Lehmann-Grube, ed.), p. 241. Springer-Verlag, Berlin and New York. Joseph, B. S., and Oldstone, M. B. A. (1975).J.E x p . Med. 142, 864. Kiessling, R., Klein, E., and Wigzell, H. (1975). Eur. J. Immunol. 5, 112. Kiessling, R., Hochrnan, P. S., Haller, O., Shearer, G. M., Wigzell, H., and Cudkowicz, G. (1977). Eur. J. Zmmunol. 7,655. Kiley, M. P., Lange, J. V., and Johnson, K. M. (1979). Lancet 6, 738. Koschel, K., and Halbach, M. (1979).J.Gen. Virol. 42, 627. Lehmann-Grube, F. (1967).Nature (London) 213, 770. Lehrnann-Grube, F. (1971).Virol. Monogr. 10, 1. Lehmann-Grube, F., Slenczka, W., and Tees, R. (1969).J.Gen. Virol. 5,63. Lehmann-Grube, F., Popescu, M., Schaefer, H., and Gschwender, H. H. (1975).Bull. W. H . 0 . 52,443. Leung, W.-C., and Rawls, W. E. (1977). Virology 81, 174. Leung, W.-C., Chosh, H. P., and Rawls, W. E. (1977).J.Virol. 22,235. Leung, W.-C., Leung, M., and Rawls, W. E. (1979).J.Virol. 30,98. Lewis, A. M., Rowe, W. P., Tumer, H. C., and Huebner, R. J. (1965).Science 150,363. Lillie, R. D., and Armstrong, C. (1945).Arch. Pathol. 40, 141. Lundstedt, C. (1969).Acta Pathol. Microbiol. Scand. 75, 139. Lundstedt, C., and Volkert, M. (1967). Acta Pathol. Microbiol. Scand. 71,471. McCullagh, P. (1972). Transplant. Rev. 12, 180. McFarland, H. (1974).J.Zmmunol. 113, 173. McMaster, G. K., and Carmichael, G. G. (1977).Proc. Natl. Acad. Sci. U.S . A. 74,4835. Maiztegui, J. I., Laquens, R. P., Cassio, P. M., Casanova, M. B., d e la Vega, M. T., Ritacco, V., Segal, A., Femandez, N. J., and Arana, R. M. (1975).J.Infect. Dis. 132, 35. Mannweiler, K., and Lehmann-Grube, F. (1973). In “Lyrnphocytic Choriomeningitis Virus and Other Arenaviruses” (F. Lehmann-Grube, ed.), p. 37. Springer-Verlag, Berlin and New York. Marker, O., and Volkert, M. (1973).J.E x p . Med. 137, 1511.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV
INFECTION
329
Merigan, T. C., Oldstone, M. B. A,, and Welsh, R. M. (1977). Nature (London) 268,67. Mims, C. A. (1966).J. Ptithol. Bacteriol. 91, 395. Mims, C. A., and Blanden, R. V. (1972). Znfect. Immun. 6, 695. Mirns, C. A., and Subrahmanyan, T. P. (1966).]. Ptrthol. Bacteriol. 91, 403. Minis, C. A., and Wainwright, S. (1968).J. Immunol. 101, 717. Murphy, F. A., and Whitfield, S. G. (1975). Bull. W. H . 0 . 52, 409. Murphy, F. A,, Webb, P. A., Johnson, K. M., and Whitfield, S . G. (1969).J. Virol. 4,535. Murphy, F. A., Webb, P. A., Johnson, K. M., Whitfield, S. G., and Chappell, W. A. (1970). J . Virol. 6, 507. Murphy, F. A,, Winn, W. C., Jr., Walker, P. H., Fleniister, M . R., and Whitfield, S. G. (1976). Lab. Incest. 34, 125. Murphy, F. A., Buchnieier, M. J., and Rawls, W. E. (1977). Luh. Inuest. 37, 502. Nathanson, N., Monjan, A. H., Panitch, H. S., Johnson, E. D., Peturssen, G., and Cole, G. A. (1975). I n “Viral Immunology and lmn~unopathology”(A. L. Notkins, ed.), p. 357. Academic Press, New York. Notkins, A. L., Mergenhagen, S . E., Rizzo, A. A,, Scheele, C., and Waldmann, T. A. (1966).J . E x p . Med. 123, 347. Oldstone, M. B. A. (1975). Prog. Med. Virol. 19, 84. Oldstone, M. B. A. (1979); Compr. Virol. 15, I. Oldstone, M. B. A,, and Dixon, F. J. (1969).J. E x p . Med. 129, 483. Oldstone, M. B. A,, and Dixon, F. J. (197O).J. E x p . Med. 131, 1. Oldstone, M. B. A., and Dixon, F. J. (1971a).J. E l p . Med. 134, 32s. Oldstone, M. €3. A., and Dixon, F. J. (1971b). Immunopathol., Znt. Symp., 6th, 1970, p. 391. Oldstone, M. B. A., and Dixon, F. J. (1972).J. E x p . Med. 135, 827. Oldstone, M. B. A,, and Peters, C. J. (1978). Hand. Clin. Neurol. 