Advances in Immunology Volume 22

CONTRIBUTORS TO THIS VOLUME ABUL K. ABBAS CHARLES B. CARPENTER HARVEYR. COLTEN ANTHONY J. F. D’APICE STEPHEN C. GREBE M...

Author: Frank J. Dixon | Henry G. Kunkel

11 downloads 895 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

DOWNLOAD PDF

CONTRIBUTORS TO THIS VOLUME ABUL K. ABBAS

CHARLES B. CARPENTER HARVEYR. COLTEN ANTHONY J. F. D’APICE STEPHEN C. GREBE MICHAELE. LAMM YA. S. SHVARTSMAN J. WAYNE STREILEIN

M. P. ZYKOV

ADVANCES IN

Immunology EDITED B Y

FRANK J. DlXON

HENRY G. KUNKEL

Scrippr Clinic and Research Foundation la lolla, California

The Rockefeller University N e w York, N e w York

VOLUME 2 2

1976

ACADEMIC PRESS New York

Sun Francisco

A Svbaidiory of Horcovrt Brace lovanovich, Publishen

London

COPYRIGHT 0 1976, BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. N O PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

111 Fifth Avenue, New York, New York 10003

United Kingdom Edition published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London N W I

LIBRARY O F CONGRESS CATALOG CARD

NUMBER:61-17057

ISBN 0-12-022422-4 PRINTED IN THE UNITED STATES O F AMERICA

LIST OF CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors' contributions begin.

ABUL K. ABBAS,Departments of Medicine and Pathology, Peter Bent Brighum Hospital and Harvard Medical School, Boston, Massachusetts (1) CHARLESB. CARPENTER, Departments of Medicine and Pathology, Peter Bent Brigham Hospital and Harvard Medical School, Boston, Massachusetts ( 1 ) HARVEYR. COLTFN,Division of Allergy, Department of Medicine, Children's Hospital Medical Center and Department of Pediatn'cs, Harvard Medical School, Boston, Massachusetts ( 6 7 ) ANTHONYJ . F. D'APICE,Departments of Medicine and Pathology, Peter Bent Brigham Hospital and Harvard Medical School, Boston, Massachusetts ( 1 ) STEPHENC. GREBE,'Departments of Cell Biology and Internal Medicine, Southwestern Medical School, University of Texas Health Science Center at Dallas, Texas (119) ~IICHAELE. LAMM, Department of Pathology, New York Uniuers-ity School of Medicine, New York, New York (223)

YA. S. SHVARTSMAN, The All-Union Research Institute of Influenza, Leningrad, USSR (291) J . WAYNESTREILEIN, Departments of Cell Biology and Internal Medicine, Southwestern Medical School, University of Texas Health Science Center at Dallas, Texas (119) M. P. ZYKOV, The All-Union Research Institute of Influenza, Leningrad, USSR (291)

'Present address: Naval Medical Research Institute, National Naval Medical Center, Bethesda, Maryland 20014. vii

PREFACE The fact that immunology was founded and first developed in close relationship to the practice of medicine may sometimes be forgotten today in view of the many important contributions made by immunologists to the basic sciences of structural chemistry, molecular and cellular biology, and genetics. However, these contributions often aid in the solution of human health problems, thus adding to the routes by which immunology makes its impact on medicine as a whole. The reviews in Volume 22 serve as good examples of immunology’s varied facets, since their authors present subjects of both practical and theoretical medical value. One of the most vexing practical, immunological problems is the rejection of allografts. While the mechanics of transplantation are reasonably well in hand, control of those immunologic processes which lead to rejection still eludes us. In the first article, Drs. Carpenter, d’Apice, and Abbas draw on their extensive clinical and laboratory experience in dealing with the role of humoral immunity in both graft rejection and enhancement. The authors present in detail the characteristics of the antigraft antibodies and the humoral and cellular mediator mechanisms they activate in inducing graft rejection. Also described are the properties of enhancing antibodies and the means by which such molecules actually can protect organ grafts from the rejection process. It would appear that knowledge in this field has now developed to the point where many of the aspects of graft rejection and perhaps the problem as a whole can be stated in modern immunologic terms, which should lead to studies providing definitive answers to the many remaining questions of transplantation rejection. With our improved knowledge of the biochemistry and function of the complement system, it has become possible to begin sophisticated studies of the biosynthesis of the various complement components. In the second paper, Dr. Colten, a leader in this field, reviews our knowledge of complement biosynthesis, placing it in the perspective of modern-day biology. While the use of in vivo and in vitro methodologies aimed at localizing and quantitating complement synthesis has produced some conflicting information, it now appears that normally components 3, 6, and 9 are produced primarily in the liver, components 2 and 4 are produced in monocytes and macrophages, and components 5 and 8 are produced in a wide variety of organs and tissues. However, there is a good bit of evidence that pathologic conditions may well alter these sites ix

X

PREFACE

of complement formation. In addition, the genetic deficiencies of specific complement components found in man and animals have provided welldefined situations for the study of the controls of gene action. Finally, the author discusses the very important nongenetic factors often found in disease which appear to influence complement formation. In the two decades since the first description of graft-versus-host reactions the importance of this phenomenon in the field of immunology has greatly increased. In the third contribution, Drs. Grebe and Streilein present a comprehensive account of how the graft-versus-host reaction, originally considered to be merely the product of unusual stimulation of immunocompetent donor cells, has now come to be recognized as the result of the interplay between donor T cells and host lymphocytes which follows the usual rules of lymphocyte interactions in immune responses. Thus, the graft-versus-host reaction provides a readily studied, somewhat exaggerated sequence of antigenic stimulation, lymphocyte activation, and finally regulation of an immune response. As such this model has taken on new importance for the transplanter, the immunologist, and the immunochemist. The complex cellular events involved in the dual role IgA plays in both humoral and secretory immunity are discussed by Dr. Lamm in the fourth article. First he considers the evidence supporting the origin of IgA secreting cells from precursors that formerly made IgM. Among the most distinctive features of the IgA system is its predominance in mucous membranes. The cellular events responsible for maintaining this concentration of IgA forming cells in the mucous membrane are well described. Finally Dr. Lamm presents the interesting story whereby two distinct cell types synthesize different components of the secretory IgA complex which are later assembled for secretion by mucous membrane epithelia. A specific example of the role of secretory antibodies in disease is presented in the last paper, where Drs. Shvartsman and Zykov discuss secretory anti-influenza immunity. The formation and/or function of secretory anti-influenza antibodies appear to depend on both general characteristics of the host and its environment and upon the special features of immunologic challenge by the virus. The possibility of prophylaxis against influenza by passive administration of antibodies into the respiratory tract is evaluated. Finally, there is a penetrating discussion of the importance of, secretory antibodies in preventing spread of influenza, i.e., in establishing a collective immunity based upon the presence of secretory antibodies.

FRANK J. DIXON HENRYG. KUNKEL

The Role of Antibodies in the Rejection and Enhancement of Organ Allografts CHARLES B. CARPENTER,' ANTHONY J. AND ABUL K. ABBAS

F. d'APICE,

Departments of Medicine ond Pothology, Peter Bent Brighom Hospital ond Harvard Medical School, Boston, Marsochuretts

I. Introduction. . . . . . . . . . . . . . 11. The Destructive Effects of Antibody in AlIotranspIantation . . . . A. The Role of Antibody in Allograft Rejection . . . . . . B. Mechanisms of Antibody Action in Allograft Rejection . . . . 111. The Protective Effects of Antibody in AIlotransplamtation . . . . A. Physicochemical and Biologic Properties of Enhancing Antibodies . B. Induction of Enhancement: Kinetics and Regulatory Factors . . . C. Effects of Enhancing Antibodies on the Immunology and Pathology of Allograft Rejection . . . . . . . . . . . D. Mechanisms of Action of Enhancing Antibodies . . . . . IV. Conclusions . . . . . . . . . . . . . . References . . . . . . . . . . . . . .

.

1 2 2 18 27 29 35 39 48 54 55

I. Introduction

The first reports of technically successful renal transplantation appeared in 1905 (Carrel1 and Guthrie, 1905; Floresco, 1905). Many subsequent experiments made it obvious that autografts could survive indefinitely whereas allografts ceased to function after several days. Allografted kidneys developed massive and progressive increases in cellular infiltrate and edema fluid with subsequent necrosis, in striking contrast to the minimal changes seen in autografts ( Wu and Mann, 1934). The immunologic basis for these differences was shown by Medawar (1944), who demonstrated the development of active immunity against the foreign tissue. Since the discovery of the two basic subdivisions of the lymphoid immune system: thymus-dependent and thymus-independent (Miller, 1961; Cooper et al., 1966), there has been considerable interest in their relative roles in allograft rejection. In general terms, thymusdependent ( T cells ) responses involve cell-mediated immunity, and Tindependent ( bone marrow-dependent or B cells) responses involve humoral antibodies. Various workers have implicated one or other of these subdivisions as being of predominant importance in transplantation (Stetson, 1963; Brent, 1971). It is now evident that an eitherlor situation

' Investigator, Howard Hughes Medical Institute. 1

2

CARPENTER, D’APICE, AND ABBAS

does not exist and that both cellular and humoral factors are involved in the rejection reaction. Their relative importance depends on the animal model, the type of allograft, the presence or the absence of specific presensitization of the recipient, and whether the recipient’s immune system has been modified by immunosuppressive agents. Previous articles have dealt with a number of these aspects (Merrill, 1967; Carpenter and Merrill, 1969). The major purpose of this review is to examine the evidence for immunoglobulin participation in the host response to organ allografts, to place in some perspective a variety of phenomena reported in experimental and clinical transplantation, and to point to those areas of investigation needed to clarify many of the incomplete and even paradoxical results to date. Clearly, aIloantibodies may have in vivo effects that are both damaging and protective to grafted tissue; furthermore, the presence in vivo of alloantigens from the graft makes possible the formation of soluble immune complexes, which may impair cellular effector mechanisms in addition to having phlogogenic properties. II. The Destructive Effects of Antibody in Allotransplantation

A. THE ROLE OF ANT~BODY IN ALLOGRAFT REJECTION This section will present evidence pertaining to the involvement of antibody in the rejection reaction. Animal and human transplantation are considered in separate subsections, and in both, the evidence presented pertains to transplantation across major histocompatibility barriers (e.g., H-2, AgB, HLA). 1. Functional and Morphologic Studies a. Animal Transpluntution. The morphology of a rejecting primary renal allograft in an unmodified dog recipient is predominantly one of gross swelling ( Simonsen et al., 1953) and cellular infiltration (Williams et al., 1964). However, plasma cells can be detected in small numbers by 4 5 days, and immunoglobulin is present in arterioles and in the media of interlobular arteries (Horowitz et al., 1965; Lubbe et ul., 1972). Rejecting rat renal allografts have been reported as showing no significant vascular lesions in the Brown Norway (BN) and DA to Lewis models (Feldman and Lee, 1967). However, other studies across the immunologically weaker (Lewis x BN) F, (LBN F,) to Lewis barrier have indicated vasculitis to be a prominent feature, although the time of onset of these lesions was at day 4 in the study of Abbas et al. (1974a) and as late as 2-3 weeks in that of Guttmann et al. (1967a). Despite these differences, the onset of vascular lesions heralded terminal uremia

REJECTION AND ENHANCEMENT OF ORGAN ALLOGRAFTS

3

in both studies. The only variation of significance between the methods of these investigators was the performance of a contralateral nephrectomy at the time of transplantation by Abbas et al. (1974a) whereas Guttmann et al. (1967a) delayed the contralateral nephrectomy for 2-5 days. It is not clear how this variation can account for the marked difference in the onset of the Iesions. It is notable that the demonstration of immunoglobulin (Ig G ) and complement ( C3 ) deposits parallels the demonstration of vascular lesions. Feldman and Lee (1967), who found no significant arteriolar or arterial vascular lesions, found no immunoglobulin deposition, yet in the model in which early vascular lesions were present, IgG and C3 were demonstrable in association with the vascular lesion (Abbas et al., 1974b). Similarly, Lindquist et al. (1971) described a series of events occurring in glomeruli of the rejecting rat renal allograft that clearly link IgG and C3 deposition with polymorphonuclear leukocyte ( PMN ) invasion and subsequent endothelial destruction and later capillary occlusion by platelets and fibrin. Vasculitis is also a prominent feature in the rejection of a primary allograft by the unmodified sheep (Pedersen and Morris, 1974a) and rabbits (Hobbs and Cliff, 1973). These studies indicate considerable variability in the incidence and severity of vascular lesions and immunoglobulin deposition in primary vascularized allografts among various animal species and even among strain combinations within a species. There is evidence, discussed in a subsequent section, that these morphologic changes are markers of humoral immunity. Although the morphology of rejection of primary transplants indicates a variable, and possibly secondary, role of humoral immunity, the rapidity of onset and morphology of the rejection reaction occurring in secondary (second set) transplants leave little doubt that, in this instance, humoral immunity is the predominant and only necessary form of immunity. Secondary transplants fail rapidly with complete loss of function within 1 3 days (Simonsen et al., 1953). The graft becomes red and swollen within 12 hours owing to intravascular stasis and interstitial hemorrhage. Fibrinoid necrosis of glomerular capillaries, arterioles, and interlobular arteries is marked by 24 hours. Cellular infiltration is not a feature for the first 24 hours and thereafter consists predominantly of neutrophils. In the rabbit, presensitization results in massive cortical necrosis that is complete within 48 hours (Klassen and Milgrom, 1971; Holter et al., 1972). However, these findings do not necessarily preclude some contribution by the presensitized cellular immune system. At one time, humoral immunity was thought to play little if any role in rejection, a belief based largely on skin graft experiments and on

4


the failure of early attempts to induce rejection by passive serum transfer (reviewed by Stetson, 1963). Later, however, there were many reports of serum transfer causing accelerated allograft rejection ( Kretschmer and Perez-Tamayo, 1961; Steinmuller, 1962; Dubernard et al., 1968; Pedersen and Morris, 1974a). Several experiments demonstrate this particularly clearly. Perfusion of a dog’s kidney in situ with alloimmune serum prepared against the recipient, results in rapid renal shut down ( Altman, 1963). The same effect has been found with similar experiments in goats (Cochrum et al., 1969) in which histologic examination showed a pattern of destruction almost identical to that seen in hyperacute rejection in humans. IgG deposits were present in the intima of renal vessels. Straus et al. (1971) performed the same experiment in rats. No pathologic changes were present 24 hours after in situ perfusion of the kidneys with specific antiserum; however, by 10 days there were marked mononuclear infiltration and fibrinoid necrosis, disruption of internal elastic lamina1 and intimal proliferation of arteries. The point has also been made by experiments that could be called passive transfer in reverse. Clark et al. (1968) and Foker et al. (1969) showed that a dog rejected its own kidney after retransplantation, if the organ had initially been transplanted for a short time into another specifically sensitized dog. The allograft could not be rejected in the intermediate host, since it had been rendered profoundly leukopenic by whole-body irradiation and hypocomplementemic by administration of cobra venom factor. These experiments suggest that rejection is brought about by the dog’s own effector mechanism after exposure to antigraft antibody in the sensitized temporary recipient. Hyperacute rejection can be induced readily in a variety of animals by the passive administration of antiserum (Klassen and Milgrom, 1971; Holter et al., 1972); however, most rat strains are peculiarIy resistant to this form of rejection. Circulating antigraft antibodies have been demonstrated both in dogs (Yamada and Kay, 1968) and in rabbits (McDonald et al., 1964); however, the titer is low until the graft is completely rejected or removed. The presence of serum antibody per se is not a reliable marker for rejection activity. SO far, we have considered evidence for the involvement of humoral immunity only in relation to primarily vascularized organ grafts. The morphology of rejecting skin allografts differs considerably from that described in relation to renal allografts; in particular, IgG deposition has not been reported in primary grafts. The question of the role of humoral immunity in skin graft rejection has been reviewed on several occasions (Stetson, 1963; Winn, 1970), and the subject is still controversial. The main arguments against antibody having an important role have been the highly variable, but usually negative, results of attempts


5

to accelerate rejection by passive immunization ( Stetson, 1963), the failure of early attempts to induce rejection of allografts placed in diffusion chambers, even in highly presensitized recipients (Algaire et al., 1954, 1957; Woodruff, 1957), and the apparent lack of effect of blood group isoantibodies against human skin allografts. These arguments have to some extent been countered by more recent work. A major difference between skin grafts and renal allografts is that the former are not primarily vascularized grafts, and vascular continuity can occur only as a result of the slow ingrowth of recipient vessels. Skin allografts placed on unmodified recipients are rejected (presumably by cell-mediated mechanisms) before vascular continuity is achieved and consequently before humoraI mechanisms can play a prominent role. Hyperacute rejection of skin allografts can be induced by passive antiserum transfer; however, it seems to be possible only when the antiserum is injected around the graft site (Stetson and Demopoulos, 1958), or if the graft is well healed in and vascular connections with the host have been established. The latter situation has been produced by Hasek (Bubenik et al., 1970; Hasek et al., 1969) by the induction of neonatal tolerance in ducks and by Baldamus (Baldamus et al., 1973; Winn et al., 1973) by thymectomy and antilymphocyte serum therapy prior to xenografting. In both experiments passive antiserum administration induced rapid graft destruction. Subsequent work with diffusion chambers indicated that enclosed grafts could, in fact, be rejected (Algaire, 1959; Gabourel, 1961). The reason for the failure of the early experiments was shown to be technical, in that antibody and complement did not easily enter the chambers (Wakefield and Amos, 1958; Amos and Wakefield, 1958, 1959). Equally convincing are the results of the reverse experiment performed in mice (Najarian and Feldman, 1962a,b) and rats (Kretschmer and PerezTamayo, 1962) in which skin allografts, placed on nonimmune animals, were rejected in an accelerated fashion if immune lymphocytes were placed in a distant diffusion chamber. On the whole, these experiments indicate that skin grafts are susceptible under certain circumstances to the destructive effects of humoral immunity, but suggest that its role in rejection of primary skin grafts is considerably less important than it is in primarily vascularized organ grafts. The principal difference appears to be related to the accessibility of the graft to antibody. Free cell grafts, such as lymphoid and bone marrow cells, are freely accessible to circulating antibody and the importance of antibody in their rejection is well established (Stetson, 1963; Winn, 1974, Moller and Moller, 1962). b. Human Transplantation. The morphology of rejection of allografts

6


in humans differs from that in animals predominantly in its timing. This is due to the influence of various modifying factors, especially those of immunosuppressive therapy and of chronic uremia-itself a potent immunodepressant. A further point of difference is the often present uncertainty as to whether a human transplant recipient is sensitized to the allograft despite negative in vitro serologic reactivity. This point will be considered in greater detail in Section II,A,2. It is customary to consider rejection under these circumstances in three clinically and pathologically distinguishable categories : hyperacute, acute, and chronic rejection. Hyperacute rejection is characterized by an early onset, often within minutes of vascularization of the graft. Histologic examination shows an intense PMN infiltrate and vascular thrombosis (Morris et d., 1970; Williams et al., 1968). This form of rejection is associated in most instances with known presensitization (Morris et al., 1970; Pate1 and Terasaki, 1969) or major blood group incompatibility of the graft (Sheil et al., 1969). Again, the speed of onset, rapidly of destruction, lack of lymphocyte infiltrate, and the fact that it can be caused by isoantibodies are indicative of its humoral mediation. It is worth mentioning the Shwartzman reaction as a cause of allograft failure. This phenomenon, classically described in the rabbit, is characterized by the abrupt onset of disseminated intravascular coagulation and is a complex and partly immunologic event induced by a variety of agents, including endotoxinemia. A localized form has been invoked in several instances as the reason for abrupt renal allograft failure. In only one of these instances can the Shwartzman reaction be considered unequivocally to be the etiologic factor. Schiff et al. (1974) described the development of acute cortical necrosis in a normally functioning HLA identical, MLC negative renal allograft 22 days after transplantation. There were neither PMN infiltrate nor IgG deposition in the graft, and antigraft antibody could not be detected in the recipient’s serum. A subsequent skin graft from the same donor was not rejected in an accelerated fashion. From these observations, it is extremely unlikely that rejection was the pathogenetic mechanism. It must be emphasized that the localized Shwartzman reaction is a rare cause of allograft failure that mimics hyperacute rejection. Acute rejection (rejection episode) is the form of rejection most commonly seen. It can best be equated with the rejection of a primary transplant in an unmodified animal recipient, and, although most commonly occurring in the first few months, it may occur at any time after grafting. The pathology can vary from a mild form of cellular infiltrate to severe destructive fibrinoid necrosis and thrombosis ( Kincaid-Smith,


7

1967; Porter, 1966). Deposits of IgG and complement are found frequently in association with vascular lesions (Lindquist et al., 196813; McKenzie and Wittingham, 1968). I t seems probable that this form of rejection can occur as a result of both first- and second-set immunologic reactions (Terasaki et al., 1971). Chronic rejection is a form seen only in modified recipients, the time taken for its development precluding its becoming evident in the unmodified animal. The brunt of the attack is borne by the medium and large arteries, where there is progressive intimal proliferation and thickening with resultant ischemia ( Kincaid-Smith, 1967; Busch et al., 1971b). Busch et al. (1969) in a large and careful study of human renal allografts have associated these obliterative lesions with IgG deposition on the vascular endothelium; they associated fibrinoid necrosis with medial IgG deposition. It seems likely, as they suggest, that the obliterative lesions are the result of repetition of the same immunologic insults to the vascular endothelium which are initially responsible for acute rejection episodes. The report of Jcannet et al. (1970) also indicates the predominant importance of humoral immunity in the pathogenesis of this lesion. They have shown that, in nearly all instances in which the lesion develops, circulating antigraft antibody can be detected, whereas the lesion is rarely found in the absence of detectable antibody. Immunoglobulin (and complement) deposits are frequently found in long-term renal allografts (Lindquist et al., 196813; McKenzie and Wittingham, 1968; McPhaul et al., 1970), are more obvious during periods of clinically evident rejection, but are also detectable during quiescent phases. Although it will be discussed in a subsequent section, it is worth noting at this stage that the glomerular lesion frequently seen in longterm renal allografts both in man and experimental animals may be a manifestation of rejection due to immune complexes in which either the antigen is organ specific or is a histocompatibility antigen. The fact that the pattern of glomerular immunoglobulin deposition is very commonly granular ( McPhaul et al., 1970) lends support to the concept that these lesions are induced bv antigen-antibody complexes. Hereditary agammaglobulinemia provides a situation in which one would expect to be able to test the importance of humoral immunity in allograft rejection. Unfortunately, the few available reports do not entirely clarify the situation. Good (1959; Good et al., 1962) has reported several examples of prolonged skin graft survival in these patients; however, Schubert et al. (1960) found only minor prolongation ( u p to 3 weeks) of graft survival. Finally, the fact that antigraft antibody is not usually detectable by the microlymphocytotoxicity assay while the graft is in situ should be

8

CARPENTER? D'APICE, AND ABBAS

mentioned, since it has been used as a point in evidence against antibody having an important role in rejection. However, more sensitive techniques, in particular mixed cell agglutination (Milgrom et al., 1966; Jeannet et al., 1970), anti-globulin consumption ( Iwasaki et al., 1967a) and the antibody-dependent lymphocyte-mediated cytotoxicity assay (d'Apice et al., 1974; Myburgh et al., 1974b) are capable of detecting antibody at this time. The rapid rise in the amount of antigraft antibody, which occurs on removal of the graft, suggests that the latter functions as a very efficient sponge for antibody (Milgrom et al., 1966), resulting in low levels in the circulation while it is in situ. In addition, liberation of free antigen from the graft resulting in circulating antigen-antibody complexes would result in free antibody being less readily detectable. c. HistoZogic Markers of Effects of Humoral Immunity. In subsections a and b above we have mentioned various histologic changes occurring in rejecting renal allografts that were considered to be markers for the involvement of humoral immunity. In this section the evidence for this link will be examined in detail. i. lmmunoglobulin deposits. The presence of these deposits of itself does not prove that the antibody is necessarily inducing allograft destruction; however, this link has been made by various observations. First, these deposits are frequently, but not always, associated with sites of active destruction of the graft. These sites, the glomerular capillaries, afferent arterioles and arteries, are also the sites of maximum density of HLA antigens ( Sybesma et al., 1974). Second, these immunoglobulins can be eluted from the organ by various methods, and such eluates have been shown to react in vitro with cells of the donor and often other third parties who share histocompatibility antigens with the donor (Hager et al., 1964; Hampers et al., 1967; Spong et al., 1968; Goldman et al., 1971; Pedersen and Morris, 1974b). Even more convincing are the studies of Spong et al. (1968), who showed that the IgG fraction of an eluate from rejecting primary rat renal allografts would produce a lesion indistinguishable from early rejection if injected into normal animals of the donor strain. The finding of immunoglobulin deposits in areas of the transplant such as the renal glomerulus, where there is little or no sign of destruction at the time, is not necessarily evidence against antibody being involved in rejection, since these deposits are most likely in the form of antigen-antibody complexes, which may cause a form of rejection nephritis that becomes morphologically evident only at a later time after grafting. ii. Fibrinoid necrosis. Vasculitis or fibrinoid necrosis is frequently found in severe forms of acute rejection in both man and animals and


9

is a prominent feature in hyperacute rejection. The evidence supporting this lesion as a marker of humoral immunity is in part dependent on its association with immunoglobulin deposition ( Abbas et al., 1974), particularly that in the muscular coats of arteries and arterioles (Busch et al., 1971b). Fibrinoid necrosis can be readily induced by perfusion of an intact kidney with immune serum (Terasaki et al., 1962; Dubernard et al., 1968; Cochrum etal., 1969; Straus et al., 1971). Further evidence indicating humoral induction of this lesion is the elegant work of Pedersen and Morris (1974a) who demonstrated a temporal relationship between the development of vasculitis in rejecting sheep allografts and the appearance of antigraft antibody in the draining lymphatics. They also showed that the same changes could be induced rapidly by the administration of antigraft antiserum (Pedersen and Morris, 1974b). iii. P M N infiltration. Infiltrates of PMN are common in early allograft rejection ( Kincaid-Smith, 1967) and the number present in the glomeruli of renal allograft biopsies taken within 1 hour of transplantation correlates with the outcome of the graft as assessed by rejection grades (Kincaid-Smith et al., 1968). Several other reports also indicate that their presence in large numbers in glomeruli at this time signifies a poor prognosis (Starzl et al., 1968; Williams et al., 1968; Weymouth et al., 1970). A link between PMN infiltration and humoral immunity is suggested by the association of PMN infiltration with hyperacute rejection. Williams et al. (1968) have reported that six of seven human renal allografts that showed heavy PMN invasion in biopsies taken 1 hour after transplantation underwent hyperacute rejection. Similarly, Gewurz et al. (1966) have reported heavy PMN infiltration in sheep to dog renal xenografts that were undergoing hyperacute rejection. iv. Platelets and fibrin. There is a considerable body of evidence indicating that platelets accumulate and aggregate during rejection and are associated with the vascular lesions (Porter et al., 1964; Lowenhaupt and Nathan, 1968, 1969; MacDonald et al., 1970; Busch et al., 1971b; Lowenhaupt et al., 1971). Subsequent studies have shown that the impairment of platelet survival originally described by Mowbray ( 1967) is due to platelet trapping and consumption by the allograft, which can be detected by the use of radiolabeled autologous platelets (Claes et al., 1970) and by serial estimation of platelet factors 3 and 4 (Anderson et al., 1974). Hobbs and Cliff (1973) have recently provided direct evidence of one pathogenetic role of platelets. In their model, a slice of renal cortex was placed in an ear chamber of a rabbit that simultaneously received an orthotopic renal allograft from the same donor. Direct observation

10


of the ear chamber graft showed that the vascular changes of rejection were associated with endothelial adherence of platelets which undergo morphologic changes known to be associated with their release of vasoactive amines. Aggregation of platelets was not a necessary prerequisite for this change, since single adherent platelets underwent the same alterations. Platelets are also clearly involved in the formation of the microthrombi commonly seen in rejection and in the more extensive thrombi occurring in hyperacute rejection ( Kincaid-Smith, 1967; Porter, 1967). Similarly, fibrin deposition is associated with these thrombi and with the subsequent chronic proliferative vascular lesions ( Porter, 1967; Porter et al., 1968). Fibrin can be detected by measuring radioactivity in the allograft after injection of radiolabeled fibrinogen ( Salaman, 1970) or by measuring urinary fibrin degradation products ( Clarkson et al., 1971). Lowenhaupt et al. (1971) have shown a clear correlation between the presence of platelet aggregates in canine renal allografts and of cytotoxic antibodies. In this model, platelet aggregates developed within 10 minutes of renal allotransplantation in presensitized dogs. The aggregates did not develop if the antibody response to the presensitizing immunization was suppressed with azathioprine. However, if the azathioprine was administered after the recipient had been sensitized, but prior to transplantation, antigraft antibodies persisted and platelet aggregation occurred with rapid loss of the graft. Wardle (1972a,b) has shown that platelet aggregation occurs following antigen-antibody interaction on endothelial cells. Platelet aggregation can also be induced by high local concentrations of ADP and thrombin and by exposure of the capillary basement membrane. The coagulation pathway can be activated, with consequent fibrin deposition, by local exposure of the basement membrane and possibly via the binding of C l q to antigen-antibody complexes with subsequent activation of Hageman factor (Austen, 1974). v. Mononuclear cell infiltration. Mononuclear cell infiltration would naturally be expected to be a good marker for the involvement of cellular immunity in allograft rejection. However, there is evidence indicating that it is of little help in differentiating between humoral and cellular mechanisms. The mononuclear infiltrate is contributed to by thymusindependent lymphocytes (Lindquist et at., 1968a), which may be involved in the rejection process through the mediation of antibody: i.e., via antibody-dependent lymphocyte-mediated cytotoxicity. It has already been noted that immune serum or antibody eluted from a rejecting organ produces a mononuclear infiltrate morphologically identical to that in a rejecting allograft, if injected into a normal animal against which the antibodies are directed. Obviously, in these instances there can be no contribution at all made by specifically sensitized thymus-de-


11

pendent lymphocytes. Consequently, many of the infiltrating mononuclear cells in a rejecting allograft may be there as a result of the action of antibody.

2. Clinical Significance of Presensitixation In clinical practice presensitization is usually assessed by repeated monthly screening of a potential allograft recipient’s serum for antibodies reactive with a panel of lymphocytes of differing HLA phenotypes, and immediately prior to transplantation by a specific “cross-match” between the recipient’s serum and his prospective donor’s lymphocytes. Various modifications of the microlymphocytotoxicity assay first introduced by Terasaki (Mittal et al., 1968) are most commonly used. In essence, the assay measures the specific cytotoxic activity of a serum, in the presence of excess complement, against lymphocyte targets. Cell death is assessed by the uptake of a supravital dye such as eosin. The use of this assay has indicated the clinical importance of presensitization and has contributed information concerning the role of antibody in rejection. It should be remembered, however, that the association of presensitization as assessed by this method and allograft rejection cannot be taken blindly as evidence for a role of humoral immunity. It is quite possible that in many circumstances humoral presensitization simply serves as a marker for that state in the whole immune system. Indeed, Garovoy et al. (1973a) have shown a high degree of correlation between humoral immunity, as assessed by the microlymphocytotoxicity assay, and cellular immunity as assessed by direct lymphocyte-mediated cytotoxicity. Having sounded this note of caution, we will turn to the subject of hyperacute rejection, which is considered to be induced by purely humoral mechanisms. There is a high incidence of hyperacute rejection if transplantation is performed in the face of a positive direct cross-match ( Kissmeyer-Nielsen et al., 1966; Terasaki et al., 1968; Williams et al., 1968; Patel and Terasaki, 1969). Despite several reports of individual instances in which specific presensitization did not result in hyperacute rejection (Heale et al., 1969; Patel and Terasaki, 1969; Terasaki et d., 1971), the overwhelming burden of evidence is that a positive crossmatch constitutes an absolute contraindication to transplantation. The most convincing evidence for the belief that humoral immunity is the predominant and only necessary form of immunity in the pathogenesis of this form of rejection is the fact that it can be induced by the naturally occurring blood group isoantibodies (Starzl et al., 1964; Williams et al., 1967; Sheil et al., 1969) and by perfusion of a kidney prior to transplantation with sera containing antigraft antibodies (Cross et al., 1974).

12


The problem of presensitization extends to a more subtle level than the presence or the absence of a positive cytotoxic cross-match. There is considerable evidence that lesser degrees of presensitization, indicated by known previous or present reactivity of a recipient’s serum with lymphocytes of various HLA phenotypes, but with a negative direct cross-match with his specific donor, are associated with more frequent and severe rejection episodes and markedly reduced graft survival (Terasaki et al., 1971; Patel et al., 1971; Opelz et al., 1972; Stenzel et al., 1974; Myburgh et al., 1974a) as compared to recipients who have never been known to be presensitized. The fact that a significant number of these grafts are rejected in a hyperacute fashion (Starzl et al., 1968; Patel and Terasaki, 1969; Myburgh et aE., 1969), estimated in a series of over 1000 transplants by Terasaki et al. (1971) to be 8%,indicates the insensitivity of the current cross-matching techniques. Other groups, most notably that of Dausset, do not find this difference in graft survival between the presensitized and nonsensitized groups; however, somewhat at variance with this, Dausset finds an inverse correlation between graft survival and the presensitizing agent, namely blood transfusions ( Dausset, 1973). The considerably poorer graft survival of this group, even excluding those developing hyperacute rejection, suggests that they were presensitized against their donors, despite the negative cross-matches. An alternative explanation is that these individuals, although not specifically sensitized against their donors, respond rapidly and vigorously to the foreign histocompatibility antigens, and these individuals are marked as “high responders” by the fact that they have previously formed cytotoxic antibodies in response to immunization by pregnancy, blood transfusions, or other histocompatibility antigen sources. There is some, although as yet incomplete, evidence that both of these explanations may be correct. Lending support to undetected presensitization as the explanation are the known and previously mentioned lack of sensitivity of the microlymphocytoxicity test, as emphasized by the occasional occurrence of hyperacute rejection in association with a negative cross-match, and the fact that antibody can often be detected by the antibody-dependent lymphocyte-mediated cytotoxicity assay when it is not detectable by the microlymphocytotoxicity test ( d’Apice et al., 1974). This highly sensitive assay for IgG antibodies has recently been reported to be of value in predicting graft outcome when performed with pretransplant recipient serum (Ting and Terasaki, 1974; Jeannett et al., 1975). Belzer et al. (1974) in a study of 200 consecutive patients given primary cadaver grafts found no significant difference in l-year graft survival between the group with preformed cytotoxic antibodies (55% l-year survival,


13

n = 56) and those without detectable antibodies (60%1-year survival, n = 144). They attribute this result to the use of a highly sensitive cross-match technique and suggest that poor graft survival in presensitized recipients in other studies is due to undetected presensitization. In a further attempt to increase the detection of presensitization, they cross-match known highly reactive sera from the past, as well as the most recent serum sample, against the donor’s lymphocytes. Van Rood (1974), Van Hooff et al. (1972), and Oliver et al. (1972) have shown that graft survival can be improved in the group who have preformed cytotoxic antibodies by close matching of the “Four,” or second B locus segregant series, of HLA antigens. This suggests that the effect of undetected presensitization can be avoided when antigens against which it may be directed are matched. The original suggestion that the difference in graft survival between the two groups of patients was related to their immunoresponsiveness was made by Opelz et al. (1972). They showed that potential renal allograft recipients, nearly all of whom received blood transfusions, could be separated into two groups, those who formed cytotoxic antibodies and those who did not, and that there was a marked difference in graft survival between the groups. The difference between the groups was most marked if the patients had been on dialysis for a year or longer, giving them sufficient time (and presumably blood transfusions) to respond to the foreign HLA antigens. Morris and Dumble (1973) have shown a high inverse correlation between pretransplant serum reactivity with a panel of continuous lymphoblastoid cell lines and allograft survival. The antigens on these cell lines, against which these sera are reactive, include HLA antigens and various Epstein-Barr virus antigens. These antigens are recognized by heat-stable antibodies. Another antigen, which they named “D,” is recognized by a heat-labile antibody (Dumble et al., 1975). All sera that were reactive with normal lymphocytes in the microlymphocytotoxicity test were shown to have heat-stable antibodies reactive with the cell lines. Subsequently, to determine whether the heat-stable reactivity present in cell line-reactive, but normal lymphocyte-nonreactive, sera was directed against HLA antigens or only against the virally determined antigens, the same group used a simple inhibition of absorption technique and detected sublytic amounts of anti-HLA antibody in some sera (Dumble and Morris, 1975). This would support undetected presensitization as the explanation, However, supporting the concept of differing immunoresponsiveness was the finding that the heatlabile antibody was present in sera of 98%of normal subjects, but was not detected in the sera of the “low responder” group (Morris and Dumble, 1973). This suggests that uremic immunosuppression was suffi-

14


cient to abolish or reduce to sublytic levels the usually present anti-“D” antibody in the “low responders” but not in the “high responders.” Consequently, this suggests that these groups may be viewed in terms of immunosuppressibility, rather than responsiveness. These various reports do not allow any firm conclusion as to which explanation for the difference in graft survival between the groups is correct; however, they are not necessarily mutually exclusive views. 3. Specificity of Antibodies Involved in Allograft Destruction

a. Antibodies against Major Histocompatibility Antigens. Little more need be said about antibodies against the major histocompatibility antigens since the first two sections have dealt exclusively with these antibodies. If the definition of major histocompatibility antigens is broadened to include all presumed antigen products of the major histocompatibility locus other than HLA, AgB, H-2, etc., namely those of the L D and I (immune response) regions, then a considerable area remains. However, practically nothing is known about the importance of antibodies directed against these antigens, except for the fact that such antibodies can produce enhancement (Davies and Alkins, 1974) and have been detected in some immune human sera (Van Rood et al., 1975). The definition of these antigens in humans is currently an area of intense activity . b. Antibodies against Minor Histocompatibility Antigens. The question of minor histocompatibility antigens has been reviewed recently ( Graff and Bailey, 1973). In most instances, their effects on allograft function and survival are either so mild as to be difficult to examine or are completely obscured by the greater influence of major histocompatibility differences. We have recently encountered a situation in which the effects of minor histocompatibility antigen( s ) are sufficient or sufficiently isolated to be examinable from the viewpoint of the destructive effects of antibody directed against it. BN and LBN F, rats possess an organspecific proximal tubular basement membrane (TBM) antigen( s ) that is not present in some Lewis substrains. It is also found in Fischer rats, which are of the same AgB type (AgB1) as Lewis rats, and in the guinea pig. Clearly, it is not linked to the major histocompatibility system of the rat, but because it can induce an immune response in rats that lack it, it can be termed a minor histoconipatibiIity antigen. LBN F, kidney transplantation to deficient Lewis recipients exposes the host to this antigen, resulting in an antibody response. If graft survival is maintained for 1 month or more by means of enhancing Lewis anti-BN antiserum, anti-TBM antibody is detectabIe in the recipient’s serum and in the allograft and is associated with significant tubular


15

injury (Abbas et al., 1975a,b). These studies confirm those of Lehman et al. (1974), who found essentially the same lesions in rejected transplants in the same strain combinations. It is probable that the tubular lesions reported in Fischer to Lewis renal allografts, but not in the reverse situation (White et al., 1969), represent an antibody response against this organ-specific minor histocompatibility antigen. Final proof that the lesions are antibody-induced awaits experiments in which the antibody is eluted and reinjected into TBM antigen-positive rats. Similar anti-TBM antibodies have been detected in human renal allografts, although they were not detected in the recipient’s own kidneys, indicating that they were formed in response to the allograft and were not due to preexisting autoimmune disease (Williams et al., 1967; Berger and Yaneva, 1973; Wilson et al., 1974). The clinical significance of lesions induced in response to minor histocompatibility antigens cannot be accurately assessed; however, it is probably minimal owing to the overshadowing influences of the immune response to major histocompatibility differences. It is likely that they will assume greater significance when methods of preventing the immune response to the major antigenic differences are perfected. This will require the development of methods of detecting and matching or preventing the response to minor antigens, particularly those that are organ specific. c. Antibodies against Extrinsic Agents That Cross-React with Histocompatibility Antigens. It is now widely appreciated that antigens present on a variety of bacteria and possibly viruses are sufficiently similar to the histocompatibility antigens of mammals to be able to induce the formation of antibodies that cross-react with the latter. Streptococci ( Rapaport et al., 1966) and staphylococci ( Rapaport and Chase, 1965) have been shown to induce allograft hypersensitivity in several animal species; the antigens responsible for this effect are contained in the bacterial cell wall ( Rapaport et al., 1966). Similarly, glycoproteins, polysaccharides, and lipopolysaccharides from a variety of bacteria and plants can inhibit the complement-dependent cytotoxic effect of anti-HL-A antisera on appropriate target cells (Mittal et al., 1973). Rapaport et al. ( 1971) have shown that preimmunization with streptococcal antigens results in accelerated rejection of renal allografts in dogs; however, specific antistreptococcal antibodies could be identified neither in the circulation nor in the transplanted organs. Since adjuvant effects are possible, the most direct evidence on this point is the demonstration by Rapaport et al. ( 1969) that antistreptococcal antibodies can induce hyperacute rejection of guinea pig skin allografts. In the clinical situation, associations between bacterial ( Simmons

16


et al., 1970) and viral infections (David et al., 1972; Briggs et al., 1972; Lopez et al., 1973) have been reported, but antibodies against these agents have not been directly identified in association with the rejection episodes. Although the foregoing indicates that antibodies formed in responses to antigens present on extrinsic agents can sometimes induce allograft rejection, their importance in the clinical situation is uncertain, particularly since the high incidence of infection in immunosuppressed patients makes it quite conceivable that the reported associations are simply coincidental. d. Antibodies against Antilymphocyte Sera. Antilymphocyte globulin ( ALG) has enjoyed a considerable trial period as an immunosuppressive agent in clinical transplantation. It is usually prepared by immunizing horses or goats with various populations of human lymphocytes and has on occasion contained antibodies that cross-react with human glomerular basement membranes (GBM) (Guttmann et at., 1967b; Iwasaki et al., 196%). The nephritogenic capacity of such antibodies is well established in experimental systems (Guttmann et at., 1967b; Katz et al., 1967). Apart from antibodies present in these preparations, the recipient may make an immune response against the foreign protein, and resulting immune complexes may induce serum sickness and allograft lesions. In the clinical situation, Thiel et al. (1971) have reported glomerular deposition of horse globulin in all of 13 patients who were treated with intravenous ALG; however, clinical nephritis did not develop in any case. The same ALG preparation was shown to induce nephrotoxic nephritis in monkeys. Deposition of horse globulin on the GBM is Iess frequent when the ALG if administered subcutaneously or intramusculady (C. B. Wilson et ul., 1971). In a study of 32 patients who received intramuscular ALG, only 2 developed the appropriate glomerular deposits, and one of these had morphologic evidence of glomerulonephritis in the graft (Busch et al., 1971a). Typical immune complex nephritis can be induced in rats with heterologous ALG preparations (Guttmann et al., 1967b; Carpenter et al., 1971). The most striking feature of the clinical studies is the relative safety, at least from the risk of inducing allograft damage, with which ALG can be used despite serious complications that theoretical considerations would predict. Screening of ALG lots for anti-GBM antibodies prior to clinical use is now considered mandatory. e. Antibodies Involved in RecurTent Glomerulonephritis. Glomerulonephritic lesions are commonly found in human renal allografts; however, there are many possible etiologies, including rejection, de novo development of nephritis, autoimmune processes, and recurrence of the recipi-


17

ent’s original disease. In order to identify those instances of recurrent nephritis, one must demonstrate not only similar clinical and morphologic disease, but also similar immunologic mechanisms operative both before and after transplantation (Hume and Bryant, 1972). On this basis, recurrent nephritis has been documented most clearly in isograft recipients (Glassock et al., 1968) of whom up to 65% develop a disease that is identical to their original immunologic renal disease. Despite this finding there have been few subsequent reports of recurrent nephritis developing in renal grafts between identical twins. It is more difficult to determine the incidence of recurrent nephritis in allografts, since glomerular alterations due to rejection may be indistinguishable from those caused by the original disease. The demonstration that original and posttransplant lesions are due to the same persisting immunologic mechanism has to date only been achieved with anti-GBM mediated disease, since elution of the antibody is possible (Lerner et al., 1967). In a series reported by Dixon et al. (1969), 7 of 11 patients with anti-GBM disease developed recurrent nephritis. The incidence of linear immunoglobulin deposits in the glomeruli of renal allografts varies from 10 to 25% (McPhaul et al., 1973), and although it is unproved, these may represent instances of recurrent anti-GBM disease. Against this interpretation is the fact that this incidence of linear deposition is much higher than that found in glomerulonephritic patients. Andres et al. (1970) have provided a possible explanation. They have found that in some instances the linear immunoglobulin deposition in allografts differs from that seen in anti-GBM nephritis (Goodpasture’s syndrome) in that it is separated from the GBM. Thus, it is likely that these instances do not represent cases of recurrent nephritis and that the antibodies may be directed against histocompatibility antigens along the capillary wall and represent rejection. The question of recurrent immunologic disease due to antigen-antibody complexes is less easily examined because technical difficulties in eluting complexes have so far prevented frequent identification of the antigens and antibodies involved. Nevertheless, morphologic and immunofluorescence studies strongly suggest that immune complexes car? cause recurrent glomerulonephritis ( Dixon et al., 1969). The prevalence and importance of recurrent immunologic disease, like the significance of minor histocompatibility antigens, will probably become evident when methods of preventing the overshadowing immune response against major histocompatibility antigens are perfected. f . Antibodies Inducing True Autoimmune Allograft Damage. Proof that autoantibodies are formed after transplantation and cause allograft damage is dependent on demonstrating that the antibodies react with

18

CARPENTER, D’APICE,AND

ABBAS

tissues of both the donor and recipient, hence are truly autoantibodies, and that they were not present in the recipient’s serum or on the target organ( s ) prior to transplantation. At present such a mechanism of allograft damage is purely theoretical since there have been no reports of cases that would satisfy the suggested criteria. Two mechanisms peculiar to the aIIograft situation could cause such a response. First, rejection whether humorally or cellularly mediated may expose and alter tissue antigens in the graft, which are shared by the host and initiate a humoral response; and second, allogeneic stimulation may induce an antibody response to autoantigens that would not normally be immunogenic (Katz, 1972) . One of the two cases reported by Klassen et al. (1973) clearly represents true autoimmune allograft damage. Anti-TBM antibodies were detected by immunofluorescence and elution studies in both the rejected renal allograft and in the patient’s own kidney. Unfortunately, the authors were unable to determine whether the autoimmune disease preexisted or developed subsequent to transplantation.

B. MECHANISMS OF ANTIBODY ACTIONIN ALLOGRAFT REJECZION So far this review has been concerned with an examination of evidence from which it can be concluded that humoral immunity plays a role in allograft rejection and that the importance of this role is variable. The mechanisms by which antibody produces these damaging effects will now be considered. Antibody on its own, although capable of binding to the target organ, produces no clearly recognizable lesion. Damage is produced as a result of activation and focusing of the secondary mediator systems by antibody. In addition to antibody bound to the target organ, antibody in the form of circulating immune complexes may induce aIIograft damage. 1 . The Complement System

Although far from fully elucidated, the contribution of the complement system is the best understood of the secondary mediators of humoral immunity in rejection. The subject has recently been reviewed by Cooper and Cochrane ( 1974) and by Carpenter (1974). Complement components, particularly, C3 and Clq, can regularly be demonstrated in rejecting allografts by immunofluorescence techniques both in animals ( Lindquist et al., 1971; Abbas et al., 1974b) and in man (Lindquist et al., 1968a; McPhaul et al., 1970; Andres et al., 1970) and are usually associated with deposits of IgG or IgM. Prevention of activation of the complement system by various mecha-


19

nisms has been shown to prevent, or at least considerably delay, hyperacute rejection. Pregraft perfusion of a kidney with donor-specific succinylated IgG ( Habal et al., 1973) or donor-specific F ( ab’), (Kobayashi et al., 1971, 1972; Habal et al., 1973; Holter et al., 1973) prevents or delays hyperacute rejection after transplantation into a specifically sensitized host. These effects have been demonstrated in monkeys and rabbits. In experiments performed in monkeys (Habal et al., 1973), the control graft survival time, which averaged 12 minutes, was increased to 44 hours after F( ab’), pretreatment of the allograft. Both of these altered forms of antibody are incapable of activating the complement system by the classical pathway, and although F( ab’) has a very weak ability to activate the alternative pathway succinylated IgG has none, since the heataggregated form fails to convert factor B of the complement system. Continuous citrate infusion, which prevents complement activation by divalent cation depletion prevents hyperacute rejection of a renal allograft for as long as the infusion is continued (Kux et al., 1971). Cobra venom factor ( Co F), which activates the alternative pathway resulting in C3 consumption and consequent depletion, has been used to evaluate the importance of complement in antibody-induced allograft rejection. Foker et nl. ( 1969) transplanted allogeneic kidneys into sensitized, leukopenic dogs treated with CoF for 4 hours, during which rejection did not occur and the kidneys functioned well. They then regrafted them into the original donors that had also been treated with COF. All 12 allografts survived for at least 24 hours and none were subsequently lost for immunologic reasons. The control group were treated identically except that CoF was omitted and 8 of the 12 allografts were uremic by 12 hours. In other experiments (Kux et al., 1971; Kobayashi et al., 1972), simple pretreatment of sensitized allograft recipients with CoF produced little, if any, delay in the ensuing hyperacute rejection. The first report of serum complement levels in human allograft rejection was that of Guiney et al. (1964), in which they showed a reduction of total hemolytic complement activity (CHSO) and C2 levels in association with renal allograft rejection. A subsequent report from the same laboratory showed a 50%reduction in C2 levels with and after rejection and suggested its use as a diagnostic test for rejection (Austen and Russell, 1966). Confirmation of these findings has been reported by Gewurz et al. (1966). C3 and C4 levels were studied in this laboratory and showed an initial rise from the baseline 2-3 days after the onset of rejection, followed by a decline to subnormal levels and finally a slow rise back to the base line (Carpenter et al., 1967; Shehadeh et al., 1970). Further studies with highly purified radiolabeled C3 and C4 indicated hypercatabolism of these components during both acute

20


and chronic renal allograft rejection compared to catabolic rates in stable patients or normal subjects (Carpenter et al., 1969). Consequently, it is clear that complement is activated at the time of rejection and is therefore likely to be part of the rejection process. Further supporting this contention is evidence that during rejection, complement is consumed in the allograft (Foker et al., 1969). Kux et al. ( 1973) measured CH50 levels across experimental canine renal allografts after revascularization, and at 9 minutes there was an average arterial-venous gradient of 30%which was associated with endothelial lesions in biopsies examined 2 hours after transplantation. Autografts showed neither CH50 gradients nor vascular lesions, A subsequent report by the same authors on using this test in human transplantation showed an inverse relationship between the presence of a 10% or greater CH50 gradient and subsequent graft function (Kux et aZ., 1974). In the only instance of hyperacute rejection in their study, there was a very large gradient. Caution should be exercised in interpreting these findings, since such gradients, although likely to be due to complement consumption in the allograft, may also be induced by release of anticomplementary substances from the allograft as a result of damage. Winn et al. (1973) have demonstrated the importance of complement in antibody-mediated rejection of rat to mouse skin xenografts. The grafts were induced to heal in by pretreatment of the recipients with antilymphocyte serum, and by thymectomy. Established grafts were rejected rapidly after intravenous administration of mouse anti-rat serum. This effect could be prevented completely by pretreatment of the recipients either with F(ab’), fragments of rabbit anti-rat serum or with chicken anti-rat serum, which was shown to be unable to activate mouse complement. Partial or complete protection of the grafts could also be obtained by depleting recipients of complement with CoF or with heataggregated 7-globulin. Further studies with congenic mice, whose only known difference was at the Hc locus, which controls C5 levels, showed that markedly larger doses of antiserum were required to induce rejection in the C5-deficient strain than in the strain with normal C5 levels, and that the tempo of rejection was slower and resulted in a higher spontaneous recovery rate in the C5-deficient mice. It is quite clear from these observations that complement can be of major importance as a mediator of the effects of humoral immunity in allograft rejection; however, how complement induces these lesions is less clear. Many possible mechanisms have been described in other systems (Henson, 1974; Cooper and Cochrane, 1974). Complement may act directly, causing cytolysis or it may induce a variety of functional vascular alterations through either the direct actions of by-products of


21

the complement cascade (such as anaphylatoxin) or the actions of such by-products on other cells (such as platelets and mast cells) which release vasoactive substances ( Ruddy, 1974). Similarly, by-products of the complement system can induce structural lesions by inducing intravascular coagulation ( Austen, 1974) or the release of destructive enzymes from PMNs (Henson, 1974). Equally, some of these by-products are directly chemotactic for macrophages and PMNs and induce phagocytosis (Ruddy, 1974). As discussed in Section II,A, when biopsies from rejecting transplants are reviewed in detail, a variety of injury patterns are evident; therefore, various primary and secondary mediators are IikeIy to be important in their production. In discussing possible mechanisms by which complement may be instrumental in the induction of allograft damage, we rely for most information on systems other than transplantation and must presume that the immunopathologic mechanisms are universal and consequently important in allograft rejection. a. Direct Cytolysis. This effect of complement is well established in in vitro systems, such as the lysis of antibody-coated erythrocytes and lymphocytes, and is a consequence of the sequential activation of the complement system through C9. This phenomenon is the basis of in vitro assays for the various complement components and the microlymphocytotoxicity assay. The correlation between vascular endothelial lesions occurring within 2 hours of renal allografting and the CH50 gradient across the graft (Kux et al., 1973), although suggestive of direct complement-induced cytolysis involving the terminal complement components, does not exclude indirect actions of complement such as may be produced by PMNs as a result of activation of the earlier complement components ( C S C 7 ) . Similarly, abrogation of the activation of complement by C3 depletion with CoF does not allow differentiation of the direct cytolytic and indirect effects of complement, since activation of C3 is the key step in all complement-mediated lesions. The studies of Winn et al. (1973) provide evidence against direct complemcnt-mediated cytolysis in skin xenograft rejection. The experimental protocol has already been outlined earlier in this section. Adequate levels of complement were unable to induce detectable graft damage after the injection of antigraft serum if circulating PMNs were drastically reduced. In short, although it is possible that complement can directly cause cytolysis in allografts, there is no direct evidence to support this belief. Further, one should be cautious in extrapolating the results of in uitro complement-mediated cytotoxicity to in vivo situations, because alloanti-

22

CARPENTER, D’APICE, AND AEBAS

bodies can rarely produce efficient lysis with autologous complement; rather, heterologous serum sources are generally required. Ferrone et al. ( 1971) suggested that it is the presence of “natural” antibodies against polyspecific target antigens in the heterologous serum that augments the lytic activity by increasing the degree of target cell sensitization. It is possible also that species differences in the efficiency of complement fixation play a role; in any event, heterologous complement is not present in uivo. b. Stimulation of PMNs. In a previous section, the evidence that PMN infiltration is a marker for the involvement of humoral immunity in rejection was discussed. There is considerable evidence from other systems that by-products of complement activation are capable of inducing their presence, In particular, C3a, C5a, and complexed C567 are chemotactic for neutrophils. The subject has been reviewed recently (Henson, 1974; Becker and Henson, 1974). Once localized by these chemotactic factors or other factors analogous to migration inhibitory factor ( M I F ) , which are demonstrable by the in vitro leukocyte migration technique, other by-products of complement activation can induce their adherence (C3b), phagocytosis ( C3b and C577), and degranulation (C3b). PMNs also possess receptors for the Fc portion of IgG, and it is possible that they may be activated by this mechanism ( Becker and Henson, 1974). Lindquist et a2. (1971) have shown that PMNs can induce allograft damage by interposition of cellular processes between and beneath endothelial cells with consequent stripping and exposure of the basement membrane. Further evidence of their importance is present in the previously mentioned report of Winn et a2. (1973). In their skin xenograft model, antigraft antiserum rapidly induced graft destruction; however, this could be prevented or considerably reduced if the number of circulating PMNs was reduced by administering anti-PMN serum or nitrogen mustard therapy. They suggest that the major role of complement lies in its ability to generate chemotactic factors for PMNs, which then induce tissue damage. c. Platelets and Intravascular Coagulation. The adherence, aggregation, and release reaction of platelets depends on the presence of a receptor for activated C3. Platelets from rabbits, mice, rats, guinea pigs, cats, dogs, and horses have C3 receptors, whereas those of humans, baboons, pigs, oxen, sheep, and goats do not, but have instead an Fc receptor that is not present on platelets of those species having C 3 receptors. The subject has recently been reviewed by Henson (1974) and by Becker and Henson ( 1974). Briefly, C 3 receptor-bearing platelets bind to bound C3-C9, resulting in their lysis and release of their contents, or they may release granule contents by a secretion process without lysis. This


23

secretion response is dependent on C3 but does not require the participation of the terminal components. Further, platelet aggregation can be induced by ADP released from platelets by either mechanism. In the human, however, aggregation and release are induced by immune complexes via the platelet Fc receptor, but complement activation appears to inhibit these functions, probably in part owing to steric hindrance of the Fc receptor by Clq, which binds to a site on the complexed IgG adjacent to the Fc site. The interrelationships of complement, coagulation, fibrinolysis, and the kinin system have recently been reviewed by Austen (1974) and Brown (1974). It would appear that complement participates via the induction of platelet factor 3 (PF3 ) release by C3 receptor-bearing platelets and by the induction of tissue damage by PMNs resulting in Hageman factor activation; furthermore, complement activation may be amplified by activation of C1 and C 3 by plasmin. Whether this information, obtained in nontransplant, largely in vitro systems, also applies to allotransplantation is unknown.

2. Antibody-Dependent Cell-Mediated Cytotoxicity This immune phenomenon is characterized by an in vitro reaction in which target cells are lysed by a nonsensitized lymphoid cell population in the presence of target cell-specific IgG antibody, the reaction being independent of complement. Many synonyms and abbreviations have been used for this system, including antibody-dependent cell-mediated cytotoxicity ( ADCMC, ADCC, or ACMC ), antibody-dependent lymphocyte-mediated cytotoxicity ( antibody dependent LMC, AbLMC ) , antibody-mediated cell-dependent immune lympholysis ( ABCIL) , and lymphocyte-dependent antibody assay ( LDA assay). ACMC has to date been demonstrated by in vitro methods only; however, these in vitro demonstrations appear highly likely to represent an in vivo effector mechanism, as will be discussed. ACMC has recently been fully reviewed (Perlmann et al., 1972; MacLennnn, 1972; Perlmann, 1975). In brief, the effector cell, called a K cell, appears to be either a lymphocyte or a nonphagocytic, non-glassadherent monocyte (Perlmann and Perlmann, 1970; Campbell et al., 1972; MacLennan, 1972; Greenberg et al., 1973). I t is non-immunoglobulin-bearing ( Greenberg et al., 1973) and thymus-independent ( Harding et al., 1971; Campbell et al., 1971; Lamon et al., 1972; Van Boxel et al., 1972; Britton et al., 1973), but has a receptor for a site on the Fc portion of IgG (Gelfand et al., 1972; Larsson and Perlmann, 1972; Moller and Svehag, 1972). The K cell is present in normal human and animal peripheral blood, lymph nodes, and spleens, but in varying numbers (Mac-

24


Lennan and Harding, 1970). The antibody that is capable of inducing K cell cytotoxicity is IgG ( MacLennan et al., 1969, 1970; Perlmann and Perlmany 1970; Wasserman et al., 1971; MacLennan and Howard, 1972), and indirect evidence suggests that all four human IgG subclasses are active (Larsson et al., 1973; MacLennan et al., 1973). The first step in the reaction is binding of antibody to the target cell which presumably results in alteration in the conformation of the Fc portion of the IgG molecule (Givol et al., 1974), such that it is recognized and bound by the K cell via its Fc receptor (Gelfand et al., 1972; Larrson and Perlmann, 1972; Moller and Svehag, 1972). The mechanism of lysis appears in all the ways so far examined to be identical to that produced by sensitized T cells (Strom et al., 1973a; Carpenter et al., 1974). The assay is a highly sensitive method of detecting antibody, and activity can be detected in some sera to dilutions as high as (Perlmann et al., 1972). There is some, although as yet incomplete, evidence that ACMC is involved in allograft rejection. As mentioned, K cells are present in normal human peripheral blood, and they are also present in reduced numbers in the blood of uremic subjects awaiting renal transplantation (dApice et al., 1974). The sera of many of these patients contain antiHLA antibodies which are capable of mediating the reaction. Ting and Terasaki (1974), Jeannet et al. (1975), and unpublished studies from our own laboratory suggest a poor transplant prognosis if specific antigraft antibodies are detected by ACMC in sera of patients pretransplantation. Antigraft antibody cannot be detected in these sera by the regular complement-dependent microlymphocytotoxicity assay. It is obvious that while suggestive for a role of ACMC in rejection, this type of study cannot provide the final answer, because such a strong correlation can be interpreted as simply reflecting a highly sensitive method of detecting presensitization in general, and that the rejection process itself is mediated by any of a variety of cellular or humoral mechanisms. More directly supporting the involvement of this mechanism in rejection is the demonstration that K cells are present in rejecting allograft tissue, both in man and rats, and that they are capable of inducing lysis of target cells similar or identical to those of the donor in the presence of antigraft antiserum (Tilney et al., 1975; Strom et al., 1975a). The same effect was obtained in three of four experiments when the human recipients’ own sera were used (T. B. Strom and C. B. Carpenter, unpublished data). The most convincing evidence of the in vivo relevance of the system has been provided by Hersey (1973a,b) using a syngeneic tumor model in rats, which can be protected against lymphoma development by passive immunization with antitumor antiserum. The


25

effect correlated well with the in uitro ACMC titer of the antiserum, but was shown to be independent of complement-mediated cytotoxic antibody or cytophilic antibody. Forman and Britton (1973) have shown that the in uitro reactivity of lymphoid cells from mice immunized against an allogeneic lymphoma varies in certain characteristics with time after immunization. The cytotoxic cells present in spleens harvested late (7-14 days) after immunization show the typical characteristics of thymus-dependent killer cells in that they are antitheta serum sensitive and their cytotoxicity is inhibited by antitarget antiserum. Spleen cells harvested early (3-5 days) after immunization are also directly cytotoxic for specific target cells; however, at this time the cytotoxic effect is not reduced by antitheta serum and complement treatment of the population and is enhanced by antitarget antiserum. Several alternative explanations for this effect are available. The authors favor the K cell, whose functional characteristics fit those found, as the early cytotoxic cell. Naturally present K cells could be induced to cytotoxic activity by antigen-antibody complexes that bind to the Fc receptor. On transfer to the in uitro system, these K cell-bound complexed antibodies bind to the target cells, either by virtue of uncomplexed antigen binding sites, or by a process of equilibration of antibody binding between the complexed non-cell-bound antigen and target cell antigens. This explanation hinges on the presumption that sufficient quantities of IgG antibody against the tumor would be present as early as 3 days after primary immunization. An unlikely but theoretically valid alternative to this interpretation was pointed out by the authors. It is possible, instead of indicating that two distinct cell types are involved, that one cell type (presumably a T cell) is involved, but it acts via different cytotoxic mechanisms at different stages of its maturation. A further alternative explanation is that the early cytotoxic effect is due to specifically sensitized B cells analogous to those described in patients with bladder carcinoma ( OToole et al., 1974). The increase in cytotoxicity with antitarget antiserum would then be explained as K-cell cytotoxicity. These cytotoxic B cells have not been fully characterized and may be found to be normal K cells activated by antibody secreted by small numbers of specifically sensitized B cells. Although Forman and Britton (1973) did not demonstrate this mechanism late after immunization, they did not exclude its continuance. It should be possible to separate the contributions of T cells and K cells at this stage by the effects of adding irrelevant antigen-antibody complexes to the in uitro assay, since these inhibit ACMC by blocking the K-cell Fc receptors, but do not affect direct lymphocyte-mediated cytotoxicity.

26


It is evident that current knowledge does not allow any definite conclusion about the role of ACMC in allograft rejection; however, intensive research in this area may yield definitive data about an extremely interesting phenomenon, and one that provides a direct link between antibodies and cellular immunity. 3. Antigen-Antibody Complexes

Immune complexes of various antigens and their antibodies may induce allograft damage. In previous sections we have considered the possibility of immune complex-induced allograft damage, both when the lesion represents a recurrence of the original disease and when it is a new development in response to immunization by extrinsic agents. There remains the possibility of immune complexes formed of graft histocompatibility antigens and host antibodies inducing allograft damage, a mechanism that may be properly considered as part of the rejection process. The suggestion that such complexes are involved in rejection stems largely from the frequent finding of granular deposits of immunoglobulin and complement in glomeruli of renal allografts. The report of Andres et al. (1970) is a typical example. Ten of 34 biopsies from renal allografts showed granular immunoglobulin deposits in glomerular capillary walls. Granular deposits have been shown in models of experimental nephritis to represent immune complexes ( Dixon et al., 1961) , The antigen present in these deposited complexes has not been definitely identified as histocompatibility antigen; however, its high incidence would make extrinsic antigens less likely, unless immunosuppressed transplant recipients are unusually susceptible to complex formation. Little work has been done that would prove the presence of circuIating immune complexes in graft recipients. Some indirect evidence is available from two sources. First, the previously discussed findings of Forman and Britton (1973) of the development of specific K-cell cytotoxicity early after grafting suggests that these cells have been activiated by complexes of histocompatibility antigen and host antibody. Second, studies in this laboratory have shown that renal allografted rats develop circulating anti-graft antibody shortly after transplantation; however, the titer falls rapidly at 8-9 days. A simple approach to this question would be to test the capacity of these sera to block the ACMC assay, which can be used as a test for immune complexes. The mechanism by which immune complexes induce damage has been studied in nontransplant systems and has been the subject of a recent review (Cochrane and Koffler, 1973).


27

Ill. The Protective Effects of Antibody in Allotransplontation

The discovery was made almost 15 years ago that antibodies were not merely effector mechanisms of the immune system, but also played a regulatory and/or suppressive role in the immune response (reviewed by Uhr and Moller, 1968). Kaliss, Snell, and their associates extended this concept to allografts of tumors, and the phenomenon of immunologic enhancement was established (Kaliss, 1969). In the last decade it has been shown that allografts of normal tissues can also be enhanced; i.e., the rejection response can be prevented by antibody directed against donor histocompatibility antigens. The hope that immunologic enhancement may be a specific form of immunosuppression for the human transplant recipient has stimulated an intense investigative effort in the field (for reviews, see Feldman, 1972a,b; Voisin, 1971). In this section, we shall review the evidence relating to the role of immunologic enhancement in prolonging the survival of normal organ allografts. The data derived from experimental work in tumor systems will not be considered in any detail, since the reviews mentioned above deal with this topic comprehensively. Our aim will be to discuss the role of enhancing antibodies in the survival of normal tissue allografts, to define their biologic and functional properties, to identify the factors that favor their development, and to discuss the mechanism(s) by which they mediate their protective influence on allografts. Stuart et al. (1968) were the first to report successful enhancement of renal allograft survival in inbred rats. LBN F, hybrid kidneys were transplanted across a major histocompatibility barrier ( AgB 1/3 to 1 / 1 ) into bilaterally nephrectomized Lewis rats. The recipients were treated with an intravenous injection of loR viable donor spleen cells one day before transplantation and with 1 ml of a hyperimmune Lewis anti-BN (spleen and lymph node) serum ( hemagglutinin titer at least 1:2048) 2 hours before and 1 hour after grafting. Graft survival was significantly prolonged in 10 out of 10 treated animals, and 5 of these were alive and well 164-231 days after transplantation (mean survival time in untreated control rats being 17.3 +- 2.7 days). French and Batchelor ( 1969) convincingly demonstrated that such a protective effect could be induced by alloantiserum treatment alone. Thus, August or (August X AS) F, kidneys transplanted into AS rats were significantly enhanced by AS anti-August serum ( 50%lymphocytotoxicity titer of 1:256) administered immediately after tranplantation (1 ml), foIlowed by 1 ml at 24 hours and 0.5 nil 48, 72, and 96 hours aftcr grafting. In our own studies ( Abbas et al., 1974a; Strom et al., 197513) indefinite prolongation of kidney

28


graft survival in the LBN F, to Lewis model has been induced by a single intravenous injection of 0.5 ml of hyperimmune Lewis anti-BN serum (50%lymphocytotoxicity titer of 1 :850) given to recipients immediately after transplantation, Although kidney tranplants in inbred rats have proved to be particularly susceptible to antibody-mediated enhancement, the phenomenon has been extended to other species and tissues as well. Renal allografts can be enhanced in outbred dogs (R. E. Wilson et al., 1971), pigs (Calne, 1971), and rabbits (McDowall et al., 1973, Holter et al., 1973); heart allografts can be enhanced in rats (Tilney and Bell, 1974; Davies and Alkins, 1974) and mice (Judd and Trentin, 1973). Also subject to enhancement are skin allografts in rats (Zimmerman and Feldman, 1969a, 1970) and mice (McKenzie and Snell, 1973; Staines et al., 1974) and liver allografts in baboons (Myburgh and Smit, 1972; Smit and Myburgh, 1974). In addition, the available evidence suggests that the prolonged survival of ovarian and testicular grafts in rats and mice may be attributable to immunologic enhancement ( Parkes, 1958; Cock, 1962; Moller, 1964). This is not a comprehensive list of all reported instances of enhancement, and additional experiments will be referred to in subsequent discussions. Primarily we have attempted, wherever possible, to quote the most recent work in which compelling evidence for enhancement as the only or major mechanism responsible for prolonged allograft survival has been presented. The general conclusion is that antidonor antibodies clearly are capable of prolonging survival of diverse tissues in a wide variety of species. Immunologic enhancement can be induced by active immunization of recipients with donor strain histocompatibility antigens or cells (“active enhancement”) or by treating recipients with antibodies directed against the donor’s histocompatibility specificities (“passive enhancement”). In other instances, prolonged graft survival occurs without antigen or antibody treatment, a phenomenon termed “spontaneous enhancement” or “autoenhancement,” when grafts are performed between rats incompatible only at the minor histocompatibility loci (White et al., 1969) or with certain major incompatibilities when recipient nephrectomy is delayed, (Morris and Lucas, 1971). Studies of active and spontaneous enhancement must, however, be interpreted cautiously, because in many instances there is no direct evidence that the long-term prolongation of graft survival is antibody mediated. Thus, the enhancement effect cannot be transferred with serum from the allograft recipient (Zimmerman, 1971 ) , and spontaneous, prolonged survival of kidneys transplanted between at least some rat strains (such as AS and AS2) may be due to classical tolerance (Salaman et al., 1971). In fact, the relationship


29

between enhancment and tolerance is a source of some controversy among transplantation biologists ( Levey, 1971, 1972). Tolerance, as originally defined by Medawar (1956), is the specific and systematic failure of the mechanism of immunologic responsiveness; such a definition may be too limited, since the evidence is increasing that specific antibody may, in certain instances, be responsible for maintaining the “unresponsive” state (Asherson et al., 1971). Although the mechanisms responsible for tolerance and enhancement are not necessarily mutually exclusive, it appears rational to consider enhancement, at least in the context of organ allografts, as an antibody-mediated protective effect that is either induced and/or transferable by serum alone, and (for reasons that are discussed below) that occurs in the presence of a positive cellular and/or humoral immune response against donor strain antigens. It is, however, important to realize that there may be instances of enhancement when no specific enhancing antibody can be identified with presently available techniques, or when the recipient does not appear to manifest a detectable immune response to the histoincompatible graft. In such cases, one can only conclude that, although true immunologic enhancement may be operative, the possibilities of cell-mediated suppression or deletion of the relevant clones of immunocompetent cells cannot be ruled out and that more evidence is required to establish the mechanism( s ) of prolonged graft survival. In any event, it is clear that antidonor antibodies can effectively prolong the survival of not only kidney grafts, but also grafts of skin, heart, and possibly endocrine organs as well. The extent to which such antibodies maintain a state of transplantation “tolerance” is not precisely defined ( Elkins et al., 1974). AND BIOLOGIC PROPERTIES OF A. PHYSICOCHEMICAL ENHANCING ANTIBODIES

1. C h s and Subclass of Antibody The detailed physicochemical characteristics of enhancing antibodies have not been defined in terms of isolation of functionally effective molecules in a pure form. Moreover, it is quite conceivable that different antibody classes and/or subclasses may possess enhancing activity in different species and experimental models. The majority of the available evidence indicates that enhancing antibodies belong to the 7 S IgG class. Thus, the IgG fraction of hyperimmune antidonor serum prolongs renal allograft survival in rats (Hildemann and Mullen, 1973a,b; Strom et al., 1975b; Ippolito et al., 1974), skin allografts in rats (Zimmerman and Feldman, 1969a) and guinea pigs ( Nelson, 1962), and liver allografts in baboons (Smit and Myburgh, 1974). Subclass analysis of enhancing

30


antibodies has been achieved mainly in mouse tumor allograft systems, and the bulk of the evidence indicates that the activity resides largely in the electrophoretically slow moving, complement-fixing IgG, subclass. The homolog of mouse IgG, antibody possess effective enhancing properties in the rat skin graft model (Jones and Feldman, 1971), but similar information for kidney enhancement is lacking, Hildemann and Mullen (1973a,b) have shown that the IgM fraction of hyperimmune Lewis anti-Buffalo serum reduces the survival of Buffalo or F, kidney grafts in Lewis rats, but this observation remains unconfirmed. In our own studies in the LBN F1 to Lewismodel, IgM anti-BN antibodies had no detectable effect on graft survival or morphology ( Strom et aZ., 1975b). The recent observation that IgM alloantibodies in the rat do not fix rat complement (M. Ruszkiewicz and J. F. Mowbray, personal communication, 1975) and that some anti-HLA IgM antibodies block in uitro mixed lymphocyte reactivity (Colombani et al., 1973) suggest that 19 S IgM antibodies cannot be ruled out as having enhancing activity. Moreover, Fuller and Win (1973) have shown an enhancing activity of IgM antibodies in a mouse tumor model. However, it is unlikely that IgM plays a major role, since it is generally agreed that enhancing antisera can be produced only by hyperimmunization, which would be expected to result in the preferential formation of high-affinity 7 S IgG antibodies. Finally, it is possible that the graft protection is mediated not by free antibodies themselves, but by soluble immune complexes formed with histocompatibility antigens released from the graft. It is known that complexes of particulate antigens, such as sheep erythrocytes (SRBC) and anti-SRBC antibody, can suppress the humoral immune response to SRBC (Rowley et al., 1973). Hellstrom and Hellstrom ( 1974) have shown that immune complexes also prolong the survival of allografted tumors by competing with effector cells. Technical problems in assaying for the presence of circulating or cell-bound immune complexes have made it difficult to demonstrate directly their relationship to enhancement of organ allograft survival. 2. Biologic Properties of Enhancing Antibodies Hyperimmune sera capable of inducing immunologic enhancement frequently also possess agglutinating and cytotoxic activities against donor cells. It is, however, obvious that all such biologic activities need not reside in the same antibody molecules. Lucas et al. (1970) found no correlation between the ability of Lewis anti-BN lymphocyte sera to enhance LBN F, to Lewis graft survival and their titers of hemagglutinating or cytotoxic antibodies. Fabre and Morris (1972a), in contrast,


31

observed some correlation between cytotoxic titers and the kidney allograft enhancing ability of Lewis anti-DA antisera, but these results were based on testing only three sera and are therefore inconclusive. The demonstration by Enomoto and Lucas (1973) that splenectomy prevents the enhancement of LBN F, kidney grafts in Lewis recipients without blocking the appearance of cytotoxic or agglutinating antibodies lends support to the idea that enhancing activity may reside in a different antibody population. In fact, unmodified Lewis recipients of LBN F, renal allografts and Lewis animals hyperimmunized with BN lymphoid cells both respond by the production of anti-BN antibodies that are lymphocytotoxic in vitro in the presence of heterologous complement ( Abbas et aZ., 1974a). However the distribution of activity among IgG and IgM classes is radically different, virtually all serum antibody associated with acute rejection being of the IgM class whereas hyperimmune serum, which induces enhancement of graft survival, is predominantly IgG. Clearly, therefore, cytotoxicity in vitro does not reflect per se the in vivo biologic activity. The presence of cytotoxic activity in IgG-containing enhancing sera implies that such sera are complement fixing, and this fact is confirmed by the majority of studies aimed at directly investigating this aspect (Feldman, 1972a). Voisin (1971) has reviewed the contrary evidence, which indicates that enhancing sera are noncomplement fixing. Furthermore, the ability of pepsin-digested F( ab’) ’’ fragments or succinylated IgG anti-donor antibody to protect allografts from immunologic injury (Habal et al., 1973; Holter et al., 1973; Smit and Myburgh, 1974) suggests that when noncomplement-fixing fragments of enhancing antibodies are bound to the graft they may be effective in prolonging graft survival. Although the question regarding complement fixation can be resolved only when purified enhancing antibodies can be studied, the available data suggest that the induction of immunologic enhancement by immunoglobulin is probably not related to the simple property of its being noncytotoxic because of an inability to bind complement. In vitro assays of enhancing sera in many instances have demonstrated that they possess “blocking” activity; i.e., they are capable of blocking undirectional mixed lymphocyte reactions between donor and recipient cells or the cytotoxic effect of recipient lymphocytes on donor strain target cells (Phillips et al., 1973). Thus a correlation has been demonstrated between activity and effectiveness in inducing enhancement of renal allografts (Ippolito et al., 1974) and cardiac allografts (Gordon et al., 1971) in rats. Rao et al. (1974) have shown that immunization of mice with soluble alloantigen results in the preferential production of sera that enhance skin graft survival in vivo and that block mixed

32

CARPENTER, D’APICE, AND AFJBAS

lymphocyte reactivity in vitro, while both activities are correspondingly low in sera prepared by immunization with intact allogeneic cells. However, blocking activity is not consistently demonstrable in all sera used to induce enhancement ( Larner and Fitch, 1973; Guttmann, 1973), but in many such instances the factor ( s ) that blocks cellular immunity against donor cells appears subsequently in the enhanced recipient’s circulation. Passively enhanced recipients contain, in their sera, a factor( s ) that completely suppresses the cytotoxicity of recipient spleen lymphocytes against Wr-labeled BN target thymocytes (Strom et al., 1975b). Biesecker et al. (1973a) have demonstrated blocking activity in hyperimmune sera as well as in sera from long surviving, enhanced graft recipients. It is worth nothing, that, in all reported instances when allograft survival has been immunologically enhanced and a search has been made for blocking factors, such factors have been detected in the recipients’ circulations and/or in the grafted kidneys ( Biesecker et al., 1973b; Stuart et al., 1971). In fact, the presence of blocking factors in the sera or kidneys of long-surviving human transplant recipients is considered suggestive evidence that spontaneous immunologic enhancement may be the mechanism responsible for survival (Quadracci et at., 1971, 1974; Garovoy et al., 197313; Hattler and Miller, 1973). Although such associations between immunologic enhancement and the presence of factors that block cellular immunity against donor alloantigens suggest a causal relationship, they do not prove that the mere presence of blocking factors is sufficient to induce and/or maintain the enhanced state. Serum blocking activity per se may be a correlate for enhancing antibodies, but definitive proof of this relationship awaits more precise identification of the antibodies that possess these in vitro and in vivo biologic properties. 3. Specificity of Enhancing Antibodies

Enhancing antibodies are clearly specific for donor histocompatibility antigens, since antisera directed against a strain protect only allografts from that strain. This has been demonstrated by many investigators in inbred rats and mice and is applicable to allografts of skin, kidney, and heart. Although cross-reactivity of enhancing sera between certain rat strains has been reported (Zimmerman and Feldman, 1969b; Fabre and Morris, 1974b), it is not clear to what extent these strains share histocompatibility antigens, especially with respect to nonserologically defined specificities. More controversial is the question of which of those antigens coded for by the major histocompatibility complex are crucial in the induction of enhancement. The major histocompatibility system of the mouse has been studied


33

extensively in the last decade and serves as a presumptive model for the corresponding systems of other mammals. Briefly, the gene complex consists of two allelic series of genes coding for antigens which are “serologically defined,” the K and D region genes, and a number of other genes whose products are often not identifiable by serologic means but are detected by “lymphocyte defined” assays (such as mixed lymphocyte and graft-versus-host reactivity, histocompatibility-linked immune responses) (Klein and Shreffler, 1971). Corry et al. (1973) have shown that incompatibilities restricted to the H-2D region, the H-2K region, and non-H-2 loci all lead to acute rejection of cardiac allografts. It is of interest that the only strain combination in which heart allografts showed prolonged survival was B1O.BR to B6AFI, and these strains are identical at the H-2K and the adjacent 1 region. Such an observation suggests that some antigenic differences may be more important than others in the induction of rejection and, by inference, of enhancement as well ( McKenzie and Snell, 1974). The I region has been the focus of considerable recent interest among immunologists and transplantation biologists ( Shreffler and David, 1975). Antigens coded for by genes in the I region of the mouse are expressed on lymphoid cells, but not on erythrocytes (David et al., 1973). Davies and Alkins (1974) have shown that absorption of antidonor serum with donor erythrocytes does not reduce its ability to enhance cardiac allograft survival in rats, even though cytotoxic antibodies are removed. This observation has been corroborated in the case of skin graft enhancement in congenic mice (Staines et al., 1974). Such studies indicate that the target antigens for enhancement may well be antigens other than the serologically defined histocompatibility antigens, but confirmation of this point must await studies with large numbers of congenic strains with precisely defined antigenic differences. Moreover, since it is clear that different tissues or organs show a variability in their expression of histocompatibility-linked antigens, it is quite conceivable that the target ( and, therefore, the inducing) antigens for enhancing antibodies may vary depending on the type of graft. Finally, it has been claimed that enhancing antibodies must be directed against all the major histocompatibility differences between graft donor and recipient ( Davis, 1973; Moller, 1963). Studies such as those quoted above using enhancing sera absorbed with erythrocytes are potent arguments against this dictum. The observation that immunization with bone marrow cells from F, and backcross animals can effectively enhance survival of LBN F, to Lewis kidney grafts (Guttmann et d., 1972) also suggests that only a few antigens (which are expressed in the F, generation and the backcrosses ) are sufficient to induce enhancement.

34


Moreover, Fabre and Morris ( 197413) have shown that (Lewis X AS2) F, to DA kidney grafts can be enhanced with either DA anti-AS2 or DA anti-Lewis serum demonstrating that antisera against only some of the graft antigens effectively protect against rejection. Obviously, there must be at least some antigenic differences between donor and recipient that are of critical importance in determining the fate of allografts, i.e., rejection or enhancement. The precise quantitation of such antigens, their cellular expression, and their definition in terms of serologically or lymphocyte-defined loci of the major histocompatibility complex will lead to a clearer understanding of the fundamental mechanism of enhancement.

4. Antireceptor Antibodies A recent development which may be pertinent to the modulation of immune responses to foreign histocompatibility antigens is the demonstration that anti-idiotypic antibodies develop after immunization and that such antibodies may have activity directed against the antigen receptors on precursors of antibody-producing cells (Ramseier and Lindemann, 1971). Since the combining regions of the lymphocyte antigen receptor and the antibody produced are the same (Makela, 1970), antiidiotypic antibodies against the immunoglobulin produced by a cell are also antireceptor antibodies ( ARA ) . McKearn ( 1974) has shown that LBN F, hybrids injected with Lewis anti-BN alloantibody produce an ARA (since the alloantibody possesses unique idiotypic determinants ) that inhibits the response of Lewis cells to BN histocompatibility antigens. McKearn et al. (1974a) have also demonstrated that multiple idiotypic specificities might exist and that antibodies to each can be produced by repeated immunization with BN alloantigens. In the allograft recipient, such ARA may suppress the responses of immunocompetent cells to the allograft antigens by directly blocking the specific antigen receptors. In fact, maximal prolongation of kidney graft survival is observed when transplantation is delayed until levels of circulating anti-idiotypic antibodies reach their peak ( McKearn et al., 197413). Although the potential significance of ARA in determining the fate of allografts is indisputable, their role in the induction and/ or maintenance of enhancement after active or passive immunization has not yet been established. Thus, the anti-BN antibody we have used to induce enhancement of LBN F, to Lewis renal allografts (Abbas et d.,1974a) does not possess antireceptor activity ( F. P. Stuart, personal communication). I t is necessary, however, to examine enhanced graft recipients at frequent intervals for titers of circulating ARA, and such studies are yet to be performed in detail.


35

B. INDUCTION OF ENHANCEMENT: KINETICSAND REGULATORY FACTORS 1. Kinetics of Active Enhancement When prolongation of allograft survival is induced by active immunization with donor alloantigens, a critical degree of timing is necessary. The majority of studies with active enhancement of renal allografts in dogs and rats indicate that antigen pretreatment of donors must be given at least 5-7 days prior to grafting (Fabre and Morris, 1972b,c; Kim et al., 1972; Ockner et al., 1970a,b; von Haefen et al., 1973; Rivard et al., 1973; Zimmerman, 1971; R E. Wilson et al, 1971; Guttmann, 1973). LBN F, to Lewis renal allografts may show modest prolongation of survival when recipients are treated with large doses of BN splenic antigen only 1 day prior to transplantation, but in such studies the evidence for enhancement is incomplete since the protective effect cannot be transferred by recipients’ sera, and no enhancing antibody is demonstrated (Kim et al., 1972; von Haefen et al., 1973). Since critical timing and dosage of antigen pretrcatment are prerequisites for inducing active enhancement, the crucial factor is how the host’s immune system has been “primed at the point at which it encounters the histoincompatible graft. Thus, one might postulate that the dose of antigen and the duration of pretreatment may specifically promote the development of enhancing antibodies rather than cellular immunity, which might accelerate rejection. The demonstration that enhancing activity is detected in the serum 2 days after immunization, and that pretransplant splenectomy prevents the development of active enhancement after antigen administration also suggest that it is not a paralysis by antigen treatment itself, but the host’s immune response to the antigen that determines the fate of the renal graft (Enomoto and Lucas, 1973). In the final analysis, it is not clear why antigen pretreatment leads to enhancement in some situations and hyperacute rejection in others, Such a problem can be resolved only when enhancing antibodies are precisely identified, enabling one to examine their evolution after immunization in comparison with the development of cytotoxic and other forms of alloantibodies. 2. Kinetics of Passive Enhancement

When enhancement of graft survival is induced by passive serum transfer, a single injection of high-titer serum immediately after transplantation is generally sufficient to induce prolonged graft survival (Abbas et al., 1974a; Fabre and Morris, 1973; French and Batchelor, 1972; Zimmerman and Feldnian, 19694. Earlier studies on enhancement employed large doses of serum and prolonged courses of administration

36


(French and Batchelor, 1969; Fabre and Morris, 1972a), but it is now evident that a single dose is equally effective. Moreover, Fabre and Morris (1973) have demonstrated that 50 ~1 of enhancing serum produces effects that are comparable to 4 ml in the (DA x Lewis) F, to DA renal allograft model, and in our own studies we have employed only 0.5 ml of an enhancing serum that has a 50% cytotoxicity titer of 1:850 against donor cells ( Abbas et al., 1974a). Since passively administered IgG, which is the functionally active component of enhancing serum, has a half-life of only about 5.5 days in the rat and 4.5 days in the mouse (Waldmann and Strober, 1969), one would not expect this IgG to be directly responsible for long-term protection even if allograft recipients had normal rates of catabolism of antidonor antibodies. Consistent with this is the observation that the sera of enhanced allograft recipients (given a single dose of enhancing serum immediately after grafting) contain no detectable antidonor antibody for the first 2-4 days, and such antibodies are detected by day 4-6 after transplantation (Strom et al., 1975b; Biesecker et al., 197313). One can therefore suggest that “passive” enhancement of graft survival may be due to endogenously synthesized antibodies, and that passive transfer of enhancing serum might somehow promote this synthesis. The presence of “blocking factors’’ in successfully enhanced animals (Strom et al., 1975b) certainly is indicative of active synthesis of some entity. Crucial to this concept is the demonstration that prolonged graft survival occurs without either antigen or serum pretreatment if recipient nephrectomy is delayed (Morris and Lucas, 1971) or if the antigenic incompatibilities are minor (White et al., 1969), suggesting that graft recipients are potentially capable of synthesizing enhancing factor(s) de novo. Ippolito et al. (1974) have confirmed this elegantly by demonstrating that 21-day sera from Lewis rats rejecting Buffalo kidney grafts can effectively prolong Buffalo to Lewis graft survival when administered to another recipient, yet 7- or 14-day sera do not possess enhancing activity. Retrospective studies of human transplant recipients have also demonstrated the development of serum blocking factors in long-term survivors (Quadracci et al., 1974; Garovoy et al., 197313). One can therefore conclude that ( 1) allograft recipients, especially with respect to kidney grafts, are capable of producing their own enhancing antibodies, provided early rejection is overcome; ( 2 ) endogenous synthesis of enhancing factor (antibody) in detectable titers is a relatively slow process, requiring days or weeks in the rat and perhaps longer in humans; and ( 3 ) passive transfer of antidonor serum (and immunization with donor antigens) may promote a more rapid synthesis of blocking antibodies, and thus favor graft survival,


37

3. Regulatory Influences on Enhancement Induction Ever since the earliest experiments on immunologic enhancement of allograft survival it has been clear that enhancement is induced more easily and consistently in some situations than in others. Although the explanations for such differences are still largely unclear, it is obviously important to identify the major regulatory factors and distinguish situations in which they appear from those which promote rejection. In the discussion that follows, we have defined the regulatory influences that are well substantiated, without making any attempt to assign an order of significance. a. Species. Studies on renal allografts have most clearly demonstrated that the inbred rat is more susceptible to enhancement induction than any other species. Thus, although kidney grafts can be enhanced in dogs (R. E. Wilson et al., 1969, 1971) and rabbits McDowall et at., 1973; Holter et al., 1973), the duration of survival is significantly less than that achieved in rats, and often small doses of immunosuppressants are required in addition. This may simply be a reflection of the greater degree of inbreeding that has been achieved in rats, so that histocompatibility differences between donor and recipient are well defined and limited. Until recently, the rat has been the only highly inbred species in which organ transplantation ( other than skin ) was technically feasible, but studies of both kidney and heart grafts in inbred mice have demonstrated that prolonged survival attributable to enhancement can also be induced in this species (Skoskiewicz et al., 1973; Judd and Trentin, 1973). It has also been suggested that rats, compared to dogs and rabbits, are resistant to antibody-mediated hyperacute rejection, possibly owing to deficiencies in their complement system ( French, 1972; Russell, 1971). However, the ability of rats to reject hyperacutely grafts from F, donors to presensitized F, animals of different strain parentage (i.e., representing a double haplotype mismatch) (Guttmann, 1974) and to “reject” in situ kidneys perfused with alloantiserum (Straus et al., 1971) indicates that rats are potentially capable of mounting antibody-mediated rejection. We have also demonstrated that unmodified renal allografts in the LBN F, to Lewis model undergo an acute rejection process that is typical of humoral rejection, with the development of immunoglobulin and complement-associated vascular injury ( Abbas et d.,1974a,b). It therefore appears unlikely that the susceptibility of inbred rats to enhancement induction is entirely due to a defect in the mechanism(s) of rejection. It is possible that the usual antibody response to allogeneic transplants in highly inbred animals consists largely of noncytotoxic antibodies that favor enhancement, and, once an equivalent degree of inbreeding is achieved in other species, we may discover that enhancement

38


is a universal phenomenon and can be induced with the same ease and consistency as in the inbred rat model. b. Degree of Histoincompatibility between Donor and Recipient. French and Batchelor (1969) were the first to demonstrate clearly that F, hybrid kidney grafts could be enhanced more effectively than parental strain grafts. Although homozygous allografts can be enhanced in some strains (Fabre and Morris, 1974a), the efficacy of enhancement is consistently greater in the heterozygous F, to parent situation. These observations have been elegantly extended by Hildemann and his co-workers (Hildemann and Mullen, 1973a,b; Mullen and Hildemann, 1971) using rats with graded degrees of incompatibilities as donors and recipients. In fact, kidney grafts between rats differing only at the minor histocompatibility loci show prolonged survival without any treatment. This work has led to the formulation of the “law” that “the weaker the histocompatibility, the greater the effectiveness of specific immunoblocking antibodies” ( Hildemann and Mullen, 1973a,b). That enhancing antibodies are more effective when the foreign antigenic density on the target is relatively low has been confirmed in vitro (Strom et al., 1973b) and also in uiuo by using kidney and heart allografts in mice with graded immunogenic diversities ( Skoskiewicz et al., 1973; Corry et al., 1973). Although no precise explanation for the role of antigenic density is known, in a later section we shall consider how smaller or weaker histocompatibility antigenic disparities might promote enhancement. c. Strain. Even among inbred rats differing at the major histocompatibility locus, some strains are relatively resistant to enhancement. For instance, (DA x Lewis) F, to Lewis kidney grafts require larger doses of enhancing serum and show a more modest prolongation of survival than LBN F, to Lewis or even (DA x Lewis) F, to DA (Fabre and Morris, 1972a, 1973). It is not known whether this reflects a “weaker” immunogenicity of DA kidneys, or a relative inability of Lewis animals to produce enhancing antibodies after exposure to DA alloantigens. Such observations, however, highlight the need for the precise definition of the immunogenetic requirements for enhancement induction. d . Type of Graft. The observation that skin grafts are resistant to enhancement induction and undergo acute rejection although kidney grafts in the same strain combinations are easily enhanced (White and Hildemann, 1968) has been confirmed by several investigators ( Barker and Billingham, 1970; White and Hildemann, 1968). The variable susceptibility of different tissues to enhancement has not been adequately explained. It may be that this is related to the ease with which enhancing antibodies can reach the graft, with skin grafts being relatively poorly vascularized and therefore more susceptible to the directly infiltrating lymphocytes that mediate rejection. Warren et al. (1973) have shown


39

that surgically anastomosed heart allografts can be easily enhanced, but heart tissue grafted in a frce manner analogous to skin is rejected acutely. However, Freeman and Steinmuller ( 1969) have observed prolonged survival of kidney but not heart allografts in rats; moreover, renal allograft survival can be prolonged far more effectively than that of heart by immunosuppressive treatment ( Van Bekkum et al., 1969), and in untreated pigs some liver allografts show anomalous prolonged survival, yet kidney and skin grafts undergo acute rejection (Calne et al., 1967). Thus, although factors related to the mode of transplanation are undoubtedly important in determining the fate of a transplant, it is also possible that different tissues contain organ-specific histocompatibility antigens that may initiate poor antibody responses but a relatively strong cellular response and thus favor rejection. The tissue distribution of I region-determined antigens ( L D ) may play a role in this regard. Organ-specific histocompatibility antigens have been identified in mouse skin (Lance et al., 1971), but their effect on the immunogenicity of organs in relation to transplantation and enhancement has not been defined. e . Zrnmunocompetence of the Host. The treatment of recipients with antithymocyte or antilymphocyte serum potentiates the effects of immunologic enhancement (Batchelor et al., 1972; Kilshaw et al., 1974). A similar synergism has also been observed with immunosuppressive drug therapy (R. E. Wilson et al., 1971) and thymectomy (Guttmann and Falk, 1974), suggesting that grafts are more susceptible to enhancement when the cellular rejection response is abrogated. It may be that such abrogation permits the recipient to initiate and complete endogenous synthesis of the required amounts of enhancing antibodies, but direct evidence for this hypothesis is lacking. From the above discussion, one general theme that emerges is that rejection and enhancement are phenomena that exist in a delicate state of balance, such that factors which promote one tend to retard the other (Russell, 1971). Thus, outbred species, strains with large histocompatibility differences, poorly vascularized grafts, and grafts in immunocompetent hosts are more prone to undergo rejection whereas all the converse situations favor the induction of enhancement. Elucidation of the mechanisms underlying these differences will undoubtedly be a major step toward the application of enhancement to the human transplant recipient.

C. E F F E ~OFS ENHANCING ANTIBODIES ON THE IMMUNOLOGY AND PATHOLOGY OF ALLOGRAFT REJECXION Before considering the nature of the changes produced by enhancing antibodies in the sequence of the complex rejection response, it is worth-

40


while to define the basic components of allograft rejection. It is well outside the scope of this review to detail the immunopathology of rejection, and we shall therefore merely outline the major events that occur after exposure of the immunocompetent host to a histoincompatible graft and how such events can be assayed. The “afferent” arc of the rejection response consists of the recognition of foreign graft antigens by circulating small lymphocytes or, especially in the case of skin allografts, by lymphoid cells in draining lymph nodes to which graft antigens travel (Gowans, 1965). Although it has been suggested that passenger leukocytes in the graft may be important in the sensitization process, it is unlikely that they play the key role in the immune response to primary allografts ( Billingham, 1971). The “central” position of the rejection response consists of a series of cooperative interactions between cells ( macrophages, immunocompetent lymphocytes) leading to the development of effector cells and antibodies. The mixed lymphocyte reaction is an in uitro reflection of the generation of effector cells following exposure to alloantigens ( Bach, 1973). The “efferent” arc consists of reactions by activated cells and antibodies with the graft to produce rejection. CelI-mediated immunity to allografts is largely a property of thymusderived lymphocytes and can be assayed in a variety of ways (Perlmann, 1972; Cerottini and Brunner, 1974). It is measured in uitro by the ability of lymphocytes from graft recipients to (1) effect release of 51Cr from appropriately labeled donor target cells ( Cerottini and Brunner, 1974); ( 2 ) induce detachment of cultured donor cells from appropriate culture dishes (Hellstrom et al., 1971; Hellstrom and Hellstrom, 1974); and ( 3 ) inhibit colony formation of cultured target cells (Perlmann and Holm, 1969). In oivo assays of cellular immunity include the ability of lymphocytes from graft recipients to mount graft-versus-host reactions in donor strain animals (Elkins, 1964; Ford d al., 1970) and the ability of graft-bearing recipients to reject second grafts. Recently, it has been suggested that non-thymus-derived lymphocytes may also play a role in graft rejection via antibody-dependent cytoxicity (see Section 11,B,2), but the relevance of this mechanism to organ transplantation is unclear. Concomitant with the development of cellular immunity, allograft recipients form antibodies that constitute another arm of the efferent arc. Humoral immunity is measured by standard techniques of assaying lymphocytotoxic and hemagglutinating antibodies (see Section II,B,l). The peripheral effects of cellular and humoral immunity on the graft are manifested largely by a destructive cellular infiltrate and immunoglobulin-associated vascular lesions, respectively ( Najarian and Foker, 1969; Dammin, 1966), the latter being more prominent in primarily vascularized organ grafts, such as kidneys. This brief


41

summary is intended merely to focus on the major detectable events following allotransplantation, so as to enable one to describe how such events are modified by enhancing antibodies. The complexities of transplantation genetics and immunology are vast and of obvious importance, and we will not deal at this point with theoretical aspects. The inff uence of enhancing antibodies on the immunopathology of rejection has been studied by a large number of investigators, generally by assaying antidonor cellular and humoral immune responses of recipients bearing immunologically enhanced allografts and by examining enhanced organs morphologically. The levels of such in uitro responses and the morphology of the grafts, when compared to corresponding assays done either in normal animals or in unmodified graft recipients, provide an indirect measure of the effects of enhancing antibodies. The use of different assays at various times has led to observations that often appear contradictory. However, several significant conclusions can be derived regarding the effects of enhancing antibodies and provide the necessary foundation to an understanding of the mechanisms by which enhancement is induced and maintained. 1. Cell-Mediated Zmmunity

When enhancement of allograft survival is induced or transferred by serum alone, recipients’ lymphocytes do not show a significant abrogation of their specific cell-mediated immune response to donor histocompatibility antigens. This crucial observation has been corroborated in a number of studies employing different models of transplantation and different assay procedures. Peak levels of lymphocyte cytotoxicity are unaltered in enhanced Lewis recipients of LBN F, kidney grafts (Biesecker et al., 1973b; Strom et al., 1975b; Stuart et al., 1971), AS recipients of (August X AS) F, kidneys (Burgos et al., 1974) and Lewis recipients of BN skin grafts (Peter and Feldman, 1972). Graft-versus-host reactivity of recipient splenic lymphocytes is essentially normal after enhancement of kidney grafts from (AS x August) F, to AS rats (French et al., 1971), LBN F, to Lewis rats (Ockner et al., 1970a,b; Mullen et d., 1973) and (DA x Lewis) F, to DA rats (Fabre and Morris, 1972b), and also in enhanced Lewis recipients of LBN F, cardiac grafts (Tilney and Bell, 1974). Furthermore, lymphocytes from enhanced Lewis recipients of LBN F, grafts react normally in mixed lymphocyte cultures and to mitogens (Lucas et aZ., 1970). Although several investigators have noted the slightly prolonged survival of donor skin grafts in animals bearing enhanced kidneys, this prolongation is not striking and probably has only marginal significance (Stuart et al., 1970; Fabre and Morris, 1972b; Tilney and Bell, 1974). Mullen et al. (1973) have observed dimin-

42


ished cytotoxicity of peripheral blood lymphocytes from Lewis recipients of enhanced LBN F, grafts; this finding contrasts with the normal graftversus-host reactivity of spleen cells from the same animals and may reflect the presence of circulating serum factors (immune complexes or blocking antibodies) either free or bound to the lymphocytes used as the aggressor cells in the cytotoxicity assay. In most of the studies in which diminution of specific cellular immunity is observed, prolongation of graft survival has been induced by antigen treatment, and definitive evidence for enhancement is lacking (von Haefen et al., 1973). Failure of passively administered enhancing antibody to diminish or abrogate cell-mediated immunity, therefore, appears to be a fairly consistent observation. It is noteworthy, however, that different assay procedures may produce inconsistent results. Biesecker et al. (1973a,b) have shown that after LBN F, to Lewis kidney transplantation, a 6-hour W r release assay

/---

FIG.1. See legend on facing page.