34, 193. Oldstone, M. B. A., Habel, K., and Dixon, F. J. (1969). Fed. Proc., Fed. Am. Soc. E x p . B i d . 28, 429. Oldstone, M . B. A,, Tishon, A., Chiller, J., Weigle, W., and Dixon, F. J. (1973).J. I m munol. 110, 1268. Oldstone, M. B. A,, Del Villano, B. C., and Dixon, F. J. (1976).J. Virol. 18, 176. Oldstone, M. B. A., Holnistoen, J., and Welsh, R. M., Jr. (1977).J. Cell. Physiol. 91,459. Oldstone, M. B. A., Buchmeier, M. J., Doyle, M. V., and Tishon, A. (1980).]. Immunol. 124, 831. Owen, R. D. (1945). Science 102, 400. Padnos, hl., Shimonaski, G., and Came, P. E. (1971).J. Cen. Virol. 13, 163. Palmer, E. L., Obijeski, J. F., Webb, B. A., and Johnson, K. M. (1977).J. Gen. Virol. 36, 541. Pederson, I. R. (1971). Nature (London) 234, 112. Pederson, 1. R. (1973).J. Virol. 11, 416. Pederson, I. R., and Konigshofer, E. R. (1976).J . Virol. 20, 14. Perrault, J., and Holland, J. (1972). Virologc~50, 148. Pfau, C. J . (1974). Prog. Med. Virol. 18, 64. Pfau, C. J., Welsh, R. M., and Trowbridge, R. S. (1973). In “Lymphocytic Choriomeningitis and Other Arenavimses” (F. Lehmann-Grube, ed.), p. 101. Springer-Verlag, Berlin and New York. Popescu, M. (1977). “Proc. Symp. Eur. Assoc. Against Virus Dis.,” 16th 1977, p. 50. Popescu, M., and Lehmann-Grube, F. (1976).J. Cen. Virol. 30, 113. Popescu, M., and Lehmann-Gmbe, F. (1977). Virology 77, 78. Popescu, M., Schaefer, H., and Lehmann-Gmbe, F. (1976).J. Virol. 20, 1. Popescu, M., Lohler, J . , and Lehmann-Grube, F. (1979).J. Cen. Virol. 42,481.
330
M. J. BUCHMEIER ET AL.
Porter, D., and Porter, H. (1971).J.Zmmunol. 106, 1264. Pulkkinen, A. J., and Pfau, C. J. (1970).Appl. Microbiol. 20, 123. Ramos, B. A., Courtney, R. J., and Rawls, W. F. (1972).J.Virol. 10,661. Rawls. W. E., and Leung, W.-C. (1980). Cornpr. Virol. 14, 157. Rawls, W. E., Banerjee, S. N., McMillan, C. A., and Buchmeier, M. J. (1976).J.Gen. Virol. 33, 421. Rivers, T. M., and Scott, T. F. M. (1936).]. Exp. Med. 63, 415. Riviere, Y., and Bandu, M. T. (1977).Ann. Microbiol. (Paris)128a, 323. Rowe, W. P. (1954).Research Report N M 005048.14.01.Nav. Med. Res. Inst., Bethesda, Maryland. Rowe, W. P. (1956). Proc. Soc. Exp. Biol. Med. 92, 194. Rowe, W. P., Black, P. H., and Levey, R. H. (1963).Proc. Soc. Exp. Biol. Med. 114,248. Rowe, W. P., Pugh, W. F., Webb, P. A., and Peters, C. J. (1970a).J. Virol. 5, 289. Rowe, W. P., Murphy, F. A., Bergold, G. H., Casals, J., Hotchin, J., Johnson, K. M., Lehmann-Grube, F., Mims, C. A., Traub, E., and Webb, P. A. (1970b).J.Virol. 5, 651. Rutter, G., and Gschwender, H. H. (1973).In “Lymphocytic Choriomeningitis Virus and Other Arenaviruses” (F. Lehmann-Grube, ed.), p. 51. Springer-Verlag. Berlin and New York. Saleh, F., Card, G. P., and Compans, R. W. (1979). Virology 93, 369. Schwartz, R., Lohler, J., and Lehmann-Grube, F. (1978).J.Gen. Virol. 39,565. Scott, T. F. M., and Rivers, T. M. (1936).J . Exp. Med. 63, 397. Sedwick, W. D., and Wiktor, T. J. (1967).J.Virol. 1, 1244. Silbeman, S. L., Jacobs, R. P., and Cole, G. A. (1978). Infect. Zmmun. 