43

of the Brunner-Cerottini type (Cerottini and Brunner, 1974) detects cell-mediated immunity as an abrupt peak starting at days 2 3 , reaching a maximum at day 5, and declining rapidly thereafter; in contrast, a 48-hour microcytotoxicity assay based on the detachment of cultured donor fibroblasts from microtiter plates reveals a cell-mediated immune response starting at day 2, peaking at days 5-6 and gradually declining after several days with significantly elevated levels maintained u p to 30 days. The results of microcytotoxicity assays in these studies correlated poorly with impairment of graft function. I t is likely that the two assays measured different effector cell populations; nevertheless, enhanced animals showed no block in cellular immunity by either assay. Strom et al. (1975b), using the “Cr release technique, also did not demonstrate diminution of cell-mediated immunity in enhanced Lewis recipients of LBN F, kidney allografts, and concomitant examination of grafts revealed comparable peak lymphocytic infiltration in enhanced and rejecting kidneys (Abbas et al., 1974a). In addition, enhanced recipients continued to demonstrate weak lymphocyte cytotoxicity against donor cells for prolonged periods (Fig. 1). Although maximum levels of lymphocyte cytotoxicity are not signifiFIG. 1. Schematic summary of serial events in (Lewis x Brown Norway) F, + Lewis inbred rat renal transplant model. Upper panel, immunology; lower panel, histology. Control unmodified animals become uremic by day 7 and die on day 9 or 10. Enhanced animals receive 0.5 nil hyperimmune Lewis anti-Brown Norway ( BN ) serum at time of transplant, become transiently and mildly uremic on days 6 9 , and then become long-term survivors. Lymphocytotoxic antibody ( IgM ) develops in both groups and declines by day 9. Lymphocyte-mediated cytotoxicity ( L M C ) to BN target cells peaks on day 5 in controls and is delayed to day 7 with enhancement, but is unchanged in magnitude. Survivors display persistent low-level LMC (for months) and their sera block the LMC and mixed lymphocyte culture ( M L C ) . Both gronps develop cellular infiltrations a t enniparable rates. As renal blood flow declines in controls, the extent of infiltration actually declines, while a further increase occiirs in enhanced animals. Most striking is the lack of arterial and glomerular vasculitis in enhancement. (From Stroni et al., 1975b.) Reproduced by permission of Williams and Wilkins, Baltimore, Md. Control Immunology Ant ibody LMC Blocking Histology Vasculitis Cellular infiltrate

~

---

~

---

Enhanced

---I - -

44


cantly altered in enhanced animals, several recent studies have demonstrated a consistent delay in the attainment of peak activity. In the LBN F, to Lewis kidney enhancement model, Biesecker et ul. (1973b) have shown that, whereas unmodified allograft recipients show onset of lymphocyte cytoxocity (51Cr release assay) at day 3 and peak at day 5, administration of enhancing serum delays the onset to day 7 and the peak to day 11. In the same stpdy, the 48-hour microcytotoxicity assay revealed no differences betweeh control and enhanced animals, but the precise implications and significance of this latter assay are uncertain. Burgos et al. (1974) in the (AS X August) F, to AS kidney transplant model, and Debray-Sachs et al. (1973) in the (Lewis x AUgust) F, to Lewis model, have reported essentially identical findings, with significant delays in cell-mediated immunity as measured by W r release and colony-inhibition assays, respectively. The delay in cell-mediated cytotoxicity has also been observed in rats receiving skin allografts and enhancing serum (Peter and Feldman, 1972). In our own studies on passive enhancement of LBN F, to Lewis renal allografts, we have observed a 2-day delay in the attainment o f peak spleen lymphocyte cytotoxicity as measured by 51Cr release in recipients given a single dose of 0.5 ml of enhancing serum immediately after grafting (Strom et al., 1975a). The delay is consistent (Fig. 1) and is accompanied by a corresponding 2-day delay in the peak lymphocytic infiltration of the grafts themselves (Abbas et aZ., 1974a). However, the maximal levels of cellular immunity are not diminished in enhanced animals. Another major finding with regard to cell-mediated immunity is the ability of sera from long-surviving enhanced animals to suppress in vitro cellular responses against donor alloantigens. In the LBN F, to Lewis kidney allograft model, all enhanced recipients surviving beyond 9 days contain in their circulation a factor( s ) that completely suppresses cytotoxicity of recipient spIeen lymphocytes, as assayed by “Cr release from labeled donor target cells (Strom et al., 1975a) (Fig. 1). This factor(s), which has been referred to as “blocking antibody,” may in fact be an immune complex of transplantation antigens from the graft and specific alloantibodies. The factor(s) in these sera does not bind firmly to either attacking or target cells and does not have lymphocytotoxic activity, features that distinguish it from free alloantibody. Grafts in long-term survivors demonstrate granular deposits of IgG by immunofluorescence ( Abbas et al., 1974c), further suggesting that circulating complexes may be present. Biesecker et ul. (1973a) also observed serum blocking activity in the same rat allograft model by the 48-hour microcytotoxicity assay; since they were unable to detect any specific cellmediated immunity in long-term survivors by the 51Cr release assay,


45

it was obviously impossible to test for blocking activity against autologous cells.

2. Humoral Immunity The humoral antibody response of allografted recipients is somewhat variable, depending on the experimental model and the assay procedure employed. AS recipients of (AS x August) F, kidney grafts exhibit a biphasic cytotoxic antibody response, the first phase (up to 10 days after grafting) being 2-mercaptoethanol ( 2-ME ) sensitive and therefore probably of the IgM class, and the second phase (which persists up to 100 days) being 2-ME resistant ( IgG) antibody (Burgos et al., 1974). In these studies, administration of enhancing serum markedly inhibits the first phase and has little effect on the second. A similar biphasic response of both cytotoxic and hemagglutinating antibodies and the inhibitory effect of enhancing autobody on the first peak have also been observed in LBN F, to Lewis renal allograft recipients (Biesecker et al., 1973b). In our own studies, we have not detected any significant differences in serum cytotoxic antibody titers between unmodified and enhanced allograft recipients (Strom et al., 197513; Abbas et al., 1974a). Both control and enhanced responses were 2-ME sensitive (Fig. 1). Debray-Sachs’ et al. (1973) used colony inhibition as an assay for complement-mediated serum cytotoxicity in allcgraft recipients and detected a single sharp peak with no significant alteration induced by enhancing serum. Although it remains to be determined what the precise sequence of the humoral immune response in allograft recipients is, and at what stages it is modified by enhancing antibody, it appears indisputable that enhanced recipients do manifest cytotoxic antigraft antibody responses in the presence of surviving grafts. Such a dissociation between humoral immunity as detected by in witro techniques and graft function has been observed earlier in immunologically enhanced recipients ( French and Batchelor, 1969; Morris and Lucas, 1971). It is often observed that recipients bearing functioning enhanced grafts for relatively prolonged periods do not contain detectable serum antibodies against donor alloantigens ( French and Batchelor, 1969; Stuart et al., 1970; Strom et al., 1975b; Abbas et al., 1 9 7 4 ~ ) .This may be due to the fact that circulating antibodies are continuously complexed with alloantigens released from the graft. Failure to detect antibodies does not reflect an inherent inability of the host to mount a specific humoral immune response, since challenge of such hosts with donor cells invariably results in the appearance of detectable antibodies (French and Batchelor, 1969; Burgos et al., 1974; Stuart et al., 1970). Thus, during the life history of an enhanced renal allograft, the host’s

46


humoral immune response undergoes a sequential alteration: in the earliest phase, there may be abrogation or delay of an antibody peak, predominantly of the IgM class, but reduction in such a titer is not essential to successful enhancement (Strom et al., 1975b); later, no circulating antibody may be detected by in vitro techniques, although the recipient remains potentiaIly capable of manufacturing antidonor alloantibodies. In our studies, the onset of the later phase corresponds with the appearance of serum “blocking” activity ( Fig. 1). 3. Pathology of Graft Rejection

Relatively few detailed analyses of the pathology of rejection in unmodified and enhanced allografts have been attempted, and the majority of the reported studies have characterized immunologically enhanced renal allografts as demonstrating diminished rejection or cellular infiltration (Lucas et al., 1970; Kin et al., 1972), or as delayed in the usual sequence of rejection (Ockner et ul., 1970a). We have studied the sequential morphologic changes occurring in unmodified and passively enhanced LBN F, to Lewis renal allografts and demonstrated that the major effect of enhancing serum is abrogation of the vascular and glomerular lesions that are characteristic of acute unmodified humoral rejection (Abbas et al., 1974a,b) (Fig. 1 ) . Enhanced grafts show a complete absence of immunoglobulin and complement deposition in the period from 1 to 21 days after grafting, despite the presence of circulating anti-BN cytotoxic antibody during the first 8-9 days; however, the antibody appears to be IgM, a class that does not fix rat complement with rat alloantibody ( M . Ruszkiewicz and J. F. Mowbray, personal communication, 1975). Grafts functioning for over 1 month show granular immunoglobulin ( IgG ) deposits in glomeruli, suggesting that small amounts of IgG are made later and form immune complexes ( Abbas et al., 1 9 7 4 ~ ) . Measurement of allograft function by blood urea nitrogen determinations reveals an early functional impairment in enhanced grafts, followed by improvement and stabilization at mildly abnormal levels. It is likely that the early impairment reflects graft injury due to cellular infiltration, which is comparable in unmodified and enhanced grafts, and that the subsequent failure to progress to total rejection is due to blockage of the development of humoral vasculitis, which is responsible for the progressively downhill course of unmodified grafts. The stabilization of function corresponds to a “steady state” during which a delicate balance is maintained between the injurious and protective immune responses to the graft. In a subsequent section we shall discuss the relevance of these possible phases to the mechanisms of enhancement.


47

4. Noninorphologic Peripheral Effects of Enhancing Antibody Crucial to any discussion of the effects of enhancing antibody on the target organ, i.e., the graft, is the controversy as to whether such antibodies are preferentially localized peripherally in the graft. By inimunoflourescence, we have been unable to detect any IgG antibody in LBN F, to Lewis kidney grafts within the first 4 days after administering enhancing serum ( Abbas et d., 1974b). Zimmerman and Feldman ( 1970), working with IgG fractions of immune sera, were unable to demonstrate the binding of radiolabeled enhancing antibody to skin grafts. Morris and Lucas ( 1971) have demonstrated a reduced uptake of radiolabeled antidonor antibody by passively enhanced kidney grafts when compared to appropriate controls, and Jones et al. (1972) have shown that iiijected antidonor IgG, previously enriched by elution from donor cells, specifically binds to rat skin grafts, a fixation that can be blocked by prior administration of enhancing serum. French ( 1973) and Fine et al. (1973) have shown that long-surviving enhanced grafts bind radiolabeled antidonor IgG to a degree comparable to that found in freshly transplanted control kidneys. Such reports are not necessarily contradictory, since different strain combinations and antibody preparations were employed. Although studies of this type suffer from the virtual tcchnical impossibilities of precisely localizing the anatomic sites where minute amounts of enhancing antibody may be bound, it remains quite possible that a sniall but critical threshold of covered antigenic sites exists. In fact, Fabre and Morris (1974b) have shown that masking of all antigenic sites is not an absolute requirement for inducing enhancement. The precise delineation of such critical sites is of obvious importance, but the quantities of alloantibodies that may bind to organ grafts make a direct approach impractical. One obvious problem relates to the likely presence of soluble antigen in graft-bearing animals, a factor that could alter results when radiolabeled IgG is administered intravenously. More potent evidence of the peripheral localization of enhancing antibodies is provided by studies that have been aimed at assessing the effects of such localization rather than measuring it directly. Thus, skin grafts dipped in specific antiserum show prolonged survival (Zimnierman and Feldman, 1970), and ex vioo perfusion of rat and primate kidney grafts with intact or pepsin-digested anticlonor antibodies prior to transplantation also protects thein from rejection (Shaipanich et d., 1971; Habal et al., 1973). Moreover, enhanced kidneys when transplanted into second hoqts undergo delayed rejection, indicating that at least partial “graft adaptation” occurs (Stuart et al., 1970; Fabre and Morris,

48


1972b) and strongly suggesting a peripheral effect of enhancing antibodies on the graft. The possibility that antibody directed against cell surface determinants may “modulate” antigenic expression (Old et al., 1968), such as on the endothelium of a graft, could provide an alternative explanation to persistent antibody coating of cell surfaces; however, direct evidence for loss of antigens is lacking because of the technical difficulties noted above. Finally, Bowen et al. (1974) have shown that it is not possible to abrogate long-term enhancement of F, kidney grafts by parabiosing the hosts with syngeneic normal partners, indicating that the enhanced grafts resist rejection despite exposure to the immune system and circulation of a normal animal. Of course, active suppression of the normal partner by factors from the enhanced partner may occur. Nevertheless, all this evidence suggests that enhancing antibodies may have some peripheral action whose major consequences are protection against rejection mechanisms. At precisely what stage of the rejection sequence such a peripheral protective effect comes into play will be discussed in the next section.

D. MECHANISMS OF ACTION OF ENHANCING ANTIBODIES The mechanism by which enhancing antibodies prolong allograft survival has been a subject of dispute ever since the phenomenon of enhancement was established. Much of the evidence has been derived from models of tumor allografts, and since this has been discussed comprehensively, it is not considered here (for reviews, see Voisin, 1971; Feldman, 1972a,b). In the context of organ allografts, it is clearly estabIished that antibodies against some or all of the histocompatibility antigens are capable of prolonging survival, either in a free form or complexed with the relevant alloantigens. We have detailed above the evidence to suggest that the protective effect is a direct result of the host’s immune response itself, and that the treatment given (i.e., donor antigens in active enhancement or antidonor serum in passive enhancement) favors the type of immune response that results in graft protection. Moreover, serum “blocking” factors may be the in vitro correlate of enhancing activity, although the precise relationship awaits clear identification of the antibodies involved. The crux of the controversy regarding the mechanism of enhancement has centered around the locus of effect on the rejection response: at the afferent, central, or efferent levels. Part of the difficulty is that this division (afferent, central, efferent) is too simplistic, since it does not clearly provide for a peripheral level of blockade against cellular effector mechanisms; that is a, “central”


49

blockade at the efferent site. I t now seems likely that this may be an important site of action for enhancing, or blocking, factors. Since prolonged allograft survival occurs in the presence of cellular and humoral immunity against donor alloantigens, enhancement does not represent a significant abrogation of the generation of the immune response; therefore, there is no major block in either the sensitization (afferent) or central phases of rejection (Fig. 1). It is possible that in vitro assays of immune responsiveness do not accurately reflect the in vivo situation, but such a possibility appears unlikely ( Cerottini and Brunner, 1974). The major problem with an efferent blockade hypothesis is the ability of minute quantities of antibodies to induce enhancement, these quantities being so small that it is unreasonable to expect them to block or modulate a significant number of antigenic sites in the graft (French, 1973). Moreover, in the parallel situation of antibody-induced suppression of cellular responses to protein and other antigens, it appears likely that the block is mediated primarily by a reaction with the immunocompetent cells themselves ( Rowley et al., 1973). The mechanisms of afferent, central, and efferent blockade need not be mutually exclusive, and one or more may be operative at different stages during the life history of the enhanced allograft. French and Batchelor (1972) have suggested that three phases, possibly with some overlap, may be distinguished during the course of an enhanced rat renal allograft, and this concept is probably applicable to other organ grafts as well. It is important, therefore, to consider the temporal dynamics of an enhanced system. As an illustration, the events occuring in our own model of renal transplantation can be divided into the following phases (Fig. 1 ) : (1) the phase of enhancement induction, which in this model corresponds to days 1 to 4 and during which the major discernible alteration is a delay in the development of cell-mediated immunity (Strom et al., 197513); ( 2 ) the phase of rejection abrogation, corresponding to days 5 to 9, during which the unmodified allograft is destroyed by acute rejection, while the enhanced graft shows a significant block in the development of humoral rejection and is thus protected despite detectable antidonor cellular and humoral immunity ( Abbas et al., 1974a); and ( 3 ) the steady-state phase, extending from day 9 onward, during which the graft continties to survive with mild to moderate functional impairment, accompanied by lesions apparently mediated by immune complexes and low levels of cellular cytotoxicity. In addition, this phase is marked by the presence of serum blocking factors and the absence of detectable antidonor cytotoxic antibody in the circulation ( Abbas et al., 1974c; Strom et at., 1975b). Although division into such phases is necessarily somewhat arbitrary, the sequence of immunologic

50


events observed in other models of enhancement of renal and skin allografts suggests that distinct events may be occurring sequentially after transplantation ( French and Batchelor, 1972; Biesecker et al., 1973b; Jones et al., 1972). During the induction phase, there is a significant delay in the onset and progression of cell-mediated immunity against donor antigens, possibly accompanied by inhibition of the early IgM antibody peak (Biesecker et al., 1973b; Burgos et al., 1974; Strom et al., 1975b; Debray-Sachs et al., 1973; Jones et al., 1972). If an afferent block exists, it can only be partial, since sensitization of the host to graft antigens does occur. Since administration of donor antigens or antidonor serum has a central effect on the generation process delaying the attainment of peak cellular immuity, it may simultaneousIy permit the endogenous de novo synthesis of antibodies that are primarily responsible for effecting the abrogation of rejection. Thus, the second phase in the life history of the enhanced graft is not directly a result of the administered treatment (i.e., donor antigen or antidonor antibody) but could be due to a protective immune response of the host itself. On the other hand, an alternative and more straightforward explanation of events following passive antiserum administration in the rat would be that rejection vasculitis does not occur because of an early failure in production of the IgG alloantibodies that produce vasculitis in the vessels of control grafts. As noted above, IgM alloantibodies do not seem to be cytotoxic with rat complement. In the enhanced LBN F, to Lewis renal allograft, immunoglobulin is not deposited in the graft vasculature, hence acute humoral rejection is prevented (Abbas et al., 1974b). Therefore, one can postulate that passive transfer interferes with T-cell and B-cell cooperation during immune induction, causing a blunted IgG response as well as a delay in T-cell effector production. More direct evidence for a deficient IgG response is necessary, but the alternative hypothesis of endogenous synthesis of a special type of antibody has even less experimental support. If there is a blocking effect by this putative antibody on cells, including cytotoxic lymphocytes, analogous to the in vitro blocking that is seen with sera of long-surviving enhanced recipients, we cannot discern it by morphologic studies; there is no detectable diminution of early lymphocytic invasion in enhanced renal allografts ( Abbas et al., 1974a). Furthermore, serum blocking activity is not detectable until 9 days after transplantation (Strom et al., 1975b). After the acute rejection process is prevented, the enhanced allograft demonstrates slowly diminishing degrees of cellular infiltration, the development of glomerular lesions, probably mediated by immune complexes, and mild or moderate functional impairment ( Abbas et al.,


51

1975b). This “steady state” is likely to represent a complex balance between rejection and protection. Thus, levels of persisting lymphocyte cytotoxicity are low, and one can surmise that the in v i m effect of aggressor lymphocytes may also be opposed by serum blocking factors (antibodies or immune complexes) that are detectable in uitro ( Strom et al., 1975b; Biesecker et al., 1973b). The possibility of peripheral alteration (adaptation of the graft or blocking and/or modulation of graft antigens ) exists. Although the topography of antigenic site density in tissue is unknown, the facts that haplotype matched F, to parental rat renal allografts are usually easier to enhance then totally mismatched grafts, and that target cell susceptibility to lysis is also dependent upon alloantigen site density (Strom et nl., 1973b), would lead us to conclude that further alteration in expression or availability of antigenic sites in the chronic “steady state” of enhancement could play some role in perpetuation of that state. Nevertheless, there is growing evidence that some sort of central mechanism may be of major importance in preventing rejection in the “steady-state” enhancement situation. clearly, the effect is not one of total nonresponsiveness, as would be obtained in true tolerance. As noted, animals have cytotoxic effector lymphocytes at a time when the grafts are not being rejected. Both passive transfer of hyperimmune serum at the time of traiisplantation and active immunization with donor antigen prior to transplantation produce similar effects. Adoptive transfer of immune cells or parabiosis with immune animals do not produce rejection of long-term enhanced grafts in the rat kidney model ( Bowen et nl., 1974), indicating the presence of positive enhancing factors in the enhanced animals. The nature of these blocking factors are still in dispute; however, the most intriguing possibilities rclate to a form of central inhibition of the responding cell by antibody complexed with donor alloantigen. In other words, a form of immune complex present both in the circulation of the animal as well as on the membranes of responding cells may modulate the immune response to prevent both the proliferative response in lymphoid tissue as well as the effector function of cytotoxic lymphocytes. There is evidence for a form of cytophilic antibody ( I g G ) , eluted from phagocytic cells in dogs immunized by renal allografts, that behaves as a blocking factor of high potency with regard to the mixed lymphocyte response (Miller et d.,1973). In any situation where an allograft is in place it would be extremely difficult to separate the effects of antibody alone from the effects of immune complexes, since small amounts of antigen are released from grafts, and one would expect small but perhaps critical quantities of immune complexes to be present at all times. The situation would result in a form of split tolerance in which some antibody

52


is formed, but cellular mechanisms are not discerned in vivo because of this form of blockade (Brent et al., 1972). Although definitions of tolerance vary from laboratory to laboratory, central inhibition of the T-cell response to antigen, whether by antigen alone or by antigen complexed with antibody is certainly a description of one form of such tolerance. The evidence for immune complexes that block cellular immunity in rat transplant situations is mostly inferential at the present time. Immunocompetent cells react with immunoglobulin via membrane receptors for the Fc portion of the immunoglobulin molecule. The presence of such Fc receptors on B lymphocytes has been clearly defined and extensively studied (Basten et al., 1972). Varieties of cells including mast cells, neutrophils, monocytes, platelets, and leukocytes have receptors that bind immunoglobulin of various classes and subclasses. Furthermore, certain subpopulations of T cells, particularly after immunologic activation, bind IgG as well. Therefore, mononuclear cells involved in the immune response, namely T cells, B cells, and macrophages, all have the potential for interaction with antibody, particularly when it has been complexed with antigen. Sinclair (1969) has shown that IgG to SRBC loses it ability to produce passive enhancement if its Fc portion has been removed by pepsin digestion. The immune response to SRBC involves a cooperative response between B cells and T cells, the precise details of which await elucidation. However, it is apparent that compatibility for products of the I region of the mouse H-2 gene complex is important in this cooperative response ( Katz et al., 1973). Since incompatibility for I-region antigens results in a positive mixed lymphocyte (MLC) response ( L D difference), and since it has been shown by Dickler and Sachs (1974) that the I-region product is identical to, or closely associated with, the Fc receptor of B cells, it is likely that the interaction of the B cell Fc receptor with the Fc portion of IgG may play an important regulatory role in the induction of a B-T cooperative immune response. Confirmation of this hypothesis comes from the in vitro findings of Feldman and Diener (1972) in which F( ab’), fragments, when complexed with flagellin, a thymus-independent antigen, were able to produce inactivation of immunocompetent B cells in vitro. In other words, in a system when B-T cooperation is not essential, the Fc component of the IgG is irrelevant to suppression by immune complexes. Yoshida and Anderson (1972) first demonstrated that T cells activated to alloantigens would react with immune compIexes in the form of antibody ( IgG) coated erythrocytes (EA). Studies of Fridman and Golstein (1974) indicate that the Fc receptor on activated T cells may be released into culture medium since the soluble supernatant material, which they


53

have called immunoglobulin binding factor ( I B F ) , has specificity for the Fc portion of IgG, as assayed by inhibition of C1 fixation and complement-mediated lysis. The effect is one of steric hindrance to fixation of C1 and is not specific for the C1 binding sites, since IBF does not react with IgM molecules. Some T cells in the immunologically activated population also form rosettes with EA (Yoshida and Andersson, 1972); however, the identity of the T-cell Fc receptor and the secreted*IBF, though likely, has not as yet been established. All the experiments showing a T-cell receptor for immunoglobulin, whether by rosette formation with EA or by binding of aggregnted IgG, indicate that the affinity of the interaction is reasonably high, since the rosettes are relatively stable and aggregates are not casily washed off. Some information exists regarding the specificity of Fc receptors of B and T cells for immunoglobulin subclasses, findings of potential importance to the enhancement problem. Basten et ul. (1972) showed that IgG, was the predominant mouse immunoglobulin binding to the B cells contained within thoracic duct lymph. IgGI, a noncomplement-fixing class, bound much more avidly than IgG,; in fact, IgG,, and IgA myeloma proteins did not inhibit binding, whereas IgG-.,, and IgM were only weakly inhibitory. In contrast, Lee and Paraskevas (1972) showed that IgG, did not bind to activated T cells, where the IgG., subclass did preferentially bind to T cells. These fiindings were confirmed by Soterides-Vlachos et al, (1974), who showed that IgG, is the major class involved in the binding to mouse T cells. Therefore, evidence exists for preferential interaction of two kinds of Fc receptors on B and T cells for different immunoglobulin subclasses. Non-T cells may have receptors also for activated C3 (C3b or C3d), but the importance of this receptor in control of immune responses is incompletely defined (Perlmann, 1975). Recently it has been shown that T cells from normal mice have such complement receptors, as well as Fc receptors for IgG. ( Soterides-Vlachos et ul., 1974). In fact, 30% of complement receptor lymphocytes ( C R L ) from normal spleens are T cells, and 10% of T cells in normal mouse spleen have Fc receptors. In alloimmunized animals the number of T cells with Fc receptors rises to a level of 70% while no T cells demonstrate complement receptors. The apparent discrepancies in these recent studies with previous work which indicated that Fc and complement receptors were the exclusive property of non-T cells are in part explained by differences in methodology (double versus single labeling of cells) and by the fact that a number of cell separation techniques are incomplete and do not necessariIy foIIow the functional classification of various subsets of both B and T cells. With immunoglobulin receptors on T, B, and macrophage classes of

54


immunocompetent cells, and with some specificity to immunoglobulin subclasses, there are a number of possible ways in which antibody and immune complexes may alter an immunologic response. One immunization has occurred, a complicated interplay of soluble antigen from the graft with immunoglobulin produced in the animal could act in a variety of ways on the ongoing “steady state.” There is little direct evidence as yet that suppression is caused in vivo by immune complexes directly bound to the surfaces of effector T cells, unless these complexes are rather loosely bound. Lymphoid cells removed from enhanced animals can show direct cytolytic activity against donor target cells and can respond in mixed lymphocyte culture after being washed and placed into neutral medium (Strom et al., 1975b). Ortiz-de Landazuri and Herberman (1972) have shown that rat spleen cells immune to a syngeneic leukemia virus require several hours of incubation before they exert cytotoxic activity in vitro but the exact mechanism of such activation is not known; possibly, it could be the result of shedding of complexes from activated T cells. It is clear, then, that a number of critical studies must be performed on the nature of blocking factors present in an animal’s enhanced immunologic state, relating to immunoglobulin composition ( class and subclass ), interaction with effector cell membranes, and specificity ( I-region LD determinants vs cytotoxic SD determinants). Similar studies on the hyperimmune sera used to promote enhancement by passive transfer may reveal some common mode of action; on the other hand, induction and maintenance may depend upon entirely different mechanisms. IV. Conclusions

When one considers all these data, it is difficult to formulate a simple set of concise conclusions regarding the role of antibody in organ transplantation. However, this review has emphasized that the fate of a vascularized organ graft can be closely dependent upon humoral factors in constant and intimate contact with graft endothelium. For example, if complement-fixing immunoglobulin with specificity for endotheliaI antigens is present and if the target antigens are of sufficient density, vascular injury results. If the immunoglobulin is complexed to soluble antigens from the graft, deposition of immune complexes in vessels and capillaries may produce injury, or similar complexes may serve to inhibit cell-mediated effector processes. It is also possible for free immunoglobulin to cover antigenic sites and to block target cell injury by effector cells, Furthermore, it is clear that the normal physiology of immunoglobulin-cell interactions and immunoglobulin-mediator activations pro-


55

vides a variety of potential mechanisms for qualitative and quantitiative alterations in induction of the immune response and in the efficiency of effector mechanisms at a later stage, Of importance to our current understanding is the fact that many of the observed phenomena and the problem as a whole can be stated in modern immunologic terms, and it is increasingly likely that studies can be devised to approach some detailed solutions. The tasks of sorting out the components of the alloimmune response in kinetic sequence, and of isolating and defining the specificity and subclass requirements of enhancing and blocking factors, should in time answer some of the key questions and provide a basis for a rational approach to modification of alloinimunity.

ACKNOWLEDGMENTS We wish to thank our colleagues, Drs. Joseph M. Corson, Gustave J. Dammin, John P. Merrill, and Terry B. Strom, for their substantive contributions to much of this work.

REFERENCES Abbas, A. K., Corson, J. M., Carpenter, C. B., Galvanek, E. G., Merrill, J. P., and Danimin, G. J. (1974a). Amer. J. Pathol. 75, 255. Abbas, A. K., Corson, J. M., Carpenter, C. B., Galvanek, E. G., Merrill, J. P., and Daniniin, G. J. (1974b). Amer. J. Pathol. 75,271. Abbas, A. K., Coison, J. M., Carpenter, C. B., Strom, T. B., Merrill, J. P., and Dammin, G. J. (1975a). Amer. J. Plrthol. 79, 255. Abbas, A. K., Corson, J. M., Washington, S. N., and Carpenter, C. B. (197513). Unpublished observations. Algaire, G. H. (1959). I. Nut. Cancer Inst. 23, 435. Algaire, G. H., Weaver, J. M., and Prehn, R. T. (1954). J. Nut. Cancer Inst. 15, 493. Algaire, G. H., Weaver, J. M., and Prehn, R. T. ( 1957). Ann. N.Y. Acad. Sci. 64, 1009. Altman, B. (1963).Ann. Roy. Coll. Sitrg. Engl. 33, 79. Amos, D. B., and Wakefield, J. D. (1958). 1. Nut. Cancer Inst. 21, 657. Amos, D. B., and Wakefield, J. D. (1959). J. Nut. Cancer Inst. 22, 1077. Anderson, M., Dewar, P., Fleniming, L. B., Hacking, P. M., Morley, A. R., Murray, S., Swinney, J., Taylor, R. h.1. R., Uldall, P. R., and Wardle E. N. (1974). Clin. Nephrol. 2, 93. Andres, G. A., Accinni, L., Hsu, K. C., Penn, I., Porter, K. A., Rendall, J. M., Seegal, B. C., and Starzl, T. E. ( 1970). Lab. Invest. 22, 588. Asherson, G. L., Zembala, M., and Barnes, R. M. R. ( 1971 ). In “Iinrniinological Tolerance to Tissue Antigens” ( N . W. Nisbet and M. W. Elves, eds.), p. 189. Orthopaedic Hospital, Oswestry, England. Austen, R. F. (1974). Transplant. PTOC.6, 39. Austen, K. F., and Russell, P. S. ( 1966). Ann. N.Y. Acad. Sci. 129, 657. Bach, F. H. ( 1973). Transplant. Proc. 5,23. Baldamus, C. A., McKenzie, I. F. C., Winn, €1. J., and Russell, P. S . (1973). J. Zmmzinol. 110, 1532. Barker, C. F., and Billinghain, R. E. (1970). Nature (London) 225, 851.

56


Basten, A,, Warren, N. L., and Mandel, T. (1972). 1. Erp. Med. 135, 627. Batchelor, J. R., Fabre, J. W., and Morris, P. J. (1972). Transplantation 13, 610. Becker, E. L., and Henson, P. M. (1974). Adoan. Immunol. 17,93. Belzer, F. O., Perkins, H. O., Fortmann, J. L., Kountz, S. L. Salvatierra, O., and Cochrum, K. C. ( 1974). Lancet 1, 774. Berger, J., and Yaneva, H. ( 1973). Ann. Inst. Pastetrr, Paris 124, 163. Biesecker, J. L., Fitch, F. W., Rowley, D. A., Scollard, D., and Stuart, F. P. (1973a). Transplantation 16, 421. Biesecker, J. L., Fitch, F. W., Rowley, D. A., and Stuart, F. P. (1973b). Transplantation 16, 432. BilIingham, R. E. (1971). Cell. Irnrnunot. 2, 1. Bowen, J. E., Batchelor, J. R., French, M. E., Burgos, H., and Fabre, J. W. ( 1974). Transplantation 18, 322. Brent, L. (1971). In “Cellular Interaction in the Immune Response.” (S. Cohen, G. Cudkowicz, and R. T. McCluskey, eds.), p. 250. Karger, Basel. Brent, L., Brooks, C., Lubling, N., and Thomas, A. V. (1972). Transplantation 14, 382. Briggs, J. D., Timbury, M. C., Paton, A. M., and Bell, P. R. F. (1972). Brit. Med. I. 4, 520. Britton, S., Perlmann, H., and Perlrnann, P. (1973). Cell. Immunol. 8, 420. Brown, D. L. (1974). In “Proceedings of the Second International Congress of Immunology” ( L . Brent and J. Holbrow, eds.), Vol. 1, p. 191. North-Holland Publ., Amsterdam. Bubenik, J., Perlmann, P., and Hasek, M. (1970). Transplantation 10, 290. Burgos, H., French, M. E., and Batchelor, J. R. (1974). Transplantation 18, 328. Busch, G. J., Braun, W. E., Carpenter, C. B., Corson, J. M., Galvanek, E. R., Reynolds, E. S., Merrill, J. P., and Darnmin, G. J. (1969). Transplant. Proc. 1, 267. Busch, G. J., Kobayashi, K., Birtch, A. G., Galvanek, E. G., Lukl, P., and Carpenter, C. B. (1971a). Hum. Pathol. 2, 299. Busch, G . J., Reynolds, E. S., Galvanek, E. G., Braun, W. E., and Dammin, G. J. ( 1971b). Medicine (Baltimore) 50, 29. Calne, R. Y. (1971). Transplant. Proc. 3, 21. Cahe, R. Y., White, H. J. O., Yoffa, D. E., Maginn, R. R., Binns, R. M., Samuel, J. R., and Molina, V. P. (1967). Brit. Med. I . 2, 478. Campbell, A. C., MacLennan, I. C. M., Snaith, M. L., and Bamett, I. G. (1972). Clin. Exp. Immunol. 12, 1. Carpenter, C. B. (1974). Transplant. Proc. 6, 83. Carpenter, C. B., and Merrill, J. P. (1969). Arch. Intern. Med. 123, 501. Carpenter, C. B., Gill, T. J., 111, Merrill, J. P., and Dammin, G . J. (1967). Amer. J. Med. 43, 854. Carpenter, C. B., Ruddy, S., Shehadeh, I. H., Muller-Eberhard, H. J., Merrill, J. P., and Austen, K. F. (1969). J . Clin. Invest. 48, 1495. Carpenter, C. B., Rashid, A., Busch, G. J., Katz, A. I., Scheffer, A. L., Valentine, M. D., Birtch, A. G., Kay, R. G., Levine, L., Murray, J. E., Austen, K. F., and Merrill, J. P. (1971). Colloq. Int. Cent. Not. Rech. Sci. 190, 641. Carpenter, C. G., Garovoy, M. R., Strom, T. B., and MerrilI, J. P. (1974). In “Lymphocyte Recognition and Eff edor Mechanisms” ( K. Lindahl-Liessling and D. Osoba, eds.), p. 351. Academic Press, New York. Carrell, A., and Guthrie, C. C. ( 1905). Science 22,473.


57

Cerottini, J.-C., and Brunner, K. T. (1974). Adoan. Immunol. 18, 67. Claes, G., Bergentz, S. E., Roos, B., and Rosengren, B. (1970). Eur. Surg. Res. 2, 48. Clark, D. S., Foker, J. E., Good, R. A,, and Varco, R. L. (1968). Lancet 1, 8. Clarkson, A. R., MacDonald, M. K., Petri, J. J. B., Cash, J. D., and Robson, J. S. (1971). Brit. Med. J. 3, 447. Cochrane, C. G., and Koffler, D. ( 1973). Aduan. I m m u t d 16, 185. Cochrum, K. C., Davis, W. C., Kountz, S. L., and Fudenberg, H. H. (1969). Trunsplant. Proc. 1, 301. Cock, A. B. (1962). lmmunology 5, 169. Colombani, J., Colombani, M., and Dausset, J. (1973). Transplantation 16, 257. Cooper, M. D., Peterson, R. D., South, M. A., and Good, R. A. (1966). J . Exp. M e d . 123, 75. Cooper, N. R., and Cochrane, C. G. (1974). Transplant. Proc. 6,51. Corry, R. J., Winn, H. J., and Russell, P. S. ( 1973). Transplantation 16, 343. Cross, D. E., Whittier, F. C., Cuppage, F. E., Crouch, T., Manual, E. L., and Grantham, J. J. (1974). Transplantation 17, 626. Daniniin, G. J. ( 1966). In “The Kidney” (I.A.P. No. 6),p. 455. Williams 8i Wilkins, Baltimore, Maryland. d’Apice, A. J. F., Tait, B. D., and Morris, P. J. (1974). Transplantatioti 18, 20. Dausset, J. ( 1973). Transplant. Proc. 5, 1751. David, C. S., Shreffler, D. C., and Frelinger, J. A . (1973). Proc. Nat. Acad. Sci. U.S. 70, 2509. David, D. S., Millian, S. J., Whitsell, J. E., Schwartz, G. H., Riggio, R. R., Stenzel, K. H., and Rubin, A. L. ( 1972). Ann. Strrg. 175, 257. Davies, D. A. L., and Alkins, B. J. (1974). Nature ( L o n d o n ) 247, 294. Davis, W. C. ( 1973). Transplant. Proc. 5, 625. Debray-Sachs, M., Soucy, P. E., and Hamburger, J. (1973). J . Immunol. 110, 661. Dickler, H. B., and Sachs, D. H. ( 1974). J. Exp. Med. 140, 779. Dixon, F. J., Feldman, J. D., and Vazquez, J. J. (1961). J. Exp. Med. 113, 899. Dixon, F. J., McPhaul, J. J., and Lerner, R. A. (1969). Transplant. Proc. 1, 194. Dubernard, J. M., Carpenter, C. B., Busch, G. J., Diethelm, A. G., and Murray, J. E. (1968). Surgery 64, 752. Dumble, L. J., and Morns, P. J. (1975). Tissue Antigens 5, 103. Dumble, L. J., Jack, I., and Morris, P. J. (1975). Clin. Exp. Immunol. (in press). Elkins, W. L. (1964). J. Erp. Med. 120, 329. Elkins, W. L., Hellstrom, K. E. (1974). Transplantation 18,38. Enomoto, K., and Lucas, Z. J. ( 1973). Transplantation 15, 8. Fabre, J. W., and Morris, P. J. (1972a). Transplantation13, 604. Fabre, J. W., and Morris, P. J. (1972). Transplantation 14, 634. Fabre, J. W., and Morris, P. J. ( 1 9 7 2 ~ )Transplantation . 14, 698. Fabre, J. W., and Morris, P. J. ( 1973). Transplantation 15, 397. Fabre, J. W., and Morris, P. J. (1974a). Transpkzntation 18, 429. Fabre, J. W., and Morris, P. J. (1974b). Transplantation 18, 436. Feldman, J. D. (1972a). Aduan. Immunol. 15, 167. Feldman, J. D. ( 1972b). Pathol. Annu. 7, 19. Feldman, J. D., and Lee, S. (1967). J. Exp. Mecl. 126, 783. Feldnian, M., and Diener, E. ( 1972). J. Immrinol. 108, 93. Ferrone, S., Cooper, N. R., Pellegrino, M. A,, and Reisfeld, R. A. (1971). J. Immunol. 107, 939.

58


Fine, R. N., Batchelor, J. R., French, M. E., and Shumak, K. H. (1973). Transplantation 16, 641. Floresco, N. (1905). 1. Physiol. Pathol. Cen. 7, 47. Foker, J. E., Clark, D. S., Pickering, R. J., Good, A. A., and Varco, R. L. (1969). Transplant. Proc. 1, 296. Ford, W. L., Burr, W., and Sinionsen, M. (1970). Transplantation 10, 258. Forman, J., and Britton, S. (1973). J . Exp. Med. 137, 369. Freeman, J. S., and Steinmuller, D. (1969). Transplantation 8, 530. French, M. E. (1972). Transpkntation 13, 447. French, M. E. ( 1973). Transplant. Proc. 5, 621. French, M. E., and Batchelor, J. R. ( 1969). Lancet. 2, 1103. French, M. E., and Batchelor, J, R. ( 1972). Transplant. Reu. 13, 115. French, M. E., Batchelor, J. R., and Watts, H. G. (1971). Transplantation 12, 45. Fridnian, W. H., and Golstein, P. (1974). Cell. Immunol. 11, 442. FulIer, T. C., and Winn, H. J. ( 1973). Transplant. Proc. 5,585. Gabourel, J. D. ( 1961). Cancer Res. 21, 506. Garovoy, M. R., Franco, V., Zschaeck, D., Carpenter, C. B., Stroni, T. B., and Merrill, J. P. (1973a). Lancet 1,573. Garovoy, M. R., Phillips, S. M., Carpenter, C. B., and Merrill, J. P. (197310). Transplant. Proc. 5, 129. Gelfand, E., Resch, K., and Prester, M. (1972). Eur. J . Immunol. 2, 419. Gewurz, H., Clark, D. S., Finstad, J., Kelley, W. D., Varco, R. L., Good, R. A., and Gabrielsen, A. E. (1966). Ann. N.Y. Acad. Sci. 129, 673. Givol, D., Pecht, I., Hochman, J., Schlessinger, J., and Steinberg, I. Z. (1974). In “Proceedings of The Second International Congress of Immunology” ( L . Brent and J. Holbrow, eds.), VoI. 1, p. 39. North-Holland Publ., Amsterdam. Glassock, R. J., Feldman, D., Reynolds, E. S . , Dammin, G. J., and Merrill, J. P. (1968). Medicine (Baltimore) 47, 411. Goldnian, B. S., Shumak, K. H., Silver, M. D., Cartier, J., LixfeId, W., Bigelow, W. G., and Crookston, J. H. (1971). Transplant. Proc. 3, 515. Good, R. A. (1959). Immunopathol., Znt. Symp., l s t , 1958 p. 41. Good, R. A., Kelly, W. D., and Gabrielsen, A. E. ( 1962). Immunopathol., Int. Symp., 2nd, I961 p. 353. Gordon, R. O., Stinson, E. B., Souther, S. J., and Oppenheim, J. J. (1971). Transplantation 12, 484. Gowans, J. L. (1965). Brit. Med. Bull. 21, 106. Graff, R. J., and Bailey, D. W. (1973). Transplant. Reo. 15, 26. Greenberg, A. H., Hudson, L., Shen, L., and Riott, I. M. (1973). Nature (London), New Biol. 242, 111. Guiney, E. J., Austen, K. F., and Russell, P. S . (1964). Proc. Soc. Erp. B i d . Med.1 115, 1113. Guttmann, R. D. ( 1973 ). Transplantation 15,594. Guttmann, R. D. (1974). Transplantation 17, 383. Guttmann, R. D., and Falk, R. E. (1974). Transplantation 17, 228. Guttmann, R. D., Lindquist, R. R., Parker, R. M., Carpenter, C. B., and Merrill, J. P. ( 1967a). Transplantation 5, 668. Guttmann, R. D., Carpenter, C. B., Lindquist, R. R., and Merrill, J. P. (1967b). J . Exp. Med. 126, 1099. Guttmann, R. D., Lindquist, R. R., Kawabe, K., and Ockner, S. A. (1972). Transplantation 13, 15.


59

Habal, M. B., Carpenter, C. B., Kobayashi, K., Busch, G. J., Misra, M., and Birtch, A. G. (1973). Transplant. Proc. 5, 601. M. P., and Wallach, D. F. (1964). Ann. N.Y. Acad. Sci. 120, Hager, E. B., DLIPLIY, 447. Hampers, C. L., Kolker, P., and Hager, E. B. (1967). J. Zrnrnunol. 99, 514. Harding, B., Pudifin, D. J., Gotch, F., and MacLennan, I. C. M. (1971). Nature (London), New Biol. 232, 80. Hasek, M., Karakoz, I., Skamene, J. C., Noriza, K., Bubenik, J., Nernec, M., and Vlasta, S. (1969). Transplant. Proc. 1, 527. Hattler, B. G . , and Miller, J. ( 1973). Transplant. Proc. 5, 655. HeaIe, W. F., Morris, P. J., Bennett, R. C., Mortensen, P. J., and Ting, A. (1969). Med. J. At&. 2, 382. Hellstrom, I., HellstrBni, K. E., and Allison, A. E. (1971). Nature (London) 230, 49. Hellstroni, K. E., and Hellstroni, I. (1974). Adcan. Zmmunol. 18, 209. Henson, P. M. ( 1974). Transplant. Proc. 6, 27. Hersey, P. (1973a). Brit. J. Cancer 28, Suppl. 1, 11. Hersey, P. ( 1973b). Nature (London), New Biol. 244, 22. Hildemann, W. H., and Mullen, Y. (1973a). Transplantation 15, 231. Hildemann, W. H., and Mullen, Y. (19731,). Transplant. Proc. 5, 617. Hobbs, J. B., and Cliff, W. J. (1973). J. Exp. Med. 137,776. Holter, A. R., McKearn, T. J., Neu, M. R., Fitch, F. W., and Stewart, F. P. (1972). Transplantation 13, 244. Holter, A. R., Neu, M. R., McKearn, T. J., Lynch, A. F., and Stuart, F. P. (1973). Transplant. Proc. 5, 593. Horowitz, R. E., Burrows, L., Paronetto, F., Dreiling, D., and Kark, A. E. (1965). Transplantation 3, 318. Hume, D. M., and Bryant, C. P. ( 1973). Transplant. Proc. 4, 673. Ippolito, R. J., Mahoney, R. J., and Murray, I. M. ( 1974). Tiatisplantation 17, 89. Iwasaki, Y., Talmadge, D., and Starzl, T. E. (1967a). Transplantation 5, 191. Iwasaki, Y., Porter, K. A,, Amend, J. R., Marchioro, T. L., Ziihlke, V., and Starzl, T. E. (1967b). Surg., Gynecol. Obstet. 124, 1. Jeannet, M., Pinn, V. W., Flax, M. H., Winn, H. J., and Russell, P. S. (1970). N . Engl. 1. Med. 282, 111. Jeannet, M., Vassalli, P., and Botella, F. ( 1975). Transplant. Proc. ( i n press). Jones, J. M., and Feldman, J. D. (1971). Transplantation 12, 114. Jones, J. M., Peter, H. H., and Feldman, J. D. (1972). J. Zmtntcnol. 108, 301. Judd, K. P., and Trentin, J. J. ( 1973). Transplantation 16, 351. Kaliss, N. ( 1969). Znt. Rev. Exp. Pathol. 8, 241. Katz, D. H. (1972). Transplant. Rev. 12, 141. Katz, D. H., Unanue, E. R., and Dixon, F. J. ( 1967). J. Immtcnol. 98, 260. Katz, D. H., Haniaoka, T., Dorf, hi. E., and Benacerraf, B. (1973). Proc. Nat. Acad. Sci. U.S. 70, 2624. Kilshaw, P. J., Brent, L., and Thomas, A. V. ( 1974). Transplantation 17, 57. Kim, J. P., Shaipanich, T., Sells, R. A,, Maggs, P., Lukl, P., and Wilson, R. E. (1972). Transplantation 13, 322 Kincaid-Smith, P. ( 1967). Lancet 2, 849. Kincaid-Smith, P., Morris, P. J., Saker, B. M., Ting, A., and Marchall, V. C. (1968). Lancet 2, 748. Kissnieyer-Nielsen. F., Olsen, S., Petersen, V. P., and Fjeldborg, 0. (1966). Lancet 2, 6'62.

60


Klassen, J,, and Milgrom, F. (1971). Transplant. Proc. 3, 598. Klassen, J., Kano, K., Milgrom, F., Menno, A., Anthone, S., Sepulveda, M., Elwood, C., and Andres, G. A. ( 1973). Int. Arch. Allergy Appl. lmmunol. 45, 675. Klein, J., and Shreffler, D. C. (1971). Transplant. Reu. 6, 3. Kobayashi, K., Hricko, G. M., Lukl, P., Hunsicker, L., Patel, R., Reisner, G. S., and Birtch, A. G. ( 1971). Surg. Forum 22,246. Kobayashi, K., Hricko, G. M., Habal, M. B., Lukl, P., Busch, G. J., Hunsicker, L., Reisner, G. S., and Birtch, A. G. ( 1972 ). Transphtation 14, 374. Kretschmer, R. R., and Perez-Tamayo, R. (1961). J. Exp. Med. 114, 509. Kretschmer, R. R., and Perez-Tamayo, R. (1962). J. Erp. Med. 116, 879. Kux, M., Boehmig, H. J., Amemiya, H., Torisu, M., Yokoyama, T., Launois, B., Popovtzer, M., Wilson, C. B., Dixon, F. J., and Starzl, T. E. (1971). Surgery 70, 103. Kux, M., Piza, F., Rauhs, R., Simma, W., Stemberger, H., Stockinger, L., and Wiedermann, G. ( 1973). Z. Immunitaetsforsch 145, 432, Kux, M., Piza, F., Steinberger, H., Wagner, O., and Wiedermann, G. (1974). Eur. Surg. Res. 6, 5. Lamon, E. W., Skurzak, H. M., KIein, E., and Wigzell, H. (1972). J. Exp. Med. 136, 1072. Lance, E. M., Boyse, E. A., Cooper, S., and Carswell, E. A. (1971). Transplant. Proc. 3, 864. Larner, B. J., and Fitch, F. W. (1973). Transplantation 16, 54. Larsson, A., and Perlmann, P. ( 1972). Int. Arch. Allergy Appl. Immunol. 43, 80. Larsson, A., Perlmann, P., and Natvig, J. (1973). Immunology 25, 675. Lee, S. T., and Paraskevas, F. (1972). J. lmmunol. 109, 1262. Lehman, D. H., Lee, S., Wilson, C. B., and Dixon, F. J. (1974). Transplantation 17, 429. Lerner, R. A., Glassock, R. J., and Dixon, F. J. (1967). J. Erp. Med. 126, 989. Levey, R. H. ( 1971). Transplant. Proc. 3, 41. Levey, R. H. (1972). Transplant. Proc. 4,543. Lindquist, R. R., Guttmann, R. D., and Merrill, J. P. (1968a). Amer. J. Pathol. 52, 531. Lindquist, R. R., Guttmann, R. D., Merrill, J. P., and Dammin, G. J. (1968b). Amer. J. Pathol. 53, 851. Lindquist, R. R., Guttmann, R. D., and Merrill, J. P. (1971). Transplantation 11, 1. Lopez, C., Simmons, R. L., Mauer, M., Park, B., Najarian, J. S., and Good, R. A. ( 1973). Transplant. Proc. 5, 803. Lowenhaupt, R. W., and Nathan, P. (1968). Nature (London) 220, 822. Lowenhaupt, R. W., and Nathan, P. (1969). Transplant. Proc. 1, 305. Lowenhaupt, R. W., Nathan, P., and Menefee, M. G. (1971). Transplant. Proc. 3, 453. Lubbe, F. H., Eastham, W. N., and van der Herik, A. (1972). Transplantation 14, 649. Lucas, Z. J., Markley, J., and Travis, M. (1970). Fed. Proc., Fed. Amer. SOC. Exp. Biol. 29, 2041. MacDonald, A. G., Busch, G. J., Alexander, J. L., Pheteplace, E. A., Menzoian, J., and Murray, J. R. (1970). Transplantation 9, 1. McDonald, J. C., Milgrorn, F., and Witebsky, E. (1964). J. Amer. Med. Ass. 190, 122.