19,533. Smadel, J. E., and Wall, M. J. (1940).J. Exp. Med. 72, 389. Smadel, J. E., Baird, R. D., and Wall, M. J. (1939).J.Exp. Med. 7 0 , s . Smadel, J. E., Wall, M. J., and Baird, R. D. (1940).J.Exp. Med. 71,43. Staneck, L. D., Trowbridge, R. S., Welsh, R. M., Wright, E. A., and Pfau, C. J. (1972). Infect. Zmmun. 6,444. Steward, M. W., Katz, F. E., and West, N. J. (1975). Clin. Exp. Zmmunol. 21, 121. Stewart, W. E. (1979). “The Interferon System.” Springer-Verlag, Berlin and New York. Theofilopoulos, A. N., Wilson, C. B., Bokisch, V. A,, and Dixon, F. J. (1974).J.Exp. Med. 140, 1230. Thomsen, A. R., Volkert, M., and Marker, 0. (1979). Acta Pathol. Microbiol. Scand., Sect. C 87,47. Traub, E. (1935). Science 81, 298. Traub, E. (1936).]. Exp. Med. 63, 847. Traub, E. (1964).Arch. Gesamte Virusforsch. 14, 65. Traub, E., and Kesting, F. (1963).Arch. Gesamte Virusforsch. 14,55. Vanze, B. E., Douglas, R. G., Betts, R. F., Bauman, A. W., Fraser, D. W., and Hinman, A. R. (1975). Am. J. Med. 58, 803. Veltri, R. W., and Kirk, B. E. (1971).J. Gen. ViroZ. 10, 17. Vezza, A. C., Card, G. P., Compans, R. W., and Bishop, D. H. L. (1977).J.Virol. 23,776. Vezza, A. C., Clewley, J. P., Gard, G. P., Abraham, N. Z., Compans, R. W., and Bishop. D. H. L. (1978).J. Virol. 26, 485. Wagner, R. R., and Snyder, R. M. (1962).Nature (London) 196,393. Walker, D. H., Murphy, F. A., Whitfield, S. G . ,and Bauer, S. P. (1975).Exp. Mol. Puthol. 23, 245. Webb, P. A. (1965).J. Trop. Med. Hyg. 14, 799. Webster, J. M., and Kirk, B. E. (1974). Infect. Immun. 10,516.
VIROLOGY AND IMMUNOBIOLOGY OF
LCMV
INFECTION
331
Weissenbacher, M. C., Coto, C. E., and Calello, M. A. (1976). Znteruirology 6, 42. Welsh, R. M., and Buchmeier, M. J. (1979). Virology 96, 503. Welsh, R. M., and Kiessling, R. (1980a). In “Natural Cell Mediated Immunity Against Tumors” (R. B. Herberman, ed.). Academic Press, New York (in press). Welsh, R. M., and Kiessling, R. (1980b). Scand. J . Zmmunol. 11, 363. Welsh, R. M., and Oldstone, M. B. A. (1977).J. Exp. Med. 145, 1449. Welsh, R. M., and Pfau, C. J. (1972).J. Gen. Virol. 14, 177. Welsh, R. M., O’Connell, C. M., and Pfau, C. J. (1972).]. Gem Virol. 17,355. Welsh, R. M., Burner, P. A., Holland, J. J., Oldstone, M. B. A., Thompson, H. A,, and Villarreal, L. P. (1975). Bull. W. H. 0. 52, 403. Welsh, R. M., Lampert, P. W., Burner, P. A., and Oldstone, M. B. A. (1976). Virology 73, 59. Welsh, R. M., Lampert, P. W., and Oldstone, M. B. A. (1977). 1.Infect. Di.7. 136, 391. Welsh, R. M., Zinkernagel, R. M., and Hallenheck, L. A. (1979).J.Immunof. 122,475. Welsh, R. M. (1978).J. E x p . Med. 148, 163. Welsh, R. M., and Zinkemagel, R. M. (1977). Nature (London) 286, 646. Wilsnack, R. E., and Rowe, W. P. (1964).J . E x p . Med. 120, 829. World Health Organization (1975). Bull. W. H. 0. 52, Nos. 4,5,6. Youngner, J. S. (1977). In “Microbiology” (D. Schlessinger, ed.), p. 451. Am. sot. M icrobiol., Washington, D .C. Zinkernagel, R. M., and Doherty, P. C. (1973).j. Exp. Med. 138, 1266. Zinkernagel, R. M., and Doherty, P. C. (1974a). Nature (London) 251, 547. Zinkernagel, R. M., and Doherty, P. C. (197411). I n “Immunologic Tolerance: Mechanisms and Potential Therapeutic Applications” (D. H. Katz and B. Benacerraf, eds.), p. 403. Academic Press, New York. Zinkemagel, R. M., and Doherty, P. C. (1979). Adu. Zmmunol. 27,51. Zinkernagel, R. M., and Welsh, R. M. (1976).J. Immunol. 117, 1495.
This Page Intentionally Left Blank
Index A
Allergic encephalomyelitis, experimental, 224-251 Antibody anti-idiotype, contact sensitivity and, 131-133 contact sensitivity and, 128-130 lymphocyte activation and, 72-78 Arenaviruses, relationship with LCMV, 298-299 Autoimmunity experimental, 184-185 allergic encephalomyelitis, 224-251 thyroiditis, 185-224 induction by bypassing T cell tolerance, 182 suppressor cells in, 183-184
B Basophils, activation of, 92-100 R cell activation and function in acute LCMV infection, 311-313 in persistent LCMV disease, 313-321
rosetting, 15-16 spontaneous and lectin-induced lymphocyte aggregation, 12-15 plasma membrane function in activation of lymphocytes, 31-33 drugs affecting cytomusculature, 33-35 drugs affecting cytoskeleton, 35-40 other factors, 40-72 structural relationships with plasma membrane, 2-5 Contact sensitivity control of anti-idiotype antibodies and, 131-133 major histocompatibility complex and, 130-131 induction and elicitation afferent limb, 123-125 antibody and, 128-130 chemical requirements for, 122-123 efferent limb, 125-128 genetic control and, 128 tolerance to in absence of suppressor T cells, 135-138 general features, 133-135 suppressive mechanisms, 138-148 suppressor factors, 148-152 Cytomusculature, drugs affecting, 33-35 Cytoskeleton, drugs affecting, 35-40
C E
Cell(s) differentiated, LCMV interaction with, 299-302 nonlymphoid, activation of, 87-100 Cell cortex control of recognition phenomena that take place at the plasma membrane level killer-target interactions, 23-31 macrophage-lymphocyte interactions, 19-23 migration of lymphocytes from blood to lymphoid tissue, 16-19
Experimentally induced tolerance, relationship to self-tolerance, 178-179 activation of competent T cells, 182-183 induction of autoimmunity by bypassing T cell tolerance, 182 polyclonal activation, 179-182 suppressor cells in autoimmunity, 183184 F
Fibroblasts, activation of, 87
333
334
INDEX G
Genetic control, of contact sensitivity, 128 H
H-2 antigens, LCMV polypeptides and, 291-292 I
Immunologic tolerance, type of acquired central unresponsiveness, 167-178 peripheral inhibition, 162-167 Interferon, production, LCMV acute infection and, 303-308 K
Killer cell(s) activation and function, in acute LCMV infection and, 308-309 interaction with targets, 23-31 1
Lymphocytes activation by anti-Ig antibodies, 72-78 cell cortex and plasma membrane functions in activation of, 31-33 drugs affecting cytomusculature, 33-35 drugs affecting cytoskeleton, 35-40 other factors, 40-72 cell cortex control of recognition phenomena that take place at plasma membrane level killer-target interactions, 23-31 macrophage-lymphocyte interactions, 19-23 migration of lymphocytes from blood to lymphoid tissue, 16-19 rosetting, 15-16 spontaneous and lectin-induced lymphocyte aggregation, 12-15 to cluster but not to cap?, 78-87 microvilli formation and shedding, 8-12 structural relationships of plasma membrane with cell cortex, 2-5 uropode formation and capping, 5-8
Lymphocytic chotiomeningitis virus infection acute immune response disease B cell activation and function in, 311-313 cytotoxic T cell activation and function in, 310-311 generalized host responses and production of interferon, 303-308 macrophage activation and function in, 309-310 natural killer cell activation and function in, 308-309 pathology of, 302-303 chronic immune response disease B cell activation and function in, 313-321 T cell activation and function in, 32 1-325 clinical features of human illness and safety precautions, 277-279 historical features and experimental model system used for immunobiologic and immunopathologic research, 279-282 virus and host cell interactions interaction with differentiated celIs, 299-302 nonassociation of viral polypeptides and H-2 antigens, 291-292 relationship with other arenaviruses, 298-299 virion structure and organization, 282-288 virus mutants and variants, 292-297 virus replicative cycle, 288-291