61

McDowall, R. A. W., Batchelor, J. R., and French, M. E. (1973). Lancet 2, 797. McKearn, T. J. ( 1974). Science 183, 94. McKearn, T. J., Stuart, F. P., and Fitch, F. W. (1974a). 1. Immunol. 113, 1876. McKearn, T. J., Finnegan, A., Neu, M. R., Fitch, F. W., and Stuart, F. P. (1974b). Surg. Forum 25, 263. McKenzie, I. F. C., and Snell, G. D. (1973). 1. Exp. Med. 138, 259. McKenzie, I. F. C., and Snell, G. D. (1974). Transplantation 17, 328. McKenzie, I. F. C., and Wittingham, S. (1968). Lancet 2, 1313. MacLennan, I. C. M. (1972). Transplant. Rev. 13, 67. MacLennan, I. C. M., and Harding, B. (1970). Nature (London) 227, 124. MacLennan, I. C. M., and Howard, A. (1972). Immunology 22, 1043. MacLennan, I. C. M., Loewi, G., and Howard, A. (1969). Immunology 17, 897. MacLennan, I. C. M., Loewi, G., and Harding, B. (1970). Immunology 18, 397. MacLennan, 1. C. M., Howard, A., Gotch, F. M., and Quie, P. G. (1973). Immunology 25, 459. McPhaul, J. J., Jr., Dixon, F. J., Brettschneider, L., and Starzl, T. E. (1970). N. Engl. I. Med. 282, 412. McPhaul, J. J., Jr., Thompson, A. L., Jr., London, R. E., Klebanoff, G., Cosimi, A. B., DeLemos, R., and Smith, R. B. (1973). J. C h . Inoest. 52, 1059. Makela, 0. (1970). Transplant. Reo. 5, 3. Medawar, P. B. (1944). J. Anat. 78, 176. Medawar, P. B. ( 1956). Phil. Tmns. Roy. Soc. London, Ser. B 239, 357. Merrill, J. P. (1967). Aduan. Immunol. 7, 276. Milgroni, F., Litvak, B. I., Kano, K., and Witebsky, E. (1966). J. Amer. Med. ASS. 198, 136. Miller, J., Rood, F., Lifton, J., and Hattler, B. G . (1973). Transplantation 16, 142. Miller, J. F. A. P. ( 1961). Lancet 2,748. Mittal, K. K., Mickey, M. R., Singal, D. P., and Terasaki, P. I. (1968). Transplantation 6, 913. Mittal, K. K., Terasaki, P. I., Springer, G. F., Desai, P. R., McIntire, F. C., and Hiroeta, A. A. ( 1973). Transplant. Proc. 5, 499. Moller, E., and Moller, G. ( 1962). J. Exp. Med. 115, 527. Moller, G. (1963). J. Nat. Cancer Inst. 30, 1153. Moller, G. (1964). Transplantation 2, 281. Moller, G., and Svehag, S. E. ( 1972). Cell. Immunol. 4, 1. Morris, P. J., and Dumble, L. J. ( 1973). Lancet 2, 16. Morris, P. J., Ting, A., and Kincaid-Smith, P. ( 1970). I n “Histoconipatibility Testing” ( P. I. Terasaki, ed.), p. 371. Munksgaard, Copenhagen. Morris, R., and Lucas, Z. J. (1971). Transplant. Proc. 3, 697. Mowbray, J. F. (1967). lntegration Intern. Med., Proc. Int. Congr., 9th 1966 Excerpta Med. Int. Congr. Ser. No. 137, p. 106. Mullen, Y., and Hildeniann, M. W. ( 1971). Trunsplant. Proc. 3, 669. Mullen, Y., Takasugi, M., and Hildemann, W. H. (1973). Transplant. Proc. 5, 611. Myburgh, J. A . , and Smit, J. A. (1972). Transplantation 14, 237. Myburgh, J. A., Cohen, I., Gecclter, L., Meyers, A. M., Abrahams, C., Furman, K. I., Coldberg, B., and Van Bleck, P. J. P. (1969). N . Engl. 1. Med. 281, 131. Myburgh, 5. A., Maier, G., Smit, J. A., Shapiro, M., Meyers, A. M., Rabkin, R., Van Bleck, P. J. P., and Jersky, J. (1974a). Transplantation 18, 206.

62


Myburgh, J. A., Maier, G., Smit, J. A,, Shapiro, M., Meyers, A. M., Rabkin, R., Van Bleck, P. J. P., and Jersky, J. (197413). Transplantation 18, 213. Najarian, J. S., and Feldman, J. D. (1962a). J. E x p . Med. 115, 1083. Najarian, J. S., and Feldman, J. D. (1962b). Ann. N.Y. Acad. Sci. 99, 470. Najarian, J. S., and Foker, J. E. (1969). Transplant. Proc. 1, 1022. Nelson, D. S. ( 1962). Brit. J. Exp. Pathol. 43, 1. Ockner, S. A,, Guttmann, R. D., and Lindquist, R. R. (1970a). Transplantation 9, 30.

Ockner, S. A,, Guttmann, R. D., and Lindquist, R. R. (1970b). Transpkmtation 9, 39. Old, L. J., Stockert, E., Boyse, E. A., and Kim, J. H. (1968). J . Ezp. Med. 127, 523. Oliver, R. T. D., Sachs, J. A., Festenstein, H., Pegmm, G. D., and Moorhead, J. F. (1972). Lancet 2, 1381. Opelz, G., Mickey, M. R., and Terasaki, P. I. (1972). Lancet 1, 868. Ortiz-de Landazuri, M., and Herberman, R. B. ( 1972). 1. Exp. Med. 136, 969. O’Toole, C., Stejskal, V., Perlmann, P., and Karlsson, M. (1974). j . Exp. Med. 139. 457. Parkes, A. S. (1958). Transplant. Bull. 5, 45. Patel, R., and Terasaki, P. I. (1969). N. Engl. J. Med. 280, 735. Patel, R., Merrill, J. P., and Briggs, W. A. (1971). N. Engl. J. Med. 285, 1974. Pedersen, N. C., and Morris, B. (1974a). Transphtation 17, 48. Pedersen, N. C., and Morris, B. (1974b). J. Exp. Med. 140, 619. Perlmann, P. ( 1972). Transplant. Proc. 4, 295. Perlniann, P. ( 1975). Clin. Immzmobiol. 3, (in press). Perhann, P., and Holm, G. ( 1969). Aduan. Immunol. 11, 117. Perlmann, P., and Perlmann, H. (1970). Cell. Immunol. 1, 300. Perlniann, P., Perlmann, H., and Wigzell, H. (1972). Transplant. Rev. 13, 91. Peter, H. H., and Feldnian, J. D. ( 1972). J. Exp. Med. 135, 1301. Phillips, S. M., Strom, T. B., Corson, J. M., Carpenter, C. B., and Merrill, J. P. ( 1973). Transplant. Proc. 5, 577. Porter, K. A. (1966). In “Pathology of the Kidney” (R. H. Heptinstall, ed.), p. 601. Little, Brown, Boston, Massachusetts. Porter, K. A. (1967). J. Clin. Pathol. 20, Suppl. 518. Porter, K. A., Calne, R. Y., and Zukoski, C. F. (1964). Lab. Inuest. 13, 809. Porter, K. A., Andres, G. A., Calder, M. W., Dossetor, J. B., Hsu, K. C., Rendall, J. M., Seegal, B. C., and Starzl, T. E. (1968). Lab. Invest. 18, 159. Quadracci, L. J., Hellstrom, I., Striker, G. E., Marchioro, T. L., and Hellstrom, K. E. (1971). Cell. Immunol. 1, 561. Quadracci, L. J., Tremann, J. A., Marchioro, T. L., and Striker, G. E. (1974). Transplantation 17, 361. Ramseier, H., and Lindemann, J. (1971). J. Erp. Med. 134, 1083. Rao, V. S., Bonavida, B., Zieghelboim, J., and Fahey, J. L. (1974). Transplantation 17. 568. Rapaport, F. T., and Chase, R. M., Jr. ( 1965). J. Exp. Med. 122, 733. Rapaport, F. T., Chase, R. M., Jr., and Solowey, A. C. (1968). Ann. N.Y. Acad. Sci. 129, 102. Rapaport, F. T . , Markowitz, A. S., and McCluskey, R. T. (1969). J. Exp. Med. 129, 623.


63

Rapaport, F. T., Chase, R. M. Jr., Markowitz, A. S., McCluskey, R. T., Shiniada, T., and Watanabe, K. (1971). Transplant. Proc. 3, 89. Rivard, J. V., Beaudoin, J. G., and Guttmann, R. D. (1973). Transplantation 16, 88. Rowley, D. A., Fitch, F. W., Stuart, F. P., Kohler, H., and Consenza, H. (1973). Science 181, 1133. Ruddy, S. ( 1974). Transplant. Proc. 6, 1. Russell, P. s. ( 1971). Transplant. Proc. 3, 960. Salaman, J. R. (1970). Brit. Med. J. 2, 517. Salaman, J. R., Elves, M. W., and Festenstein, H. ( 1971). Transplant. Proc. 3, 577. Schiff, M., Goffinet, J. A., Hendler, E., Ruddle, N., Kashgarian, M., and Lytton, B. ( 1974). Transplantation 17, 580. Schubert, W. K., Fowler, R., Jr., Lester, W. M., and West, C. D. (1960). Transplant. Bull. 26, 125. Shaipanich, T., Vanwjick, R. R., Kim, J. P., Lukl, P., Busch, G. J., and Wilson, R. E. (1971). Surgery 70, 113. Shehadeh, I. H., Carpenter, C. B., Monterio, C. H., and MerriII, J. P. (1970). Arch. Intern. Med. 125, 850. Sheil, A. G . R., Stewart, J. H., Tiller, D. J., and May, J. (1969). Transplantation 8, 299. Shreffler, D. C . , and David, C. S. (1975). Adunn. Immunol. 20, 1.25. Simmons, R. L., Weil, R., Tallent, M. B., Kjellstrand, C. M., and Najarian, J. S. (1970). Transplant. Proc. 2, 419. Simonsen, M., Beumann, J., Gammeltoft, A,, Jensen, F., and Jorgensen, K. (1953). Acta Pathol. Microbial. Scand. 32, 1. Sinclair, N. R., St. C. (1969). J. Exp. Med. 129, 1183. Skoskiewicz, M., Chase, C., Winn, H. J., and Russell, P. S. (1973). %znspEant. Proc. 5, 721. Smit, J. A., and Myburgh, J. A. (1974). Transplantation 18, 63. Soterides-Vlachos, C., Gyongyossy, M. I. C., and Playfair, J. H. L. (1974). Clin. E r p . Immunol. 18, 187. Spong, F. W., Feldman, J. D., and Lee, S. (1968). J . Immunol. 101, 418. Staines, N. A., C u p , K., and Davies, D. A. L. (1974). Transplantation 18, 192. Starzl, T. E., Marchioro, T. L., Holmes, J. H., and Waddell, W. R. (1964). Surg., Gynecol. Obstet. 118, 819. Starzl, T. E., Lerner, R. A., Dixon, F. J., Groth, G. G., Brettschneider, L., and Terasaki, P. I. (1968). N . Engl. J. Med. 278, 642. Steinmuller, D. (1962). Ann. N . Y . Acad. Sci. 99, 629. Stenzel, K. H., Whitsell, J. C., Cheigh, J. S., Riggio, R. R., Stubenbord, W. T., Sullivan, J. F., Rubin, A. L., and Fotino, M. ( 1974). Transplant. Reo. 6, 89. Stetson, C. A. ( 1963). Aduan. Immunol. 3, 97. Stetson, C. A., and Demopoulos, R. (1958). Ann. N.Y. Acad. Sci. 73, 687. Straus, A., Straus, F., 11, and Roseman, D. (1971). Transplant. Proc. 3, 570. Strom, T. B., Garovoy, M. R., Carpenter, C. B., and Merrill, J. P. (1973a). Science 181. 171. Strom, T . B., Carpenter, C. B., Phillips, S. M., Garovoy, M. R., and Merrill, J. P. ( 1973b). Transplantation 16, 103. Strom, T. B., Tilney, N. L., Carpenter, C. B., and Busch, G. J. (1975a). N . Engl. J. Med. 292, 1257.

64


Strom, T. B., Carpenter, C. B., Garovoy, M. R., Abbas, A. K., Corson, J. M., and Soulillou, J. P. ( 197513). Transplantation (in press). Stuart, F. P., Saitoh, G., and Fitch, F. W. (1968). Science 160, 1463. Stuart, F. P., Fitch, F. W., and Rowley, D. A. (1970). Transplant. Proc. 2, 483. Stuart, F. P., Fitch, F. W., Rowley, D. A., Biesecker, J. L., Hellstrom, K. E., and Hellstrom, I. ( 1971). Transplantation 12, 331. Sybesnia, J. P., Kater, L., Borst-Eilers, E., dePlanque, B. A,, Van Soelen, T., and Tuit, T. (1974). Transplantation 17, 576. Terasaki, P. l., Akiyana, T., McCleland, J. D., and Cannon, J. A. (1962). Ann. N.Y. Acad. Sci. 99, 645. Terasaki, P. I., Thrasher, D. L., and Hauber, T. H. (1968). In “Advances in Transplantation’’ (3. Dausset, J. Hamburger, and G . MathB, eds.), p. 15. Munksgaard, Copenhagen. Terasaki, P. I,, Mickey, R. M., and Kreisler, M. (1971). Postgrad. Med. I . 47, 89. Thiel, G., Moppert, J., Mahlich, J., Biihler, F., Vischer, T., Enderlin, F., Weber, H., and Zollinger, H. U. (1971). Transplant. PTOC.3,741. Tilney, N. L., and Bell, P. R. F. (1974). Transplantation 18, 31. Tilney, N. L., Strom, T. B., MacPherson, S. G., and Carpenter, C. B. (1975). Transplantation ( in press ) . Ting, A., and Terasaki, P. I. (1974). Transplantation 18,371. Uhr, J. W., and Moller, G. (1968). Aduan. Immunol. 8, 81. Van Bekkum, D. W., Heystek, G . A., and Marquet, R. L. (1969). Transplantation 8, 678. Van Boxel, J. A., Stobo, J. D., Paul, W. E., and Green, I. (1972). Science 175, 194. Van Hooff, J. P., Schippers, H. M. A., van der Steen, G . J., and Van Rood, J. J. ( 1972). Lancet 2, 1385. Van Rood, J. J. ( 1974). Transplant. Proc. 5, 1747. Van Rood, J. J., van Leeuwen, A., Keuning, J, J., and Termijtelen, A. (1975). Transplant. Proc. (in press). Voisin, G. A. ( 1971). Progr. Allergy 15, 328. von Haefen, U., Wren, S. F. G., Shaipanich, T., Martins, A. C. P., Busch, G . J., and Wilson, R. E. ( 1973). Transplantation 16, Q95. Wakefield, J. D., and Amos, D. B. (1958). Proc. Amer. Ass. Cancer Res. 2, 354. Waldmann, T. A., and Strober, W. (1969). Progr. Allergy 13, 1. Wardle, E. N. ( 1972a). J . Roy. Coll. Physicians London 7,5. Wardle, E. N. (1972b). In “Microcirculation,” Vol. 2, BibI. No. 12, p. 64. Karger, Basel. Warren, R. P., Lofgreen, J. S., and Steinmuller, D. (1972). Transplantation 16, 458. Wasserman, J., PackalBn, T., and Perlniann, P. (1971). Int. Arch. Allergy Appl. Immunol. 41, 910. Weymouth, R. J., Seibel, H. R., Lee, H. M., Hume, D. M., and Williams, G. M. ( 1970):Amer. J . Pathol. 58, 85. White, E., and Hildemann, W. H. (1968). Science 162, 1293. White, E., Hildemann, W. H., and Mullen, Y. (1969). Transplantation 8, 602. Williams, G. M., Lee, H. M., Weymouth, R. F., Harlan, W. R., Tr., Holden, K. R., Stanley, C. M., Millington; G . A., and Hume, D. M. (1967).-Surgery 62, 204. Williams, G . M., Hume, D. M., Hudson, R. P., Jr., Morris, P. J., and Milgrom, F. (1968). N . Engl. 1. Med. 279, 611.

REJECrION AND ENHANCEMENT O F ORGAN ALLOGRAFTS

65

Williams, M. A., Morton, M., Tyler, H. M., and Dempster, W. J. (1964). J. Erp. Pathol. 45, 235. Wilson, C. B., Dixon, F. J., Fortner, J. G., and Cerilli, J. ( 1971). J. Clin. Znoest. 50, 1525. Wilson, C. B., Lehman, D. H., McCoy, R. C., Gunnells, J. C., Jr., and Stickel, D. L. (1974). Transplantation 18, 447. Wilson, R. E., Rippin, A., Dagher, R. K., Kinneart, P., and Busch, G. J. (1969). Transplantation 7, 360. Wilson, R. E., Maggs, P. R., Vanwjick, R., Shaipanich, T., Holl-Allen, R. T. J., Lukl, P., and Sinionian, S. J. (1971). Transplant. Proc. 3, 705. Winn, H. J. (1970). Transplant. Proc. 2, 83. Winn, H. J. (1974). Ann. N.Y. Acad. Sci. 101, 23. Winn, H . J., Baldamus, C. A,, Jooste, S. V., and Russell, P. S. (1973). J. Exp. Med. 137, 893. Woodruff, M. F. A. (1957). Ann. N. Y. Acad. Sci. 64, 1014. Wu, P. P. T., and Mann, F. C. (1934). Arch. Stirg. (Chicago) 28, 889. Yamada, T., and Kay, J. H. (1968). Surgery 63, 637. Yoshida, T. O., and Andersson, B. ( 1972). Scand. J. Immtinol. 1, 401. Zimmerman, B., and Feldnian, J. D. ( 19G9a). I. Zmmzlnol. 102, 507. Zimmerman, B., and Feldman, J. D. ( 1969L). J. Kmmunol. 103,383. Zimmerman, B., and Feldman, J. D. ( 1970). 1. Immunol. 104, 628. Zimmerman, C. E. (1971). Transplant. Proc. 3, 701.

Biosynthesis of Complement

.

HARVEY R COLTEN’ Division o f Alfergy. Department o f Medicine. Children’s Hospital Medical Center and Deportment of Pediatrics. Harvord Medical School. Boston. Morrochuretts

I . Introduction . . . . . . . . I1. Historical . . . . . . . . . I11. Methods for Studies of Complement Synthesis A . Tissue Culture Conditions: Media and Cells . B. Assay Systems . . . . . . . C . Biosynthesis in Vioo . . . . . . D . Criteria . . . . . . . . IV. Sites of Synthesis of Complement Components . A . Biosynthesis of the First Component . . B. Biosynthesis of the Second Component . . C . Biosynthesis of the Fourth Component . . D . Biosynthesis of the Third Component . . E . Biosynthesis of C5 and C6 . . . . F. Biosynthesis of C7 and C8 . . . . G . Biosynthesis of C9 . . . . . . H. Complement-Associated Proteins . . . V. Ontogeny of Complement . . . . . VI . Genetic Regulation of Complement Synthesis . A . C4 Deficiency . . . . . . . B. C5 Deficiency . . . . . . . C. C2 Deficiency . . . . . . . D. Deficiency of C1 Inhibitor . . . . VII . Nongenetic Control of Complement Biosynthesis VIII . Concluding Remarks . . . . . . References . . . . . . . . .

.

.

. .

.

.

.

. .

.

.

.

.

.

.

.

. .

. .

. .

.

.

.

. . .

. . .

. . .

. . .

. . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

. . . .

. . . .

. . . .

. . . .

. . . .

. .

. .

. .

. .

. .

.

.

.

.

.

67 70 74

74 77 80 81 82 82 85 87 89 92 94 95 95 96 97 97 101 102 104

105 112 113

I . Introduction

In the past two decades. considerable progress has been made in the study of complement biosynthesis with the isolation and characterization of the individual components of complement. the design of suitable immunochemical and functional assays for complement proteins. and improvements in tissue and organ culture techniques . These advances have made it possible to identify the sites of synthesis of most of the complement proteins . Moreover. with methods developed in the course of biosynthesis studies. an approach to questions of more general biological interest was made possible . Although the genetic heterogeneity of * Recipient of a United States PubIic Health Service Research Career DeveIopment Award 5 KO4 HD70558 . 67

68

HARVEY R. COLTEN

complement by no means approaches that of immunoglobulins, genetic variants of complement proteins have been recognized and well defined deficiencies of complement have been described in man and experimental animals, providing excellent models for investigations of genetic control of protein synthesis. In addition, changes in serum complement resulting from the “acute phase response” or in specific diseases, such as systemic lupus erythematosis and some forms of nephritis, have raised important general questions about the control of plasma protein metabolism. Some of these questions can be approached with current methods. One purpose of this chapter is to emphasize these actual and potential applications of studies of complement biosynthesis. In addition, a review of the historicaI aspects of work on this problem will be presented in an attempt to credit the insights of some of our predecessors and to point out the many problems that, to this date, confuse and inhibit progress in this area. It is hoped that such exercise may enhance the former and reduce the latter. Considerable attention will therefore be given to a discussion of modern methodology and criteria for demonstrating net biosynthesis by isolated cells and tissues in oitro. After a discussion of the localization of the sites of complement biosynthesis, we will turn to a consideration of the factors that affect, at a genetic and microenvironmental level, rates of complement production. For many years, progress in complement research and the interest of most immunologists in the complement system were limited by a rather confusing nomenclature. In 1968, a standardized nomenclature was proposed (Austen et al., 1968). According to this convention, the individual complement components are designated numerically in the order of their reaction sequence as follows: C1, C4, C2, C3, C5, C6, C7, C8, C9. Since an understanding of the functional properties of the complement components was initially derived from studies of immune hemolysis, a convention was established for designating the reaction sequence and intermediate complexes on cell surfaces. E represents erythrocytes; A signifies antibody directed against the erythrocytes; and the intermediates are indicated by EAC, followed by the numbers corresponding to the individual components that have reacted with EA. For example, EAC142 describes an intermediate consisting of erythrocytes sensitized with antibody and containing the first, fourth, and second components of complement. Activated forms of the components are designated by a bar above the symbol; e.g., activation of C1 to Ci. Fragments of the components are indicated by a lowercase letter; e.g., C3a, C3b, C ~ CC3d. , The complement sequence and the biological activities generated from the complement system are outlined in Figs. l a and b. Only tentative agreement has been established on nomenclature

BIOSYNTHESIS OF COMPLEMENT S A + Cl

CaZ

*

69

* saci

c 4

SAC 1 4 b Z a 3 b + C 3 a

C3 INACTIVATOR c5

I

SAC 14 b 2a3b5b t C 5 a

c6

/

SAC 14bZa3b5bS

C 6 INACTIVATOR L’

C789

SAC I

t

L CHEMOTAXIS HISTAMINE RELEASE SMOOTH M U S C L E C O N T R A C T I O N INCREASED VASCULAR PERMEABILITY

C3a

142356789

1

I\

*c1423

CHEMOTAXIS

C6.7.0.9

C(-9

(b) FIG. 1. ( a ) The complement sequence: S indicates site on a cell surface; indicates point at which inhibitor acts in the sequence. ( b ) Complement-dependent biological activities.

70

HARVEY R. COLTEN

for the proteins of the properdin or alternative pathway of complement activation, and few studies of biosynthesis of these proteins have been performed. 11. Historical

Complement was discovered in the late 19th century, when it was recognized that the bactericidal activity of fresh serum required the participation of at least two factors: a heat-stable substance, specific for the particular organism (antibody); and a second, nonspecific, heatlabile factor, now designated complement ( Buchner, 1889; Pfeiff er and Issaeff, 1894; Bordet, 1896). Within 5 years after its discovery, several studies designed to determine the site of complement production or release had been performed. Initial interest in determining the origin of complement was stimulated by attempts to integrate this newly discovered activity into emerging and competing theories of the cellular or humoral basis of resistance to microbial infections. These early experiments and conclusions drawn therefrom were limited by at least two considerations. Only fragmentary data were then available on complement and its biological activities. Indeed evidence for a multicomponent system was not forthcoming until more than 15 years after the first attempts to study complement synthesis ( Ferrata, 1907). Second, discovery of a soluble bacteriolytic activity fanned the controversy between adherents to the cellular and humoral theories of immunity. As a consequence, studies by each school were designed with substantial bias, in an effort to support their conflicting claims. In 1901 Metchinoff published a monograph (English translation, Metchnikoff, 1905) on his phagocytic theory of immunity, in response to criticism raised at the International Congress in Paris in the preceding year. In this account he surveyed complement in the context of the phagocytic theory, citing his own work (Metchnikoff, 1889) which indicated that at least a portion of the bactericidal activity of serum was derived from substances that escaped from leukocytes during clotting. At about the same time Hankin (1892) suggested that these cytolytic factors were products of eosinophilic leukocytes, but this view was soon abandoned owing to lack of definitive experimental observations. Buchner (1894) agreed with Metchnikoffs view that leukocytes were the source of this lytic agent but again experimental data were lacking. Later work by Gengou (1901) supported Metchnikoffs theory as he claimed that bacteriolytic activity was present only in serum and was not found in plasma. It is testament to the force of Metchnikoffs prestige that although between the appearance of Gengou’s studies and those of Addis in 1912 no fewer than eight papers appeared that contradicted

BIOSYNTHESIS OF COMPLEMENT

71

Gengou’s findings; nonetheless, Metchnikoffs theory on the origin of complemcnt was still widely held. In Addis’ (1912) experiment, fowl plasma was carefully collected in the absence of anticoagulants. Addis demonstrated a lack of thromboplastic activity, but at the same time complement activity in the plasma. H e argued that if cells had bccn damaged during handling of the blood, the plasma would have coagulated owing to release of thromboplastin. Since abundant complement activity was present in the thromboplastin-free plasma, leukocytes could not be the source of complement. It should be noted that Metchnikoff did not suggest that complement was synthesized in leukocytes. Although he speculated that macrophages might secrete complement during their life-span, he favored the vicw that cell death was required for release of the lytic acitivity. This controversy subsided after presentation of direct evidence obtained in experiments such as those of Addis and studies in vivo (Ecker and Rces, 1922), only to be resurrected in modern studies of Complement synthesis. No doubt some of the confusion in interpretation of thew early experiments was due to uncertainty as to whether each investigator was studying the same substance. Later findings of lysozyme synthesis in monocytes (Gordon et al., 1974) and other lytic activities (Hirsch, 1956) in white cell extracts makes this possibility likely. . Interest in the liver as a site of complement production was first generated by experiments of Ehrlich and Morgenroth (1900), which demonstrated a clear distinction between cells that synthesized antibody from those that synthesized complement, based on evidence that only complement, not antibody, was depressed in sera of animals poisoned with phosphorus. Shortly thereafter, Miiller ( 1911) found that perfusion of liver with serum led to an increase in complement activity in the effluate and concluded that the liver was the site of complement synthesis. Dick (1913), in an attempt to refine these cxperinients, perfused dog organs with saline until free of blood, thcn homogenized each and sought complement activity in the tissue cxtracts-but found none. H e then turned to in vivo experiments, which revealed no change in serum complement associated with surgical rernovnl of splccn, pancreas, adrenals, or a portion of the small intcstine, nor was there any effect of mercuric chloride poisoning on complement levels although extensive damage was done to kidney and colon. On the other hand, administration of chloroform led to a marked decrease in serum complement activity which paralleled histologic evidence of liver damage. He also trcated dogs with hydrazine, which did decrease complement levels, but he did not know, nor did he do experiments to show, that hydrazine inactivated preformed complement, as was demonstrated later by Ecker and Pillimer (1942). At the

72

HARVEY R. COLTEN

time of Dick's experiments at least two, and possibly three, components of complement were recognized. Dick expected to find a significant decrease in the euglobulin fraction of complement since Whipple and Hurwitz (1911) had shown that chloroform destruction of dog liver resulted in decreased plasma levels of another globulin, i.e., fibrinogen. This expectation reflected the incomplete understanding of the multiplicity of distinct plasma proteins. Experimentally, midpiece ( complement components in euglobulin fraction) was unaffected by chloroform destruction of liver whereas endpiece (the pseudoglobulin complement components ) was significantly depressed. This perhaps reflected the marked decrease in C4 observed in later experiments of this type. Several years later Olsen (1922) also attempted to repeat a variant of Muller's experiment, but with refinements made possible by the ability to detect three distinct complement activities. He established that the liver was a source of several complement components. Perfusion of guinea pig liver with a solution containing antibody-sensitized sheep erythrocytes ( E A ) resulted in lysis of the cells, particularly if small amounts of either midpiece or endpiece were added to the perfusate. He further suggested that the liver was a source of a heat-labile complement component since maximal lysis was obtained when the liver was perfused with EA in the presence of heated (55OC) serum, and erroneously concluded from this that the liver did not synthesize C3. The importance of fixed elements of the reticuloendothelial system in synthesis of complement was evidenced from the work of Jungeblut and Berlot ( 1926). These investigators studied changes in serum-complement levels as a function of time, following the intravenous injection of India ink. They found a marked decrease in total complement activity within 15 minutes of the injection and maximal depression at approximately 3 hours. Complement levels returned to normal at 24 hours and were well above normal on the third to fourth day after injection of the ink. Controls demonstrated that in vitro complement activity was removed, presumably by adsorption, when India ink was mixed with fresh serum. However, they demonstrated that the carbon particles disappeared from the circulation in a few seconds and reasoned that, since the maximum depression in complement activity was observed at 3 hours, the results were not simply due to adsorption of complement to the carbon particles. A concomitant change in metabolic acitvity of the reticuloendothelial cells of liver and spleen was demonstrated with methylene blue and nitroanthraquinone. This experiment revealed markedly reduced respiratory activity of the reticuloendothelial cells within 1-8 hours after injection of India ink and increased metabolic activity within 1 day after the injection. These results were not inconsistent with the


73

previously proposed theory that mobile phagocytic cells were capable of synthesizing complement, and in fact later experiments using modern techniques would indicate that both fixed and mobile phagocytic cells were capable of synthesizing at least some of the complement components. Moreover, they anticipated the subsequent studies that demonstrated changes in rates of complement synthesis with phagocytosis. It is surprising that a direct examination of the tissues synthesizing complement was not forthcoming for at least 30-35 years, since Carrel and Ingebrigtsen had shown synthesis of antibody in vitro as early as 1912. They incubated guinea pig bone marrow or lympyh node slices in the presence of goat erythrocytes and normal guinea pig serum. An increase in hemolytic activity, i.e., lysis of the goat cells, was noted as a function of time of incubation. A slight modification of this approach would have permitted in vitro studies of complement biosynthesis. The four classical components of complement had been separated and methods were available for measuring each, but from the late 1920s through the early 1950s little progress was made in further defining the complement system. Interest in studies of complement biosynthesis likewise declined until in 1951 Rice and her co-workers undertook a series of experiments on the effects of diet and several liver toxins on complement and coagulation activities ( Rice et al., 1951a,b; Rice, 1954). They showed that animals fed a protein-deficient diet had serum-complement levels that were only one-fifth to one-eighth those of controls. The most striking change was in the fourth component of complement, with lesser changes in levels of the third and second components. There was no effect on the first component of complement. Using carbon tetrachloride, they confirmed Dick's observation that administration of hepatotoxins resulted in marked depression of serum complement levels (Rice et al., 1951b). Again the fourth component and to a lesser extent the third, second, and first components of complement were decreased from pretreatment levels. Evidence that the liver was a site of complement synthesis also was derived from short-term experiments ( Gordon, 1955) in which evisceration of rats resulted in more than 90% decrease in complement activity 24 hours following the operation. Interesting as these experiments might have been, all were limited by the unsuitability of the methods for measuring complement and by the indirect metabolic effects on the experimental animals of removal or destruction of one or several organs. After the introduction of radiolabeled amino acids for studies of antibody synthesis in vitro (reviewed by Stavitsky, 1961), a technical problem reported by Keston and Katchen (1956) ultimately led to the first modern studies of complement biosynthesis (Hochwald et al., 1961). It was recognized in these early

74

HARVEY R. COLTEN

studies of antibody synthesis that immunoprecipitation of culture media containing 14C-labeled amino acid led to the adsorption of “nonspecific” radioactive material. This material was nondialyzable, increased in amount with increasing time of incubation, and was diminished in amount or absent in cultures of damaged tissues or incubation media deficient in amino acids. It was apparent that this “nonspecific” material was a protein or proteins apparently being synthesized by tissues under the conditions of these experiments. The development of methods for obtaining highly purified complement components and the production of antibodies against these specific proteins ( Miiller-Eberhard et al., 1960; Miiller-Eberhard and Biro, 1963) made it possible to exploit this observation for studies of complement biosynthesis in uitro. In a parallel development, Siboo and Vas (1965) attempted to measure appearance of complement activity in media harvested from cultures of guinea pig tissues. These experiments were not entirely successful owing to significant limitations in assay systems but they did indicate a direction of work in this area that followed further development and refinement of methods for measuring complement activities. It is unfortunate that, for some time, none of the workers studying complement biosynthesis followed the leads offered by the elegant experiments of Peters and Anfinsen (1950a,b), who showed that the liver was capable of synthesizing albumin in vitro. These experiments were highly instructive, since they were able to satisfy most of the currently accepted criteria for demonstration of net synthesis of protein in vitro. They showed that liver tissues incorporated radiolabel into a protein with the physicochemical characteristics of albumin and also showed a net increase in total protein and an increase in specific activity of the recovered albumin. With this background, the pertinent literature on complement biosynthesis of the past 15 years will be discussed in detail in Section IV as it applies to the sites of synthesis of the individual complement proteins. I l l . Methods for Studies of Complement Synthesis

A. TISSUE CULTURECONDITIONS: MEDIAAND CELLS The development of suitable tissue culture techniques for studies of protein synthesis by eukaryotic cells in vitro has progressed largely on a trial and error basis. Outside of requirements for nutritionally balanced media with optimal concentrations of ionic species and adequate buffering capacity there are few data that would permit a rational choice of the best medium for any specific experiment. The large variety of commercially available media testify to this generalization. In addition, sev-


7s

era1 supplements (e.g., calf serum, hormones, vitamins) are often added to improve longevity of cells and tissues in culture, frequently without complete understanding of the specific effects of these constituents. Additional considerations must be taken into account in the choice of medium, particularly if functional assay systems are used to monitor protein synthesis. For example, the protein must be stable under the conditions of culture. In studies of synthesis of the early-acting complement components C4 and C2, complex media such as Medium 199 (M199) or Minimal essential medium ( M E M ) with fetal calf serum have been used successfully because, in M199 or MEM, these components remain functional for up to 96 hours at 37OC (Colten, 1972a). For detection of biologically active C1, C3, and C5-9, two media may be substituted for M199 or MEM in short-term (72 hours) experiments (Colten, 1973). One is a commercially available preparation, Neuman and Tytell's serumless medium (Grand Island Biological Co. ), and the other is a special minimal medium designed for studies of biosynthesis of the late-acting complement components which is prepared from stock reagents to contain the following components: NaCl, 0.147 M ; dextrose, 0.012 M ; KCI, 0.4 mg/ml; NaHCO,, 3.96 mg/ml; MgCI, 10 M , CaCL, 1.5 X 10.' M ; NaH,PO,, 0.94 mglml; essential and nonessential amino acids; and stock vitamins. These reagents are dissolved in sterile pyrogen-free water; the solution is Millipore-filtered and then stored at -2OOC. Immediately before use, 10 M glutamine, 2 pg/ml of hydrocortisone succinate, 1% heated fetal calf serum, 50 U of penicillin per milliliter, and 50 pg of streptomycin are added. Studies of primary cell cultures or organ cultures require considerable preliminary preparation since all tissues contain large quantities of preformed complement components. In order to demonstrate synthesis of relatively small amounts of complement, in vitro baseline levels must be reduced without damaging the tissues. In studies of C1 synthesis (Colten et al., 1966), for examplc, this was accomplished by incubating the tissues in media containing ethylenediamine tetraacetate ( EDTA) to reduce baseline levels of CI from many thousands to less than one effective C1 molecule per tissue cell. This approach has the potential difficulty of differences in sensitivity of various cell types to damage from EDTA. Some methods of cell and tissue preparation avoid this problem. For instance, perfusion of organs with medium, extensive washing of cell suspensions, and incubation of tissues or cells in the presence of proteolytic enzymes have all becn used successfully to decrease baseline levels of preformed complement proteins. Established cell lines have also been used in many studies of complement synthesis. The use of permanent cell lines has many advantages in that large numbers of

76

HARVEY R. COLTEN

cells are easily obtained and, aside from problems of genetic drift, the cells are identical from time to time (i.e., reproducibility of conditions is maximized). However, interpretation of results of studies in which these cells are used requires caution since the cell lines generally must be maintained in media supplemented with serum and may be capable of adsorbing or ingesting proteins from the media. Release of these proteins must be distinguished from de nouo synthesis of protein. Furthermore, one cannot draw conclusions regarding the normal site of synthesis from the origin of the tissue or cell giving rise to permanent cell lines, although these cells may be particularly valuable for investigations of the regulation of inducible and noninducible cell functions. Identification of the specific cellular sites of complement biosynthesis has been approached by two methods; fluorescent antibody staining of individual cells with anticomplement component antibodies and a modification of the hemolysis in gel technique used successfully in studies of antibody synthesis. A third method, exploited to a limited degree in antibody synthesis studies, is immunoprecipitation in gel of the product of a single cell; it has not yet been applied to studies of complement producton. Fluorescent antibody techniques have the principal advantage of simplicity but are limited for biosynthetic studies by a number of considerations. First, the technique is qualitative; second, even with appropriate controls it is difficult or impossible to establish that the cell in question is actively synthesizing the specific protein; finally, the number or types of cells synthesizing the protein may be grossly underestimated since there is some evidence to suggest that under normal conditions plasma proteins are rapidly secreted after synthesis and assembly (Colten, 1974b). A method for detection of antibody production by single cells was introduced by Jerne and Nordin (1963) and Ingraham and Bussard (1964). This method depends on local production of hemolysin by antibody-forming cells suspended in a gel containing a lawn of indicator erythrocytes. After the addition of complement, plaquelike zones of hemolysis surrounding the antibody-forming cells are easily detectable. The basic method has undergone extensive refinement to increase sensitivity and to extend it to detect synthesis of virtually any specific antibody so long as antigen can be coupled to the indicator cells. At the Ciba Symposium on Complement, Lepow (1965) suggested that a modification of this method might permit studies of complement synthesis by individual cells so as to take advantage of the sensitivity and specificity of the hemolytic assays. Lepow’s suggestion has been picked up by several investigators and has been particularly useful in studies of synthesis of the early-acting complement components, CI, C4, and C2 (Colten et al., 1968b; Littleton et at., 1970; Wyatt et at., 1972).


77

Although many technical problems still remain, recent studies of the cell-free biosynthesis of C3 and C4 indicate that fundamental problems in molecular biology may yield to this approach (R. Hall and H. R. Colten, unpublished, 1975; R. Hall, P. Hansen, and H. R. Colten, unpublished, 1975). This is particularly significant when one considers that, in preliminary experiments, it has been possible not only to detect the synthesis of immunochemically identifiable C3 and C4 polypeptides but also the production of biologically active C4 in cell-free systems. Basically, methods in use at present consist of preparation of polysomes from liver, spleen, or peritoneal cells, priming with initiation factors from wheat germ or rabbit reticulocyte lysates, and monitoring of synthesis with immuiiochemical and functional assays for the individual complement proteins. The size of newly synthesized protein is estimated on polyacrylamide gels, and the presence of uniform incorporation of radiolabeled amino acids is detected with radioautography of peptide maps. Problems inherent in this approach are magnified by the relatively small amounts of specific complement protein synthesized in relation to total protein synthesis. This stands in contradistinction to the relatively large proportion of hemoglobin or immunoglobulin to total protein synthesized by polysomes prepared from reticulocytes and myeloma plasma cells, respectively.

B. ASSAYSYSTEMS Ideally one would wish to monitor production of complement in vitro by organs, cells, or subcellular components with a detection system that is sensitive, specific, and reproducible, is capable of detecting active and inactive complement proteins, and can distinguish newly synthesized from preformed protein. No single method satisfies all these criteria, but it is possible to approach the ideal condition with the use of several methods in combination. As noted earlier, the hemolytic assays for individual complement components are extremely sensitive and specific (Rapp and Borsos, 1970). Nanogram quantities of the biologically active component can be detected in complex mixtures-a decided advantage for measurements of complement in tissue culture media. In earlier studies of complement synthesis, and in occasional papers appearing more recently, synthesis of the active forms of the four classical components of complement, C1, C4, C2, C3 (now known to consist of C3 and C5-9), were detected with the so-called R reagents. These reagents designated R1, R2, R3, and R4 for the corresponding component that can be measured, are prepared by destroying or removing the appropriate component in whole serum. The practical limitations of use of these reagents are significant and now widely recognized (outlined by Mayer, 1961),

78

HARVEY R. COLTEN

so that they have been largely replaced by specific hemolytic assays, which depend on preparation of stable cell intermediates and the availability of functionally purified complement components. For example, in the modern hemolytic assay for C4, a sample is mixed with a cell intermediate consisting of erythrocytes sensitized with antibody and containing the first component of complement (EAC1). After a suitable incubation period, the cells are then exposed to the remaining complement components (C2, C3, C5-9) in excess, resulting in lysis of the erythrocytes. The extent of lysis ( a function of the hemoglobin released) is measured spectrophotometrically, and from this the absolute number of active C4 molecules in the sample can be calculated. Because of the specificity of these measurements, the biologically active complement components can be detected in complex mixtures of proteins without fractionating or concentrating the sample. The molecular assays for individual complement components are based on the one-hit theory of immune hemolysis. The one-hit theory states that a single effective site, resulting from the sequential action of antibody and the nine components of complement, is a necessary and sufficient condition for lysis of the erythrocyte. As a consequence, from the Poisson distribution, it is possible to calculate on a molecular basis the concentration of hemolytically active specific complement components present in a given sample. More detail on the theoretical and practical aspects of these assays can be found in Rapp and Borsos ( 1970). The hemolytic assays for individual complement components vary with respect to efficiency. For example, whereas a single molecule of C1 is sufficient to generate a hemolytic site (Colten et al., 1967), as many as 300 C3 molecules on the red cell surface are required for one lytic site ( Colten and Alper, 1972). Therefore, even discounting other potential problems of stability of the component in various tissue culture media and efficiency of uptake of the components on the indicator cells, certain components are more easily detected with this method. Direct comparison of rates of synthesis of the individual components must take these differences into account. Moreover, these measurements do not detect hemolytically inactive complement proteins. Nor do they permit a distinction between preformed and newly synthesized protein. Other biological activities of complement components, such as immune adherence and chemotactic activity, may be used for quantitation of biosynthesis, but these have had only limited use in synthesis studies because of a lack of specificity and/or sensitivity. Biosynthesis of complement components in uitro was first detected by means of imniunoprecipitation of specific proteins by Hochwald et al. ( 1961) . With this method, the incorporation of radiolabeled amino


79

acids into specific proteins may be cstiniated with a variety of techniques including scintillation counting and radioautography of immunoprecipitin bands in gels. Although considerably less sensitive than the functional assay described above, if proper controls are included, the immunochemical method is capable of differentiating newly synthesized from preformed protein. Briefly, tissues or cells are incubated in media containing either "C- or 'H-labeled amino acids. After an appropriate incubation period the media are harvested, dialyzed, and concentrated. Generally because of the extremely small amount of labeled protein produced in short-term experiments, carrier unlabeled protein must be added to detect immunoprecipitation of radiolabeled protein. The effect of inhibitors of protein synthesis on incorporation and suitable controls for nonspecific binding must bc included. In spite of these controls, the limitations of this method are considerable, particularly if it represents the only evidence of complement synthesis. In an attempt to control for the most significant of theqe limitations, binding of an unlabeled component to a radiolabeled antigen-antibody complex, several methods have been employed. The following prelimiilaiy experiment with human liver indicates one approach to resolving this dilemma. The experiment depends on the well known polymorphic forms of C 3 ( Alper et at., 1968), detectable on prolonged agarose electrophoresis. Liver tissue from an individual with known C3 type was cultured in the presence of "Clabeled amino acids; the medium was harvested, dialyzed, and concentrated. The medium was then mixed with serum from an individual with a rate variant C3 type, and the mixture was subjected to electrophoresis and immunofixation with anti-C3. Radioautography was then performed in the usual manner. In liver, this method demonstrated net synthesis of C3 since radiolabeled bands correspond only to the C3 type of the cell donor, not to that of the carrier rare C3 type; i.e., labeled protein was not incorporated nonspecifically into each of the precipitin bands. We chose to study C3 production by liver, since abundant independent evidence (cited below) indicates that this is a site of C3 synthesis, but this approach is applicablc to any of the other lcss well studied components for which genetic polymorphic forms have been described, e.g., C4 (Bach et al., 1971), C6 (Hobart et al., 1975), and properdin factor B ( Alper et al., 1972). Additional controls to test for incorporation or radiolabel into complement protein can be performed by taking advantage of the specificity of interaction of a component with a suitable erythrocyte intermediate. For example, in studies of the biosynthesis of C2 by guinea pig tissues in vitro (Rubin et al., 1971), it was shown that immunoprecipitable radiolabeled C2 was removed by adsorption with EAC4 but not by EA and that adsorption of the labeled protein

80

HARVEY R. COLTEN

was associated with the uptake of hemolytically active C2 since the EAC4 could then be lysed by addition of C1, C3, and C5-9. In another experiment (Strunk et al., 1975), decay of radiolabel from EAC42* was accompanied by a concomitant loss of lytic C2 activity. In several studies (Rommel et al., 1970; Strunk et al., 1975; Bing et al., 1975) it has also been possible to show, using a variety of physicochemical and immunoabsorbent column methods, that radiolabel was incorporated into authentic complement proteins. That is, the physicochemical characteristics of the newly formed proteins corresponded to those of the native proteins in whole serum. It is obvious from the foregoing that several methods must be employed in a given set of studies of synthesis in vitro in order to offset the limitations of each. Thus, demonstration of net increase in biologically active protein, reversible inhibition of that increase in the presence of well known inhibitors of protein synthesis, and incorporation of radiolabeled amino acids into protein immunochemically, functionally, and physicochemically identical with the component are all required to establish synthesis in vitro.

C. BIOSYNTHESIS in V i m While many questions regarding biosynthesis can be tested with in vitro methods, it is apparent that control of plasma protein concentration in vivo is dependent not only on synthesis rates but also on immune and nonimmune catabolic rates. Moreover, the possibility that synthesis rates are affected by circdating factors, produced at sites remote from the primary site of synthesis, cannot be studied in vitro without prior evidence obtained from metabolic studies in the intact organism. As was indicated above, some of the first investigations of sites of complement biosynthesis were performed in experimental animals, but these required destruction or ablation of organ systems. More sophisticated techniques permit a study of metabolism of several of the complement components using trace-labeled purified components in viuo. From the rate of disappearance of radiolabeled protein in plasma, one can calculate the catabolic rate as a percentage of the plasma pool, the synthesis rate, and the extravascular to plasma pool ratio, using the analytic method of Matthews (1957) (Fig. 2 ) . The validity of this analysis depends on two critical assumptions. First, that catabolism occurs in a pool in equilibrium with the plasma pool; i.e., sequestration of labeled protein in an inaccessible pool does not occur. Second, that a steady-state condition for the protein under study is maintained throughout the period of observation. These conditions have been satisfied in most of the published accounts of complement metabolism in man. Two additional approaches to investigations of complement synthesis in vivo have

81


'EV

HOURS

FIG.2. Observed and calculated metabolism of intravenously administered tracelabeled plasma protein. TB, total body; P, plasma; EV, extravascular radioactivity. Resolution into exponential curves with intercept A and slope h from plasma disappearance curves.

been applied to both clinical conditions and studies in experimental animals. First, advantage has been taken of organ transplantation, performed in a clinical setting, to determine sites of synthesis of plasma proteins (Merrill et al., 1964; Kashiwagi et al., 1968) including specific complement components ( Alper et al., 1969). Similar studies have been performed in experimental animals ( Phillips and Thorbecke, 1965; Phillips et al., 1969). Second, the availability of genetically determined deficiencies of complement proteins (reviewed by Alper and Rosen, 1971; Stroud and Donaldson, 1974) has made it possible to study fetal synthesis of complement and to test for transplacental transfer of complement in humans and experimental animals. The results of these studies will be described at length in subsequent sections.

D. CRITERIA As has been discussed elsewhere, many studies of complement biosynthesis are difficult to interpret because they failed to control one or more of several problems in experimental design. That is, there was a failure to distinguish between release of preformed protein and synthesis, to demonstrate incorporation of radiolabeled amino acids into biologically active protein, and frequently use was made of assay systems with significant qualitative and quantitative limitations. The recognition of these problems has led to a definition of minimum criteria for estab-

82

HARVEY R. COLTEN

lishing that a given cell or tissue is a site of synthesis of a complement protein: 1. The tissues and cells must be well characterized and wherever possible available in pure culture or highly enriched in cell type. 2. Tissue culture media and culture conditions must be chosen for optimal maintenance of the cells and for stability of the complement components. 3. Reduction of the baseline levels of preformed complement without damage to the cell must be accomplished. 4. Production of biologically active protein should be detected with the use of one or more quantitative measures of complement activity. Optimally the kinetics of synthesis and secretion should be monitored. 5. Incubation of the cells in medium containing radiolabeled amino acids should result in incorporation of label into immunochemically identifiable complement protein. It must be established that the radiolabeled protein has the functional and physicochemical characteristics of the complement component. Demonstration of labeled peptides on mapping of proteolytic digests of the newly formed protein will assure uniform incorporation of radiolabel. 6. The appearance of functional protein and incorporation of radiolabel must be reversibly inhibited by well known inhibitors of protein synthesis or by incubating the cells at low temperature. The effects of inhibitors on total protein synthesis should be assessed as well. 7. Results of studies with e.g., fluorescent antibody, organ transplant, or extirpation, should be consistent with results of in vitro studies. IV. Sites of Synthesis of Complement Components

A. BIOSYNTHESIS OF THE FIRST COMPONENT The first component of complement ( C l ) is a macromolecule of molecular weight approximately 1 x loo which can be dissociated in the presence of sodium EDTA into three distinct subcomponents designated Clq, Clr, and C l s (Lepow et d.,1963). Clq, the largest of the three, with a molecular weight of about 400,000, an electrophoretic mobility in the y-globulin region, and a sedimentation constant of 11S, consists of 6 globular subunits joined to a central fibrillar core (reviewed by Cooper and Reid, 1974). C l q is the portion of the C1 macromolecule that contains sites capable of binding with a region in the Fc portion of most IgG and IgM immunoglobulins; but not with IgA, IgD, or IgE. C l q is unique among the complement proteins inasmuch as its amino acid composition and portions of its sequence are similar to that


83

of collagen. Clr, a 7s protein with p mobility, normally exists in serum noncovalently bound to C l q which activates it by a mechanism as yet poorly understood; and the Clr, in turn, activates Cls, altering its electrophoretic niobility and converting the proesterase to an active esterolytic enzyme. Cls, an (Y? protein,niolecular weight 8O,OOO-1OO,OOO, possess the active enzymatic site( s ) of the C I molecule which facilitate the cleavage of the natural substrates of this enzyme, C4 and C2. In 1966 a series of experiments designed to identify the site of synthesis of biologically active C1 indicated that, in guinea pig (Colten et al., 1966) and man (Colten et al., 1968a), macromolecular C l was synthesized primarily, if not exclusively, in the small and large intestines. This conclusion was based on observations that isolated segments of intestine in short-term organ culture were the only tissues capable of producing hemolytically active C1. Production of C1 in uitro was temperature dependent, was reversibly inhibited by actinomycin D and pyromycin, and was accompanied by incorporation of "C-labeled amino acids into a molecule functionally similar to C1. In subsequent studies ( Colten et al., 1968b), using a modification of the Jerne plaque technique capable of detecting production of hemolytically active C1 by individual cells, evidence was presented that in guinea pig the coluninar epithelial cell was the site of C1 synthesis. At about the same time, Thorbecke and her co-workers (Stecher and Thorbecke, 1967a; Stecher et al., 1967) reported production of C l q in uitro by human and monkey liver, spleen, bone marrow, and lung as well as in macrophages isolated from peritoneal exudates and lung washings. Their conclusions were based solely on detection of radiolabeled C l q precipitin lines developed with an anti-Clq antiserum in the presence of carrier serum or euglobulin. Active C l q protein was not monitored. In a study of the ontogenetic development of C l q in the piglet, Day and her colleagues (1970) failed to find hemolytic C1 in any of the embryonic tissues examined, Therefore, they used a modification of Thorbecke's method for the detection of incorporation of labeled amino acids into C l q protein. Whole pig serum served as carrier and the amount of label incorporated into C l q was estimated by the intensity of the line on radioautography. Portions of the fore, mid and hind gut were the first tissues to synthesize C l q protein as estimated by this technique. Radiolabeled C l q was detectable in the 48-day embryo, the earliest tissue examined, at which time no incorporation of labeled amino acid into C l q was detected in cultures of other tissues, including thymus, spleen, liver, kidney, heart, or lymph node. Lymph node, spleen, liver, and lung cultures obtained from fetuses later in gestation also labeled Cl q. Labeling was significantly reduced in cultures incubated in medium

84

HARVEY R. COLTEN

containing chloramphcnicol. Although the cellular site of Clq synthesis was not identified, these workers suggested that since nonintestinal tissues, particularly those rich in lymphoid cells, also produced Clq, a mesenchymal cell, not an epithelial cell, was the site of synthesis. Others (Lai A Fat and van Furth, 1975) have also suggested that lymphoid cells or macrophages are sites of C l q synthesis. Many studies since have effectively ruled out the lymphocyte as a site of C l q synthesis (Stecher and Thorbecke, 1967a; Stecher et al., 1967). In none of the studies suggesting synthesis of C l q by macrophages was biologically active C1 or C l q detected, nor were attempts made to characterize the apparently newly synthesized protein. However, support for the theory that C1 was synthesized by mononuclear cells came from two other sources. On the one hand, Moore and Vas (1968) found C1 release from peritoneal cells using a modification of the Jerne plaque technique, but two limitations of this study diminished the validity of their conclusions, First, R reagents were used to develop the hemolytic plaques, and, second, de novo synthesis was not established; i.e., there were no controls to rule out release of preformed C1. On the other hand, several clinical studies, indicating independent control of Clq, Clr, and Cls concentrations in sera of some patients with severe combined immunodeficiency ( OConnell et al., 1967; Gewurz et al., 1968; Stroud et al., 1970) and in a patient with selective deficiency of C l r (Pickering et al., 1970) suggested the possibility that the known subcomponents of C1 might be synthesized in different cell lines with assembly of the macromolecule in one of the cell types. Moreover, bone marrow transplantation for severe combined immunodeficiency resulted in an apparent reconstitution of C l q (Ballow et al., 1973; Yount et d., 1974). From these and other studies (Kohler and Miiller-Eberhard, 1969), it was postulated that the synthesis of C l q and IgC were linked. Two observations make this an untenable hypothesis at the present time. Metabolic studies of trace-labeled C l q in patients with agammaglobulinemia revealed accelerated catabolism, not depressed synthesis ( Kohler and MiillerEberhard, 1972), and serum levels of C l q and IgG vary coincidentally with exogenous ?-globulin administration in patients with agammaglobulinemia (5. R. Pickering, unpublished, cited in Gabrielson et al., 1974). Renewed interest in epithelial cells as a site of C1 synthesis has been generated by a series of elegant experiments by Bing and his colleagues (1975). They demonstrated synthesis of macromolecular C1 and C l s in long-term primary suspension cultures of normal human colon, adenocarcinoma of the colon, and transitional epithelial cells of bladder, urethra, and renal pelvis. It is believed that the epithelial cells of the


85

urogenital tract and colon may be of common (endodermal) origin since early in embryonic life the hind gut and mesonephric duct are continuous with the cloaca. Synthesis of biologically active C1 was inhibited by cycloheximide and was accompanied by incorporation of radiolabeled amino acids into macromolecular C1 and its subcomponent Cls. Moreover the newly formed protein was isolated from the tissue culture medium by means of specific affinity column chromatography. Cultures of cells from prostatic adenoma, renal cell carcinoma, and fibroblasts did not synthesize C1. Subcloning experiments, now in progress, should establish finally whether, as it appears, all the subcomponents of C1 are synthesized in a single cell type. Wyatt (1974) has also provided preliminary evidence that the HEK (human embryonic kidney) and Ma177 (human embryo intestine) cell lines produce C1 in uitro, but the precise cells of origin and homogeneity of these lines are not known. A single report (van Zeipel, 1970) has appeared of synthesis of a C1-like molecule by HeLa cells (derived from a cervical carcinoma), but examination of at least two other HeLa lines has failed to reveal synthesis of either C l q (Stecher and Thorbecke, 1967c; Colten and Wyatt, 1972), Cls, or intact C1 (Colten, 1972b). These contradictory findings can probably be accounted for by extensive genetic drift during many years in culture and the resulting heterogeneity of “HeLa” cell lines.