M
Macrophage( s) activation and function, in acute LCMV infection, 309-310 activation of, 87-92 interaction with lymphocytes, 19-23 Major histocompatibility complex, contact sensitivity and, 130-131 Mast cells, activation of, 92-100 Microvilli, formation and shedding, 8-12
335
INDEX P
Plasma membrane cell cortex control of recognition phenomena at, 12-31 cell cortex function in activation of lymphocytes and, 31-72 structural relationships with cell cortex, 2-5 Polyclonal activation, tolerance and, 179-182 Polymorphonuclear neutrophils, activation of, 87-92
S Self-tolerance mechanism of, 161-162 relationship to experimentally induced tolerance, 178-179 activation of competent T cells, 182183 induction of autoimmunity hy bypassing T cell tolerance, 182
polyclonal activation, 179-182 suppressor cells in autoimmunity, 183-184 Suppressor cells, autoimmunity and, 183- 184 Suppressor T cells, contact sensitivity and, 135-138
T
T cell(s) activation and function, in persistent LCMV disease, 321-325 competent, activation of, 182-183 cytotoxic, activation and function in acute LCMV infection, 310-311 T cell tolerance, bypassing of, 182 Thyroiditis, experimental, 185-224
U
Uropode, formation and capping, 5-8
This Page Intentionally Left Blank
CONTENTS
OF PREVIOUS VOLUMES Antibody Production by Transferred Cells CHARLESG . COCHRANE AWD
Volume 1 Transplantation
Immunity and Tolerance
FRANK J. DIXON
M. HASEK, A. LENGEROVA,AND T. HRABA
Phagocytosis
DERRICKROWLEY
I tnmunological Tolerance of Nonliving Antigens
Antigen-Antibody Infections
RICHARD T. SMITH
Reactions i n Helminth
E. J . L. SOULSSY
Functions of the Complement System
ABRAHAMG. OSLER
Embryological Development of Antigens
REED A. FLICKINGER
I n Vitro Studies of the Antibody Response
ABRAM8.STAVITSKY
AUTHOR INDEX-SUBJECT INDEX
Duration of Immunity in Virus Diseases
J. H . HALE Volume 3
Fate and Biological Action of AntigenAntibody Complexes
I n Vitro Studies of the Mechanism of Anaphylaxis
WILLIAM 0. WEIGLE Delayed Hypersensitivity to Simple Protein Antigens P. G. H. GELLAND B. BENACERRAF
K. FRANK AUSTEN AND JOHN H. HUMPHREY The Role of Humoral Antibody in the Homograft Reaction
The Antigenic Structure of Tumors P. A. CORER
AUTHOR INDEX-SUBJECT INDEX
CHANDLER A. STETSON Immune Adherence
D. S. NELSON Reaginic Antibodies
D . R. STANWORTH
Volume 2 Immunologic Specificity a n d Molecular Structure
Nature of Retained Antigen a n d its Role i n Immune Mechanisms
DAN H. CAMPBELL AND JUSTINE S. CARVEY
FREDURUSH Heterogeneity of y-Globulins
JOHN L. FAHEY The Immunological Significance of the Thymus
J. F. A. P. MILLER, A. H . E. MARSHALL,A N D R. G. WHITE
Blood Groups i n Animals Other Than M a n
W. H. STONE AND M. R . IRWIN Heterophile Antigens a n d Their Significance i n the Host-Parasite Relationship
C . R. JENKIN
Cel Iu la r Genetics of Immune Responses
G. J . V. NOSSAL
AUTHOR INDEX-SUBJECT INDEX
337
338
CONTENTS OF PREVIOUS VOLUMES
Volume 4
Volume 6
Ontogeny a n d Phylogeny of Adaptive Immunity ROBERT A. G O O D AND BEN PAPERMASTER
Experimental Glomerulonephritis: Immunological Events a n d Pathogenetic Mechanisms EMILR. UNANUE AND
w.
Cellular Reactions i n Infection EMANUELSUTER AND HANSRUEDY RAMSEIER Ultrastructure of Immunologic Processes JOSEPH D. FELDMAN Cell W a l l Antigens of Gram-Positive Bacteria MACLYNMCCARTYAND
STEPHEN I. MORSE Structure and Biological Activity of Immunoglobulins
SYDNEYCOHENAND RODNEY R. PORTER Autoantibodies a n d Disease
H. G. KUNKEL AND E. M. TAN Effect of Bacteria a n d Bacterial Products on Antibody Response J. MUNOZ
AUTHOR INDEX-SUBJECTINDEX
FRANK J, DIXON Chemical Suppression of Adaptive Immunity ANN E. GABRIELSON AND ROBERT A. G O O D Nucleic Acids as Antigens O T T O J . PLESCIA AND WERNER BRAUN I n Vitro Studies of Immunological Responses o f Lymphoid Cells