B. BIOSYNTHESIS OF THE SECOND COMPONENT C2 is a p-globulin with a sedimentation constant of approximately 6 S and molecular weights of 117,000 and 130,000 for human and guinea pig C2, respectively. The subunit structure of this trace plasma protein is not known. When C2, one of the natural substrates of C1, is cleaved, one of the two fragments is bound to a fragment of C4 generating a C3 cleaving enzyme. The C a complex is unstable and liberates an inactive C2 fragment and a C4 site that is capable of accepting fresh C2 to regenerate the C3 cleaving enzyme. Several authors, using a variety of methods, have concluded that the macrophage is the site of C2 biosynthesis. Early studies of C2 biosynthesis (Siboo and Vas, 1965; Rubin et al., 1971; Colten, 1972a) indicated that production of hemolytically active C2 could be detected in many organs, including lung, bone marrow, spleen, and in some species, liver. C2 synthesis by guinea pig liver was not established although production of C2 was observed in primary cultures of rat (J. Breslow, P. Rothman, and H. R. Colten, unpublished, 1975) and human liver. In several of these studies it was shown that C2 production was in fact due to de novo synthesis, inasmuch as the appearance of C2 in culture was reversibly inhibited by cycloheximide and was accompanied by incorporation

86

HARVEY R. COLTEN

of radiolabeled amino acids into a protein inimunochemically and functionally identical with C2. Subsequently, using a modification of the Jerne plaque technique to detect guinea pig C2 synthesis (Wyatt et al., 1972) and isopycnic density-gradient centrifugation of fetal liver cells to partially purify cells that synthesize human C2 (Colten, 1972a), it was shown in each case that C2 was produced by large mononuclear cells, probably macrophages. For example, it was found that approximately 10-15% of the mononuclear cells in a guinea pig peritoneal exudate produced C2-specific hemolytic plaques in gel and C2 production was restricted to the glass-adherent mononuclear phagocytic cell population. Human C2 production was observed only in fractions rich in mononuclear cells high in the albumin gradient. Recently, a significant advance in methodology, permitting maintenance of human peripheral blood monocytes in primary culture for prolonged periods (10-12 weeks), has made it possible to demonstrate C2 synthesis by circulating cells of the monocyte/macrophage series (L. P. Einstein and H. R. Colten, unpublished, 1975). These cells maintain the capacity to synthesize C2 and lysozyme, to phagocytize large latex particles, to kill Listeria monocytogenes, and to rosette with IgG-sensitized erythrocytes. Within the limits of the methods empIoyed, these appear to be pure monocyte cultures (Table I). A cell line developed by Wyatt (1974) should also be useful for long-term studies of C2 synthesis in uitro. This cell obtained from a peritoneal exudate induced in strain 2 guinea pigs has a mean generation time of 40 hours and an epithelioid appearance; it adheres firmly to surfaces, resisting attempts to dislodge it from plastic tissue culture dishes with EDTA-trypsin. C2 production by this cell line has continued through more than 22 passages over a 15-month period. Wyatt (1974) has also described an interesting change in two other cell lines (1 and 10) derived from a chemically induced ( diethylnitrosamine ) guinea pig hepatoma. When first tested, neither line was synthesizing C2. However, after passage in ascites form, back into strain 2 guinea pigs, and retested more than one year later, both synthesized substantial quantities of C2. These findings have been confirmed in our laboratory, and experiments underway are designed to explore this intriguing observation. Several reports of C2 synthesis by other hepatonia cell lines have appeared (Levisohn and Thompson, 1973; Strunk et al., 1975). Some of these cells are highly differentiated and have retained many functions characteristic of hepatic parenchymal cells. In spite of these findings, since in primary culture only monocytes or macrophages synthesize C2, it is likely that in the liver, the Kupffer cell, not a parenchymal cell, is the site of C2 production,

87


TABLE I CHARACTERIZATION O F MONONUCLEAR CELLSFROM H U M A N BLOOD" -~

Lymphocyte Property

Monocyte

Topology E Rosette E-IgG Rosette E-IgM-C3 Rosette Surface IgG Protein synthesis Peroxidase Lysoz yme CZ IgG Other Glass adherence (serum) Phagocytosis (5.7pm diameter beads) Kill Lzsteria inonocylogenes Response to mitogens

++ + + + ++ +

T

+

B

+ +

+ +

+

Properties of peripheral blood monocyte cultures are indicated by bold plus signs; known characteristics of T and B cells by light plus signs (not detected in monocyte cultures). Peroxidasc-positive granules were present only in early cultures; other rhararteristics pcrslst at least 8 weeks in vitro.

C. BIOSYNTHESEOF THE FOURTH COMPONENT The fourth component of complement has a molecular weight of 209,000 and consists of two (J. 0. Minta, unpublished, 1975) or three ( Schreiber and Muller-Eberhard, 1974) subunits linked by disulfide bridges. Its electrophoretic mobility ( plE globulin) is altered after cleavage by Cls. One of the fragments links target cell surfaces and C2 to yield the C3 cleaving enzyme C a . Several independent lines of evidencc indicate that the macrophage is a site of C4 biosynthesis. As indicated in an earlier section, interest in the liver as a primary site of C4 synthesis depended on experiments showing marked depression of serum C4 levels in animals exposed to hepatotoxins or after hepatectomy. Jensen et al. (1971) in the first study of synthesis of functional C4 in vitro demonstrated synthesis by liver and concluded that the parenchymal cell was the likely site of C4 production. In the modern era of complement biosynthesis research, Thorbecke and her colleagues (1965; Stecher and Thorbecke, 1967c) studied incorporation of radiolabeled amino acids into C4 protein by a variety of murine

88

HARVEY R. COLTEN

and primate cell types. They concluded that the macrophage was the principal site of C4 synthesis; but these studies were incomplete, as no studies of biologically active C4 were performed. Indirect evidence that macrophages produce biologically active C4 was also obtained in experiments that showed lysis of guinea pig macrophages in the presence of a rabbit anti-guinea pig macrophage antiserum and R4 (Chan, 1970; Fust and Surjin, 1971). Neither reagent alone was capable of lysing the cells. In addition C4-containing cells have been identified in guinea pig marrow and spleen with fluorscent-labeled anti-C4 antibodies ( Chan and Cebra, 1966). Work by at least two groups (Littleton et al., 1970; Wyatt et al., 1972) has confirmed these observations and extended the initial experiments to include direct measurements of production of biologically active C4. The adherent cells in guinea pig peritoneal exudates, spleen cells (Ilgen et al., 1974), and cells from upper fractions of discontinuous density gradient ultracentrifugation of human fetal ( Colten, 1972a) or guinea pig liver (Ilgen and Burkholder, 1974) are capable of synthesizing biologically active C4 in short-term tissue culture. Moreover, it has been demonstrated that at least some of the cells capable of synthesizing C4 also synthesize C2 (Wyatt et al., 1972). It was estimated, using a hemolytic plaque method, that, of a starch-induced peritoneal exudate or peritoneal wash, approximately 10-15% of the mononuclear cells synthesized C4. As noted above, a comparable percentage of the cells yield hemolytic plaques when assayed for C2 production. With a considerably less sensitive plaque assay, designed to detect only those peritoneal cells capable of producing both C2 and C4, approximately one plaque per 1000 cells was detectable. Thus, it appears that at least some mononuclear phagocytes synthesize both C2 and C4. At least two sources of mononuclear phagocytes, human peripheral blood and lung macrophages, yield cells that synthesize C2 but produce no hemolytically active C4 (Fig. 3 ) . l'C-Labeled amino acids are apparently incorporated into C4 protein, but attempts to detect the appearance of active C4 in these cultures have thus far been unsuccessful (L. P. Einstein and H. R. Colten, unpublished, 1975; R. Laguarda, G . L. Huber, and H. R. Colten, unpublished, 1975) . Studies underway should indicate whether this finding represents heterogeneity of macrophages, production of a fragment of C4, or, least likely, a technical problem in the detection system. Biosynthesis of C4 by established cell lines has also been demonstrated (Stecher and Thorbecke, 1967c; Wyatt, 1974; Strunk et al., 1975) using both functional and immunochemical methods. Several lines derived from normal rat liver or hepatomas produced significant quantities of C4 in culture. One of these lines has been subcloned and has many

89


45r

/

FIG.3. Production of hemolytically active C2( 0 ) and C4( m ) by human lung In medium 199 macrophages. Complement production at 37°C in medium 199; 0: with cycloheximide ( 2 pg/nil); at arrow, cells in cycloheximide were washed, and one set was incubated in medium alone, another in rnedium with cycloheximide.

characteristics of liver parenchymal cells. Three human fibroblast cell lines seemed to incorporate lC-labeled amino acids into C4 protein whereas a variety of other cell lines including HeLa cells did not (Stecher and Thorbecke, 1 9 6 7 ~ )De . nouo snythesis of C4 by the fibroblasts was not established unequivocally in these experiments.

D. BIOSYNTHESIS OF

THE

THIRDCOMPONENT

The third component of complement is also a p globulin; of all the complement proteins, it is present in highest concentration in normal serum. Thus, it was the first of the complement components isolated in chemically pure form. It has a molecular weight of approximately 190,000 and consists of two polypeptide chains with molecular weights of 120,000 and 75,000 linked by disulfide bridges (Nilsson et al., 1975). C3 is a highly polymorphic protein (Alper et al., 1968) as revealed by electrophoresis and inimunofixation in agarose. There are two common genetically determined forms, designated S for the slower or cathodal type and F for the faster. A number of rare variants of the S and F alleles have been described. It appears that all variants are functionally identical ( Colten and Alper, 1972). Perhaps because of the central role of C3 in the expression of the biologic effects of the complement system and the early development

90

HARVEY R. COLTEN

of methods for immunochemical identification of C3, many investigators have studied synthesis of this component using a variety of in viuo and in Vitro techniques. The first studies of C3 biosynthesis in uitro suggested that this component was synthesized by many tissues and cell types. Claims were made by Thorbecke and her colleagues (1965) that lymph nodes, spleen, liver, lung, bone marrow, adult thymus, and the fibria of the fallopian tube were sites of synthesis of C3. This conclusion was based on detection of radiolabeled C3 precipitin arcs on immunoelectrophoresis of culture media. The authors recognized the limitations of these studies, namely, that the net synthesis of complement activity or of individual proteins was not demonstrated and that data based solely on incorporation of 'T-labeled amino acids must be interpreted with care. They cited their previous studies (Hochwald et at., 1964) which indicated that other proteins may apparently be labeled by virtue of interaction of the protein with another labeled protein, and not internally labeled. Nonetheless, this series of investigations (Hochwald et al., 1961; Asofsky and Thorbecke, 1961), the first studies of complement synthesis in vitro, provided considerable stimulii for further work. One group (Glade and Chessin, 1968) has suggested that the lymphocyte was the cellular site of C3 synthesis, but several authors (Stecher et al., 1967; Lepow, 1965) have provided both direct and indirect evidence that this is not the case. Others (Stecher and Thorbecke, 1967a; Lai A Fat and van Furth, 1975) have indicated that macrophages synthesize C3 but again this is based only on immunochemical detection of I4C-labeledamino acid incorporation into PIC protein. In this regard an interesting set of experiments should be noted (Phillips and Thorbecke, 1965, 1966). Rat bone marrow was injected into irradiated mice and the recipient mouse serum was examined for the presence of donor rat protein. I n addition, recipient mice were killed at intervals of 2-14 weeks, and liver, spleen, mesenteric lymph nodes, peritoneal macrophages, and thymus were placed in culture in the presence of 'C-labeled lysine and isoleucine for estimates of production of both rat and mouse protein. Examination of the sera of recipient animals indicated that rat transferrin was detectable within 1-2 weeks whereas significant amounts of IgG and IgM of donor origin were not detectable until about 2-3 weeks after bone marrow transplantation. C3 was not detected in the sera of recipient animals. Incubation of sp1een.h uitru, however, revealed labeling of rat IgG, IgM, C3, and transferrin. The latter two proteins were detected in tissues obtained from animals transplanted 1-3 weeks earlier whereas immunoglobulin production was not detected until 4 weeks after transfer of the marrow cells. Label was present in both rat and mouse a2-macroglobulin even


91

when tissues were taken from control or nonchimeric animals, demonstrating once again the limitations of radioimmunocheniical methods as sole evidence for synthesis. Cultures of peritoneal macrophages showed significant labeling of transferrin and C3 in cultures obtained from animals 4 weeks after bone marrow transplantation, but liver cultures produced no rat serum proteins. A significant difference was noted between postirradiation chimeras induced with bone marrow and neonatal chimeras undergoing graft-vs-host reaction induced with rat spleen. Namely, production of rat transferrin and C3 was clearly present in the postirradiation chimeras and, in fact, preceded the production of immunoglobulin whereas neonatal chimeras showed rat immunoglobulin production but no evidence of other rat protein synthesis (Phillips and Thorbecke, 1966). These provocative findings should clearly be pursued by means of modern methods for detection of complement biosynthesis. Alper et al. (1969) performed an ingenious experiment making use of the technique of prolonged agarose electrophoresis to detect polymorphic forms of C3 and showed that the liver was the principal, if not the only, site of C3 synthesis. In this study, a patient with a hepatoma received a liver transplant following total hepatectomy. After the transplant the patient’s serum C3 concentration declined from over 2 mglml to 0.4 mg/ml and then increased to 1.35 mg/ml by the fourteenth postoperative day. The recipient had a rare C3 type (FS,,,) which, on the basis of family studies, was almost certainly inherited and not the result of production of an abnormal C3 due to malignant transformation of liver cells. The donor was of C3 type SS. Within 20 hours after transplantation, the C 3 type changed almost completely from that of the recipient to donor type, and by 6 weeks postoperatively no C3F nor So.6type could be detected. A concomitant change in haptoglobin type (another protein synthesized by liver ) was also noted, confirming the observations of Merrill et al. (1964) and Kashiwagi et al. (1968). The possibility that the observed change in C3 type was due to the immunosuppressive regimen itself was ruled out by observations that no change in C3 type accompanied the immunosuppression required for kidney transplantation. Additional evidence that the liver is a major site of C3 synthesis was obtained in experiments (Johnson et al., 1971b) showing “staining of hepatic parenchymal cells with fluorescent-labeled anti-C3 antiserum. In addition, production of hemolytically active C3 has been detected in vitro in cultures of human fetal and postnatal liver (Colten, 1972a). Extrahepatic production of hemolytically active C3 has been detected in synovial tissues from patients with rheumatoid arthritis, but not from patients with traumatic or degenerative joint disease (Ruddy and Colten, 1974). This and other studies would indicate the necessity to study

92

HARVEY R. COLTEN

C3 synthesis under a variety of conditions before excluding the possibiIity that a given cell type is a site of C3 synthesis. As with many of the other complement proteins, production of C3 by long-term, established cell lines has been reported (Stecher and Thorbecke, 1967c; Strunk et al., 1975). These cell lines have been particularly useful in studies of the microenvironmental control of complement synthesis. Several investigators have studied the metabolism of C3 in v i v ~in normal subjects and in patients with genetic or acquired abnormalities of the complement system ( Alper and Rosen, 1967; Sliwinski and Zvaifler, 1972; Peters et al., 1972; Carpenter et al., 1969). Estimates of normal C3 synthesis rates were between 0.45 and 2.7 mg/kg per hour. This corresponds to a rate approximately ten times the rate of C3 synthesis by isolated liver fragments in vitro (Colten, 1972a). These studies will be discussed in detail below in the context of control of complement synthesis.

E. BIOSYNTHESIS OF C5 AND C6 The fifth component of complement, also a globulin with p mobility, has a molecular weight close to that of C3. Its similarity to C 3 in biological activity is striking in that cleavage of C5 also yields products with chemotactic, anaphyIotoxic, and perhaps opsonic properties. Preliminary data on the subunit structure of C5 are now available (Nilsson et al., 1975). In early studies of C5 biosynthesis, Phillips et al. (1969) made use of the observation that several mouse strains were known to be deficient in C5 or MuBl (the immunochemically defined C5 protein) (Rosenberg and Tachibana, 1962; Terry et al., 1964; Cinader and Dubiski, 1963).Bone marrow cells from allogeneic or apparent congenic strains were injected into X-irradiated recipients. Two strains thought to be congenic-B10D2 old line, a C5-deficient strain, and BlOD2 new line, a strain with an intact complement system-were used in some of these experiments. In others, irradiated A/ Jax animals served as recipients of marrow cells from C57B16xA F,. Total complement activity was detected in sera of recipients of the marrow within 2-3 days and persisted for almost 3 weeks after the transplant, indicating production of biologically active C5 by bone marrow cells. Controls consisting of irradiated A/ Jax mice injected with A/ Jax marrow remained C5 deficient. Examination of sera of each of the recipient mice immunoelectrophoretically did not reveal MuBl protein after transplantation, This discrepancy between the immunochemical and hemolytic data was no doubt due to differences in sensitivity of the assay systems. In vitro studies indicated that the C5-producing marrow cells were present in the recipient spleen since cultures of this organ prepared 3-4 weeks after establishment of

93


TABLE I1 SITESO F SYNTHESIS O F COMPLI~MJ:.NT PI~OTEINS~ Component

c1 C h Clr Cls c2 C3 c4 C5 C6

c7 C8 c9 C1 inhibitor

Species

Organ*

Cell

Human, guinea pig

Intestine/GU Epithelial (columnar and (excluding kidney) transitional) Human Intestine/GU Epithelial (columnar and transitional) (excluding kidney) ? ? Human In testine/GU Epithelial (columnar and transitional) (excluding kidney) Macrophage/monocyte Human, guinea pig Wide distribution Parenchymal Human Liver Primate, murine Extrahepatic ? Macrophage Guinea pig, human Wide distribution Macrophage Human Wide distribution ? Macrophage Murine Wide distribution ? liabbit Liver ? ? 9 Pig Wide distribution Itat Liver Probably parenchymal Human Liver Parenchymal

a Wherever possible, observations were based on st,udies that satisfy minimum criteria for establishing net synthesis. * GU, genitourinary tract.

chimerism showed labeling of MuBl when incubated in the presence of 'fXabeled amino acids. Restoration of hemolytic C5 was quantitatively greater in the allogeneic transfer than in the transfer of marrow between more closely related strains, raising interesting questions as to the effect of a graft-vs-host reaction on complement biosynthesis. Much later, Levy et al. (1973) demonstrated that the adherent cells from the spleen of B10D2 new line mice were the only cells of those examined that were capable of synthesizing C5, as judged by the appearance of hemolytic C5 activity in media harvested from spleen cell cultures. Kideny, brain, fibroblast, liver and nonadherent spleen cells failed to produce C5 under these conditions even when twenty times as many cells of these types were compared to the adherent spleen cell population. Puromycin, albeit, at relatively high concentrations ( 50 g / m l ) , inhibited C5 production, indicating active synthesis. These experiments suggested that the macrophage was a site of C5 synthesis. This suggestion was not confirmed in studies of the adherent cell population isolated from guinea pig peritoneal exudates ( Colten, 1974b ) . Although hemolytically active C5 was present in medium harvested from peritoneal cell cultures,

94

HARVEY R. COLTEN

there was no net increase in CS activity, nor effect of temperature inhibitors of protein synthesis on the appearance of C5. This indicated that substantial quantities of preformed C5, either bound to or within the peritoneal cells, were released under the culture conditions. The reasons for this finding are not clear at the present time. In any case, it is likely that a widely distributed cell type is the site of C5 synthesis as evidenced by studies of biosynthesis of this component in man. TWO studies of human C5 biosynthesis have been reported. In one of these (Colten, 1973), production of biologically active C5 was detected in lung, liver, spleen, and fetal intestine. In the other study (Kohler, 1973), C5 production as judged by incorporation of radiolabeled amino acids, was detected in thymus, placenta, peritoneal cells, and bone marrow as well. Thus far the specific cell type that synthesizes human C5 has not been identified. Several clonal strains of rat hepatoma including a strain that also synthesized C3, C9, C1 inhibitor, albumin, and tyrosine aminotransferase, synthesize biologically active C5 (Strunk et al., 1975). As with other studies of protein synthesis by long-term cell lines, it would be imprudent to conclude on the basis of this study that hepatic parenchymal cells are a site of C5 synthesis. However, primary rat liver cell cultures, rich in parenchymal cells and relatively depleted in Kupffer cells, produce C5 in culture ( J . Breslow, P. Rothman, and H. R. Colten, unpublished, 1975). A difference in species is an unlikely explanation to account for the different observations with respect to C5 synthesis by mouse and rat or guinea pig cell types. Final conclusions must therefore await additional studies. The availability of rabbits deficient in C6 (Rother et al., 1966) and a recent report of C6 deficiency in humans (Leddy et al., 1974) suggest the need for more extensive studies of C6 synthesis. Thus far there is only one reported study of C6 biosynthesis (Rother et al., 1968). In that study it appeared that the liver was the major site of C6 production, but synthesis of C6 by other tissues could not be excluded. The precise cellular site of C6 synthesis was not defined. In fact, many technical problems encountered in that work have subsequently been resolved so that a reexamination of C6 synthesis in vitro might now be fruitful.

F. BIOSYNTHESIS OF C7 AND C8 The biosynthesis of C7 has not been studied, and only isolated reports of C8 synthesis are available. For example, Geiger et al. (1972) have investigated the ontogeny and synthsis of C8 by tissue obtained from fetal pigs. Tissues from fetal pigs were examined between gestational


95

ages 47 and 112 days of intrauterine life. The hemolytic assay was used to detect C8 in dialyzed and concentrated tissue culture medium harvested from the culture. Significant C8 production was noted in spleen, liver, lung, intestine, and kidney. No production was observed in lymph nodes, thymus, or bone marrow. Estimates of C8 production in gut were performed in short-term experiments because of the problems of bacterial contamination, particularly with adult tissues. In these specimens from adult pigs, definite evidence of C8 synthesis by intestine could not be obtained. Ontogenetically the liver, lung, and intestine were the first organs capable of synthesizing C8. Production was detectable in the earliest samples obtained (gestational age 47 days). By 100 days gestation, the colon, kidney, and spleen were also producing significant amounts of C8. In the lung, reversible inhibition of C8 production was noted in the presence of cycloheximide and actinomycin D. Production of C8 was temperature dependent in cultures of fetal kidney, spleen, liver, and lung. Thus the distribution of organs synthesizing C8 is similar to that of several other complement proteins but the cell of origin has not been established. G. BIOSYNTHESIS OF C9 Synthesis of C9 by a long-term rat hepatoma cell strain was demonstrated several years ago (Rommel et al., 1970). This line is highly differentiated, having many features characteristic of normal hepatic parenchymal cells. For example, it is capable of synthesizing albumin C3, C1 inhibitor, and C5 but not C l , C4, or significant amounts of C2 (Strunk et al., 1975). Moreover, in this cell, the glucuronide conjugating system is intact. C9 production was temperature dependent and reversibly inhibited by cycloheximide and actinomycin D. More recently (J. Breslow, P. Rothman, and H. R. Colten, unpublished, 1975) primary rat liver cell cultures have been prepared by dissociating rat liver cells with collagenase several days after partial hepatectomy. Low speed centrifugation provided cell suspensions enriched in parenchymal cells. These cells produced substantial amounts of hemolytically active C9 in culture; production was reversibly inhibited by cycloheximide. On the basis of these experiments, it is likely, but not yet firmly established, that C9 is synthesized by liver parenchymal cells.

H. COMPLEMENT-ASSOCIATED PROTEINS The synthesis of C1 inhibitor by liver has been demonstrated in shortterm cultures by both functional (Colten, 1972a) and immunochemical ( Gitlin and Biassuci, 1969; Colten, 1972a) assay systems. Production was inhibited by cycloheximide and was temperature dependent. Ap-

96

HARVEY R: COLTEN

proximately 5-10% of normal human liver parenchymal cells fluoresced with labeled anti-Cl inhibitor (Johnson et al., 1971a). Fluorescence of Kupffer or other liver cells was not noted. Synthesis of the C3b inactivator and the proteins of the properdin system have not yet been studied in any detail. V. Ontogeny of Complement

Three approaches to the study of fetal synthesis of complement have been employed: (1) synthesis of specific complement components in uitro by isolated fetal tissues; ( 2 ) demonstration of a maternal-fetal discordance of genetic type in those complement proteins exhibiting genetic polymorphisms [a method adapted from methods first used to establish fetal synthesis of haptoglobin (Rausen et al., 1961) and GC globulin (Hirschfield and Lunell, 1962)l; and (3) a related method showing the presence of a complement component in sera of fetuses borne by genetically deficient mothers. From these studies it can be concluded that most of the complement proteins are synthesized early in fetal life, and in several species (man, rodents, and unguIates) the placenta is an effective barrier to passage of complement either from or to the fetal circulation. In general, all the complement proteins, with the exceptions of C7 and C9 which have not been studied, are synthesized by the fetus at or shortly before the onset of immunoglobulin synthesis. Synthesis of C1 (Colten et al., 1968a,c) and C l q (Kohler, 1973; Day et al., 1970) has been demonstrated in tissues obtained early in gestation. Neither thymectomy, nor antigenic stimulation leading to antibody formation, nor homograft rejection significantly affected the C1 content of fetal serum (Colten et al., 1968~).This is somewhat surprising in view of the apparent effects of IgG concentration on C1 metabolism ( Kohler and Muller-Eberhard, 1969,1972). Synthesis of biologically active human C2 and C4 has been detected in fetal tissues obtained as early as 8 weeks of gestation (Colten, 1972a), but incorporation of radiolabeled amino acids into C4 protein was only barely detectable in tissues obtained at 9 weeks (Adinolfi and Gardner, 1967; Adinolfi et al., 1968; Gitlin and Biasucci, 1969; Kohler, 1973). The rates of C2 synthesis in fetal liver remain approximately constant throughout gestation, but C4 synthesis rates in the first trimester are only one-fifth to one-fourth that observed in fetal liver at 22 weeks of gestation (Colten, 1972a). Whether, in the more mature fetal liver, this represents an increased rate of production per cell or recruitment of more cells that synthesize C4 is not known. Additional evidence for C4 synthesis in the fetus and for lack of transplacental passage of C4


97

or C2 was obtained in studies of paired maternal cord serum C4 types (Bach et al., 1971) and cord blood of a homozygous C2-deficient child (Ruddy et uZ., 1970), respectively. Gitlin and Biasucci ( 1969) obtained preliminary evidence that human fetal liver at approximately 4 weeks of gestation synthesized C3. This stage in development is approximately coincident with the appearance of hepatic cells in the liver lobes (Bloom, 1926). In another study, production of biologically active C3 by fetal liver was detected at 14 weeks, the earliest specimen available (Colten, 1972a). The protein is detectable in fetal sera between week 5 and week 10 of gestation (Adinolfi and Gardner, 1967; Gitlin and Biasucci, 1969; Adinolfi, 1972) and is undoubtedly of fetal origin, as no evidence for transplacental passage of C3 could be demonstrated by comparing C3 types in paired maternal cord blood samples (Propp and Alper, 1968). Lack of transplacental passage and fetal synthesis of C5 and C6 were established by demonstrating C5 protein in the sera of fetuses bred from C5-deficient female mice by C5-sufficient males (Tachibana and Rosenberg, 1966) and in a similar experiment using C6-deficient rabbits (C. Biro and T. Borsos, unpublished, 1975). Isolated human fetal tissues synthesize hemolytically active ( Colten, 1973) and immunochemically identifiable C5 ( Kohler, 1973). Likewise C8 synthesis has been shown in fetal pig tissues early in gestation (Geiger et d., 1972). Finally, fetal synthesis of C1 inhibitor (Gitlin and Biasucci, 1969; Colten, 1972a; S. Ruddy, A. Sheffer, and K. F. Austen, unpublished, 1975) and properdin factor B (Alper et al., 1972) has also been established. VI.

Genetic Regulation of Complement Synthesis

The plethora of genetically determined complement deficiencies in man and experimental animals (reviewed by Alper and Rosen, 1971; Stroud and Donaldson, 1974) has provided material for several promising lines of investigation using methods developed in modern studies of complement biosynthesis. A. C4 DEFICIENCY After the discovery and development of a strain of guinea pigs genetically deficient in C4 (Ellman et al., 1970), Colten and Frank (1972) studied the biosynthesis of C2 and C4 by tissues and cells isolated from affected animals. The study of C 2 biosynthesis by tissues from C4-deficient animals was of particular interest, since it had been shown that in some, but not all, homozygous C4-deficient animals the C2 serum levels were approximately one-half normal ( Frank et al., 1971). Measurements of complement synthesis in short- and long-term cell cultures

98

HARVEY R. COLTEN

indicated that homozygous deficient animals produced no detectable C4 and that the rate of C2 synthesis was approximately 40% of normal. Tissues and cells from heterozygous deficient animals produced C4 at a rate intermediate between that of the normal and homozygous deficient. In each case, the relative reduction in rates of in uitro C2 and C4 synthesis by heterozygous and homozygous deficient cells correlated well with the reduced serum levels of the corresponding component. In an attempt to define the genetic lesion responsible for C4 deficiency, peritoneal exudate cells from a homozygous C4-deficient guinea pig were fused in uitro with a cell line of human origin (Fig. 4). The resulting hybrid cells, derived from parental cells each incapable of C4 biosynthesis by themselves (Stecher and Thorbecke, 1967c; Colten and Frank, 1972) synthesized functionally active human C4 but no detectable guinea pig C4 (Colten and Parkman, 1972). Several explanations for these results were considered. Among these was the possibility that, in the hybrid cells, a product of the C4-deficient genome was capable of derepressing the human C4 gene but that the defective guinea pig C4 gene could not respond to the signal. In order to test this possibility, attempts were made to identify the putative derepressor. Peritoneal cells isolated from C4-deficient guinea pigs, although incapable of synthesizing C4, produced a factor that specifically induced synthesis and secretion of functionally active C4 by HeLa cells (Colten, 1972b). Operationally this substance has been designated a regulator factor, since the site of action of the factor in the indicator (HeLa) cells is not known. It is possible HUMAN (HELAI CELL

C 4 DEFICIENT PE CELL

c 4 DEFICIENT C2 c4

GP 0

HYBRID

HELA

0

0 0

H

FIG. 4. Somatic cell hybrids obtained by fusion of C4-deficient peritoneal ( P E ) cells and HeLa cells. 0 indicates confluent cultures. Production by parent and hybrid cells of C2 and C4 and species ( H = human, GP = giunea pig) of gene product.


99

that the C4-deficient cells produced and accumulated a true derepressor of C4 gene function. However, the factor that was detected in these studies may be acting at any of several steps in the control of specific protein synthesis, not on the C4 gene itself. The results suggested that active protein synthesis by the C4-deficient peritoneal cells and by HeLa cells was required to produce and detect the factor. Two other studies supported this concept. The first experiment ruled out the possibility that the medium harvested from C4-deficient cells enhanced the stability of C4 or affected the assay for C4. In the second experiment, HeLa cells were incubated in medium 199; the medium was removed, centrifuged, and then mixed with an equal volume of medium collected from a C4-deficient cell culture. This mixture was incubated at 37OC, and, at timed intervals, aliquots were removed to assay for C4. No C4 was detected, ruling out the possibility that the C4-deficient and HeLa cells each produced fragments that could combine in the cell-free fluid to yield a functionally active C4 molecule. The factor from the C4-deficient cells was found to be heat stable (at 56°C for 1 hour), nondialyzable, and partially inactivated by RNAase and trypsin, but not by DNAase. Chromatography on Sephadex G-100 demonstrated that the activity eluted in a position corresponding to a 45,000 MW marker (Colten, 19744. In the initial experiments, a substantial lag was noted between the time of addition of the factor and the onset of detectable C4 production by HeLa cells. This observation suggested the possibility that the response to the factor might be a function of the cell cycle. This possibility was tested with HeLa cells synchronized by shaking them from actively growing cultures. At timed intervals after synchronization, medium containing the regulator factor was added to the HeLa cell preparations. The results indicated that the duration of the lag period before C4 synthesis was detectable was apparently a function of the phase of the cell cycle at which the factor was added to the cells. If the factor was added shortly after synchronization, i.e., while the cells were in GI, there was a lag period of approximately 17-18 hours before the appearance of detectable C4 production. On the other hand, if the factor was added at approximately the time of onset of DNA synthesis (as detected by the incorporation of tritiated thymidine), the lag period was somewhat less than 8 hours. Additional experiments, with short exposures to the factor, indicated that the HeLa cells were responsive to the factor in early S phase. A number of possible interpretations for this observation have been considered and are currently under investigation. One possibility seems likely, in that HeLa cells are more responsive to the factor when adherent than when in suspension cultures ( D . Stacey and H. R. Colten, unpublished, 1975), so that the response may depend pri-

100

HARVEY R. COLTEN

marily on the physical state of the plasma membrane and only indirectly on the cell cycle. The interpretation of these experiments may be even more complicated by the observation that some but not all fetal sera are also capable of inducing C4 production in HeLa cells, Evidence has been obtained that the activity in fetal serum is quite different from the factor produced by the C4-deficient cells. First, not all HeLa cells respond to the serum but, thus far, all the HeLa cells that have been examined do respond to the C4-deficient factor. Second, when the serum is subjected to Sephadex chromatography, there is a diffuse elution pattern with no clear-cut peak of activity in the 45,000 MW region. Third, normal guinea pig peritoneal cells are responsive to the serum but are not responsive to the C4-deficient factor. Finally, the C4-deficient factor is effective on HeLa cells grown in medium containing rat serum (which has no activity by itself), and its effect is reversible. On the other hand, the effect of fetal serum is not reversible. Thus, the stimulating factors in serum and C4-deficient cells can be separated. The final interpretation of the original experiments, however, must await further work since it appears that induction of complement synthesis in HeLa cells may be accomplished by several different factors. Recent studies on the cell-free synthesis of C4 should make it possible to resolve s0m.e of these interesting questions. Nevertheless, a few preliminary generalizations about the regulator are warranted. (1) The regulator factor appears to initiate biosynthesis of a specific protein without affecting total protein synthesis in the responsive cell. ( 2 ) The effect of the regulator is reversible. ( 3 ) The regulator is not species specific; i.e., a human cell is capable of recognizing and responding to guinea pig regulator. ( 4 ) Synthesis of the regulator and the response to it are both sensitive to inhibition by relatively low concentrations of actinomycin D and other inhibitors of protein synthesis. ( 5 ) The amount of regulator recovered from genetically deficient peritoneal cells was about 5-10 times that recovered from primary cultures of normal peritoneal cells. This could be the result of increased synthesis or decreased degradation of the factor in genetically deficient cells. These observations suggest the possibility that in deficient cells the regulator of C4 gene function accumulates because an interruption in the normal C4 biosynthetic pathway prevents normal feedback control. It has been shown that, in hybrid cells, activation or expression of a specialized gene function may depend on gene dose (Davidson and Benda, 1970; Fougkre et al., 1972) and that it is possible to demonstrate cross-species induction of gene function in the hybrid cell when one of the parent cells has a greater than normal chromosome complement (Peterson and Weiss, 1972). The results of the experiment with C4-deficient cells suggest the possibility of a somewhat similar


101

mechanism of “activation” of a silent gene. In hybrid cells, an increased gene dose may supply regulator factor or a similar mediator in sufficient quantity to initiate synthesis of a specialized gene product; whereas, in genetically deficient cells, defective feedback control may lead to increased production or accumulation of a regulator. A search for other specific regulator factors seems warranted, particularly in those genetic lesions that result in total absence of gene product.

B. C5 DEFICIENCY In a similar study of the molecular basis of C5 deficiency, Levy and co-workers (Levy and Ladda, 1971; Levy et al., 1973) have also made use of the technique of somatic cell hybridization in uitm. The two closely related mouse strains, B10D2 old (C5 deficient) and new (C5 sufficient) line, were used as sources of parent cells for the hybridization studies. As mentioned in an earlier section of this review, splenic macrophages from new line mice were the only cells capable of synthesizing C5. Hybridization of B10D2 new line kidney cells with B10D2 old line splenic macrophages and btween B10D2 old line macrophages and adult chicken erythrocytes indicated that, whereas none of the parent cells were capable of producing C5 by themselves, the hybrids produced C5 in tissue culture. Injection of the hybrid cells into old line mice resulted in the appearance of hemolytic complement activity in their sera. This activity peaked between 3 and 6 days after injection, gradually declined and was undetectable at 21 days after the injection, coincident with the appearance of anti-C5 antibody (Levy and Ladda, 1971). Appropriate controls ruled out a nonspecific effect ( a ) of Sendai virus (used to increase the frequency of fusion) and ( b ) of fusion itself, with the use of cells that would not provide adequate gene complementation; e.g., B10D2 old line kidney cells by B10D2, old line macrophages. Appearance of chemotactic activity in the sera of recipient animals paralleled the appearance of hemolytically active C5. Likewise appearance of MuBl (C5) antigen corresponded to the aforementioned activities. Several interpretations of these experiments were considered: ( 1 ) If C5 deficiency is due to a structural gene deletion, production of C5 in mouse hybrid cells might be explained by a simple complementation of the structural gene derived from a kidney cell and the regulator from the splenic macrophages thereby permitting expression of the structural gene. Results with the mouse by chicken erythrocyte hybrids were inconsistent with this interpretation since mouse C5 was produced. A limitation of this thesis is that it is probable that the chicken C5 equivalent would not be detected by the methods used in these experiments. Nonetheless, the problem of explaining production of mouse C5 remains even if there was concomitant production of chicken C5. ( 2 ) The possi-

102

HARVEY R. COLTEN

bility that some subunits of C5 were supplied by the chicken and other subunits by the deficient mouse cells, but neither immunochemical nor structural data are available to rule in or rule out this possibility. ( 3 ) C5 deficiency might be due to lack of a derepressor or presence of an abnormal repressor in the CSdeficient mouse cells. The experiments argue against these possibilities but do not rigorously exclude them. Levy et al. (1973) suggested that their experiments may provide a prototype for genetic repair in humans in spite of the rejection of the hybrid cells or inactivation of the gene product 21 days after restoration of C5 activity. They argue that in many of the human deficiency states a small amount of the protein is present and therefore the recipient would not be likely to make antibodies against the restored gene product. They do point out, however, that histocompatibility differences between the donor of the deficient and normal parental cells would lead to rejection of hybrids that expressed cell surface histocompatibility antigens. They cite Puck's experience (Puck et al., 1971) in which he was able to select against certain histocompatibility determinants by incubating hybrid cells in the presence of antisera raised against cell surface antigens. In fact, Puck's antisera were not HL-A specific but were heterologous ( rabbit) antisera against cell surface determinants. The principal difficulty with this argument is that it does not take into account observations that C2 (Fu et aZ., 1974), C4 (B. Benacerraf, unpublished, 1975), properdin factor B (Allen, 1974), and perhaps a regulator of C3 gene function (Ferreira and Nussenzweig, 1975) are closely linked to histocompatibility loci. Therefore hybrids carrying the chromosome for these gene products would also be likely to express cell surface histocompatibility antigens. C. C2 DEFICIENCY With the use of a method for prolonged culture of human peripheral blood monocytes it was shown that monocytes from three homozygous CZdeficient individuals synthesized no detectable C2 in uitro, even after 8 weeks in culture, whereas normal and heterozygous-deficient monocytes synthesized C2 throughout this period ( Einstein et uZ., 1975). Although monocytes from some heterozygous deficient individuals synthesized C2 at slower rates than monocytes from normals, there was no direct relationship between rates of C2 synthesis in vitro and the serum C2 concentrations. Perhaps this was due to stimuIation of C2 synthesis in heterozygous deficient cells by in vitro factors (e.g., fetal calf serum, glass surfaces) to a proportionately greater extent than in normal monocytes. Alternatively, release of the cells from in vivo regulatory factors may have been sufficient to account for these findings. Of note is an earlier


103

observation that C2 synthesis by peritoneal macrophages from homozygous C4-deficient guinea pigs was stimulated to a greater extent than was C2 synthesis by normal macrophages when both were exposed to heat-killed pneumococci ( Colten and Frank, 1972). The possibilities that C2-deficient serum or a product of C2-deficient monocytes might inhibit the biosynthesis or affect the stability or detection of C2 in vitro were tested and ruled out by showing that normal monocytes synthesized C2 at the same rate in medium supplemented with C2-deficient serum, or medium preincubated with C2-deficient monocytes; and medium from the C2-deficient monocyte cultures had no effect on the assay for or stability of preformed C2. Monocytes from normal and heterozygous deficient individuals incorprated radiolabeled amino acids into protein immunochemically identified as C2; there was no labeled C2 in the medium harvested from homozygous deficient monocytes. Incorporation of radiolabel into total protein was similar in monocytes from homozygous and heterozygous C2-deficient individuals and normals. Moreover, monocytes from the C2-deficient pateints could not be distinguished from the monocytes of other family members or unrelated normal subjects on the basis of morphology. They were equally capable of phagocytizing latex particles 5.7 pm in diameter, forming rosettes with IgG or C 3 coated erythrocytes, and killing L . monocytogenes, indicating that the defect was apparently specific for C2 synthesis. Fusion between C2-deficient monocytes or normal monocytes and undifferentiated cells, and monitoring C2 production and HL-A cell surface determinants, should make it possible to define the genetic lesion in C2 deficiency and to test whether both functions ( C 2 production and HL-A antigens) are invariably expressed coincidentally in hybrid cells. Hybrids between mouse macrophages and L cells express several macrophage-specific markers including H-2 antigens ( Gordon et al., 1971). Unfortunately, C2 synthesis by these hybrids was not studied. Others have shown that, in hybrid cells, C2 synthesis may be expressed (Levisohn and Thompson, 1973) and persist for many months in culture (R. Parkman, R. L. Davidson, and H. R. Colten, unpublished, 1975). In one of these studies (outlined in Fig. 5 ) some hybrids between human fetal liver and mouse 3T3 cells continued to synthesize and secrete human C2 up to one year after hybridization, although, as is characteristic of mouse-human hybrids, most of the human chromosomes had been lost. It appeared, based on examination of several clones, that persistence of a group C human chromosome correlated with the capacity to synthesize C2. This group C chromosome seemed to correspond to chromosome 6, but fluorescence-banding techniques were not sufficiently refined to determine this with certainty, These data, however, are not incompatible

104

HARVEY R. COLT’EN HUMAN FETAL LIVER

MOUSE FIBROBLAST

1

HYBRID CELLS

CLONE

SELECTIVE MEDIUM

e-

SUBCLONE Hernolytlc assay “‘C-labeled amino acids, radiotmmunodtffusion

FIG.5. Outline of experiment to determine the human chromosome coding for C2 synthesis.

with the localization of HL-A loci to chromosome 6 (van Someren et al., 1974; Lamm et al., 1974) and the close linkage between C2 deficiency and HL-A type (Fu et at., 1974).

D. DEFICIENCY OF C1 INHIBITOR Among the complement deficiency states, hereditary angioneurotic edema (Donaldson and Evans, 1963) is unique in its pattern of inheritance and is one of few genetic abnormalities of the complement system that results in significant clinical symptoms. In about 85% of affected individuals, the disease is due to deficiency of the natural inhibitor of C1, an a,-glycoprotein. In the balance of the pateints, a dysfunctional but immunochemically normal C1 inhibitor molecule replaces the normal protein (Rosen et al., 1971). Each form of the disease is inherited as an autosomal dominant. This raises an interesting question regarding genetic mechanism inasmuch as affected individuals are heterozygous for the defect, yet in the common variant of the disease the average concentration of C1 inhibitor is about 17%of normal rather than the expected 50%of normal. Likewise, in the rarer form of the disease all the immunochemically detectable C1 inhibitor protein is of the abnormal type. Evidence has been obtained to support the theory that C1 inhibitor deficiency is due to a defect in synthesis (Johnson et al., 1971a). On the basis of the model proposed for bacterial genes, a multicomponent system consisting of a structural gene, an operator site, and a pro-


105

moter site contiguous with the structural gene plus regulator genes, which need not be adjacent to the structural loci, may be postulated for the synthesis of C1 inhibitor. The regulator gene may code for a diffusible repressor substance which binds with the operator site and reversibly blocks transcription of the structural gene. Unbinding of the repressor by interaction with the inducer would permit gene expression. It has been proposed that hereditary angioneurotic edema may be due to a mutation arising in the regulator locus resulting in production of an abnormal repressor (Shokeir, 1973). The repressor may be modified so that it loses its affinity for the inducer and therefore the system remains constantly repressed. Since the repressor is diffusible, both genes would fail to function. An alternative hypothesis would propose that, in the normal, both genes are required for synthesis of C1 inhibitor. That is, full expression of the gene is inherited as though it were a recessive trait. This mechanism has been suggested for expression of some cell surface histocompatibility genes in mice (Shreffler and David. 1975). The problem of explaining the rare variant form of hereditary angioneurotic edema is much more difficult. The suggestion has been made in the past that this is a mutation which affects the attachment of thc carbohydrate moiety to the C1 inhibitor protein, but direct evidence for this is lacking. In any case two experiments, one of which is in progress, may differentiate among the possible genetic mechanisms that account for defective C1 inhibitor synthesis. In one, liver cells from a patient with hereditary angioneurotic edema have been fused in uitro with mouse fibroblasts (P. Rothman, L. P. Einstein, V. H. Donaldson, and H. R. Colten, unpublished, 1975). In the other, fibroblasts from pateints with hereditary angioneurotic edema will be fused with a rat hepatoma that is known to synthesize C1 inhibitor. These hybrids may make it possible to test for the presence of the postulated abnormal repressor. The mechanisms of other genetic diseases of the complement system may also yield to this type of analysis. The advantage of using complement as a model system for genetic studies is based primarily on the ease of detecting the gene product, the large number of well characterized genetic abnormalities, and the availability of cell lines capable of synthesizing individual complement proteins. VII. Nongenetic Control of Complement Biosynthesis

The recognition of the importance of Complement in host defenses has prompted many studies of complement levels in patients with a variety of diseases, particularly in those with acute infections. Although

106

HARVEY R. COLTEN

all such studies were limited by the fact that a static measure of serum complement concentration cannot reveal the dynamics of synthesis, catabolism, and distribution of a plasma protein, many interesting and provocative observations have resulted from this approach. In some fulminant infections such as meningococcal meningitis (Ecker et al., 1946) and dengue fever with shock (Bokisch et al., 1973) levels were often reduced largely as a result of increased consumption, but perhaps also due to depressed synthesis. On the other hand, it was found that during the course of most acute infections with pyogenic organisms there is often a marked increase in complement levels (Dick, 1912; Ecker et al., 1946). These clinical observations have been explored experimentally, but as yet no unifying control mechanisms have been uncovered to account for the striking changes in complement metabolism resulting from environmental pertubations. For example, staphylococcal bacteremia leads to a transient decline and then a marked increase in total complement activity in the sera of experimental animals (Baltch et al., 1962) as well as a change in apparent synthesis rates (Williams et al., 1963). Viable microorganisms are apparently not required for this effect as shown by the experiments of Jungeblut and Berlot (1926) many years before, indicating a similar effect with India ink. Ample evidence exists for a stimulatory effect of endotoxin on synthesis of certain plasma proteins including C-reactive protein and C3 (Hurlimann et al., 1966; Thorbecke et al., 1965; Stecher and Thorbecke, 1967b). This response may be quite prompt; that is, an increase in plasma protein synthesis by liver can be detected within 30 minutes after injection of the endotoxin. Whether these effects are mediated by similar or distinct mechanisms is not known. In view of a recent series of investigations, the possibility must be considered that phagocytosis or “activation” of cells of the reticuloendothelial system may account in part for the aforementioned findings, Early in the study of C2 biosynthesis by guinea pig peritoneal macrophages, it was noted that the material used to induce the exudate affected the proportion of cells synthesizing C2 and the amount produced per cell (Colten and Wyatt, 1972). Only 2 4 % of peritoneal cells obtained from oil- or casein-induced exudates produced C2 whereas 10-152 of cells from a starch-induced exudate or a peritoneal wash synthesized C2. This observation suggested either a qualitative difference in the macrophage populations obtained or that phagocytosis of certain particles might lead to a change in rate of secretion and/or synthesis of C2. Direct studies, in which normal guinea pig macrophages were exposed in vitro to heat-killed pneumococci, revealed up to a 10-fold increase in C2 and C4 production over baseline rates (Colten, 1974b). The extent


107

of stimulation was highly variable owing largely to significant variations in baseline rates of C2 production. In these experiments, the cells were maintained in a serum-free medium because normal serum also contains a factor or factors that stimulate C2 and C4 biosynthesis in vitro (Stecher and Thorbecke, 1967b; Colten, 1974a,b). The mechanism by which serum or phagocytic activity affects the rates of complement biosynthesis is not clear at the present time, but the effects are strikingly similar to those of identical stimuli on intracellular enzyme activities (Cohn and Benson, 1965). It is unlikely that the serum stimulatory factor is a complement component inasmuch as serum from guinea pigs with a genetic deficiency of C4 was as effective as normal guinea pig serum in stimulating C4 synthesis by guinea pig peritoneal macrophages (Colten and Frank, 1972). The signal for a change in rate of complement synthesis and the details of the cellular response to this signal have therefore not been defined. However, in the case of a “phagocytic” stimulus, there was suggestive evidence that increased secretion of C2 by cells exposed to particles in suspension required new protein synthesis, not merely accelerated release of preformed protein. These experiments raised the possibility that control of local production of complement at a site of inflammation may be of importance in affecting the balance between host defenses and an invading microorganism. A similar, perhaps related, finding was noted in studies of complement synthesis by syiiovial tissues (Ruddy and Colten, 1974). Previous studies of rheumatoid arthritis indicated that the involved synovial tissue synthesizes immunoglobulins. This conclusion was based on two lines of evidence: in vivo metabolism of IgG (Sliwinski and Zvaifler, 1970) and the in vitro biosynthesis of immunoglobulin by isolated rheumatoid synovium (Smiley et nl., 1968). A similar finding in the study of the metabolism of lZ5I-labeledC3 provided indirect evidence for local synthesis of this component of Complement by the synovium of a patient with rheumatoid arthritis. The ratio of C3-specific activities in synovial fluid and plasma indicated that approximately 50%of the intraarticular C3 was not derived from,the plasma pool. Direct evidence for synthesis of C3 and several other complement components by rheumatoid synovium was obtained with in vitro studies. Synovial tissue obtained at surgery from patients with rheumatoid arthritis produced biologically active C2, C3, C4, and C5 in short-term culture. Production of these complement components was temperature dependent and was inhibited in the presence of cycloheximide, a well known inhibitor of protein synthesis. When incubated in the presence of ‘C-labeled amino acids, the synovium incorporated label into immunoprecipitable C4, C3, and properdin factor B. Differences in rates of

108

HARVEY R . COLTEN

complement-component synthesis, among multiple samples of synovium obtained from the same joint, were consistent with the irregular distribution of the inflammatory lesions in rheumatoid synovial tissue. The frequency of synthesis of complement components was higher among cultures obtained from patients with rheumatoid arthritis than among those from patients with degenerative or traumatic arthritis, but, in contrast to the findings for immunoglobulin synthesis, production of C4 and C5 by synovial tissue did not appear to be restricted to rheumatoid arthritis. However, C3 synthesis was detected only in synovium from rheumatoid joints. Evidence has been obtained that sterile inflammatory reactions in U ~ U O may affect the capacity of mononuclear cells at remote sites to synthesize C2 and C4. The findings emerged from a study of the effects of carcinogens on macrophage functions (Colten and Borsos, 1974). In earlier experiments it appeared that in uitro several chemical carcinogens inhibited the biosynthesis of guinea pig C2 and C4. Among these compounds were some that affected complement biosynthesis at concentrations of M or less. In general, the capacity to inhibit C2 and C4 biosynthesis correlated with the carcinogenic activity of the compounds. A significant exception to this conclusion was noted in the case of some carcinogenic nitrosamines. This exception was of particular interest inasmuch as the studies were initially prompted by the observation that administration of diethylnitrosamine, a water-soluble carcinogen, to guinea pigs resulted in a selective decrease in the serum concentration of C4. Production of C2 and C4 by normal guinea pig cells and tissues in culture were unaffected by diethylnitrosamine, although two other carcinogenic nitrosamines inhibited biosynthesis of these components in uitro. Furthermore, diphenylnitrosamine, a noncarcinogenic analog of diethylnitrosamine, inhibited C2 and C4 production in oitro. Based on these observations, the possibility was considered that some nitroso compounds might be effective inhibitors of complement synthesis in uiuo but not in uitro, perhaps owing to metabolism of the compound to an active agent. As a corollary, it seemed possible that diphenylnitrosamine was metabolized to an inactive compound with respect to carcinogenic activity and capacity to inhibit complement production. In order to test this possibility, several nitroso compounds were administered to normal guinea pigs, and at timed intervals cells and tissues were removed for studies on the in uitro C2 and C4 biosynthesis. The results (H. R. Colten and T. Borsos, unpublished, 1975) indicated that two of the carcinogenic nitroso compounds, dimethylnitrosamine and diethylnitrosamine, were effective inhibitors of C2 and C4 biosynthesis in uiuo but not in uitro. Diphenylnitrosamine, on the other hand, inhibited C2 and C4 biosynthesis only in vitro but enhanced biosynthesis in duo.


109

In order to explore the surprising effects of diphenylnitrosamine on C2 and C4 synthesis, several other agents unrelated to carcinogenic compounds (turpentine, complete Freund's adjuvant) were injected intramuscularly to induce inflammatory reactions. The preliminary results of these experiments suggested that an inflammatory reaction at a remote site will lead to a 5- to 10-fold increase in C2 and C4 synthesis rates by peritoneal cells even when the PE cells are maintained in culture for 72 to 96 hours. These results are quite similar to those of Hartveit et al. ( 1973), who demonstrated in mice a stimulatory effect of an inflammatory response on C3 synthesis. In addition, they found that the extent of this response was genetically determined. Developmental changes also must be taken into account when one considers that, for example, adult mouse or rat liver failed to synthesize C3 in oitru except after stimulation of the animal with endotoxin (Thorbecke et al., 1965) whereas C3 synthesis was easily demonstrated in liver from unstimulated juvenile animals. Perhaps the latter situation is analogous to the striking increase in plamsa protein synthesis by liver, including C3 (Thorbecke et al., 1965) and C4 (Fust et al., 1972), several days after partial hepatectomy. That is, the stimulus for proliferation of liver cells and control of protein synthesis in each case (normal development and post hepatectomy) may be similar. Both functions persist in short-term tissue culture and therefore do not require a continuing external stimulus. Perhaps these factors in addition to stimulation by microbial agents account for the rapid rise in complement levels in the early months of postnatal life ( Fireman et al., 1969). An association between liver disease and changes in serum complement has been noted by many investigators. Cirrhosis, even in the absence of liver failure, may be accompanied by decreased serum complement levels (Jordan, 1953; Ashershon, 1960). Levels of C2, C3, and C4 and principally affected, but C1 concentration is not ( Kourilsky et d., 1973). Experimental studies of the effect of a low-protein diet on complement levels have revealed significant depression of serum C4 and to a lesser extent C3 and C2, but no effect on CI (Rice et al., 1951a). These findings are for the most part substantiated by clinical observations (Sirisinha et al., 1973) and may be related to some of the changes in complement levels in cirrhosis. Changes in complement associated with viral hepatitis are more complex. Reports of increased levels (Kosmidis and Leader-Williams, 1972; Kourilsky et al., 1973) and of decreased levels particularly when arthritis is present ( Alpert et al., 1971) have appeared. Consistent increases, particularly in C3, have been reported in cases of obstructive jaundice suggesting the possibility that C3 may normally be secreted into the biliary system. In none of these studies has complement synthesis or catabolism been investigated, SO

110

HARVEY R. COLTEN

that interpretations of these observations must await further investigations. Several investigators have studied the metabolism of Clq (Kohler and Muller-Eberhard, 1972), C3 (Alper et al., 1966), and C4 (Carpenter et al., 1969) in vivo in normal subjects and in patients with acquired abnormalities of the complement system using radiolabeled purified complement proteins. Metabolic studies of C3 in patients with membranoproliferative glomerulonephritis ( MPGN ) have suggested that decreased C3 biosynthesis often contributes to depressed serum levels of C3 ( Alper et al., 1966; Peters et al., 1972). Other authors (Hunsicker et al., 1972) have claimed that the decrease in serum C3 associated with MPGN is solely a consequence of increased C3 catabolism. As indicated above, recent work (Colten, 1972a) has confirmed that the liver is a principal site of C3 biosynthesis in that short-term cultures of liver produced biologically active C3, production of C3 was temperature dependent and reversibly inhibited by cycloheximide, and 14C-labeled amino acids were incorporatea into C3 protein. Liver biopsies were obtained from patients with MPGN, patients with other hypocomplementemic renal diseases, and controls ( pateints without renal diseases). In general, rates of biosynthesis of C3 by liver specimens in uitro corresponded to synthesis rates calculated from metabolic turnover studies (Colten et al., 1973). Liver samples from two pateints with MPGN failed to produce detectable C3 in vitro, although they were capable of synthesizing C2 and C5. Studies of the metabolic turnover of radiolabeled C3 in these two patients also indicated a depression of C 3 biosynthesis. The rates of C3 synthesis did not correlate with the serum concentration of C3, i.e., some patients with low serum C3 levels had greater synthesis rates than those with higher serum C3 levels, and vice versa. Preliminary data suggested two possible mechanisms for decreased C3 biosynthesis in MPGN. One mechanism is a relative deficiency of a normal heat-stable serum factor that stimulates C3 biosynthesis. A comparable effect of serum on C3 and transferrin synthesis by murine cells has also been observed (Stecher and Thorbecke, 196713). The other is perhaps an inability of the liver from patients with MPGN to respond to this C3-stimulating factor. In short, both increased catabolism and decreased synthesis account for depression of C3 levels in MPGN. A similar finding was noted in studies of C 3 metabolism in systemic lupus erythematosis (SLE), but results were somewhat more variable (Alper and Rosen, 1967; Hunsicker et al., 1972; Sliwinski and Zvaifler, 1972). For example, in one study, two of three patients with untreated SLE and low serum C3 levels had depressed synthesis rates whereas the synthesis rate was normal in the third patient. Clq and C4 turnover


111

were markedly increased, and in another study C l q synthesis rates were found to be significantly increased as well (Kohler and Muller-Eberhard, 1972). Treatment with between 40 and 120 mg of prednisone per day resulted in an increase in C 3 synthesis rate in each of the patients and a variable effect on C 3 catabolic rate (Sliwinski and Zvaifler, 1972). Apparently contradictory data were obtained in a study of C3 levels after administration of cortisone to experimental animals ( Atkinson and Frank, 1973). Guinea pigs treated with cortisone acetate ( 5 mglkg) demonstrated a 20% and a 51%elevation of CH,, and C1 titers, respectively. C4, C2, and C3-9 complex were all somewhat higher but not statistically different from controls. At 20 mg/kg, C1 was 137%of control values and the other components were essentially unchanged. At 100 mg/kg, C1 and C9 were the only components not significantly depressed. During the first 2 weeks of high-dose cortisone treatment there was a progressive fall in CH5", C2, C4, and (23-9. After discontinuation of therapy these levels remained low for approximately 2 weeks. Factor B of the properdin system was also significantly reduced in animals treated with 20 mg or 100 mg/kg per day. Periorbital abscesses secondary to bleeding procedures abolished the depression of CH,,, C4, and C3-9 complex, again emphasizing the effect of inflammation on complement levels. The authors indicated that it is impossible to say whether these effects are due to altered synthesis and/or catabolic rates. Direct experiments, on the other hand, showed that in tissue culture the presence of hydrocortisone at a concentration of approximately M stimulated synthesis of C3 and transferrin as judged by semiquantitative methods ( Stecher and Thorbecke, 196713). In experiments designed to quantitate the magnitude of this effects of cortisone it was shown that incubation of a well differentiated rat hepatoma in the presence of hydrocortisone succinate ( 4 x lo-' M ) increased rates of C3 production u p to %fold over baseline but had no effect on C5 synthesis (Strunk et al., 1975). The posssibility that this effect was restricted to malignant cells was considered and ruled out by showing a comparable effect of hydrocortisone on C3 synthesis by primary cultures of normal rat liver (J. Breslow, P. Rothman, and H. R. Colten, unpublished, 1975). Suggestive evidence has been obtained that hormones other than cortisol may also affect complement synthesis, but these preliminary observations have also not been pursued to any significant degree. For example, Urbach and Cinader (1966) first noted that the concentration of MuBl (C 5) protein in sera of normal male mice was approximately twice that in normal female sera. Churchill et al. (1967) found that late-acting complement components ((23-9 complex) in sera from male mice were eight to ten times higher than the corresponding components in sera

112

HARVEY R. COLTEN

from female mice. Evidence was presented that this difference is due to the effect of androgens and estrogens on late-acting complement components tentatively identified as C5 and C6. The sex hormones had a greater effect on C6 than on C5. In humans it is known that total complement activity (Ecker and Rees, 1922) and C3 (Propp and Alper, 1968) are significantly increased late in pregnancy. Studies of complement metabolism and of the effects of sex hormones on complement synthesis in vitro should elucidate the mechanisms responsible for these phenomena. The effects of malignant transformation on complement synthesis have received little attention. However, Stecher and Thorbecke ( 1967c) have shown that cells of 3T3, a mouse fibroblast cell line that is apparently capable of incorporating 14C-labeled amino acids into C3 protein, are unable to do so 1week after viral transformation of the cells with MSV40. In a related experiment Green et al. (1966) reported studies of collagen synthesis by 3T3 and SV40-transformed 3T3 in the presence of low and of high ascorbate concentrations. Maximal collagen synthesis occurred in cells grown in medium containing high concentrations of ascorbate, but under these conditions the transformed cells produced less collagen than the parental cell. However, at low ascorbate concentrations the transformed cells appeared to synthesize more collagen than the parent cell, suggesting that collagen synthesis in transformed cells is less dependent on exogenously supplied ascorbate. These interesting observations have not been pursued in respect to complement biosynthesis but clearly offer promise of a fruitful line of investigation. VIII. Concluding Remarks

When embarking on a review of even a limited area of scientific investigation, one is first struck by the extraordinary volume of published information. As time passes it becomes more obvious that the promising leads offered by isolated observations or contradictory data far outnumber the firmly established facts. With these thoughts in mind, I have attempted in this review to suggest some of the areas in which future studies of complement biosynthesis may cIarify several fundamental problems in medicine, biology, and biochemistry. It is apparent that we have not even established unequivocally the sites of synthesis of each of the complement and complement-related proteins. Such studies may not be particularly intellectually challenging now that methods for investigations of complement synthesis in vitto have been well defined. They are, on the other hand, clearly a prerequisite for providing answers to the more challenging questions. The


113

complement deficiency diseases and genetic polymorphisms of some of the complement proteins provide excellent models for studies of genetic control of protein synthesis in eukaryotes. Application of the technique of somatic cell hybridization and cell free synthesis systems have already proved useful and provocative. The interpretation of these experiments may be more difficult than was anticipated at first, but there is little doubt that, eventually, similar studies will make it possible to dissect the genetics of plasma protein production at a molecular level much as has been done in studies of bacterial genetics. A description of the nongenetic molecular control of plasma protein synthesis and its role in normal host defenses against microbial infection will soon be available if, as anticipated, the preliminary findings are followed by more sophisticated study of this problem. Inasmuch as recent work has begun to reveal the fine structure and details of function of the complement proteins, it is not surprising that these goals now look close. These studies and the development of exquisite immunochemical and functional assay systems for quantitative measurements of complement protein have provided the tools for a direct attack on the essential issues.

ACKNOWLEDGMENTS I thank Ms. Barbara Caruso, Rita Callan, and Prudence Hansen for assistance in the preparation of this manuscript.

REFERENCES Addis, T. (1912). J. Znfec. Dis. 10, 200. Adinolfi, M. ( 1972). In Ciba Symposium “Ontogeny of Acquired Immunity,” p. 65. Elsevier, Amsterdam. Adinolfi, M., and Gardner, B. (1967). Acta Paediat, Scand. 56, 450. Adinolfi, M., Gardner, B., and Wood, C. B. S. (1968). Nature (London) 219, 189. Allen, F. H., Jr. (1974). V o x Sang. 27, 382. Alper, C . A., and Rosen, F. S. ( 1967). J. Clin. Inuest. 46, 2021. Alper, C. A,, and Rosen, F. S. ( 1971). Aduan. Immunol. 14, 252. Alper, C. A., Levin, A. S., and Rosen, F. S. (1966). Science 153, 180. Alper, C. A., Propp, R. P., and Watson, L. (1968). J. Clin. Inuest. 47, 2181. Alper, C. A., Johnson, A. M., Birtch, A. G., and Moore, F. D. (1969). Science 163, 286. Alper, C. A,, Boenisch, T., and Watson, L. (1972). J. Exp. Med. 135, 68. Alpert, E., Isselbacher, K. J., and Schur, P. H. ( 1971). N . Engl. J. Med. 285, 185. Ashershon, G. L. (1960). Australus. Ann. Med. 9, 57. Asofsky, R., and Thorbecke, G. J. (1961). J . Exp. Med. 114, 471. Atkinson, J. P., and Frank, M. M. (1973). J. Immunol. 111, 1061.

114

HARVEY R. COLTEN

Austen, K. F., Becker, E. L., Borsos, T., Lachmann, P. J., Lepow, I. H., Mayer, M. M., Miiller-Eberhard, H. J., Nelson, R. A., Rapp, H. J., Rosen, F. S., and Trnka, Z. (1968). Bull. W. H. 0. 39, 935. Bach, S., Ruddy, S., MacLaren, J. A., and Austen, K. F. (1971). Immunology 21, 869. Ballow, M., Day, N. K., Biggar, W. D., Park, B. H., Yount, W. J., and Good, R. A. (1973). Clin. Immunol. lmmunopathol. 2, 28. Baltch, A. L., Osborne, W., Canarile, L., Hassirdjian, A., and BUM, P. (1962). J. Immunol. 88, 361. Bing, D. H., Spurlock, S. E., and Bern, M. M. (1975). C h . Immunol. Immunoputhol. (in press). Bloom, W. (1926). Amer. J. Anat. 36, 451. Bokisch, V. A,, Top, F. H., Russell, P. K., Dixon, F. J., and Miiller-Eberhard. H. J. (1973). N. Engl. J. Med. 289, 996. Bordet, J. ( 1896).Ann. Inst. Pasteur, Paris 10, 193. Buchner, H. (1889). Zentralbl. Bakteriol., Parusitenk. 6,561. Buchner, H. (1894). Muenchen. Med. Woehenschr. 41,717. Carpenter, C. B., Ruddy, S., Shehadeh, I. H., MiiIler-Eberhard, H. J., Merrill, J. P., and Austen, K. F. ( 1969). J. Clin. Inuest. 48, 1495. Carrel, A., and Ingebrigtsen, R. ( 1912 ). J. E x p . Med. 15, 287. Chan, P. C. Y. (1970). Expetientia 26, 189. Chan, P. C. Y., and Cebra, J. J. (1966). Immunochemistry 3, 496 (abstr.). Churchill, W. H., Weintraub, R. M., Borsos, T., and Rapp, H. J. (1967). J. Exp. Med. 125, 657. Cinader, B., and Dubiski, S. ( 1963). Nature (London) 200,781. Cohn, Z. A., and Benson, B. (1965). J. Exp. Med. 121,835. Colten, H. R. (1972a). J. Clin. Inuest. 51, 725. Colten, H. R. (1972b). Proc. Nat. Acad. Sci. US.69,2233. Colten, H. R. ( 1973). Clin. Immunol. Immunopathol. 1, 346. Colten, H. R. (1974a). In “Somatic Cell Hybridization” ( R . L. Davidson and F. F. de la Cruz, eds.), p. 197. Raven Press, New York. CoIten, H. R. (1974b). In “Proceedings of The Second International Congress of Immunology” (L. Brent and J. Holbrow, eds.). Vol. 1, p. 183. North-Holland Publ., Amsterdam. Colten, H. R., and Alper, C. A. (1972). J. Immunol. 108, 1184. Colten, H. R., and Borsos, T. ( 1974). J. Immunol. 112, 1107. Colten, H. R., and Frank, M. M. (1972). Immunology 22,991. Colten, H. R., and Parkman, R. (1972). Science 176, 1029. Colten, H. R., and Wyatt, H. V. (1972). In “Biological Activities of Complement” ( D. H. Ingram, ed. ), p. 244. Karger, Basel. Colten, H. R., Borsos, T., and Rapp, H. J. (1966). Proc. Nut. Acad. Sci. US. 56, 1158. Colten, H. R., Borsos, T., and Rapp, H. J. (1967). Science 158, 1590. Colten, H. R., Gordon, J. M., Borsos, T., and Rapp, H. J. (1968a). J . Exp. Med. 128, 595. Colten, H. R., Cordon, J. M., Rapp, H. J., and Borsos, T. (1968b). J. Immmol. 100, 788. Colten, H. R.,Silverstein, A. M., Borsos, T., and Rapp, H. J. ( 1 9 6 8 ~ )Immunology . 15, 459. Colten, H. R., Levy, R. H., Rosen, F. S., and AIpers, C. A. (1973). J. Clin. Inuest. 52, 20a (abstr.).


115

Cooper, N. R,, and Reid, K. B. M. ( 1974). In “Proceedings of the Second International Congress Immunology” (L. Brent and J. Holbrow, eds.), Vol. l , p. 291. North-Holland, Publ., Amsterdam. Davidson, R. L., and Benda, P. (1970). Proc. Not. Acad. Sci. U.S. 67, 1870. Day, N. K., Gewurz, H., Pickering, R. J., and Good, R. A. (1970). J. Immunol. 104, 1316. Dick, G. F. (1912). J. Infec. Dis. 10, 383. Dick, G. F. (1913). 1. Infec. Dis. 12, 111. Donaldson, V. H., and Evans, R. R. (1963). Amer. J. Med. 35,37. Ecker, E. E., and Pillemer, L. (1942). Ann. N.Y. Acad. Sci. 43, 63. Ecker, E. E., and Rees, H. M. (1922). J. Infec. Dis. 31, 361. Ecker, E. E., Seifter, S., Dozois, T. F., and Barr, L. (1946). J. Clin. Inoest. 25, 800.

Ehrlich, P., and Morgenroth, J. (1900). Berlin Klin. Wochenschr. 37, 453. Einstein, L. P., Alper, C. A., Bloch, K. J., Herrin, J. T., Rosen, F. S., David, J. R., and Colten, H. R. (1975). N . Engl. J . Med. 292, 1169. Ellman, L., Green, I., and Frank, M. M. ( 1970). Science 170, 74. Ferrata, A. (1907). Berlin. Klin. Wochenschr. 44, 366. Ferreira, A., and Nussenzweig, V. (1975). J. Exp. Med. 141, 513. Fireman, P., Zuchowski, D. A,, and Taylor, P. M. (1969). J. Immunol. 103, 25. FougBre, C., Reinze, F., and Ephrussi, B. (1972). Proc. Nat. Acad. Sci. U.S. 69, 330. Frank, M. M., May, J., Gaither, T., and Ellman, L. (1971). J. Exp. Med. 134, 176. Fu, S. M., Kunkel, H. G., Brusman, H. P., Allen, F. H., and Fotino, M. (1974). J. Exp. Med. 140, 1108. Fiist, G., and Surjin, M. (1971). Boll. 1st. Sieroter. Milan. 50, 488. Fiist, G., SurjAn, M., and Keresztes, M. (1972). Boll. 1st. Sieroter. Milan. 51, 304. Gabrielsen, A. E., Linna, T. J., Wertekamp, D. P., and Pickering, R. J. (1974). Immunology 27, 463. Geiger, H., Day, N., and Good, R. A. (1972). J. Immunol. 108, 1092. Gengou, 0. (1901). Ann. Inst. Pasteur, Paris 15, 68. Gewurz, H., Pickering, R. J., Christian, C. L., Snyderman, R., Mergenhagen, S. E., and Good, R. A. ( 1968). Clin. E x p . Immunol. 5,437. Gitlin, D., and Biasucci, A. (1969). J. Clin. Inuest. 48, 1433. Glade, P. R., and Chessin, L. N. (1968). Int. Arch. Allergy. Appl. Immunol. 34, 181. Gordon, P. (1955). Proc. SOC. Exp. Biol. Med. 89, 607. Gordon, S., Ripps, C. S., and Cohn, Z. (1971). J. Exp. Med. 134, 1187. Gordon, S., Todd, J., and Cohn, Z. A. (1974). J. Exp. Med. 129, 1228. Green, H., Todaro, G. J., and Goldberg, B. (1966). Nature (London) 209, 916. Hankin, E. H. (1892). 2. Bakteriol. Parasitenk. 12, 809. Hartveit, F., BZrve, W., and Thunold, S. (1973). Acta Pathol. Microbid. Scand. Suppl., Sect. A 236, 54. Hirsch, J. G. (1956). J. Exp. Med. 103, 613. Hirschfeld, J., and Lunell, N. 0. ( 1962). Nature (London) 196, 1220. Hobart, M. J., Lachmann, P. J., and Alper, C. A. (1975). In “Protides of the Biological Fluids” (H. Peters, ed. ), p. 575. Pergamon, Oxford. Hochwald, G. M., Thorbecke, G. J,, and Asofsky, R. (1961). J. Exp. Med. 114, 459. Hochwald, G. M., Jacobson, E. B., and Thorbecke, G. J. (1964). Fed. Proc., Fed. Amer. SOC. Exp. Biol. 23, 577 ( abstr. ).

116

HARVEY R. COLTEN

Hunsicker, L. G., Ruddy, S., Carpenter, C. B., Schur, P. H., Merrill, J. P., MullerEberhard, H. J., and Austen, K. F. (1972). N . Engl. I. Med. 287, 835. Hurlimann, J., Thorbecke, G . J., and Hochwald, G. M. (1966). J. Exp. Med. 123, 365. Ilgen, C . L., and Burkholder, P. M., ( 1974). Immunology 26, 197. Ilger, C. L., Bossen, E. H., Rowlands, D. T., Jr., and Burkholder, P. M. (1974). Immunology 26, 659. Ingraham, J. S., and Bussard, A. ( 1964). J . E r p . Med. 119, 667. Jensen, J. A,, Garces, M. C., and Iglesias, E. (1971). Infec. Immunol. 4, 12. Jerne, N. K., and Nordin, A. A. ( 1963). Science 140,405. Johnson, A. M., Alper, C. A., Rosen, F. S., and Craig, J. M. (1971a). Science 173, 553. Johnson, A. M., AIper, C. A., Rosen, F. S., and Craig, J. M. (1971b). 1. Clin. Invest. 50, 50a (abstr. ). Jordan, F. L. J., (1953). Act. Med. S c a d 144, 268. Jungeblut, C. W., and Berlot, J. A. (1926). 1. Erp. Med. 43,797. Kashiwagi, N., Groth, C. G., and Starzl, T. E. (1968). Proc. SOC. Exp. Biol. Med. 128, 247. Keston, A. S., and Katchen, B. ( 1956). J. Zmmunol. 76,253. Kohler, P. F., (1973). J. Ckn. Invest. 52, 671. Kohler, P. F., and Muller-Eberhard, H. J. (1969). Science 163,474. Kohler, P. F., and Muller-Eberhard, H. J. (1972). J . Clin. Inoest. 51, 868. Kosmidis, J. C., and Leader-Williams, L. K. ( 1972). Clin. Erp. Immunol. 11,31. Kourilsky, O., Leroy, C., and Peltier, A. P. (1973). Amer. 1. Med. 55, 783. Lai A Fat, R. F. M., and van Furth, R. (1975). Immunology 28,359. Lamm, L. U., Friedrich, U., Peterson, G . B., Jorgensen, J., Nielsen, J., Therkelsen, A. J., and Kissmeyer-Nielsen, F. ( 1974). Hum. Hered. 24,273. Leddy, J. P., Frank, M. M., Gaither, T., Baum, J., and KIemperer, M. R. (1974). I. Clin. Invest. 53, 544. Lepow, I. H. (1965). Complement, Ciba Found Symp., 1964 p. 117. Lepow, I. H. Naff, G . B., Todd, E. W., Pensky, J., and Him, C. F. (1963). J . Exp. Med. 117, 983. Levisohn, S. R., and Thompson, E. B. (1973). J . Cell. Physiol. 81, 225. Levy, N. L., and Ladda, R. L. (1971). Nature (London), New Biol. 229, 51. Levy, N. L., Synderman, R., and Ladda, R. L. ( 1973). Proc. Nut. Acad. Sci. U.S. 70, 3125. Littleton, C., Kessler, D., and Burkholder, P. M. (1970). Immunoolgy 18, 691. Mathews, C. M. E. (1957). Phys. Med. Biol. 2, 36. Mayer, M. M. (1961). In “Experimental Immunochemistry” (C. A. Kabat and M. M. Mayer, eds. ), 2nd ed., p. 133. Thomas, Springfield, Illinois. Merrill, D. A., Kirkpatrick, C. A., Wilson, W. E. C., and Riley, C. M. (1964). Proc. SOC. Exp. Biol. Med. 116, 748. Metchnikoff, E. (1889). Ann. Zmt. Pasteur, Paris 3,664. Metchnikoff, E. (1905). “Immunity in Infective Diseases,” p. 591. Cambridge Univ. Press, London and New York. Moore, D. L., and Vas, S. I. (1968). Immunology 15, 185. Muller, L. (1911). Zentralbl. Bakteriol. Parasitenk, Infertiomkr., Abt. I: Orig. 57, 577. Muller-Eberhard, H. J., and Biro, C. ( 1963). J. Erp. Med. 118, 447. Miiller-Eberhard, H. J., Nilsson, U., and Aronsson, T. (1960). J. Exp. Med. 111, 201.


117

Nilsson, U. R., Mandle, R. J., Jr., and McConnell-Mapes, J. A. ( 1975). 3. Immunol. 114, 815. O’Connell, R. M., Enriquez, P., Linman, J. W., Gleich, G. J., and McDuffie, F. C. (1967). J. Lab. Clin. Med. 70, 715. Olsen, 0. (1922). Biochem. 2. 133, 24. Peters, D. K., Martin, A., Weinstein, A., Cameron, J. S., Barrett, T., Ogg, C. S., and Lachmann, P. J. (1972). Clin. Exp. Immunol. 11, 311. Peters, T., Jr., and Anfinsen, C. B. (1950a). J. Biol. Chem. 182, 171. Peters, T., Jr., and Anfinsen, C. B. ( 1950b). J. Biol. Chem. 186, 805. Peterson, J. A., and Weiss, M. C. (1972). Proc. Nut. Acad. Sci. U.S. 69, 571. Pfeiffer, R., and Issaeff, R. (1894). 2. H y g . Infektionskr. 17, 355. Phillips, M. E., and Thorbecke, G. J. (1965). Nature (London) 207, 376. Phillips, M. E., and Thorbecke, G. J. (1966). Int. Arch. Allergy Appl. Zmmunol. 29, 553. Phillips, M. E., Rother, U. A., and Rother, K. 0. (1969). Immunology 17, 315. Pickering, J. R., Naff, G. B., Stroud, R. M., Good, R. A., and Gewurz, H. (1970). J. Exp. Med. 131, 803. Propp, R. P., and Alper, C. A. ( 1968). Science 162, 672. Puck, T. T., Wuthier, P., Jones, C., and Kao Fa-Ten (1971). Proc. Nut. Acad. Sci. US.68, 3102. Rapp, H. J., and Borsos, T. (1970). “Molecular Basis of Complement Action.” Appleton, New York. Rausen, A. R., Gerald, P. S., and Diamond L. K. (1961). Nature (London) 191, 717. Rice, C. E. ( 1954). 1. Immunol. 73,375. Rice, C. E., Boulanger, P., and Plummer, P. J. G. (1951a). Can. J. Med. Sci. 29, 13. Rice, C. E., Boulanger, P., and Plummer, P. J. G. (1951b). Can. J. Med. Sci. 29, 48. Rommel, F. A., Goldlust, M. D., Bancroft, F. C., Mayer, M. M., and Tashjian, A. H., Tr. (1970). J . Zmmunol. 105, 396. Rosen, F. S., Alper, C. A., Pensky, J., Klemperer, M. R., and Donaldson, V. H. ( 1971). J. Clin. Znoest. 50, 2143. Rosenberg, L. T., and Tachibana, D. K. ( 1962). J. Immunol. 89,861. Rother, K., Rother, U., Miiller-Eberhard, H. J., and Nilsson, U. R. (1966). J. Exp. Med. 124, 773. Rother, U., Thorbecke, G. J., Stecher-Levin, V. J., Hurlimann, J., and Rother, K. ( 1968). Immunology 14, 649. Rubin, D. J., Borsos, T., Rapp, H. J., and Colten, H. R. (1971). J. Immunol. 106, 295. Ruddy, S., and Colten, H. R. (1974). N . Engl. J. Med. 290, 1284. Ruddy, S., Klemperer, M. R., Rosen, F. S., Austen, K. F., and Kumate, J. (1970). Immunology 18, 943. Schreiber, R. D., and Miiller-Eberhard, H. J. (1974). J. Exp. Med. 140, 1324. Shokeir, M. H. K. (1973). Clin. Genet. 4,494. Shreffler, D. C., and David, C. S. (1975). Adoan. Immunol. 20, 125. Siboo, R., and Vas, S. I. ( 1965). Can. J. Microbiot. 11,415. Sirisinha, S., Suskind, R., Edelman, R., Charuptana, C., and Olsen, R. E. (1973). Lancet 1, 1016. Slikwinski, A. J., and Zvaifler, N. H. (1970). J. Lab. Clin. Med. 76, 304. Sliwinski, A. J., and Zvaifler, N. H. (1972). Clin. Erp. Immunol. 11, 21.

118

HARVEY €3. COLTEN

Smiley, J. O,, Sachs, C., and Ziff, M. (1968). J. Clin. Invest, 47, 624. Stavitsky, A. B. (1961). Advun. Immunol. 1, 211. Stecher, V. J,, and Thorbecke, G. J. (1967a). J . lmmunol. 99,643. Stecher, V. J., and Thorbecke, G. J. (1967b). J. Immunol. 99, 653. Stecher, V. J., and Thorbecke, G. J. ( 1 9 6 7 ~ )J.. lmmunol. 99, 660. Stecher, V. J., Morse, J. H., and Thorbecke, G . J. (1967). Proc. SOC.Exp. Biol. Med. 124, 433. Stroud, R. M., and Donaldson, V. (1974). In “Proceedings of The Second International Congress of Immunology” (L. Brent and J. Holbrow, eds.), Vol. 1, p. 288. North-Holland Publ., Amsterdam. Stroud, R. M., Nagaki, K., Pickering, R. J., Gewurz, H., Good, R. A., and Cooper, M. D. (1970). Exp. Immunol. 7,133. Strunk, R. S., Tashjian, A. H., and Colten, H. R. (1975). 1. Immunol. 114, 331. Tachibana, D. K., and Rosenberg, L. T. (1966). J. lmmunol. 97,213 Terry, W. D., Borsos, T., and Rapp, H. J. ( 1964). 1. Immunol. 92, 576. Thorbecke, G. J., Hochwald, G . M., van Furth, F., MiilIer-Eberhard, H. J., and Jacobson, E. B. (1965). Complement, Cibu Found. Symp., 1964 p. 99. Urbach, G., and Cinader, B. (1966). Proc. SOC. E x p . Biol. Med. 122, 779. van Someren, H., Westerveld, A., Hagemeijer, A., Mees, J. R., Meera Khan, P., and Zaalberg, 0. B. (1974). Proc. Nut. Acad. Sci. US.71, 962. van Zeipel, G. (1970). Actu Puthol. Microbiol. Scund. 78, 258. Whipple, G. H., and Hurwitz, S. H. (1911). J . Exp. Med. 13, 136. Williams, C. A,, Jr., Asofsky, R., and Thorbecke, G. J. (1963). J. Exp. Med. 118,

315.

Wyatt, H. V. (1974). Eur. J . Immunol. 4, 34. Wyatt, H. V., Colten, H. R., and Borsos, T. (1972). J . Immunol. 108, 1609. Yount, W. J., Utsinger, P. D., Catti, R. A., and Good, R. A. (1974). J. Pediut. 84, 193.

Graft-versus-Host Reactions: A Review'

.

STEPHEN C. GREBE' AND J W A Y N E STREllElN' Departments of Cell Biology ond lnternol Medicine. Southwestern Medical School. University o f Texos Health Science Center ot Dallas. Texas

.

I Introduction . . . . . . . . . . . . . I1. Graft-versus-Host Reactions (GVHRs) . . . . . . . A . General Characteristics . . . . . . . . . . B . Systemic GVH Processes and Assays . . . . . . . C . Local GVHRs and Assays . . . . . . . . . D . GVH Disease in Man and Other Primates: Lessons from the Clinic I11. The Cellular Basis of Graft-versus-Host Reactions . . . . . A . Identification of Immunocompetent Cells . . . . . B . Studies in Cell Cooperation . . . . . . . . . . . . . . C . Alteration of GVHRs by Suppressor T Cells . IV. Donor Lymphoid Cell Participation in Graft-versus-Host Reactions A . Donor Lymphocyte Traffic, Homing. and Distribution . . B. Donor Lymphocyte Proliferation . . . . . . . C . Role of Lymphokines in CVHRs . . . . . . . D . Generation of Specific Effector Cells . . . . . . . E . Fate of Donor Histocompatibility Antigen-Reactive Cells . . V. Host Participation in Graft-versus-Host Reactions . . . . A . Immunogenic Stimulus . . . . . . . . . . B. Development of Lesions . . . . . . . . . . C. Alteration of Host Inimunologic Capability . . . . . . VI . Experimental Modification of Graft-versus-Host Reactions . . A . General Considerations . . . . . . . . . B Reduction of GVH Potential by Manipulation of Donor and/or Donor Lymphocytes . . . . . . . . . . C . Attempts to Alter GVHRs by Treating the Host . . . . . VII . Immunoregulation and the Graft-versus-Host Reaction . . . . A . Spontaneous Alterations of GVHRs as Manifestations of Imniunoregnlation . . . . . . . . . . . . . B Cellular and Molecular Aspects of Immunoregulation in GVHRs C . The Imniunoregulatory Role of the Spleen . . . . . . VIII . Summary: Perspectives and Prospects . . . . . . . References . . . . . . . . . . . . .

.

. . . . . .

.

.

.

.

.

120 121 121 124 127 137 144 145 147 152 154 154 156 157 158 160 161 162 163 166 171 171 171 180 183 184 187 201 204 205

'This work was supported in part by United States Public Health Service Grant A1 l R O l 10678. * Present address: Naval Medical Research Institute. National Naval Medical Center. Bethesda. Maryland 20014 . Markle Scholar in Academic Medicine .

119

120

STEPHJ3N C. GREBE AND J. WAYNE STREILEIN

I . Introduction

Graft-versus-host reactions (GVHRs) are at once one of the most intriguing facets of transplantation immunology as well as the most baMing. While they have reached clinical importance recently by virtue of the increased number of allogeneic bone marrow transplants being carried out in man, they continue to offer a fertile arena for fundamental immunologic research, It is now recognized that GVHRs represent a unique complex of systems for studying the dynamics of cellular events that proceed from the specific stimulation of antigen-reactive lymphocytes. The intricate cellular interactions that underpin GVHRs and give rise to the several distinctive clinical syndromes have come under increasing experimental scrutiny, As a consequence, the perhaps too simplistic view of the pathogenesis of GVHRs that prevailed a decade ago is obviously inadequate to account for the phenomena now known to be associated. The use of this genre of immune reactivity to dissect the means by which T and B lymphocytes communicate is but one brilliant example of the opportunities for unraveling immunologic issues offered by the use of experimental GVHRs. It is over 20 years since the first formal descriptions of GVH disease were published (Billingham et al., 1955; Trentin, 1956, 1957; Billingham and Brent, 1957; Simonsen, 1957). By definition, it was demanded that a GVHR could be initiated only by a “graft” of immunologically competent lymphocytes introduced into a “host” that confronts the graft with a major histocompatibility difference yet is unable to mount a similar immunologic attack against the intrusive donor lymphoid cells ( Billingham, 1968). The early series of experiments led to the identification of the “immunologically competent cell” predicted and described by Medawar (1963) as the cell responsible for the initiation of GVH disease. Subsequent work from a number of laboratories demonstrated unequivocally that a cell identified morphologically as a small lymphocyte is the immunologically competent cell that initiates the GVH process. The same lymphocyte is now recognized as the cellular keystone of all immunologic responses, and a considerable portion of that understanding has come from studies of GVHRs. Immunologists have gathered an extensive body of information bearing upon the initial events of antigen recognition, the classes of cells (lymphocytes, ancillary cells, macrophages ) involved in the primary interaction with antigen, the differentiation events that follow, and finally, the expression of immunocompetence, whether cell-mediated or antibodymediated. However, relatively little has been learned about the fate of antigen-reactive cells following their specific stimulation, differentia-

GRAFT-VERSUS-HOST REACTIONS

121

tion, and commitment. Comparatively little is known about the complex mechanisms that regulate the course of an ongoing immunologic reaction and eventually direct its resolution. As a class of immunologic reactivities, GVHRs regularly display a sequence of evolution from initiation, through differentiation and effectuation, to ultimate resolution, and thus they offer an opportunity to study at first hand the potential mechanisms of immunoregulation. And while the studies bearing on these mechanisms are recognized as being crucial to developing ideas in contemporary inimunobiology, there is in addition the persistent awareness that discoveries at this basic level might carry significant “take home” back to the clinic, where the menace of severe, lethal GVH disease as a regular complication of allogeneic bone marrow transplants is a major problem. Thus, for both the basic scientist and the clinical transplanter, GVHRs are unresolvcd enigmas, which present, collectively, a model for the study of immunoregulation for its own sake as well as for its impact in clinical medicine. This review is an attempt to probe the mechanisms of GVHRs, to codify and evaluate some of the recent studies that have utilized GVH systems to investigate pertinent problems in immunology, to consider the GVH disease associated with clinical transplantation in light of these recent studies, and finally, to suggest certain horizons that lie ahead for students of the GVHR, especially in the realni of immunoregulation. II. Graft-versus-Host Reactions (GVHRs)

A. GENERALCHARACTERISTICS Among the most widely acknowledged forms of GVHRs are runt disease, in which an immunologically immature animal is placed at risk by competent lymphoid cells from a histoincompatible adult; secondary disease, in which an animal initially protected from the lethal effects of ionizing irradiation by engraftment of hematopoietic stem cells from an allogeneic or xenogeneic donor is subsequently subjected to the lethal effects of mature immunocompetent cells; parabiosis intoxication, produced by the chronic cross-circulation of peripheral blood between two immuiiologically competent, but antigenically disparate, individuals; and F , hybrid disease, which occurs following the injection of parental strain lymphoid cells into an F, hybrid animal derived from two inbred strains which differ at the major histocompatibility locus of their species (Billingham, 1968; Elkins 1971a). Studies involving parental strain lymphoid cells injected into adult F, hybrid animals constitute some of the most elegant and far-reaching

122

STEPHEN C. GREBE AND J. WAYNE STREILEIN

work carried out in GVHRs. Perhaps the most legitimate criticism against GVH studies with newborn or irradiated animals is that these hosts do not display normal immunologic competence with respect to grafted tissues bearing nonparental transplantation antigens or to other potentially immunogenic challenges from whatever source. Thus, conclusions regarding the pathogenesis of GVH disease and the attendant lymphoid interactions and immunologic capabilities of these animals may be inappropiate when applied to normal, adult animals; the impact of such results may have only limted importance in the understanding of immune processes in general, pertaining, rather, to the phenomenon of the GVHR as an end in itself. The genetic tolerance of F, hybrids toward parental strain tissues constitutes a specific deficit, and in all other regards adult F, hybrid animals derived from inbred strains manifest normal immunocompetence. One might argue that intense scrutiny of the course of secondary disease in laboratory animals would provide a more realistic picture of clinical GVH disease in man, which may arise after therapeutic bone marrow engraftment of a patient who has received a potentially lethal dose of X-irradiation. Likewise, the recent findings of reproductive immunobiologists suggest that maternaI lymphocytes may gain access to the fetus across the placental barrier and incite demonstrable GVH disease when there is a sufficient histoincompatibility between mother and fetus, or else active maternal sensitization might prompt a vigorous and intensive reevaluation of the runt disease induced by the injection of immunocompetent lymphocytes into immunologically immature neonatal rodents (Beer and Billingham, 1971, 1973; Harrison, 1972; Morse et al., 1974). Perhaps one of the most impressive pieces of data in the studies of maternally induced GVH disease is the finding that Fischer female rats grafted with Ag-B-incompatible DA skin 27-34 days before delivery were able to incite GVH disease in their (Fischer x DA)F, offspring that resulted in 94% of the infants dying by 4 weeks after birth. It is an interesting historical postscript to the original description of “runt” disease in newborn mice (Billingham and Brent, 1959) that this GVH disease of maternal origin was procured under circumstances in which maternally induced fetal tolerance was being sought (Beer et al., 1972). The GVH disorder could be achieved in the progeny of Fischer strain mothers mated to DA fathers by the adoptive transfer of 100 million lymphoid cells obtained from Fischer donors specifically sensitized to DA alloantigens. Nearly twice that number of cells could accomplish virtually 100% lethality in (Fischer X Lewis)F, neonates whose mothers were the recipients of Fischer anti-Lewis cells. That maternal leukocytes can gain access to the fetal circulation and function as immunocompetent aggres-


123

sor cells raises the specter of maternally induced mechanisms of disease induction in humans. It has been demonstrated that reactions across even a weak histocompatibility barrier may lead to a 2- to 4-fold increase in the incidence of lymphoma in C3H(H-12) and C3H.K(H-lb) mice injected at birth with adult H-1 incompatible spleen cells (Walford, 1966). It is not unreasonable to speculate that the high incidence of lymphoma in children may have a maternal component via GVHRs (Green et al., 1960; Burnet, 1970; Armstrong et nl., 1973; Nemirovsky and Trainin, 1973). While the entire range of cell-mediated and humoral immunities against paternal antigens in multiparous females has not been surveyed, it has been shown that adoptive transfer of maternal cells into naive, syngeneic recipients equips the latter to reject paternal geneotypc skin grafts in an accelerated fashion, and that increased GVH reactivity of lymphocytes from multiparous females directed toward paternal alloantigens occurs. Thus, by these criteria, a potent, if not exaggerated state of cell-mediated immunity may be a common and regular feature of pregnancy ( Maroni and Parrott, 1973). Another system in which GVHRs have been studied in uiuo involves the use of animals rendered tolerant of allogeneic histocompatibility antigens by the injection of lymphoid cells at birth. By using lymphoid cells that do not provide runt disease, either because of their anatomic origin or genetic credentials, it is possible to render these animals subsequently nonreactive to those alien antigens (Billingham and Brent, 1959; Billingham and Silvers, 1962). The animals readily accept orthotopic skin allografts, and when these tolerant adults are then injected with specific antigen-reactive cells of the same genotype as the toleranceinducing inoculum, a fulminant GVH disease may ensue (Billingham and Silvers, 1961; Stastny et al., 1963, 1965). Although much of the work in these animals has focused on the autoimmune nature of GVHRs based on the postulate that they may represent a model for autoimmunity (Stastny et al., 1963, 1965), some aspects of these studies provide insight into the mechanisms of the immune system. Nevertheless, because of the difficulty with which tolerance is procured and the wide-ranging differences in the degree of chimerism of many of the hosts as adults, the ability to obtain reproducible levels of GVH reactivity in these tolerant recipients is limited. Therefore, as a simple expedient, F, hybrid animals not only can be used as normal, adult recipients whose immunologic behavior is not biased by the onset of immunologic maturity, as are the newborns, or prejudiced by the progressive rehabilitation of immunocompetence, as are the irradiated animals, but they can also be used as hosts in which

124


the GVH process may be consistently and reproducibly incited. It has been traditionally accepted that F, animals do not mount a specific immunologic attack on the parental strain cells, and thus these cells are provided with security of tenure. However, there are two important deviations from this "law of transplantation"-antirecognition structure and hybrid resistance, the basis of each and implications of both will be discussed later. GVH PROCESSES AND ASSAYS B. SYSTEMIC In addition to those forms of GVH disease already described, there are several other important assay systems for systemic GVHRs. The epidermoIytic syndrome ( Streilein and Billingham, 1970b,c), procured by the intracutaneous inoculation of 200 x 10" sensitized parental lymph node cells into an adult F, hybrid Syrian hamster, is a fatal disorder. The primary target is host lymphoid and hematopoietic tissue, and as an obligate consequence of leuckocyte destruction in the peripheral vasculature of the skin, toxic epidermal necrolysis ( T E N ) ensues, death intervening within 21 days (Streilein, 1971b, 1 9 7 2 ~ ) .This syndrome is well suited for the study of donor lymphocyte activation and effector mechanisms, for it is now apparent that there is a significant amount of both specific and nonspecific tissue destruction (Singh et al., 1972, 1973). Interestingly, when the same dosage of cells is delivered to the hamsters by another route, there is no subsequent epidermolysis or significant mortality ( Streilein and Billingham, 1 9 7 0 ~ )By . contrast, when adult F, hybrid rats are given GVH-inducing inocula by the intracutaneous route, epidermolysis is only an inconstant feature, but it frequently occurs in a limited manner when parental lymph node cells are administered by the intraperitoneal route (J. W. Streilein and S. C. Grebe, unpublished observations). Another system, recently described, measures the inhibition of syngeneic hematopoietic colony-forming units ( CFU ) in lethally irradiated adult F1mice given parentaI lymphoid cells in addition to 2 x 10" syngeneic bone marrow cells, which are the precursors of CFU (Blomgren and Anderson, 1972a, 1974; Boggs et al., 1973). The singular advantage of this system is the virtual elimination of the host lymphoid cell influences. As in other systems involving lethally or sublethally irradiated recipients, measurements involving organ weight, lymphoid cells numbers, or mitotic indices more nearly approximate donor lymphocyte activity alone, as opposed to a combination of donor and host contributions which are omnipresent in normal, unirradiated hosts. By means of this assay, it was possible for the authors to demonstrate specific GVH reac-

GRAFT-VERSUS-HOST

REACTIONS

125

tivity with as few as 60,000 normal lymph node cells and 125,000 normal thymocytes (fewer sensitized cells from either source were also effective). The extent of the erythroid cell growth was assessed by the amount of “Fe incoporated into the spleens of the recipients 6 days after injection of both the syngeneic bone marrow cells and the allogeneic GVH-inducing inoculum. A somewhat simpler and more direct modification of the “’Fe incorporation assay has been presented by Strong and his associates (1975). In this system, the proliferative activity of donor lymphoid cells, which were injected intravenously and harvested 4 days later from the spleens of lethally irradiated allogeneic mice, was assessed by culturing them with “H-labeled thymidine. Interestingly, a t doses of 25 to 50 x lo6 donor lymphocytes, the syngeneic control cells were nearly as active in DNA synthesis as the allogeneic cells by day 5 , suggesting that the specific immunologic proliferative response very nearly coincided in tempo with the cell division related to the putative stem cell reconstitution. Probably the most widely used GVHR assay system is the Simonsen splenic weight assay (Simonsen and Jensen, 1959) or its modification, the “discriminant spleen” assay, The constant occurrence of splenomegaly as a part of the GVH syndrome in chickens, mice, rats, and other species studied ( Simonsen, 1962a) prompted its codification as a reproducible sign of the disease. In general, splenic hypertrophy of neonatal recipients reaches a peak around 8-10 days after donor cell inoculation (Simonsen and Jensen, 1959) ; by 20-30 days after parental cell inoculation, however, the host spleen may be either atrophic or enlarged (Howard, 1961). As a formal definition, the “splenic index” is determined by the quotient of the experimental spleen weight divided by the host’s body weight over the control spleen weight (from an uninjected or syngeneic cellinjected littermate) divided by its body weight. Generally, a splenic index greater than 1.30 indicates a significant degree of GVH reactivity. It is necessary to determine spIenic indices prior to the onset of significant weight loss, which is frequently associated with GVH disease, At this point, it is appropriate to mention that frequently there are discrepancies between various GVH assays, even within a given experimental system. For example, it may be difficult to correlate splenomegaly with mortality, perhaps because spleiiomegaly may represent a number of different responses, not only immunologic (specific and nonspecific), but also hematologic. Likewise, the amount of ”-labeled thymidine incorporated by donor cells in response to allogeneic antigens (Gershon et al., 1972) may not accord strictly with mortality of the hosts. Similarly, Nisbet

126


and Simonsen ( 1967) have demonstrated a discrepancy between the proliferation of donor lymphocytes and the capacity of those cells to induce splenomegaly. Nevertheless, the spleen weight assay is valid since it is genetically specific, shows a logarithmically linear relationship to the dosage of lymphocytes injected, and is demonstrable only when the grafted cells are derived from an immunologically competent organ ( Simonsen, 1962a; Billingham, 1968). Generalized lymphoid hypertrophy has also been observed and seems to represent the nonspecific accumulation of host lymphocytes within the sites of GVH involvement (Fox, 1962; Billingham, 1968). As the disease progresses, there is a marked atrophy of the thymus, even during the phase of generalized lymphadenopathy, terminating in a widespread lymphoid atrophy, which encompasses not only lymph nodes and spleen, but myeloid tissues as well. Simonsen ( 1962a), McBride ( 1966), Billingham ( 1968), and Elkins (1971a) have all reviewed extensively other systemic GVH assays, and it is sufficient simply to acknowledge them here in summary form: 1. Phagocytic Index Assay Howard (1961) showed that, 14-15 days after the injection of nearly 100 x lo6 parental lymphoid cells into an F, hybrid, the recipients could clear intravenously injected colloidal carbon with increased vigor, as high as 10 times normal efficiency. This assay complied with all the necessary prerequisities for a GVHR and was accompanied by splenomegaly, hepatomegaly, and often, lethal “runting.” 2. Focal Periportal lnfiltration In F, hybrids that were injected intravenously with parental lymphocytes, microscopic foci of cellular infiltration could be seen around branches of the portal vein in the liver (Miller et al., 19e3). Miller et uZ. showed that the numbers of foci were related directly to the dose of donor lymphocytes injected and depended upon thymic integrity and radiosensitive cells in the donor. The assay has been used in a number of studies (Gorer and Boyse, 1959; Bain and Alton, 1964; Uyeki and Palotay, 1965; Levine, 196th). 3. Chorioallantoic Membrane Assay

For details of this assay, see Burnet and Boyer (1960), Burnet and Burnet (1960), Szenberg and Warner ( 1962), Warner and Szenberg (19641, Killby et d. (1972), Longenecker et al. (1973), and Walker et al. (1973).