RICHARDW. DUTTON Developmental Aspects of Immunity JAROSLAV STERZL AND ARTHUR M. SILVERSTEIN Anti-antibodies
PHILIP G. H. CELLAND S. KELUS
ANDREW
Conglutinin a n d lmrnunoconglutinins
P. J. LACHMANN AUTHOR
INDEX-SUBJECT
INDEX
Volume 5 Natural Antibodies a n d the Immune Response
STEPHENV. BOYDEN Immunological Studies with Synthetic Polypeptides
MICHAEL SELA Experimental Allergic Encephalomyelitis and Autoimmune Disease PHILIP Y. PATERSON The Immunology of Insulin C. G. POPE Tissue-Specific Antigens
D. C. DUMONDE AUTHOR INDEX-SUBJECT INDEX
Volume 7 Structure a n d Biological Properties of Immunoglobulins
SYDNEYCOHENAND CESAR MILSTEIN Genetics of Immunoglobulins in the Mouse
MICHAEL POTTERAND ROSE LIEBERMAN Mimetic Relationships between Group A Streptococci and Mammalian Tissues JOHN B. ZABIUSKIE Lymphocytes a n d Transplantation Immunity
DARCYB. WILSONAND n. E. BILLINGHAM
CONTENTS O F PREVIOUS VOLUMES Human Tissue Transplantation
339
Phylogeny of Immunoglobulins
HOWARDM. GREY
JOHN P . MERRILL AUTHOR INDEX-SUBJECT INDEX
Slow Reacting Substance of Anaphylaxis
ROBERT P. ORANGE AND K. FRANK AUSTEN Volume 8 Chemistry and Reaction Mechanisms of Complement
HANSJ. MULLER-EBERHARD Regulatory Effect of Antibody on the Immune Response
JONATHAN W. UHR GORAN MOLLER
Some Relationships among Hemostasis, Fibrinolytic Phenomena, Immunity, and the Inflammatory Response
OSCARD . RATNOFF Antigens of Virus-Induced Tumors
KARL HABEL
AND
Genetic and Antigenetic Aspects of Human Histocompatibility Systems
The Mechanism of Immunological Paralysis
D. W. DRESSERAND N . A. MITCHISON
AUTHOR INDEX-SUBJECT INDEX
In Vitro Studies of Human Reaginic Allergy
ABRAHAM G . OSLER, LAWRENCE M. LICHTENSTEIN, AND DAVIDA. LEVY AUTHOR INDEX-SUBJECT INDEX
Volume 1 1 Electron Microscopy of the Immunoglobulins
N . MICHAEL GREEN Genetic Control of Specific Immune Responses HUGH0. h4CDEVlTT AND
Volume 9
BARUJ BENACERRAF
Secretory lmmunog lobu lins
THOMAS B. TOMASI, JR., JOHN BIENENSTOCK
D . BERNARDAMOS
AND
Immunologic Tissue Injury Mediated by Neutrophilic Leukocytes
CHARLES G. COCHRANE The Structure a n d Function of Monocytes and Macrophages
ZANVILA. COHN The Immunology a n d Pathology of NZB Mice
J. B . HOWIE AND B . J . HELYER AUTHOR INDEX-SUBJECT INDEX
The Lesions i n Cell Membranes Caused by Complement
JOHN H. HUMPHREY AND ROBERT R. DOURMASHKIN Cytotoxic Effects of Lymphoid Cells In Vifro PETER PERLMANN AND GORAN HOLM Transfer Factor
H. S. LAWRENCE Immunological Aspects of M a l a r i a Infection
IVOR N. BROWN AUTHOR INDEX-SUBJECT INDEX
Volume 10 Cell Selection b y Antigen i n the Immune Response GREGORY W. SISKIND AND
BARUJBENACERRAF
Volume 12 The Search for Antibodies with Molecular Uniformity
RICHARD M . KRAUSE
340
CONTENTS OF PREVIOUS VOLUMES Volume 14
Structure and Function o f yM Macroglobulins
lmmunabiology of Mammalian Reproduction ALAN E. BEER AND
HENRYMETZCER Transplantation Antigens
R. A. REISFELT
AND
B. D. KAHAN
The Role of Bone Marrow in the Immune Response
NABIH I. ABDOU AND MAXWELL RICHTER Cell
Interaction
in
Antibody
R. E. BILLINCHAM Thyroid Antigens a n d Autoimmunity
SIDNEY SHULMAN Immunological Aspects of Burkitt's Lymphoma
Synthesis
D. W. TALMAGE,J. h D O V I C H , H. HEMMINGSEN
AND
The Role of Lysosomes in Immune Responses GERALD WEISSMANN AND
PETER DUKOR Molecular Size a n d Conformation of Immunoglobulins KEITH J. DORRINGTON AND
GEORGE KLEIN Genetic Aspects of the Complement System CHESTER A. ALPER AND
FREDS. ROSEN The Immune System: A Model for Differentiation in Higher Organisms L. HOODAND J. PRAHL
INDEX AUTHOR INDEX-SUBJECT
CHARLESTANFORD AUTHOR INDEX-SUBJECT INDEX
Volume 13 Structure and Function of Human Immunoglobulin E HANS BENNICH AND
S. GUNNAR 0. JOHANSSON Individual Antigenic Specificity of Immunoglobulins JOHN E. HOPPERAND
ALFRED NISONOFF In Vitro Approaches to the Mechanism o f Cell-Mediated Immune Reactions
BARRY R. BLOOM Immunological Phenomena i n Leprosy a n d Related Diseases J. L. TURK AND
A. D. M. BRYCESON Nature a n d Classification of ImmediateType Allergic Reactions
ELMERL. BECKER AUTHOR INDEX-SUBJECT INDEX
Volume 15 The Regulatory Influence of Activated T Cells on B Cell Responses to Antigen DAVIDH. KATZ AND
BARUJ BENACERRAF The Regulatory Role of Macrophages i n Antigenic Stimulation
E. R. UNANUE Immunological Enhancement: A Study of Blocking Antibodies
JOSEPH D. FELDMAN Genetics a n d Immunology of Sex-Linked Antigens DAVIDL. GASSER AND
WILLYS K. SILVERS Current Concepts of Amyloid
EDWARD C. FRANKLIN AND DOROTHEA ZUCKER-FRANKLIN AUTHOR INDEX-SUBJECT
INDEX
CONTENTS OF PREVIOUS VOLUMES Volume 16
341
Cell-Mediated Cytotoxicity, A l l o g r a f i Rejection, a n d Tumor Immunity
Human Immunoglobulins: Classes, Subclasses, Genetic Variants, a n d ldiotypes
J . €3. NATVIGA N D H. G. KUNKEL Immunologico I Unresponsiveness
WILLIAM0. WEICLE Participation of Lymphocytes in Viral Infections E. FREDERICK WHEELOCKAND
STEPHENT. TOY Immune Complex Diseases in Experimental Animals a n d M a n
C. G. COCHRANE AND D. KOFFLER The lrnmunopathology of Joint Inflammation in Rheumatoid Arthritis
JEAN-CHARLES CEROTTIhY AND K. THEODOREBRUNNER Antigenic Competition: A Review of Nonspecific Antigen-Induced Suppression
HUGHF. PROS AND DAVIDEIDINGER Effect of Antigen Binding on the Properties of Antibody
HENRYMETZGER Lymphocyte-Med iated Cytotoxicity a n d Blocking Serum Activity to Tumor Antigens KARL ERIK HELLSTROM AND
INGEGERDHELLSTROM AUTHOR INDEX-SUBJECT INDEX
NATHAN J. ZVAIFLER AUTHOR INDEX-SUBJECTINDEX Volume 19 Molecular Biology of Cellular Membranes with Applications to Immunology
Volume 17
S. J. SINGER
Antilymphocyte Serum
EUGENEM. LANCE,P. B. MEDAWAR, AND ROBERTN. TAUB I n Vitro Studies of Immunologically Induced Secretion of Mediators from Cells a n d Related Phenomena
ELMERL. BECKERA N D PETERM. HENSON Antibody Response to Viral Antigens
KEITH M. COWAN
Membrane Immunoglobulins a n d Antigen Receptors on B a n d T Lymphocytes
NOEL L. WARNER Receptors for Immune Complexes on Lymphocytes
VICTOR NUSSENZWEIG Biological Activities of Immunoglobulins of Different Classes a n d Subclasses
HANSL. SPIEGELBERC
Antibodies to Small Molecules: Biological a n d Clinical Applications \‘INCENT P. BUTLER,JR., AND
SUBJECTINDEX
SAMXI. BEISER AUTHOR INDEX-SUBJECTINDEX
Hypervariable Regions, Idiotypy, and Anti body-Combi ni n g Site
Volume 18 Genetic Determinants of Responsiveness
Volume 20
Immunological
DAVIDL. GASSERAND WILLYSK. SILVERS
J. DONALD CAPHA A N D 1. M~CHAEL KEHOE Structure and Function of the J Chain
MARIAN ELLIOTTKOSHLAND
342
CONTENTS O F PREVIOUS VOLUMES
Amino Acid Substitution and the Antigenicity of Globular Proteins
MORRIS REICHLIN The H-2 Major Histocompatibility Complex a n d the I immune Response Region: Genetic Variation, Function, and Organization
DONALDc. SHREFFLER AND CHELLAS. DAVID Delayed Hypersensitivity i n the Mouse
ALFRED J. CROWLE
SUBJECTINDEX
Cellular Aspects of Immunoglobulin A
MICHAEL E. LAMM Secretory Anti-Influenza Immunity YA. SHVARTSMAN AND
s.
M. P. ZYKOV SUBJECTINDEX Volume 23 Cellular Events in the IgE Antibody Response
KIMISHIGE ISHIZAKA Chemical a n d Biological Properties of Some Atopic Allergens
Volume 21 X-Roy Diffraction Studies of Immunoglobulins
ROBERTOJ. POLJAK Rabbit Immunoglobulin Allotypes: Structure, Immunology, a n d Genetics
THOMAS J. KINDT Cyclical Production of Antibody as a Regulatory Mechanism i n the Immune Response WILLIAM 0. WIECLE Thymus-Independent B-Cell Induction a n d Paralysis ANTONIO COUTINHO AND
GORANMOLLER SUBJECTINDEX
T. P. KING Human Mixed-Lymphocyte Culture Reaction: Genetics, Specificity, a n d Biological Implications Bo DUPONT,JOHN A. HANSEN, AND EDMOND J. YUNIS lmmunochemical Properties of Glycolipids a n d Phospholipids
DONALDM. MARCUS AND GERALDA. SCHWARTING SUBJECTINDEX Volume 24 The Alternative Pathway of Complement Activation 0. GOTZE AND H. J.