GRAFT-VERSUS-HOST

HEACTIONS

127

In 1916, Murphy described a very simple form of the GVHR, although he was unable to recognize its immunologic basis and, indeed, could not offer a satisfactory explanation for it. He layered spleen and bone marrow fragments onto the chorioallantoic membrane (CAM) of a chick embryo and observed the resulting changes. Not only did splenic enlargement occur, but the tissue fragments themselves developed into whitish nodules ( Billingham, 1968), which proved to be primarily comprised of leukocytes. Over the years, the etiology of this phenomenon has been disputed, but the reaction can be shown to be immunologic ( a GVH process), and the “pocks” that appear on the CAM represent individual foci of clonally expanded antigen-reactive cells. In fact, Burnet and Burnet ( 1960) used serial dilutions of adult chicken blood leukocytes to demonstrate the limited specificity (or “unipotentiality”) of antigenreactive cells; when 5 x 10‘ leukocytes were administered, only 10-100 pocks appeared, indicating the restricted immunologic specificity of the immunocompetent cells. 4. Splenic Explant Assay

When parental lymphoid cells are layered in single cell suspension over small “diced” explants of F, spleens, then one may establish a semiquantitative GVH assay 3-5 days later, by visually assessing an increase in the physical mass of the explanted tissue (Globerson and Auerbach, 1967). A more refined version of this assay has been developed in which the incorporation of 3H-labeled leucine into the stimulated spleen fragments reflects quantitatively GVH reactivity of the parental cells ( Auerbach and Shalaby, 1973). Finally, a variety of xenogeneic GVHRs O C C I I ~ , some locally (see Section II,C), but all possible predictors of clinical GVH disease in human bone marrow transplant patients. These “human-into-xenogeneic-host” GVHRs are less vigorous than their allogeneic counterparts (Asantila et al., 1974), and it is difficult to provide adequate controls. However, injection of human lymphocytes into mouse skin (Rees and Symes, 1973) or under kidney capsules of rats (Asantila et al., 1973; Shohat et al., 1974; Mottironi and Meuwissen, 1973) does lead to measurable GVH lesions. C. LOCALGVHRs

AND

ASSAYS

1. lntracutaneous GVHRs

Brent and Medawar (1966a,b) have executed an extensive series of studies on the transfer reaction in guinea pigs, which, in hosts unable to react against the donor inoculum, is a typical GVHR. Whether the

128


donor lymphocytes are derived from unsensitized animals (normal lymphocyte transfer reaction) or from specifically sensitized guinea pigs (immune lymphocyte transfer reaction), the time course for this delayed hypersensitivity reaction is essentially the same and is triphasic: ( a ) the inflammatory episode occurs during the first 3 days after injection of the skin and consists primarily of the confrontation of alien transplantation antigens and donor lymphocytes; ( b ) the flare-up phase occurs over the next 3 days and is marked by the process of extensive cell division, during which there is an empirical increase in the size of the skin reaction; ( c ) the fade-out stage occurs after 6 days, during which the palpable skin reactions subside, their mean diameter is reduced, and the host recovers. Other investigators have studied these delayed, cutaneous reactions in mice (Dekaris and Allegretti, 1968; Streilein and Billingham, 1970a; Streilein et al., 1970), rats (Streilein and Billingham, 1967), hamsters ( Ramseier and Streilein, 1965; Ramseier and Billingham, 1966), guinea pigs (Zakarian and Billingham, 1972), and dogs ( Streilein and Barker, 1967). Perhaps one of the most versatile of these systems is the irradiatedhamster test. By virtue of their thin, easily mobilized skin and vigorous, radioresistant inflammatory responses, hamsters given 1500 R to suppress specific rejection of allogeneic or xenogeneic grafts may then be injected with equivalent numbers of histoincompatible lymphocytes from a variety of rodent sources, or from other species, and monitored for the next 24-96 hours for manifestations of graded skin reactions. With this system, it has been possible to demonstrate specific allograft reactivity, even against H-Y antigens present on C57 male mouse lymphocytes ( Ramseier and Streilein, 1965)-a specificity which traditionally has been thought too weak to elicit an overt GVHR except under conditions of male/ female parabiosis ( Billingham, 1968), or overlying non-H-2 incompatibility ( Cornelius and Aponte, 1974). Two studies have focused attention on the intracutaneous GVHR in F, hybrid rats. In the first (Ford, 1967), peak intensity of cutaneous lesions was observed 5 days after the inoculation of 2.5 to 20 x lo6 thoracic duct lymphocytes into Ag-B incompatible F, rats. The specificity of the phenomenon was demonstrated in three ways: (1) syngeneic cells were unable to elicit any response after 2 days, and the initial response was modest; ( 2 ) cells from specifically tolerant donors were unable to cause skin reactions; and ( 3 ) presensitization of the donors against the recipient by a single skin graft increased the reactive capacity ol' the cells 1.5- to 2-fold when the receipient was Ag-B incompatible and more than 4-fold when the F, was compatible at the major histocom-

GRAFT-VERSUS-HOST

REACTIONS

129

patibility locus. Although there was histologic evidence of “blast transformation” within the lesion, there was relatively little tissue destruction, compared with either intracutaiieously induced GVHR in hamsters ( Streilein and Billingham, 1970b,c) or GVHR attendant to intraperitoneally or intravenously supplied parental lymphoid cells ( Oliiier et d., 1961; Gorer and Boyse, 1959; Stastny et ul., 1963; Billingham, 1968). The other study ( Streilein and Billingham, 1967), while principally concerned with the manifestations of cutaneous hypersensitivity in rats, not only confirmed the findings of others that presensitization of the donor increased the intensity of the skin reactions, and that major histocompatibility differences were essential, but offered some substantive evidence that isoantiseruni directed against the host could intensify the GVH reactivity of the donor cells. Another noteworthy feature of these experiments is the requirement for relatively high numbers (40 x lo6) of donor lymph node cells to incite even moderately impressive skin reactions, even when the donors me sensitized parental strain animals.

2. lntraocular GVHRs Recently it has been shown that the injection of parental lymph node cells into the anterior chambers of F, hybrid rat eyes produces a progressive, diffuse lesion, comprised principally of leukocytes (Kaplan et al., 1975a,b). The time course of the reaction is more intense and prolonged than the same reaction in allogeneic hosts, presumably because of some inhibitory facility of the allogeneic host, which suppresses the GVH reactivity by means of a concomitant “host-versus-graft” ( HVG ) reaction. Corroboration of this assertion stems from the evidence that specific presensitization of the allogeneic host before introduction of the lymphocytes into the anterior chamber further curtails the local inflammatory reaction. In order to place this phenomenon in perspective and to provide some insight into the “discriminant privilege” of the anterior chamber, it would be useful to know more about the histologic and genetic background of the cells within the lesion. 3. liitrcirenal GVHRs

Very thorough descriptions of local GVH reactivity within the renal parenchyma of F, hybrid rats have been published by Elkins (1964, 1966, 1970, 1971b; Clancy et al., 1973) aiid Volknian (1972). This “invasive-destructive” reaction follows the introduction of Ag-B incompatible lymphoid cclls under the renal capsule aiid is readily assayed by comparing the increased weight of the inoculated kidney (peaking on day 14) to that of the uninjected contralateral kidney. This weight gain

130


is preceded by a burst of donor lymphocyte mitotic activity, maximal by day 6 (Elkins, 1970), and results in histologically distinct tissue damage similar to that seen during rejection of a renal allograft (Elkins, 1964). Host lymphocytes do not play an active role in the development of the lesion, either through trapping or proliferation; nevertheless, they are the primary immunogenic stimulus, as was demonstrated by the use of chimeric animals and F,s into which parental strain kidneys were grafted. In such instances, when the renal transplants were injected with cells syngeneic with the kidney at risk, these kidneys developed lesions typical of a GVHR ( Elkins and Guttmann, 1968; Elkins, 1971b).

4. 1.n Vitro Correlates of GVHRs Before turning to the popIitea1 lymph node assay, another “local” GVHR, two noteworthy in vitro analogs should be mentioned. The first, the mixed leukocyte reaction (MLR), represents the quintessential event of antigen recognition in the GVHR: the confrontation of immunocompetent donor lyniphocytes by allogeneic, lymphocyte-borne histocompatibility antigens ( in the presence of macrophages ) and the subsequent proliferation of these specifically reactive thymus-dependent cells ( T cells) (Rodey et al., 1974). Although the MLR has been applied to the study of lymphocytes from numerous species, perhaps the most exhaustive work has been carried out by using rat lymphocytes (Wilson, 1967; Wilson et al., 1967, 1968; Wilson and Nowell, 1970, 1971; Wilson and Fox, 1971). The most salient findings from this series of experiments are as follows: a. Proliferation peaks between 4 and 6 days of culture. b. Disparity at the major histocompatibility locus is essential. c. When parental and F1 cells are mixed, only the parental cells proliferate, as measured by chromosomal markers. d. Consistent with in vivo experiments, lymphocytes from specifically tolerant and neonatally thymectomized parental donors do not proliferate in response to F, antigens; on the other hand, cells from specifically presensitized donors display a truncated proliferative pattern, suggesting that perhaps some of the target F, lymphocytes may be destroyed, causing a reduction of the immunogenic stimulus. e. Kinetic studies revealed that, after a lag period of nearly 2 days, dividing cells start exponential proliferative expansion, which lasts for approximately 4 days. f. As many as 1 3 %of peripheral blood lymphocytes enter the proliferative process in the MLR, representing cells that enter the mitotic cycle for the first time, not cells that are the progeny of lymphocytes that have already divided in culture.

GRAFT-VERSUS-HOST

REACTIONS

131

g. The high proportion ( 1 3 %of ) normally circulating peripheral blood lymphocytes that proliferate in the MLR in response to a particular set of alloantigens is a true reflection of the actual number of antigenreactive cells; it does not represent nonspecific activation or recruitment of other parental lymphocytes. Furthermore, this “alloantigenic response” is greater than a “xenoantigenic response” when normal rat lymphocytes are stimulated by either allogeneic rat or human lymphocytes, respectively. h. Consistent with Simonsen’s proposal that the numbers of lymphocytes responsive to alloantigens are not substantially expanded by specific sensitization (the “factor of immunization”-Nisbet and Simonsen, 1967; Simonsen, 1970; Ford and Simonsen, 1971), the kinetics of the MLR revealed that, in contrast to other kinds of antigens, reactivity against major histocompatibility alloantigens invoIves a large number of existing cells capable of specific, antigen-dependent proliferation. Nevertheless, lymphocytes from specifically immunized animals do respond more rapidly than normal lymphocytes (peak of proliferation at 3 days vs 5 days). i. Finally, in experiments involving germfree and conventional rats, it was shown that the reactivity to xenogeneic antigens may be due to cross-reactivity with environmental determinants ( e.g., bacterial flora), but that the impressive response of specific lymphocytes to major alloantigens is not: peripheral blood lymphocytes from both germfree and conventional rats respond equally well to Ag-B incompatible rat lymphocytes. If the MLR reflects the afferent limb and central processing of the in vivo GVHR, then cell-mediated cytotoxicity ( C M C ) provides a controlled model for studying the cfferent, or effector phase, of the allograft reaction (Cerottini et al., 1971; Cerottini and Brunner, 1974). This system represents a true cell-mediated immunity in that it is complement-independent and there must be thetn-positive, thymus-dependent lymphocytes present within the killer cell population (Lonai et al., 1971). Furthermore, there is now evidence that this form of in vitro cytotoxicity can be wholly mediated by a T-cell population depleted of bone marrowdependent cells ( B cells) (Golstein et nl., 1972; Golstein and Blomgren, 1973). Although the in vitro killing of allogeneic tumor target cells (Cerottini et al., 1971; Singh et al., 1971; Tigelaar and Gorczynski, 1974) has been widely demonstrated, the use of lymphocytes as targets may well be the most rcpresentative test system in tissue culture cytotoxicity assays, since these cells are most frequently encountered by donor lymphocytes within the animal suffering from GVH disease, and thus are the prime targets of “killer” cells (Streilein, 1971a).

132


A number of alloantibody-mediated cytotoxicity assays have been developed, and at this time cell-bound antibodies cannot be excluded as a major spearhead in the GVH attack. There are, however, many unresolved questions about the pathogenic role of antibodies in the GVH process ( Billingham, 1968; Elkins, 1971a); therefore antibody-mediated cytotoxicity probably does not represent a very realistic in vitro analog of the typical GVHR-essentially a manifestation of cell-mediated immunity.

5. The Popliteal Lymph Node ( P L N ) Assay Although many investigators have documented the generalized adenopathy of lymph nodes that accompanies the splenomegaly and thymic involution in the course of GVHRs (Simonsen, 1962a; Billingham, 1968), there has only recently emerged a method for quantifying GVHRs by the degree of lymph node enlargement, in a manner similar to that of splenomegaly. Levine ( 1968b ) described the immunologically specific hypertrophy of the PLN in a (Lewis x BN) F, hybrid rat after the injection of Lewis spleen cells directly into either the PLN or the hind foot pad. Although the details in this report are only sketchy, and some of the techniques seem unacceptably archaic ( n o quantitation of the donor spleen cells was made), nevertheless there are a number of important and provocative results: a. Within 7 days after foot-pad injection, some of the draining PLNs had increased in mass as much as 50-fold over the original 5-6-mg lymph node. A comparable amount of splenic hypertrophy in a rat 6-8 weeks old, assuming that the spleen contained almost exclusively white pulp and was devoid of red pulp, would produce a spleen weighing 20,000 mg (20 gm), or 7.5%of the body weight of the animal. b. AIthough there was a modest increase in spleen weights of these Fls, the similar “splenomegaly” in syngeneic recipients suggested a nonspecific effect, generally substantiating the claim that this reaction was a local phenomenon. c. When Lewis spleen cells were injected into allogeneic BN recipients, the PLN hypertrophy was about one-third of that demonstrable in (Lewis x BN) F,s. The impact of this work upon the students of GVH phenomena was apparently not immediate, but two years after its publication, two groups independently formalized Levine’s technique into what has now become known as the “PLN assay.” Ford et al. (1970) have provided the most elegant and detailed description of the assay, using it, like the Simonsen splenic weight assay (Simonsen, 1962a), in order to relate the degree of PLN hypertrophy


133

to the dose of lymphocytes injected, expressed linearly on a double log scale. To compare the GVH-inducing capacity of normal thoracic duct lymphocytes from rats of Ag-B compatible or incompatible backgrounds with respect to the F, hybrid recipients, they constructed doseresponse curves for each parental/F, hybrid system. It was then possible to measure relative potency by comparing mathematically the number of thoracic duct lymphocytes from different sources necessary to raise the weight of the PLN to 10 mg. In all cases, nearly 100 times the number of cells was required in the Ag-B compatible donor/ host combination to produce a GVHR equivalent to that in Ag-B incompatible rats. The authors were able to show that not only do more cells reach the PLN when injected by the subcutaneous route as opposed to the intracutaiieous route of foot pad inoculation, but also the subcutaneous route resulted in a more vigorous PLN reaction at each dose tested, and there was a lower variance associated with these mean values. Finally, a quantitative comparison of the PLN assay with Elkins’ renal weight assay system (Elkins, 1964) revealed, in their hands, that the PLN weight assay was nearly 10 times more sensitive in detecting GVH reactivity. At about the same time, Ha3kovA and GansovA (1970, 1972), injecting lymph node cells from adult inbred rats into Ag-B incompatible F, recipients, also showed a peak of PLN hypertrophy at 7 days and verified the specific GVHR basis of the hyperplasia, since irradiated parental cells were unable to incite any enlargement. They expressed PLN weight increases in several ways, so that it is difficult to compare the results of their different experiments, but the magnitude of the reactions seems to be roughly equivalent to that reported by Ford et al. (1970). In the past 4 years, many experiments involving the PLN assay have been published, and they may be classified in one of two categories: (1) those that purport to explain the mechanism of expression of this phenomenon, and ( 2 ) those that utilize the assay to study some more general aspect of GVH reactivity or transplantation immunology, primarily with regard to the immunocompetence of donor lymphocytes. Within the first class, the assay has been extended to mice (Hardt and Claesson, 1971) in which it has been shown that ( 1 ) the majority of the lymphoid proliferation in the PLN at its peak of enlargement takes place in the cells of host origin, not in the donor cells (Moniit and Everett, 1974); ( 2 ) PLN h) pertrophy in the mouse is not as vigorous as in the rat, but that HVG reactivity, after injection of F, spleen cells into parental hind foot pads, produces as much PLN hypertrophy as GVH, although the maximum falls earlier ( 4 days) than the GVHR ( 7 days) (Twist and Barnes, 1973); and ( 3 ) both GVH and HVG responses leading to significant PLN hypertrophy may occur in H-2

134

STEPHEN C. GFlEBE AND J . WAYNE STREILEIN

compatible mice, providing there is a difference at the M locus (Salaman et al., 1973). Work from three other groups has provided considerable insight into the mechanism underlying the GVH expansion of the PLN, especially with regard to the participation of the genetically tolerant host lymphocytes. One study has shown that radiosensitive host lymphocytes are necessary for the full expression of PLN hypertrophy and that this host contribution is immunologically nonspecific ( Bonney and Feldbush, 1973); another supports the idea that circulating F, host Iymphocytes are trapped within the GVH-involved PLN and may contribute in some way to the hyperplasia of that node (Emeson and Thursh, 1973). Studies conducted in our laboratory have highlighted the fact that the PLN reaction in rats, traditionally regarded as a local response, is indeed a systemic response in many aspects. Not only are host radiosensitive lymphoid cells brought to the PLN by the peripheral blood and trapped there as a consequence of the local GVHR, but even relatively small numbers of parental lymph node cells injected into the hind foot pads of adult F, hybrid rats become systemically disseminated, setting up multiple, local, metastatic GVHRs ( Grebe and Streilein, 197513). The development of local PLN hypertrophy is heavily dependent on host radiosensitive lymphoid cells (Grebe and Streilein, 1975a), half of which are of systemic as opposed to indigenous, or local, origin (Grebe and Streilein, 1975b). Radiosensitive host cells constitute the major cellular bulk of the PLN lesion but are not significant as a source of immunogenic stimulus. Rather, host radioresistant cells appear to be the primary source of alloantigenic stimulation ( Grebe and Streilein, 1975a). The specificity of the impressive splenomegaly and brachial lymph node hypertrophy, which are systemic manifestations of the locally induced GVHR, was verified by the presence within those host organs of "Cr-labeled, donorgenotype histocompatibility antigen-reactive cells ( H-ARCS) which could provoke an additional GVHR when tested by adoptive transfer to a secondary F, receipient, syngeneic with the first ('Grebe and Streilein, 1975b). Studies involving the immunoregulatory control of both local and systemic GVHRs, which can be performed both simultaneously and in tandem in this system, suggested that there may be a common mechanism underlying the alteration or modification of all ongoing GVHRs. The significance of this common denominator to the study of immunoregulation in general will be discussed in Section VII. By manipulation of donor and host with respect to parental or F, genotype, it is possible to detect important differences between GVH and HVG phenomenon. When F, lymphoid cells are injected into parental rats (Meyer and Heron, 1973; Dorsch and Roser, 1974; Grebe and


135

Streilein, 1975a) or mice (Twist and Barnes, 1973; M o d and Everett, 1974), PLN hypertrophy as an expression of HVG reactions reaches a peak by day 4-three days before the GVH peak, This “pure” HVG response may also be readily quantified, as a monitor of allograft transplantation immunity ( Meyer and Heron, 1973; Dorsch and Roser, 1974). The injection of untreated lymphoid cells into the foot pad of an allogeneic recipient results in a combined GVH and HVG responses, producing significantly less PLN enlargement than the “one-way” GVHR (Meyer and Heron, 1973, Grebe and Streilein, 1975a). However, when the allogeneic recipient is given sublethal doses of whole-body irradiation 24 hours before donor lymph node cell injection, there is apparently a suppression of the HVG reaction and a concomitant “derepression” of the GVHR (Grebe and Streilein, 1975b). This model, by using sublethally irradiated allogeneic recipients, provides a novel assay for GVH reactivity which does not require the use of F, hybrid recipients. The majority of reports citing the PLN assay have simply considered it a direct, isolated, in vivo means of quantifying the GVH reactivity of donor lymph node cells, as an updated version of the Simonsen spleen weight assay. Thus, Ford and Simonsen (1971) have repeated some of Simonsen’s earlier work on the “factor of immunization” ( Simonsen, 1962b, 1970), which postulates that presensitization of imniunocompetent cells against a weak histocompatibility antigen produces increased GVH reactivity ( b y clonal expansion of the T-cell population), while prior immunization of those lymphocytes that are reactive to strong histocompatibility specificities does not augment their GVH reactivity. Similarly, the PLN assay has been used to catalog the relative GVH-inducing capacity of lymphoid cells from various organs, such as bone marrow, marrow fractions, thymus, spleen, mesenteric lymph node, and peripheral blood (Yoshida and Osmond, 1971). Several other groups have sought some insight into the mechanism of tolerance and enhancement maintenance by critical use of the PLN assay (Bilds@eet al., 1971; Atkins and Ford, 1972; Heron, 1973). A major controversy in transplantation immunobiology today concerns the role of blocking factors and/or enhancing antibodies in classical neonatally induced transplantation tolerance. The results of recent GVHR studies bearing on this issue have consistently failed to demonstrate any “blocking factors” in the sera of highly tolerant, fully chimeric mice, after their neonatal injection with semiallogeiieic lymphoid cells, thus confirming the parallel investigations by Brent and his colleagues in similar assay systems (Brent et al., 1972; Beverly et al., 1973). Lymphocytes from highly tolerant rodents are incapable of inducing GVHRs (Ford and Atkins, 1971; Bilds@e et al., 1971), MLRs (Wilson et al.,

136


1967), or both (Heron, 1973; Brent et al., 1972; Beverly et ai., 1973). This state of nonresponsiveness is apparently not due to the presence of serum blocking factors, which could be found in some, but not all, partially tolerant animals ( Brent et al., 1972). Collectively, these studies on tolerance stand in conflict with the results of studies by the Hellstroms, that purport to show blocking factors in the sera of tolerant animals (Bansal et al., 1973; Hellstrom e t al., 1973). Furthermore, neither cells nor sera from tolerant rats could suppress PLN hypertrophy in F, rats induced by normal cells (Ford and Atkins, 1971). Abolition of tolerance by adoptive transfer leaves cells that can suppress GVH reactivity of cells from normal donors (Elkins, 1972). At the present, the conflict remains unresolved but is clearly not simply a semantic issue. An important, but perplexing, contribution to the investigations on neonatally induced tolerance has been advanced through a series of studies on the nature of “self-tolerance.” When rat lymphocytes were cultured for several hours on a monolayer of syngeneic fibroblasts, a significant proportion (44%) adhered to the monolayer (Lonai et al., 1973a). These adhering lymphocytes apparently shed their tolerance to self-antigens and in fact became specifically “sensitized toward them. They could kill syngeneic target cells in a CMC assay (Cohen and Wekerle, 1973) and cause PLN hypertrophy after foot-pad injection into syngeneic recipients (Lonai et al., 1973a). When the lymphocytes, but not the fibroblast monolayer, were pretreated with syngeneic normal mouse serum, the “self-recognition” adherence phenomenon was inhibited, suggesting that tolerance to self-antigens might be maintained by a serum factor directed against a T cell surface antigen or receptor ( Cohen and Wekerle, 1973).However, a series of experiments comparing the PLN reaction of autosensitized T lymphocytes and a true GVHR revealed an interesting difference. In the syngeneic ( “autosensitized T cell”) system, PLN enlargement was shown to be dependent on host radiosensitive T cells, but not dependent upon the proliferation of donor cells. This curious finding presents a picture consistent with a typical HVG response, in contradistinction to the true GVHR (Cohen, 1973). Taken at face value, in the absence of the CMC data, these results suggest that “autosensitized” adhering lymphocytes when cultured with syngeneic fibroblasts might acquire a novel antigenic specificity and that the reaction in the PLN of the syngeneic host is a response to these evocative plasma membrane determinants. While this view does not take into account all the data, it poses a real question about the potential changes that might occur to lymphoid cells during in vitro culture. The convenience of the PLN assay has opened the way for several investigators to determine the efficacy of immunosuppressive agents.

GRAFT-VERSUS-HOST

REACTIONS

137

A more detailed examination of immunosupprcssion in GVHRs is contained in Sections VI and VII. Pretreatment of donor lymphocytes with a “thymic chalone” depresses the PLN reaction, as assesed by the reduced uptake of tritiated thymidine in the involved node (Kiger et nl., 1973). From a massive study on the effect$ of nearly 30 different pharmacologic agents, it was concluded that antiproliferative drugs reduced PLN hypertrophy by killing dividing cells when given to the recipient. They had little cffect, however, when administcred to the donor in uiuo (Beck et nl., 1973). Rabbit anti-rat lymphocyte serum ( Ha6kovli et al., 1973a) and pig anti-rat lymphocyte serum (Ha4kovi et al., 1973b) decreased the GVH-inducing capacity of parental peripheral blood leukocytes when evaluated in the PLN assay. The abrogation was most effective if the antilymphocyte serum ( ALS ) treatment was coordinated with specific sensitization of the putative donor cells to alloantigens by skin graft ( Hagkov6 et al., 197313) . The versatility of the PLN assay is amply demonstrated by the wide variety of experimental uses that have been reported. These studies are a true potpourri, having little in common with each other. Guinea pigs, which display unusually vigorous delayed hypersensitivity reactions, have been used in the PLN assay. Lymphocytes from the spleen and Peyer’s patches have been shown to elicit significant PLN hypertrophy in the appropriate F, hybrid recipients (Levin et ul., 1974). T lymphocytes in animals which have suffered 18-23% burns have increased GVH reactivity as assessed by the PLN (Munster and Gressitt, 1973). In order to evaluate the impact of environmental antigens, such as conventional bacterial flora, Nielsen (1972) measured the GVH reactivity of lymphocytes from germfree and conventional rats in the PLN assay and found no significant difference. Suspensions of lymphocytes from multiparous female mice evoked significantly higher PLN responses than normal lymphocytes, if the mothers had borne 3-5 litters (Maroni and Parrott, 1973). Finally, although it did not involve the PLN assay, a study of brachial lymph node enlargement in F, hybrid mice, grafted in the lateral thoracic flank with parental skin, showed that full-thickness skin is an immunologically competent organ, since it provoked greater lymph node hypertrophy than syngeneic F, skin (Barker and Billingham, 1972).

D. GVH DISEASEIN MAN AND OTHERPRIMATES: LESSONSFROM THE CLINIC For the clinician, all the animal studies relevant to GVH phenomena have led ineluctably to a concern for the prominent occurrence of GVH disease in patients given allogeneic bone marrow cells to reconstitute

138


their hematopoietic deficit, which may be due either to aplastic anemia (Storb et al., 1975) or to lethal whole-body irradiation in the treatment for malignant neoplasia of hematogenous origin ( Bortin, 1970). Similarly, engraftment of bone marrow into patients with immunodeficiency disorders frequently leads to GVH disease ( DeKoning et al., 1969; Amniann et d., 1970; Billingham and Streilein, 1971; Buckley et aZ., 1971a,b; Jose et al., 1971; Levey, 1972; Streilein, 197213; Stiehm et al., 1972; Huang et d., 1973). The protocol for such bone marrow transplantation procedures has been outlined by Thomas and Storb (1970), and many of the conditions for which this therapy is normally indicated have been reviewed recently by Thomas (1974) and Math6 et al. (1974). As a consequence of the increasing frequency of clinical bone marrow transplants and the experience gained from treating these patients, there have been some unexpected insights into the pathogenic mechanisms of GVHRs. Perhaps the most sobering aspect of these engraftments has been tho prevalence of moderate-to-severe GVHRs, which arise in recipients despite “full-house” HL-A identify and MLR negativity of the matched sibling bone marrow donors (Meuwissen et al., 1971; Graw et al., 1971, 1972; Glucksberg et al.. 1974; Speck et al., 1971, 1973). Discordant data from two laboratories using the canine model have suggested that normal DL-A matched donor and recipient dogs either are likely (Storb et al., 1973) or are not likely (Rapaport et al., 1972) to suffer symptoms of GVH disease after immunosuppression and engraftment of the allogeneic marrow. This discrepancy may be attributable to differences in genetic backgrounds of the respective dog colonies. Similarly, RhL-A matched rhesus monkeys, given 900 rads of total body X-irradiation and ALS prior to bone marrow grafting, all died (median survival time = 35 days), but at least one RhL-A mismatch resulted in a more severe disease and a survival time of only 14 days (Neefe et al., 1973). These clinical findings that phenotypic identification of components of the human major histocompatibility complex ( M H C ) does not preclude the development of severe GVH disease strongly suggest either of two possibilities: (1) as yet unknown factors dictated by other genetic loci within this chromosomal region are operative in GVH induction and/ or ( 2 ) disparity at non-MHC determined histocompatibility loci is able in its own right to evoke GVH disease in the appropriate clinical and experimental setup. Klein’s experiments in congenic strains of mice have demonstrated the presence of a “GVH locus” within the H-2 complex, which is capable of stimulating a GVHR in congenic mice; similarly, the work of Salaman et nl. (1973) has shown that M locus disparate lymphocytes may provoke a GVHR in H-2 identical mice. There is ample experimental evidence gathered in the past 5 years that weak immuno-


139

genetic disparity (i.e., non-MHC) is sufficient to induce a GVHR if the recipient is particularly vulnerable, i.e., non-MHC ) is sufficient to induce a GVHR if the recipient is particularly vulnerable, i.e., in an adult after X-irradiation or a fetus in utero (Beer and Billingham, 1973). I n a series of bone marrow transplants in monkeys, which were performed without cognizance of RhL-A compatibilities, Merritt et al. (1972) provided the basis of the previously cited work and were able to show significantly prolonged survivals of irradiated, reconstituted recipients given rabbit anti-rhesus thoracic duct lymphocyte serum, prior to irradiation and grafting. Nevertheless, the animals receiving ALS lost weight and developed the classical TEN (positive Nikolsky sign) (Streilein and Billingham, 1970b) and gastrointestinal mononuclear cell infiltrates. An interesting prospect that has developed from the studies in monkeys is the demonstration that cyclophosphamide treatment and infusion of frozen and stored autologous bone marrow given 3 and 4 days, respectively, after allogeneic bone marrow into irradiated rhesus recipients were sufficient to rescue 5 out of 8 of them from “fatal” GVH disease-5 additional animals died of aplasia (Merritt et al., 1973). While some of the elements of peripheral blood recovered less promptly than those in monkeys given fresh bone marrow, the hematopoietic reconstitution was impressive enough to warrant possible clinical consideration in certain highly selected situations. Although the target of immunologic attack in GVH disease is principally host lymphoid cells ( Streilein, 1971a, 1 9 7 2 ~ clinically )~ procured GVHRs regularly elicit destructive reactions in a wide variety of tissues (Slavin and Santos, 1973). Severe GVHRs inevitably involve damage to the spleen and liver and also to the gastrointestinal tract, with such manifestations as anorexia and diarrhea ( Glucksberg et al., 1974). Cutaneous lesions constitute one of the most impressive features of systemic GVH disease in humans and serve as a qualitative, if not quantitative, gauge of the severity of the response. Skin involvement ranges from erythrodermia to TEN. Indeed, the survival of a patient who has received an allogeneic bone marrow transplant may often be predicted during the first few weeks after engraftment by an estimate of the clinical severity of the skin lesions, with the more extensively involved skin lesions serving to predict severe problems for the patient later (Glucks. berg et al., 1974). The underlying etiology of TEN is a somewhat controversial issue. While some investigators claim a direct causal relationship between the GVHR and the epidermal injury (Boranik et al., 1972), others insist that the TEN either is drug induced or is a direct toxic expression of a disseminated, permissive Staphylococcus aureus infection ( Kruegar,

140

STEPHEN C. GREBE AND J. WAYNE STREZLEIN

1973; Peck et al., 1973). However, in one instance of a GVH-induced TEN syndrome, the investigators were unable to demonstrate the presence of staphylococcal infection (Amman et al., 1972). Although it has not been widely investigated, one report (Merritt et al., 1971) has documented the existence of a toxic factor in the serum of a patient with GVH disease. This factor was cytotoxic for epidermal cells not only from the patient, but for other cells of a different genotype. It appeared to have a broad specificity for epithelial cells in general. Despite the fact that histocompatibility alloantigens are represented on the surface of virtually all cells in the body, not all cells are at equal risk to the attack of GVHR. Lymphatic, hematopoietic, and reticuloeiidothelial cells are almost always involved, and skin, gastrointestinal tract lining epithelium, and liver are frequently at risk. Muscle, bone, endocrine gland, and cells of the nervous system are almost never involved. Rased on this differential susceptibility to GVH attack, it has been postulated that the primary attack is against cells of the lymphoid system with such secondary tissues as the skin and gastrointestinal tract suffering second-hand as a reflection of innocent bystander killing (Streilein, 1971a, 1 9 7 2 ~ ) .GVHR-induced consumptive coagulopathy may be an important pathogenic mechanism ( Streilein and Tomar, 1972). The severity of the GVH disease has become a cardinal indicator of the patient’s prognosis. A scoring system based on the involvement of the skin, liver, and gut is used by at least one center to predict the severity of GVH disease (Glucksberg et al., 1974). In a survey of 61 marrow graft recipients at their center, 43 developed clinically recognizable GVH disease. The follow-up period ranged from several months to 5 years. In the patients without GVH disease or with only skin involvement, over haIf were alive at the termination of the study. Alternatively, only 13%of patients with severe GVH disease survived that same period. Likewise, 4.5 times more patients with severe GVH disease succumbed to infection than those with little or no GVH disease. The management of GVH disease in most patients is carried out with the aid of such immunosuppressive agents as ALS, methotrexate, cyclophosphamide, cytosine arabinoside, and procarbazine (Gray et al., 1968; Gray, 1973; Thomas, 1974; Math6 et al., 1974; Storb et al., 1970a,b). While often very effective in controlling the incipient GVH process, they may act like a two-edged sword. The relative nonspecificity with which these agents strike may very well abrogate the patient’s capacity to stem the tide of any nascent infections. Most prominent among the infections seen in thcse patient3 are interstitial pneumonia, chicfly caused by cytoniegalovirus (CMV) or Pneumocystis (Graw et al., 1972; Neiman et al., 1973; Glucksberg et al., 1974). It was noted that patients who

GRAFT-VERSUS-HOST

REACTIONS

141

died from interstitial pneumonia did not show seroconversion of complement-fixing anti-CMV antibodies or significant increase in antibody titer, whereas those patients free of clinical manifestations of CMV infection displayed significant titers (Neiman et al., 1973). Although it is likely that agents other than CMV may be responsible for the interstitial pneumonia, nevertheless this specific immunologic defect in large measure reflects the overall depression of immunocompetence or failure to recover immunoconipetency by these patients (see Section V,C). In the case of leukemic patients who received bone marrow transplants, the immunosuppressive drugs that rendered them susceptible to fatal infections were given not only to prepare the recipient for graft acceptance, but as part of a chemotherapeutic attack on their malignant cells (Streilein, 197213; Glucksberg and Fefer, 1973; Gengozian et al., 1973; Thomas, 1974). In this regard, there is a potentially valuable immunotherapeutic relationship between controlled GVHRs and cancer. Math6 first suggested that GVHRs might be directed against leukemia (Math6 and Amiel, 1964). It is possible that the tumor-specific antigens could evoke a vigorous immunologic response from the injected allogeneic cells when presented in concert with the host’s own histocompatibility repertoire (Streilein and Streilein, 1973). In a mouse model, it has been shown that adoptively passaged leukemia may be arrested in irradiated mice by the injection of GVH-inducing allogeneic cells (Boranic, 1971; Bortin, 1974). Of a variety of strain combinations tested, it was found that DBA donor cells provoked a greater graft-versus-leukemia reaction than GVHR in irradiated AKR mice injected with a leukemia derived from AKR mice (Bortin et al., 1973). This finding suggests that it might be feasible to apply such immunotherapy to patients in whom the tumor load has first been reduced by chemoradiotherapy, although the successful “rescue” of these patients from the adverse effects of GVHRs remains the major constraint to its direct clinical application (Bortin, 1974). Bach et al. ( 1973) have proposed a potent antitumor system in which GVH reactivity is removed from the inoculum by 5-bromo-2-deoxyuridine ( BUdR ) and light during proliferation in an MLR, then the remaining cells are sensitized to the tumor (see Section V1,B). Although it might be possible to apply the cytotoxic effects of the GVHR to the management of leukemia, at the other end of the spectrum are reports that GVHRs may result in an increased incidence of cancer (Armstrong et al., 1967; Cole and Nowell, 1970; Cornelius, 1972a,b; Gleichman et al., 1972). Experimentally, the incidence of tumors is generally higher after the induction of mild rather than severe GVHRs (Cornelius, 1972a). The presence of a GVHR is as important as the inherent susceptibility of

142


these animals to the spontaneous development of tumors (Gleichmann et al., 1972). It is apparent that the GVHR causes the activation of hitherto quiescent leukemia viruses ( Hirsch et d.,1970; Cornelius, 1972b). H-2 serotyping has revealed that the tumors may arise in both donor and host lymphoid cells (Cornelius, 1972a). The results of one study suggest that it is the interruption of T- and B-cell cooperation that provokes the development of the vinis-induced lymphomas even in the absence of significant immunosuppression (Solnik et al., 1973). The suggestion made earlier (see Section I1,A) that the transplacental passage of maternal leukocytes might be responsible for inciting runt disease in newborn infants has been developed into a full-fledged theory that attributes the induction of cancer, autoimmunity, and immunologic deficiency to the consequences of a GVHR. The vectors of this process are “Trojan horse lymphocytes” ( Schwartz, 1974), which gain access to crucial host organs undetected and subsequently incite a GVHR with the above-mentioned sequelae. The only exogenous source of these Trojan horse lymphocytes would be the maternal circulation. They could be transferred during gestation and might remain dormant for prolonged periods of time within the infant, ultimately becoming the progenitors of malignant or autoimmune lymphocytes or inducing autochthonous aberrant cells of host genotype. If maternal lymphocytes do persist within the postpartum offspring, it is likely that they are present in such low proportions as to be undetected by conventional tests for chimerism. Furthermore, their continued existence is dependent upon either the tolerance of the offspring to maternal histocompatibility antigens or a prolonged homeostatic balance between offspring and maternal lymphocytes of an unusual character. In contrast to the possible sub rosa interplay of maternal-infant lymphocytes within a normal child, the inadvertent administration of relatively large numbers of immunologically competent cells to an immunologically deficient child may result in a devastatingly acute GVHR (Buckley, 1971; Buckley et al., 1971a,b). The infusion of whole blood into such a child may provoke a systemic allogeneic disease in which the infant presents HL-A incompatible antigens to the transfused blood lymphocytes. Furthermore, intrauterine and exchange transfusions as thereapy €or hemolytic disease of the newborn have led to severe GVH disease in neonates whose immunocompetence was as yet incomplete ( Parkman, et al., 1974). Similar difficulties have been encountered when bone marrow transplantations have been used to alleviate severe combined immunodeficiency disease ( SCID), The indication for such treatment, the consistent results of marrow engraftment into SCID patients, and some of the


143

techniques used to abrogate GVH disease have been discussed and reviewed in depth elsewhere (Buckley, 1971; van Bekkum, 1972; Dupont et al., 1974; Good and Bach, 1974). The devastating and often fatal effects of systemic GVH disease have prompted an almost unending variety of attempts to alter or modify the pathogenic consequences of this form of cell-mediated injury. GVH reactivity has been modified through techniques that either eliminate GVH-inducing T cells or suppress their deleterious antihost cytotoxic activity. In the search for clinically applicable means of suppressing GVH reactivity, experiments have been performed in laboratory animals with ( a ) physical agents. ( b ) chemical agents (drugs, mitogens, etc. ) ; ( c ) specific antibody and other humoral factors; ( d ) cells with suppressor activity. Through the attempts to develop techniques that might avert GVHRs in humans, much has been learned about cellular immunology, especially at the level of T- and B-cell interaction. The quest for solutions to important clinical problems has underscored heavily the subtle and complex nature of the immune system. In summary, while the cardinal rules for the management of GVH disease resulting from bone marrow transplantation are emerging slowly, there remain some crucial problems and inconsistencies. For example, when donor and recipient are HL-A identical and MLR negative matched siblings, there is still the possibility of a severe GVHR. The apparent dominance of MLR determinants was demonstrated in one instance by the successful transplantation of completely HL-A mismatched but MLR negative bone marrow from the pateint’s uncle (Good and Bach, 1974). Another unresolved feature of clinical bone marrow transplantation is the importance of creating an essentially germfree environment for the graft by the elimination of the patient’s normal flora. I n animal experiments, germfree recipients may have markedly reduced incidences of infection and significantly prolonged survivals (Keast and Walters, 1968). However, in human recipients attempts to promote and maintain germfree conditions have met with only equivocal success (Good and Bach, 1974). An additional factor, and one that serves in a pivotal role, is the use of drugs as immunosuppressants. In broadest terms there is a fine line between suppression of the GVHR and abrogation of the host’s own immunologic competence. As our understanding of the mechanisms of the immune response become more finely tuned, we may be able to discern important differences in the susceptibility of donor GVH-inducing cells and host immunity with regard to the effects of immunosuppressant drugs. Finally, as though to acknowledge the Utopian metaphor provided

144


through inbred strains of rodents, 16 patients have received bone marrow grafts from syngeneic identical twin siblings (Thomas et aZ., 1971; Fass et al., 1973; Fefer et al., 1973; Rudolph et al., 1973; Thomas, 1974). All patients showed impressive survival rates and, those with nonmalignant diseases showed immediate and permanent hematologic reconstitution and survived 5-12 years (Thomas, 1974). There remain to be resolved a number of very important issues concerning GVHRs. At the clinical level, GVHR is viewed as a menace, a powerful foe that must be destroyed or circumvented in order for the clinical benefits of bone marrow transplantation and lymphopoietic reconstitution to be realized. Thus, at a pragmatic level, efforts are being made to avoid the emergence of GVH disease by improving tissue matching and histocompatibility testing of donors and recipients of hematopoietic cells. The less well this is accomplished, the more we must rely upon techniques of subverting and abrogating the systemic GVHR without destroying the patient at the same time. The experimentalist can adopt a more kindly view of GVHRs; they allow a unique means of studying the complex cellular interactions that determine immune responses to antigens. In addition, there is the growing realization that immunoregulation-the mechanism by which a conventional immune response is moderated in intensity and limited in duration consistent with the well-being of the host-is operative in GVHRs and may be illuminated by studying the postinduction phase during which the GVHR is contained and resolved. I l l . The Cellular Basis of Graft-versus-Host Reactions

In the preceding section the evidence was reviewed which established that all GVHRs, both in uivo and in uitro, involve the interaction of lymphocytes as the pivotal event. The MLR best exemplifies this primary and inductive aspect of GVHRs. Indeed, GVHRs in uiuo reflect this “donor/host” lymphocyte interaction at several levels, and most if not all of the manifestations of the clinical GVH syndrome may be directly traced to these lymphoid cell confrontations. The following sections of this review will deal with the role of donor lymphocytes in the induction and prosecution of GVHRs, the dual roles of host lymphoid cells, which are both the primary source of antigenic stimulus and the primary target of donor antihost effector cells, and finally with the probable consequences of donor-host interactions. Several assumptions are made about the nature of immune responses in general. These suppositions provide a useful perspective for a survey of the early as well as recent work on GVHRs.

GRAFT-VERSUS-HOST

REACTIONS

145

1. There are two analogous, sometimes complementary, systems in the immune apparatus. One is commonly referred to as thymus-dependent ( T ) and the other bursal- or bone marrow-dependent ( B ) , and their respective influences provide the basis for what has become an essential duality of immunologic dogma, described empirically as delayed hypersensitivity ( cell-mediated immunity) and immediate hypersensitivity ( antibody-mediated immunity), respectively. 2. The genetic disparity between donor and host necessary for the initiation of GVH is determined by the MHC (Bach, 1973) of the species. Traditionally, the requirement was thought to be simply due to a difference at the major histocompatibility locus of the species (H-2, mice; Ag-B, rats; B, chickens; HL-A, man) ( Billingham, 1968). Better resolution of the constituents of the MHC has revealed that the essential determinant is an immune response ( I r ) gene product difference, or a closely linked gene (Festenstein and Dkmant, 1973); GVH and MLR can be incited in mice that are identical at the so-called “serologically detectable” loci (H-2D and H-2K), but congenic with respect to differences in the Ir locus only (Klein, 1973; Klein and Park, 1973; Livnat et al., 1973; Widmer et al., 1973). Other genetic factors may also influence GVH reactivity (Gasser and Fischgrund, 1973), although it appears to be independent of high and low antibody response (Byfield and Howard, 1972) in at least one instance. 3. The capacity of a T cell population to respond to an appropriate alloantigen is probably determined by an allele of the Ir locus. Throughout the remainder of this paper, the cellular constituents of this special population of T lymphocytes which responds to a specific “major” alloantigenic determinant will be referred to as histocompatibility antigenreactive cells ( H-ARCS) (Elkins, 1971a).

A. IDENTIFICATION OF IMMUNOCOMPETENT CELLS During the period of initial characterization of GVH disease, it was shown that, under the appropriate conditions of genetic disparity and immaturity of the recipients, dissociated suspensions of adult spleen cells could produce “runt disease” both in hatched chicks which had been injected in ovo (Billingham et al., 1956) or several weeks after birth in mice which had been injected during the neonatal period (Billingham and Brent, 1959). Although it could be shown that donor cells from a variety of tissue sources could incite the disease, it soon became apparent that the common cellular denominator in all instances was the lymphocyte. The work of Gowans (1962) in adult F, hybrid rats strongly suggested that within a suspension of thoracic duct lymphocytes, which had been depleted of large lymphocytes by incubation overnight

146


at 37”C, GVH reactivity was still retained. The only morphologically identifiable cell type left was the small lymphocyte. Despite some conflicting evidence obtained in chickens (Szenberg and Warner, 1962), the evidence is overwhelming that the small lymphocyte is the immunologically competent cell postulated by Medawar ( 1963). Gowans and his associates have been able to construct partially the sequence of events that follows specific stimulation of donor lymphocytes by strong alloantigens : small lymphocytes are transformed into large, pyroninophilic blast cells, full of RNA, which then begin a sequence of differentiating mitoses. Ultimately a population of committed “effector” and/ or “memory” cells, from that may then be capable of causing immunologically mediated damage to the host (Gowans, 1962). This step-by-step analysis has been applied to the events of lymphocyte stimulation in other kinds of antigenrecognition systems. Thus, small lymphocytes have emerged as the sine qua non of GVH reactivity. But there are also other important characteristics of these cells that are not readily identified by morphological criteria. It soon became apparent that there was a hierarchy of GVHR potency related to the source of small lymphocytes. Regardless of the assay system used, it has been consistently shown that peripheral blood lymphocytes, thoracic duct lymphocytes, lymph node cells and to a lesser degree, spleen cells (although they are the most commonly used) are potent initiators of GVH reactivity, while thymocytes and bone marrow seem relatively nonreactive, except in the most sensitive tests (Billingham et al., 1962; Cantor and Asofsky, 1970; Cantor et al., 1970b; Streilein and Billingham, 1970b; Yoshida and Osmond, 1971; Cantor and Mosier, 1972; Heim et al., 1972; Dyminski and Argyris, 1973). Perhaps one of the most significant features of this work was the recognition that thoracic duct lymphocytes (Gowans, 1962) and peripheral blood ( Hildemann et al., 1962; Hildemann, 1964) were both rich sources of immunocompetent leukocytes capable of mounting GVH reactivity. The impact of this discovery in terms of normal, systemic immunity, apart from GVHRs, turns upon the fact that both of these lymphoid compartments represent recirculating lymphocytes and therefore provide a mechanism for systemic dissemination of those “memory” cells which had been generated after local sensitization. Lymphocytes do not arise ontogenetically within the peripheral lymphoid tissues, so there has been considerable interest in the genesis of the GVH-initiating cells ( Section I1,A). Because of its increasingly obvious importance in the expression of the normal immunologic responsiveness (Miller and Osoba, 1967), the thymus was soon implicated as an essential component in the development of specific immunologic competence in the GVH process. When


147

jymphoid cell suspensions derived from thymectomyzed mice ( Good et al., 1962; Dalmasso et al., 1962; Miller and Mitchell, 1967) or rats (Riecke, 1966) were injected into appropriate recipients, there was no demonstrable GVH activity. These observations created a paradox that has been resolved, in part, by further studies on GVH phenomena. Although it has been shown that thymocytes themselves are a poor choice for starting a GVH reaction, nevertheless the presence of a thymus is essential for the development of immunologic competence within the donor. Now, a distinction can be made between “immature,” or cortisonesensitive, thymocytes and “mature,” or cortisone-resistant, thymocytes (Blomgren and Andersson, 1971, 197217; Tigelaar and Asofsky, 1973a), the latter representing medullary thymocytes that are essentially identical with typical peripheral T cells. Because these immunocompetent T cells make up only about 4 5 % of the normal thymocyte population (Tigelaar and Asofsky, 1973a), it is not surprising that most classical experiments involving normal thymocytes were unable to link them to GVH reactivity, even when heroically large doses of cells were injected. Now, by more subtle assay techniques (Sosin et al., 1966), which may include preirradiation of the recipients (Hilgard, 1970a; Blomgren and Andersson, 1972a; Sprent and Miller, 1972a), more detailed studies of thymocyte capabilities have been possible (Pirofsky et al., 1973). Furthermore, with regard to GVH- competent lymphocytes in the rat, McGregor (1968) was able to show that bone marrow cells, although incapable of mounting GVHRs directly, are the source of stem cells, which ultimately become GVHinducing T-cells after the imposition of thymic influence. Similarly, fetal liver cells could give rise to inimunocompetent cells when subjected to thymic influence, but the thymus was not wholly essential in this system, for the complete expression of competence (Umiel, 1971, 1973).