MULLER-EBERHARD Volume 22 The Role of Antibodies i n the Rejection a n d Enhancement of O r g a n Allografts
CHARLESB. CARPENTER, ANTHONY J. F. D’APICE,AND ABUL K. ABBAS Biosynthesis of Complement
HARVEY R. COLTEN Graft-versus-Host Reactions: A Review
STEPHEN c. GHEBEAND J. WAYNESTREILEIN
Membrane a n d Cytoplasmic Changes i n B lymphocytes Induced b y Ligand-Surface Immunoglobulin Interaction
GEORGEF. SCHREINERAND EMIL R. UNANUE Lymphocyte Receptors for Immunoglobulin
HOWARDB. DICKLER Ionizing Radiation a n d the Immune Response ROBERTE. ANDERSON AND
NOEL L. WARNER
SUBJECTINDEX
343
CONTENTS OF PREVIOUS VOLUMES Volume 25 Immunologically Privileged Sites
CLYDEF. BAKER AND K. E. BILLINCHAM Major Histocompatibility Complex Restricted Cell-Mediated Immunity
GENE M . SHEARER AND ANNE-MARIE SCHMITT-VERHULST Current Status of Rat lmmunogenetics
DAVIDL. GASSER Antigen-Binding Myeloma Proteins of Mice
MICHAEL POTTER Human lymphocyte Subpopulations
L. CHESSA N D S. F. SCHLOSSMAN
MHC-Restricted Cytotoxic T Cells: Studies on the Biological Role of Polymorphic Major Transplantation Antigens Determining T-cell Restriction-Specificity, Function, a n d Responsiveness
ROLF M. ZINKEHNACEL AND PETERC. DOHERTY Murine lymphocyte Surface Antigens
IAN F. C . MCKENZIE A N D TERRY POTTER
The Regulatory a n d Effector Roles of Eosinophils PETER F. WELLER AX11
EDWARDJ. GOETZL SUBJECTINUEX
SUBJECT INDEX Volume 28 Volume 26 Anaphylatoxins: C3a a n d C5a TONYE. kIUCLI A N D HANS
J . ~KJLLER-EBERHARD H-2 Mutations: Their Genetics a n d Effect on Immune Functions
JAN KLEIN The Protein Products of the Murine 17th Chromosome: Genetics a n d Structure
ELLENS. VITETTA AXD 1. DONALD CAPRA Expression a n d Function of ldiotypes on Lymphocytes
K. EICHMANN The B-Cell Clonotype Repertoire
NOLAN H. SIGALAND NORMAN n. KLINMAN SUBJECT INDEX
The Role of Antigen-Specific T Cell Factors in the Immune Response
TOMIO TADAA N D KO OKUMURA The Biology a n d Complexes
Detection of
Immune
ARCYH~OSN. THEOFILOPOULOS AND FRANK J. DIXON The Human l a System
R. J. WLNCHESTERA N D H. G. KUNKEL Bacterial Endotoxins a n d Host Immune Responses
c. h40RRISON L. RYAN
IIAVID JOHN
AND
Responses to Infection with Metazoan a n d Protozoan Parasites i n Mice
GRAHAM F. MITCHELL SUBJECT INDEX
Volume 27
Volume 29
Autoimmune Response to Acetylcholine Receptors in Myasthenia Gravis and I t s Animal Model
Molecular Biology and Chemistry of the Alternative Pathway o f Complement HANS J. MULLER-EBERHARD AND
JON LINDSTROM
ROBERT D. SCHREIBER
344
CONTENTS OF PREVIOUS VOLUMES
Mediators of Immunity: Lymphokines and Monokines
Ross E. ROCKLIN, KLAUS BENDTZEN,AND DIRKGREINEDER Adaptive Differentiation of Lymphocytes: Theoretical Implications for Mechanisms of Cell- Cell Recognition and Regulation of Immune Responses
DAVIDH . KATZ Anti body-Mediated Destruction of Virus-Infected Cells
J . G . PATRICK SISSONS AND MICHAEL B. A. OLDSTONE
Aleutian ~i~~~~~ of Mink
DAVIDD. PORTER,AUSTIN E. HELEN G. PORTER
h R S E N , AND
Age Influence on the Immune System
TAKASHI MAKINODANAND MARGUERITEM. B . KAY SUBJECTINDEX