B. STUDIESIN CELL COOPERATION Following the startling revelation that thymus-derived and bone marrow-derived lymphocytes could interact “synergistically” when confronted with sheep red blood cells ( SRBC ), producing many times more antibody-forming cells together than the sum of the two alone (Claman et al., 1966; Claman and Chaperone, 1969), evidence was sought in a number of laboratories to document a similar kind of interaction in GVHRs. In these experiments, lymphocytes, bone marrow cells, or mixtures of the two were injected into F, recipients. Unlike the SRBC system, these experiments gave results that were either negative (Stuttman and Good, 1969; Davis et al., 1970) or highly equivocal (Barchilon and Gershon, 1970; Hilgard, 1970b), the difference being primarily

148


semantic. In these last two papers, the “synergism” was essentially due to an additive effect, in which there was a response just exceeding a rather arbitrary threshold. In Hilgards (1970b) system, the bone marrow cells simply played a nonspecific role in the expansion of the spleen. Likewise, mixtures of dissociated thymus and fetal liver have been used to test for synergism. Tyan (1969) was unable to demonstrate GVH reactivity using fetal liver cells alone or to show synergism between fetal liver and thymocytes in a sensitive assay system. Alternatively, while not a clear example of synergism, fetal liver and fetal thymocytes (Blessing, 1972) or neonatal thymocytes plus neonatal bone marrow ( Tridente et al., 1971; Blessing, 1973) produced demonstrable splenomegaly in lethally irradiated newborn F, mice, despite the inability of the thymocytes, liver, or bone marrow cells to do so on their own. At this time, there are only scant data available regarding the T-cell-precursor content of fetal liver, neonatal bone marrow and yolk sac (Hoffman and Globerson, 1973), but it is likely that any specific GVH-inducing competence must proceed from a T-dependent component of these lymphohematopoietic tissues (Lemmel and Good, 1971; Lonai et aZ., 1973b). Perhaps the failure of these attempts to find T-cell/ B-cell synergism is not surprising, since GVH reactivity is primarily a cell-mediated immune process. Accordingly, when Cantor et al. (1970a) first demonstrated synergism in the GVH reactivity of lymph node cells from young and old NZB mice, and then between thoracic duct and other lymphocytes (Cantor et al., 1970b), it was apparent that this “synergistic” interaction was occurring between two populations of T cells, which they have designated as TI (primary constituent of T cells in the thymus and spleen; recirculate slowly; precursor of the cell responsible for immunologic injury) and T, (present in excess in the peripheral blood and lymph nodes; recirculate rapidly; amplify activity of thymic lymphocytes) ( Asofsky et al., 1971; Cantor and Asofsky, 1972). Comparison of the kinetic profiles of splenomegaly, when using either thymocytes or lymph node cells alone or in concert, revealed that the time course of the synergistic response most closely followed that of thymocytes alone (Tigelaar and Asofsky, 1972b). These findings provide a solid basis for the contention that the T, lymphocytes, located primarily in the thymus, are the initiators (or their precursors) of the GVHR, and the recirculating T, lyniphocytes are assigned the role of amplifier cells. With regard to the thymic origin of the TI cells, conclusions of initial studies projected a l-to-1 correspondence between the initiatorprecursor cells and the so-called “cortisone-resistant,” medullary thymocytes (Blomgren and Anderson, 1969, 1971). In later studies using both

GRAFT-VERSUS-HOST

REACTIONS

149

the GVH model (Blomgren, 1973; Tigelaar and Asofsky, 1973a) and the MLR model (Cantor and Mosier, 1971), it has become apparent that there may be additional subpopulations of cortisone-resistant cells that differ in their capacity either to provoke GVHRs alone or to interact synergistically. For example, cortisone-resistant thymocytes did not produce synergistic reactions with either peripheral lymph node cells or normal thymocytes ( Tigelaar and Asofsky, 1973a). Similarly, during the period after cortisone administration, parental cortisone-resistant thymocytes displayed differential recoveries, and there was a reciprocal relationship between GVH-inducing capability and cooperative potential: when one activity was high, the other was decreased and vice versa (Blomgren, 1973). It is possible that within the thymus there may be a modulation of T, amplifier cells into TI inducers, perhaps under the selective influence of cortisone treatment, A proportion of TLnegative thymocytes, which constitute the majority of GVH-inducing lymphocytes in the thymus (Leckband and Boyse, 1971), are probably the cortisone-resistant cells. They may represent a maturational stage between TL-positive, thetu-positive precursors found only in the thymus, and the mature, TL-negutiue, theta-positive peripheral T lymphocytes ( Aoki et al., 1969; Stobo et al., 1973). Because of the widely acknowledged disparity between various GVH assay systems ( Billingham, 1968; Elkins, 1971a), Tigelaar and Asofsky (1972a) compared the synergistic effects of spleen cells and peripheral blood lymphocytes ( representing predominantly TI and T, populations, respectively) in the production of splenomegaly and in their capacity to kill neonatally injected F, recipients. They found that spleen cells were relatively less lethal than capable of causing splenomegaly, by parallel analysis, whereas lymph node cells and peripheral blood lymphocytes were more active in killing than in causing splenomegaly. Furthermore, blood lymphocytes were more lethal in F, recipients than would be predicted by summing up the individual reactivities of the two cell types. The shape of the mortality curve for thymocytes and peripheral bIood cell mixtures most nearly corresponded to that of thymocytes alone, which further corroborated the assertion that TI cells were the precursors of the effector cells. In order to demonstrate a functional heterogeneity and differential recirculating capacity of parental lymphoid cells, spleen cells were injected into lethally irradiated syngeneic mice, the lymph nodes and spleens of which were used 24 hours later to induce GVH disease in newborn F,s (Tigelaar and Asofsky, 197313). Although either population alone was deficient in GVH-inducing capacity, certain mixtures of the two did produce normal reactivity. Furthermore, when thymocytes, but not lymph

150

STEPHEN C. GREBE AND J . WAYNE STREILEIN

node cells, are in excess in an MLR-inducer cytolysis assay, there is synergistic killing ( Hayry and Andersson, 1974). In contrast to the synergistic, or cooperative, interaction between subpopulations of T cells, there have also been reports of antagonistic, “suppressor” activities ( discussed more fully in Section II1,C ) demonstrable in an afferent manifestation of the GVHR, namely, the proliferation of donor lymphocytes (Gershon et ul., 1972; Gershon and Liebhaber, 1972). However, the most perplexing aspect of this work is the gross variability of much of the data expressing the relative amount of incorporation of 3H-labeled thymidine in the spleens and lymph nodes of lethally irradiated mice that were then injected with allogeneic thymocytes. The label uptake presumably is a direct measure of the mitotic activity of the injected T cells (devoid of B cells by design) in response to H-2 alloantigens. An important criticism is that the measurement of DNA synthesis represents only the second event ( following specific engagement of antigen by the surface receptors on the T cells) of the GVH process and gives little or no hint as to the clinically relevant manifestations of the ensuing donor cell activities. The work of Asofsky and his colleagues encompasses a more impressive range of classical GVH signs, such as splenomegaly and mortality ( Tigelaar and Asofsky, 1972a,b). Perhaps when appropriate correlations have been made between specifically triggered DNA synthesis of donor cells, elaboration of lymphokine products, and the incidence of widely recognized pathogenic features of GVH disease, the now tenuous evidence for the existence of “suppressor T cells” can be more substantially documented. On the wave of interest generated by the revelation of positive interactions between subpopulations of T cells in the expression of GVHR in uivo, there have been a number of studies documenting a parallel mechanism in other animal models of GVH in uitro, both in MLR and in CMC assays. The GVH-induced suppression of syngeneic erythroid cell proliferation, assessed by reduced incorporation of 59Fewhen lethally irradiated adult F, mice were injected with allogeneic lymphocytes ( Blomgren and Andersson, 19724, may be quantitiatively amplified by the admixture of thymus and lymph node cells (Blomgren and Jacobsson, 1974b). The parallelism between the GVHR and the MLR-as an “in uitro microcosm” portraying the afferent aspects of the GVHR-has been reinforced by the demonstration that thymus and peripheral lymph node cells may produce a supra-additive response when challenged jointly in culture with mitomycin C-treated allogeneic lymphocytes ( Tittor and Walford, 1974; Tittor et al., 1974). Although normal spleen and lymph node cells did not respond synergistically in MLRs with allogeneic


151

cells, admixtures of spleen and lymph node cells from lethally irradiated mice reconstituted with syngeneic spleen cells 20 hours earlier were able to interact in a cooperative manner (Tittor et a!., 1974). This finding confirms the studies which established that T,, or GVH precursor, cells are spleen-seeking while the TL amplifiers are lymph node-seeking cells (Tigelaar and Asofsky, 1973b). At the other end of the immunologic reflex arc, studies on synergistic interactions in the execution of CMC have shown that there is a cooperative interplay between thymocytes ( T , ) and lymph node cells (T,) (Wagner, 1973; Wagner et al. 1973, 1974). A survey of the studies of cell cooperation in MLR and CMC (Cohen and Howc, 1973; Hayry and Andersson, 1973; Tigelaar and Feldmann, 1973), in addition to those already mentioned, firmly establishes this cooperative interaction as an important immunologic phenomenon. While the mechanisms of CMC are in themselves fascinating, they are not germane to this review, and the reader is referred elsewhere (Goldstein et al., 1973; Henney, 1974; Wagner and Rollinghoff, 1974). In ~nanyrespects, the synergy that can be demonstrated in assays of CMC is closely allied with the induction of MLR. Indeed, it appears that the synergy in the CMC may occur at the sensitization or afferent phase of the responsc, not during the effector phase (Hodes et al., 1974). However, in this study, the “synergizing” cell was resistant to lysis l)y rabbit anti-moLise brain serum plus complement and had adherent properties, although it could not be replaced functionally by adherent spleen or pcritoncal exudate cells. Thus, it docs not seem to be either T, or T,. One of the mo5t curious aspects of the synergistic T, and T, interaction in CMC is that, unlike in the GVH and 14LR svstems, it may be the T, cell that is a “precursor of the killer lymphocyte, and its activity is amplified, perhaps by a soluble mediator that is produced by TI cells (Wagner et uZ., 1974). Although splenomegaly and CMC represent essentially divergent processes, there is clearly a fundamental interplay betwcen the T, and T, cclls. The is\rie of their respective cellular genealogies, developmental tempos, and homing properties with respect to splcnomegaly and allograft cytolysis might lxst be resolved if we regard splenomegal? as a phenomenon oichcstratcd b y the production of mediators of both low and high specificities and amplified by the activity of “pc~iiphcral”cells. Alternativcl~~, cc11-mediated lympholysis resembles a poised spearhead, triggercd I n specific antigen and augmented to the point of visibility and completion by the reinforcement of T , soluble mediators. An all-encompassing hypothesis to evplain the composite pathogenic features of GVH disease would nrcessarily embrace both thcsc models of T, and T, interaction.

152


In addition, the vigor of the response to alloantigens results from the simultaneous stimulation by two sets of antigens, the lymphocytedefined ( L D ) and the serologically defined (SD) antigens (Alter et al., 1973; Bach, 1973). While LD differences alone appear to be sufficient to initiate an MLR (Schendel et al., 1973) or a GVHR (Klein and Park, 1973), both LD and SD differences are required for cell-mediated lympholysis (Alter et al., 1973). Based on this information, it has been postulated that T, cells have a receptor for alloantigens which are the product of the LD locus, and SD antigens stimulate the TI cells via their specific receptor (Wagner et al., 1974). This intriguing hypothesis, if borne out by further experimental data, could provide more than just a theoretical basis for bringing together what have been regarded as two unrelated observations: the existence of at least two functional subpopulations of thymus-derived H-ARCS and two sets of lymphocyte-borne alloantigens. A fresh and unbiased approach to the marriage of these concepts might ultimately lead to a real understanding of the teleological underpinnings for the major histocompatibility complex and an explanation for the apparently high proportion of immunocompetent cells bearing receptors for such histocompatibility alloantigens. C. ALTERATIONOF GVHRs

SUPPRESSOR T CELLS In a number of experimental systems in contrast to the synergistic, or cooperative, interaction between subpopulations of T cells, the presence of certain T cells may suppress the response (e.g., Baker et al., 1970; Okumura and Tada, 1971, 1973, 1974; Folch and Waksman, 1973; Rich and Pierce, 1973; Feldman, 1974; Gershon et al., 1974b; Ha et d.,1974). These studies include effects ranging from suppression of B cells responding to dinitrophenyl (DNP) derivatives (Okumura and Tada, 1973, 1974) and bovine 7-globulin (Ha et ab., 1974) to suppression of T-cell stimulation by alloantigens either in uivo (Gershon and Liebhaber, 1972; Gershon et al., 1972, 1974a-d; Hardin et al., 1973) or in uitro (Rich and Pierce, 1973; Peavy and Pierce, 1974; Rich and Rich, 1974). We will discuss primarily the evidence for the existence of suppressor T cells in GVH and MLRs. In one GVH study, it was concluded that (1) suppressor T cells must be syngeneic with the donor; ( 2 ) they do not function properly if DNA synthesis in the suppressor cells is blocked with mitomycin C; and ( 3 ) the cells may not require specific exposure to the alloantigens ( Blomgren and Jacobson, 1974a,b). In contrast, other GVH studies have demonstrated that ( a ) suppressor cells may be syngeneic with the recipient as well as with the donor (Gershon et al., 1972, 1974a), ( b ) even higher doses of mitomycin C do not abrogate the suppressor activity (Folch and Waksman, 1974a,b; Rich and Rich, 1974) [although protein synthesis is apparently essential and cycloBY


153

heximide interferes with suppressor capability ( Folch and Waksman, 1974b)], and ( c ) splenic suppressors are prominent only after pwexposure to antigen (Rich and Rich, 1974). Neverthelezs, there are many concrete points of agreement. The suppressor cells are found in either thymus (Blomgren and Jacobsson, 1974a; Gershon et al., 1972, 1974a,b,c) spleen (Folch and Waksman, 1974a,b; Rich and Rich, 1974), or both (Hardin et al., 1973) and can be removed by anti-theta serum and complement ( Peavy and Pierce, 1974; Rich and Rich, 1974). Interestingly, when mice are immunized with allogeneic lymphoid cells injected into the foot pads, 4 days later cells in the draining lymph node heighten the reactivity of responder cells in the MLR, while the spleen contains suppressor cells ( Rich and Rich, 1974 ) . Unlike the cortisone-sensitiue thymocytes, which interact synergistically in inciting GVHRs ( Blomgren, 1973; Tigelaar and Asofsky, 1973b), suppressor T cells appear to be at least partly resistant to cortisone (Gershon et al., 1972, Ha et d., 1974; Blomgren and Jacobsson, 197413). In at least one system, it appeared that the suppressor function was associated with an adherent cell in the spleen. It was apparently a T cell or at least thymus-dependent, since suppressor activity was absent in thymectomized, irradiated rats that were reconstituted with bone marrow fron syngeneic rats that were also thymectomized, irradiated, and bonc marrow reconstituted ( Folch and Waksman, 1974a,b). The mechanism of suppression is not clear at this time. In the MLR system in uitro, the regulatory cells do not kill the responder cells, as evidenced by tiypan blue exclusion and Wr-release studies. However, they appear to suppress nonspecifically; when suppressor cells were elicited by “immunizing” one strain of mice, these regulatory cells could also inhibit the response to third-party alloantigens in another strain (Rich and Rich, 1974). Considering humoral immunity, the suppressor T cell which truncates the antibody response to DNP-Ascaris acts through a protein mediator (MW 35,000-60,000), probably an a- or a p-globulin (Okumura and Tada, 1974). This factor, which could be extracted from thymocytes of specifically hyperimmune rats by freezethaw lysis, reduced the titer of hapten-specific, homocytotropic IgE when given to antibody-producing rats during the late stages of immunization, suggesting that the regulatory effect probably does not influence the afferent process of antigen recognition (Okumura and Tada, 1974). The activity of this factor can be removed with an immunoabsorbent column of antithymocyte serum ( ATS) (Tada et al., 1973), raising the possibility that it may be expressed physically as a determinant on the T-cell membrane. While the allusion to T-cell suppression of antibody production is not directly relevant to a discussion of suppressor T cells in GVHRS,

154


it points out that not only are there important discrepancies between the two systems, but relatively little is known about the molecular mediation of suppressors of the GVHR. Further elucidation could have farreaching consequences in terms of our understanding of immunoregulation in cell-mediated immunity and could benefit the clinical handling of GVH disease in humans. Furthermore, studies of the age-dependent reduction of suppressor T cells in mice (Hardin et al., 1973) and rats (Folch and Waksman, 1974a,b; Nielsen, 1974) may offer a new insight into the pathogenesis of autoimmune disease and the impact of age on immunologic surveillance. IV. Donor Lymphoid Cell Participation in Graft-versus-Host Reactions

The implication to be drawn from a consideration of the cellular basis of GVHRs is that donor and host both bring to the response unique contributions. In the next section of this review, to the extent that it is possible, the donor component will be described alone; this is followed by a description of the host component. Obviously, in GVHRs the whole is greater than the sum of its parts, but this approach allows certain important observations and distinctions to be made among the various pathogenic mechanisms enlisted in the effectuation of these responses. A. DONOR LYMPHOCYTE TRAFFIC, HOMING, AND DISTRIBUTION The route of donor cell inoculation is a major determinant of thc outcome of a GVHR, all other factors being equal. The epidermolytic syndrome in hamsters, incited by the intracutaneous injection of 200 x 10" specifically sensitized lymph node cells, could not be reproduced by giving the same dose of cells intravenously; these animals did not develop the cutaneous lesions, and most of them were alive 30 days later, when all the animals injected intracutaneously with 200 x 10" lymphoid cells were dead (Streilein and Billingham, 1970b,c). When "Cr-labeled donor lymph node cells were given intravenously to allogeneic recipients, after 24 hours the organ containing the greatest amount of label was the liver (Heslop and Hardy, 1971). However, there was more label localized in the spleen than in the lymph nodes, yet the opposite was true of labeled lymph node cells injected into syngeneic recipients. In the syngeneic (Hall et al., 1972) and allogeneic animals, there was essentially no label in the thymus, and only a modest amount could be found in the lungs (Heslop and Hardy, 1971). In a somewhat more exacting system (Bainbridge et al., 1966), more label could be recovered in syngeneic recipients injected intravenously than intraperitoneally, and although the greatest net reactivity ( "Cr label ) was recovered from spleens and livers, the greatest amount of label/


155

weight of organ was demonstrated in spleens and lymph nodes. In a system somewhat more relevant to the study of GVHRs, Haller (1964) showed that removal of spleens within 24 hours after birth of rats that had been injected intravenously at birth with allogeneic lymphoid cells circumvented the lethal effects of GVH disease, which ravaged control animals splenectomized either before the injection of the allogeneic cells or several days after. This work suggested that the major portion of H-ARCS were filtered out of the peripheral blood by the spleen initially, but then gradually left the spleen, inciting reactivity elsewhere in the host. This splenic accumulation and redistribution process has been verified in another context. In a study of T-cell and B-ccII distribution after intravenous administration of thoracic duct cells, it was shown that a large proportion of both T cells and B cells could be identified within T-dependent and B-dependent areas, respectively, of the spleen 4 hours after injection. However, by 24 hours, many of the T cells had left the spleen and could be identified in the lymph nodes, although B cells did not redistribute appreciably (Sprent, 1973). Nevertheless, both T cells and R cells can recirculate in rats (Howard, 1972; Howard et al., 1972). Treatment of donor cells with trypsin (Woodruff, 1974) or neuraminidase (Gesner and Woodruff, 1969) alters their homing, as does pretreatment of lymphoid cell donors with ALS (Morse and Asofsky, 1974), the former by modifying surface receptors, the latter by direct cytotoxicity. The abundance of data relating to the distribution of donor cells after intravenous injection is unfortunately not matched by the quantity of information on the kinetics and distribution after intracutaneous injection-a somewhat less natural or physiologic situation, but one eminently relevant to many immunologic studies. After intracutaneous inoculation, cells localize in the regional draining lymph node in significant, if somewhat low, numbers (Ford et al., 1970). Studies in sheep suggest an afferent pathway from the injection site through the draining node or chain of nodes and eventually into the thoracic duct, from which the efferent lymph empties into the vena cava (Hall and Morris, 1962, 1965, 1967, 1969). These cells may then become part of the recirculating Iymphoid pool, but there is a dearth of critical evidence about the composition of such cells with regard to their T / B makeup and their potential for homing to the different lymphoid organs of the host. One aspect of the homing phenomenon that bears mentioning is the specificity with which injected donor H-ARCS are removed from the recirculating pool by the co~ifrontationwith host alloantigens ( Gowans and McGregor, 1965). In an ingenious experiment, Ford and Atkins (1971) injected adult F1 hybrid rats intravenously with parental thoracic

156


duct lymphocytes, then obtained from the injected host (12-36 hours later) recirculating lymphocytes by means of thoracic duct drainage. These “filtered cells, containing both F, cells and some of the injected parental lymphocytes, were then tested in the PLN assay for the ability to mount a GVHR against a specific F, hybrid (syngeneic with the “filter” animal) and against a third-party F, hybrid (semiallogeneic with the original parental strain, but bearing a different alloantigen than the filter animal). The results indicated that the specifically “anti-host” reactive cells were filtered out, but that other H-ARCSwere allowed to recirculate freely and could incite appreciable GVH reactivity against the third-party F, hybrid. Recently, these authors have investigated the migratory properties and specific proliferative potential of parental thoracic duct lymphocytes injected into sublethally irradiated F, rats ( Atkins and Ford, 1975; Ford et al., 1975). They found that radiolabeled parental lymphocytes when compared with suitable control cells were consistently in excess in the spleens and reduced in the thoracic ducts of F, rats, with intermediate values found in the lymph nodes ( Atkins and Ford, 1975). Control experiments revealed that this selective homing pattern was dependent upon specific antigen-recognition and Ag-B disparity.

B. DONORLYMPHOCYTE PROLIFERATION After Gowan’s observation that donor lymphocytes do indeed proliferate in the GVHR many investigators studied, principally by means of chromosomal markers, the proportion of donor cells that were proliferating in the spleen, lymph nodes, and other lymphoid organs of appropriate recipients at various times after injection. Fox (1962) observed a burst of donor cell mitotic activity in the spleens of F, mice injected intravenously with CBA (T6) chromosomally marked spleen cells. This donor mitotic barrage peaked at 2-3 days and persisted for about 12 days at reduced levels. When evidence for donor proliferation in the lymph nodes was sought, there was even greater donor lymphocyte proliferation in them than in the spleen (50% versus 15%,averaged during the period from 7 to 21 days). Likewise, in the CAM assay in chicken embryos there were between 12 and 51% donor mitoses within the pocks 4-5 days after inoculation of the CAM with donor lymphocytes, and the spleen showed 0-9% donor mitoses ( Weber, 1970). Sublethally irradiated F, rats were given parental thoracic duct lymphocytes to study the proportion of cells within a lymphoid compartment which is responsive to histocompatibility alloantigens. An assessment of donor lymphocyte proliferation within 24 hours after injection led to

GRAFT-VERSUS-HOST

REACTIONS

157

the conclusion that 4 5 6 . 0 %of the thoracic duct lymphocytes of unprimed parental rats can “recognize” and “react” to host Ag-B alloantigens (Ford et al., 1975). If this extraordinarily high estimate of the proportion of H-ARCS is legitimate, then it may be that there are many antigenic specificities (greater than lo3) that are in fact allogeneic histocompatibility determinants, or that H-ARC receptors are uniquely expressed and/or activated in comparison with other ARCS-an oversimplification of the hypothesis advanced by Ford and his colleagues. In heavily irradiated allogeneic animals, the proliferation of donor cells has been measured in a number of ways (Bennett, 1971; Blomgren, 1971; Gershon et al., 1972; Sprent and Miller, 1972a,b; Cheers et al., 1974). While these animals apparently provide a sufficient immunogenic stimulus for donor lymphocytes, the subsequent GVHR in them does not involve host cells to as great an extent as in normal recipients. It is interesting to note that the donor cell proliferative period is of somewhat limited duration, which suggests that the clonal expansion of T cells creates a population of differentiated T cells that are now committed effectors. Although there have not been any definitive studies on the absolute requirement that donor cells divide in order to incite a GVHR, one report suggests that donor lymphocytes treated with the mitotic inhibitor mitomycin C arc unable to cause a GVHR (Lemmel and Good, 1969). Several lines of evidence suggest that donor cell proliferation may be essential for the development of thc full-blown GVH syndrome, but nevertheless some components of the phenomenon are intlependgnt of donor cell division. It is probable that these effects are mediated principally by a class of substances called lymphokines, which are produced by T cells either by specific Ftimulation (antigen) or by nonspccific stimulation (e.g., mitogen).

C. ROLEOF LYMPHOKINES IN GVHRs Of all the various T-ccll products that have been described (Pick and Turk, 1972), perhaps the most relevant to a consideration of the early phases of GVH rcactivitv are: inaciophage-migration inhibition factor ( M I F ) ; blastogenic factor ( B F ) ; lymphotoxin ( L T ) . As in other systems of delayed hypersensitivity, the release of MI F has been demonstrated in response to allogeneic histocompatibility antigens in the rat (Falk et al., 1969). An impressive feature of MIF release is that it occurs within 6-8 hours after antigenic stimulation, and furthermore, it can continue for as long as 4 days after removal of the triggering antigen (Bennett and Bloom, 1967; Pick and Turk, 1972; Rocklin, 1973, 1974). An interesting study of BF revealed that macrophages, or at least adherent

158

STEPHEN C. GREBE AND J. W A Y N E STREILEIN

cells, must be present for the responding population of lymphocytes to divide. The blastogenic activity was nonspecific, since it could be absorbed by using cells of genetically different origin than the test cells. Finally, there appeared to be a certain degree of seleztivity within the responding cell sample, suggesting that not all lymphocytes are equally susceptible to the actions of BF (Chan and Gordon, 1971). Work on LT, a lymphokine which is cytotoxic for target cells, has indicated that neither blast transformation nor DNA synthesis is necessary for its production (Williams and Granger, 1969), and that in mice and rats, a nonspecific stimulus, such as phytohemagglutinin (PHA), provokes its release (Granger and Kolb, 1968; Peter and Dawkins, 1971). Curiously, target cells that are equally susceptible to Iysis by antibody and complement may have a differential susceptibility to LT, a finding that may relate to the differential vulnerability of tissues to the ravages of GVHR independently of histocompatibility type (Kramer and Granger, 1975). In summary, many of the effects of GVH disease may be attributed to the nonspecific activity of various lymphokines. Indeed, it is possible to argue that most of the pathogenic sequelae associated with GVH disease could be due to lymphokines: the proliferation of both donor and host lymphocytes under the influence of BF; the subsequent lymphopenia and generalized inhibition of mitosis (Billingham, 1968) due to the late release of LT, which affects the host lymphoid cells most critically because they are contiguous with other tissues suffering as “innocent bystanders” when the dying lymphocytes release additional LT; the generalized suppression of mitosis, which may result from proliferation inhibitory factor (Badger et al., 1974); and the death of the host, ravaged by destruction of critical organ systems and susceptible to infections. In any case, it is possible that the dissociation between the proliferation of GVH-inducing activity and the proliferation of donor cells (Nisbet and Simonsen, 1967), or the discrepancies between various assay systems of GVH reactivity (Tigelaar and Asofsky, 1972a), may be due to the inconstant superimposition of lymphokine activity upon the specific action of cell-mediated effectors or antibody. OF SPECIFICEFFECTORCELLS D. GENERATION

The activation of specific H-ARCS leads to the production of killer

T cells or specific antibody, or both. In the GVH process, this afferent

arm of the immunologic reflex arc depends primarily upon host lymphocytes, as the principal source of offending antigen. The dual nature of this allogeneic response is perhaps best characterized by the work of Cerottini et al. ( 1971), who have described the generation of effector


159

cells in thc spleens of heavily irradiated mice. They give good evidence that thc cells are of donor genotype, and they distinguish experimentally between alloaiitibody plaque-forming cells ( PFC ) , which are complement dependent, and cytotoxic-effector lymphocytes ( C L ), which can kill target cells in the absence of complement. Furthermore, they could abolish CL activity with anti-thetu serum (Haff, 1969), but PFC levels wert’ unaffected hy siich trcatmrnt. Tho T-cell nature of cclls capablc of killing allogeneic target cells has been confirmed by the use of purified populations of “educatcd T cells that do not contain B cells (Goldstein et al., 1972). Wagner and his associates (Wagner, 1972; Wagner et d.,1972; Wagner and Feldmann, 1972; Rouse and Wagner, 1972, 1973) have also been able to assay CLs and show thc requirement for T cells. In addition, thev have used a technique of in vitro sensitization that correlates the proliferativc response with the cytotoxic phase: the proliferative activity peaked at 4 days of culture, yet the maximum number of CLs was generated at 6 days (Wagner, 1972). One of the astonishing results was that cortisone-resistant thymocytes, sensitized to alloantigens in culture, could kill 50%of the target cells in a C L assay at killerltarget ratios of 1: 1-several orders of magnitude fewer killers than in virtually any other system (Wagner et al., 1972). Thus, under appropriate conditions, it might be possible to produce an cxtremely potent donor cell inoculuni in terms of the potential for tissue destruction of the host. However, it was also shown that, like the synergism that induces splenoinegaly and mortalitv ( TigeIaar and Asofsky, 1972a), synergistic interactions between peripheral T cells and thymocytes could occur that led to more potent cytotoxicity than that of either population alone (Wagner, 1973). Wagner concluded that peripheral T cells are the precursors of CL, although thymocytes iiiay function as amplifiers (see Section II1,B). The only critical studies of the specific and nonspecific cytotoxicity associated with GVH disease have come from Singh et ul. (1971, 1972, 1973). They presented data to support the idea that specific antihost cytotoxicity was donor-cell mediated, although the host cells, under the influence of an intense GVHR could be activated to cause nonspecific cytotoxicity (Singh et a[., 1972). Thev had prcviously shown that spleen cells from mice in the throes of a GVHR could nonspecifically kill cells syngeneic, allogeneic, or xenogeneic with respect to the donor strain (Singh et al., 1971 ). Thus, although the specific cell-mediated effector component of the donor lymphocytes probably incites the greatest amount of antihost tissue destruction (Dyininski and Argyris, 1973), both donor and host nonspecific factors may play decisive roles in the pathogenesis of GVHRs.

160

STEPHEN C. GREBE *AND J. WAYNE STREILEIN

E. FATEOF DONORHISTOCOMPATIBILITY ANTIGEN-REACTIVE CELLS In the most severe instances of GVH disease, the donor H-ARCS may be responsible for the host’s demise, like an imbalanced pathogen which recklessly destroys that which sustains it. Of far greater interest is what happens to specifically antihost reactive donor lymphocytes in hosts that survive a GVHR. Theoretically, donor H-ARCS should proliferate in an unbridled fit following specific immunologic stimulation by alloantigen. Relatively little attention has been given to this fundamental problem in the area of GVH research, but the possible fates of the cells may be cited as follows: 1. Death by allergic suicide: The continued stimulation of a donor H-ARC by high doses of antigen may induce a state of hyperresponsiveness in that cell which exceeds its capacity to regulate its normal metabolic activities, and so it is killed in the line of duty. There has been some evidence to suggest this possibility (Congdon and Urso, 1957; Gorer and Boyse, 1959), but the prevailing dogmas of immunology would tend to exclude it. 2. Death from old age: It has been postulated that immunocompetent cells may become “terminally differentiated (Sterzl, 1966). Recently, it has been shown that antibody-forming cells have a finite life-span and that the production of a specific antibody by a given clone of plasma cells eventually ceases (Willamson and Askonas, 1972). If the H-ARCS were to stop proliferating, then it is reasonable that they might reach a homeostatic balance with the host, and ultimately they might not be able to maintain the critical mass necessary to inflict further damage. Despite the possibility that they could be “long-lived cells,” nevertheless, they would be expected to die out slowly. 3. Death at the hands of the host: In F, hybrids, it is possible to raise an antibody against the “recognition structure” on parental strain donor lymphocytes; this receptor is capable of recognizing the allogeneic F, antigen (Ramseier and Lindenmann, 1971). Because the depressed immunologic competence of most animals suffering from GVH disease (Lapp and Moller, 1969; Lapp, 1973) would prejudice against the host’s mounting a vigorous response to any antigen, it seems unlikely that the recovery from a severe GVHR could be attributed to this mechanism. 4. Acquisition of tolerance: One might postulate that by some mechanism analogous to the induction of tolerance in newborn animals, H-ARCS may lose their specific reactivity (Simonsen, 1960; Dineen, 1961). The proponents of this theory point to the fact that it is possible to show the existence of alloantigen-reactive cells to other histocompati-

CRAFT-VERSUS-HOST REACTIONS

161

bility alloantigens in hosts that have survived an initial GVHR despite the inactivity of the H-ARCS directed at the determinants of the host (Atkins and Ford, 1972). The question is still unresolved as to whether or not there is any such thing as a “tolerant cell,’’ or whether tolerance simply reflects the state of nonresponsiveness, presumably due to the deletion of reactive cells. 5. Control by suppressor cells: While the evidence for the existence of suppressor lymphoid cells (see Section II1,C) is somewhat controversial, it is nevertheless possible that they might become a dominant force in the recovery phase of a GVHR. Their presence would account not only for the failure of donor H-ARCS to continue their antihost assault, but might also provide an explanation for the onset of generalized immunologic incompetence that pervades the animal recuperating from GVH disease. However, it is difficult to understand how these suppressor cells might come to prominence in the wake of a turbulent GVHR, when other lymphoid cells arc destroyed. 6. Deflection of immunologic attack: If donor H-ARCS are responsible for the production of antihost antibodies, then these immunoglobulins might effectively mask host antigens, thus thwarting the CL destructive process. These “blocking factors” have been implicated in a number of instances (Voisin et al., 1968; Voisin, 1971a,b), and there have been references in the literature to the possible existence of enhancing antibodies in GVHRs ( FieId and Gibbs, 1966). Likewise, it is conceivable that antigen-antibody complexes could bind to surface receptors of effector T cells, thus inactivating them (see Section IV,D ) . Many features of host survival can be ascribed to the activities of donor cells, still it is appropriate to mention another aspect that rests with the host itself, apart from the “antirecognition structure” hypothesis. 7 . Exhaustion of appropriate host antigenic target tissue: If there were no appreciable antigens for the H-ARCS to respond to, then it is likely that the donor lymphocytes would be unable to continue their antihost campaign ( Streilein, 1972a). Although the host-at-large bears transplantation antigens, the best source is found on lymphocytes; in an animal rendered leukopenic by GVH disease, there would be relatively little of the “best antigen” around. Thus, in such an animal, with no overt signs of GVHR, the administration of new lymphocytes (syngeneic with the host) might reactivate the disease. V.

Host Participation i n Groft-versus-Host Reactions

While there is a vast collection of information about the role of donor lymphoid cells in the GVHR, comparatively little is known about the

162


role of host lymphoid cells. Some evidence already cited underlines the importance of host lymphocytes, both as a major source of alien transplantation antigens and as participants in the GVHR as it gains momentum. A third relevant feature of host lymphocytes stems from the effect of GVHRs upon them and the attendant changes in the immunologic status of the host. A. IMMUNOGENIC STIMULUS In experiments involving skin reactions, the development of lesions is dependent primarily upon the presence of radiosensitive host leukocytes ( Ramseier and Billingham, 1966; Streilein and Billingham, 1967; Zakarian and Billingham, 1972). The addition of F, lymphoid cells to an inoculum of parental lymphocytes allows normal expression of skin reactivity in F, animals that have been irradiated. Several studies have followed from the work by Steinmuller (1967) that demonstrated the existence of sensitizing passenger leukocytes within the vasculature and parenchyma of isografts which had served a short tenure grafted onto the flanks of tolerant allogeneic mice, before being returned to the original donor. During that interval on the chimeric tolerant mice, the grafts were infiltrated by a sufficient number of leukocytes to cause their original donors to reject subsequently a “first” allograft in second-set fashion, when that graft was of the same genotype as the intermediate host. Passenger leukocytes have been implicated as a mechanism by which neonatal skin grafts may survive on H-2 cornpatible allogeneic hosts for prolonged periods and even induce tolerance to subsequent grafts of adult skin ( Wachtel and Silvers, 1972). The role of passenger cells in transplantation immunology has been reviewed by Billingham ( 1971), and the opinion of that author is that passenger cells may indeed play a demonstrable role in the rejection of allografts, but under conditions in which there are clearly no such leukocytes within the graft, allograft rejection may still occur. Similarly, when F, hybrid rats are rendered leukopenic by irradiation prior to the injection of parental lymphocytes under the renal capsule, there are no typical invasive-destructive lesions ( Elkins, 1966). Alternatively, when the kidney is syngeneic with the donor lymphocytes, but there are allogeneic leukocytes percolating through the kidney, GVH reactivity occurs as a manifestation of the “innocent bystander” effect ( Elkins, 1970). It remains to be shown conclusively whether the important immunogenic lymphocyte in the GVHR is a T cell or a B cell, or both. Recently, however, it has been shown that B cells in relatively pure form may stimulate allogeneic T cells to proliferate in the MLR (Plate and McKenzie, 1973; Cheers and Sprent, 1973; von Boehmer, 1974a,b). In con-

CRAFT-VERSUS-HOST

REACTIONS

163

trast to peripheral lymphocytes, thymocytes have relatively less dense surface alloantigens, as measured by the labeling of the thymocytes or lymph node lymphocytes with specific anti-H-2-serum and idcntification in electron micrographs by using southern bean mosaic virus (Stackpole et al., 1971) or ferritin conjugated to the antibody (Aoki et al., 1969). Accordingly, one would anticipiate that thymocytes themselves are relatively poor immunogenic stimuli. In experiments in which the hosts have been given leukopenia-inducing doscs of irradiation, but have still managed to sustain donor cell-mediated GVH attacks, it is possible that the source of iinmunogenic stimulus could coine from eosinophilic Icvtkocvtes, which have a surface concentration of H-2 alloantigens roughly equivalent to peripheral lymphocytes, or from reticular cells of the thymus, spleen, and lymph nodes, which have the “most extcwsive H-2 coats of any cells so far examined (Aoki et al., 1969). On the other hand, although dissociated suspensions of epidermal cells may be capable of stimulating lymphoid cells injected into the skin of an irradiated hamster sufficiently to evoke a vigorous skin rcactioii (Hamseier and Streilein, 1965), it is unlikely that donor lymphocytes would normally encounter epidermal cells in situ in precisely the same geometric configuration. Thus, there is abundant evidence that leukocj~tcsare the essential stimulatory target for donor lymphoid cells. Another important parameter of inimunogcnicity is reflected by the phenomenon of gene dosage or, simply, thc difference between the concentration of surface antigens presented by a homozygous, allogeneic Ivniphocyte, and a heterozygous, “s~iiiiallogeneic”F, stimuIus. In GVH experiments involving both allogeneic and semiallogeneic recipients, it was observed that the onset of symptoms and chronology of mortality were faster in allogeneic than in F, recipients (Lubaroff and Silvers, 1970), and splenic indices and liver hypertrophy were greater in allogeneic newborn mice injected at birth with spleen cells, than in the F, hybrid animals (Simonsen and Jenscn, 1959). While it is not possible to exclude the effects of heterosis, or hybrid vigor, which might have some beneficial effect on the welfare of the F, hybrids, it is apparent that the quantity of antigcns that the donor lymphocytes encounter has some impact on the severitv of the ensuing GVHR.

B. DEVELOPMENT OF LESIONS In a number of studies, host lymphoid cells have been implicated in histologic and clinical manifcstations of GVH disease in three major ways: (1) contributing to the bulk of the lymphadenopathy by the genc~ralized trapping mechanism in lymphoid organs; ( 2 ) proliferating

164


extensively and apparently nonspecifically-not by specific antigen recognition on the part of the dividing host cells, and ( 3 ) serving as the principal target of immunologic attack by donor cells. Even though it is difficult to dissociate these three facets of host lymphoid cells activity, they will be considered separately. The study of lymphocyte trapping has only recently come into vogue as a unique component of immunologic research. It has been recognized for some time that lymphocytes within a node engaged in immunologic activation with an antigen are retained within that node and unable to leave via the efferent lymph for several days, whereas an unstimulated node readily releases its complement of normally recirculating cells (Hall and Morris, 1965; Ford, 1969; Ford and Marchesi, 1971). Lymphocyte trapping depends on the route of introduction of the antigen. After intravenous injection of antigen, increased numbers of isologous 51Crlabeled lymph node cells in the blood localize in the spleen, but the lymph nodes draining subcutaneously administered antigen or a skin graft show the greatest amount of label (Zatz and Lance, 1971). In this system, trapping is a transient phenomenon, lasting only about 24 hours. The authors were able to demonstrate an acceleration in the time of maximal trapping when second-set allografts were applied. Additional documentation of specificity was shown by the adoptive immunization of mice to two different red cells, chicken and sheep, by using lymphocytes that were specifically sensitized to one or the other antigen and then labeled with different isotopes. When the animal was challenged in the foot pad with one of the antigens, its draining lymph nodes selectively trapped the appropriate labeled cells ( Thursh and Emeson, 1972 ) . In two different systems, it has been demonstrated that lymph nodes and spleens involved in GVHRs may also promote the trapping of recirculating lymphocytes (Emeson and Thursh, 1973; Zatz and Gershon, 1974). Although the stimulus is clearly immunologically specific, in GVHRs of this type it is apparent that the trapped host cells are not retained in the involved organ because of their antigen-reactive nature. The implications of this will be fully discussed later. Lymphoid cell proliferation is perhaps one of the most universal features of GVHRs. In animals that are not immunosuppressed, but have normal inflammatory responses and lymphocyte levels at the time of donor lymphocyte injection, the ability of the host’s lymphocytes to divide at a higher than normal rate is remarkable, even in genetically tolerant hosts, That the dividing cells within the spleens of newborn mice with GVH disease are predominantly of host genotype, especially late in the course of the disease, is a startling but consistent finding


165

(Davies and Doak, 1960; Fox, 1962; Zeiss and Fox, 1963; Nowell and Defendi, 1964). Hilgard ( 1970a) has shown that the splenomegaly in newborn F, hybrid mice can be abolished by 500 R of X-irradiation, but may increase with 300 R. However, the animals receiving 500 R and parental strain lymphocytes developed GVH disease sooner than those given 300 R, and it was concluded that host proliferative responses (e.g., splenomegaly ) are not essential for the development of GVH disease. By the preirradiation of newborn recipients, Howard et al. (1961) were able to increase the reactivity of spleen cells from the “primary” F, host when they were injected into a “secondary” F, host (isologous with the first). They reasoned that there were proportionately more of the original donor cells in the second inoculum, since they had suppressed the host’s nonspecific proliferative response, effectively diminishing the overall host cell contribution. While it remains an unresolved facet of GVH reactivity, host lymphoid cell proliferation apparently exceeds the expanse of virtually all proliferative responses that are immunologically specific. Although it has been suggested that, even in F, hybrids, there is a response to the donor lymphoid cells ( Ramseier and Lindenmann, 1971), it is unlikely that such a putatively minor antigenic stimulus could evoke a response of host cell proliferation as massive as that of the host in GVHRs (see Section VII,B,4,c). There is little apparent specificity to the local accumulation of host mononuclear cells in response to donor cell-induced GVHRs, and it is likely that these host cells are of bone marrow origin (Bonney and Feldbush, 1973). Indeed, the mononuclear cells that normally constitute the majority of the phvsical lesions of delayed hypersensitivity reactions are bone marrow derived and do not have specific immunocompetent status (Volkman and Gowans, 1965; Lubaroff and Waksman, 1968; Spector and Willoughby, 1968). Streilein ( 1971a) has proposed that the immunologic destruction of host lymphoid and hematopoietic cells is one of the cardinal features of GVH diseasc. This common aspect of the disease seems to be an especially reasonable correlate since, host lymphocytes are not only superlative samples of cell-bound transplantation alloantigens, but they are the cells most likely to be encountered by donor lymphocytes. The activation/ destruction of lymphocytes might lead to the release of “lymphotoxini’ and other nonspecific agents of tissue destruction ( Elkins, 1971a ), which could then mediate the “innocent bystander” effect. The anemia and leukopenia associated with many forms of GVH disease (Oliner et al., 1961; Billingham et al., 1962; Streilein and Streilein, 1972) may arise not only from the direct destruction of circulating red and

166


white cells, but also from the suppression of the hematopoietic stem cells that normally maintain the physiologic levels of blood cells (Garver and Cole, 1960; Davis and Cole, 1967; Blomgren and Anderson, 1972s; Streilein and Streilein, 1972). As discussed before, the release of toxic effector molecules by both donor and host lymphocytes could be expected to cause tissue destruction of cells within the immediate vicinity of the confrontation. This might well account for the damage to skin in hamsters ( Streilein and Billingham, 1970b,c). C. ALTERATION OF HOSTIMMUNOLOGIC CAPABILITY In many respects, GVHRs are replete with apparent paradoxes. NO where is this more evident than in studies on the effect of the GVHR on host immunologic capability. Like a novice in a handball court, the animal with GVH disease can anticipate being bombarded from all directions, manifested by increased or decreased immunologic responsiveness and, depending on the circumstances, sometimes both simultaneously. For example, it has long been known that F, mice in the throes of a severe GVHR display severely impaired cellular and humoral responsiveness (Howard and Woodruff, 1961; Blaese et al., 1964; Lapp and Moller, 1969). However, at a time when an F1 mouse displays critically suppressed antibody responses to SRBC and a lowered delayed hypersensitivity reaction to keyhole limpet hemocyanin (KLH), as a result of injection with parental lymphoid cells, it is also highly resistant to an intravenous inoculation of Listeria monocytogenes ( Blanden, 1969) or to Diplococcus pneumonine and Salmonella typhimurium ( Cooper and Howard, 1961). Although this dichotomy can be explained by the preferential activation of macrophages, especially peritoneal cells ( Blanden, 1969), and the possible depletion of a thymic mediator (see below) ( Lapp, 1973), there are clearly more complex shifts in immunoreactivity that can be discerned when cellular interactions are examined beyond the level of in vivo models.

1 . The Allogeneic Efect From a chronological point of view, during the course of GVH disease there is a fundamental biphasic change in the immunologic status of the host. Initially, there is a prominent stimulation of host B cells by donor T cells, which is as yet incompletely understood despite a truly impressive collection of experiments. This fascinating phenomenon, called the “allogeneic effect” because it was first described in guinea pigs given allogeneic lymphocytes ( Katz et al., 1971), has established a new partnership between studies of GVHRs and of cellular interactions. The result of this “graft T cell-versus-host B cell” reaction is that host


167

cells primed with specific hapten synthesize specific antibody in considerable quantity. The relevant literature has been discussed in great depth elsewhere (Katz, 1972). If all aspects of alteration of host immune status by allogeneic lymphocytes in a GVHR might be drawn together under the generic umbrella of “allogeneic effect,” then it might be possible to approach the issue of syngeneic T-cell and B-cell interactions in general by an investigation of the mechanism by which T cells stimulate (or suppress) allogeneic R cells. Whether the allogeneic effect factor and the T-cell helper factor are the same is not clear. Nonetheless, it has at least been established that an allogeneic effect does not interfere with the “normal” cooperation between syngeneic T and B lymphocytes ( D . H. Katz et al., 1974a). Subsequent to the initial demonstration of the allogeneic effect in guinea pigs, it has been further studied in mice (Osborne and Katz, 1973a,b; Katz and Osborne, 1972; Katz, 1972) and rats (McCullagh, 1972; Ornellas and Scott, 1974; Ornellas et al., 1974; Scott and Ornellas, 1974) and perhaps even demonstrated in man (Daguillard et al., 1973). Although the allogeneic effect was first reported in primed animals, it may also provoke primary sensitivity against antigens that are either nonimmunogenic ( Hamilton and Miller, 1973) or tolerogenic ( Osborne and Katz, 1973a). While the allogeneic effect increases only the IgM (but not the IgG) response to DNP coupled to polysaccharide and to protein backbones in unprimed animals, it apparently triggers diverse B cells, since isoelectric focusing of serum reveals considerable antibody heterogeneity ( K l a ~and McMichael, 1974). This heightened reactivity may have a kind of specificity, but it is not the augmentation of true “immunocompetence” in the normal sense. Nevertheless, it does constitute a form of “turned-on” immunity that may account for the increased immunoglobulin found in some (Streilein and Stone, 1973), but not all (Koltay et al., 1965), GVH sera. The allogeneic effect can also overcome the requirement for a carrierspecific T cell in so-called “nonresponder” mice (Ordal and Grumet, 1972). In an extensive series of experiments entailing the abrogation of unresponsiveness to SRBCs in rats, it has been shown that the “tolerant state” is maintained by the suppression of reactive immunocompetellt cells. These cells cannot be derepressed by the injection of tolerant rats with normal syngeneic lymphocytes, but only with allogeneic cells, even if the allogeneic cells are tolerant of SRBCs ( McCullagh, 1972). Similarly, GVHRs may at least partially block the induction of unresponsiveness in both T cells and B cells in rats (Ornellas et al.,1974). Collectively, these experiments suggest that lymphocytes stimulated by alloantigens in a GVHH may produce a soluble mediator that can

168

STEPHEN C. GREBE AND J. WAYNE STFIEILEIN

stimulate host B cells and, in some cases, host T cells (Ellman et al., 1972; Katz et al., 1972). The allogeneic effect is contingent on the presence of theta-positive allogeneic cells and does not occur in the context of a HVG reaction (Cheers, 1973). It may operate, at least in one system, when mitomycin C-treated allogeneic lymphocytes are used ( McCullagh, 1972), suggesting that the factor may have many of the characteristics of a lymphokine. In fact, one study led its authors to conclude that IgE B cells and IgG B cells are influenced differently by the allogeneic effect, and the relative hapten-specific stimulation or depression of IgE B cells or of IgG B cells varies with the numbers of allogeneic cells used (D. H. Katz et al., 1974b). The variant kinetics indicate that the two subclasses of B cells used have certain intrinsic differences, according to the authors, and do not point either to more than one T-cell population or more than one T-cell factor in the stimulation of the two B-cell subclasses. In vitro studies of the allogeneic effect have shown that soluble factors may be involved. Supernatants from allogeneic lymphoid cell mixtures may replace T cells or T-cell factors in nude mice (Kettman and Skarvall, 1974 ) , adult thymectomized, irradiated, and bone marrow-reconstituted mice (Britton, 1972) and anti-theta-treated spleen cell cultures (Arnierding and Katz, 1974). Such supernatants may reconstitute both primary and secondary B-cell antibody responses in vitro ( Armerding and Katz, 1974). These authors used a soluble, heat-labile, and somewhat trypsinsensitive factor, with a molecular weight between 30,000 and 40,000, which they have designated as the “allogeneic effect factor.” It is obtained by the in vitro admixture of irradiated allogeneic spleen cells and syngeneic thymocytes, the latter having been activated in vivo in an irradiated, syngeneic recipient by irradiated, allogeneic spleen cells. It is interesting to note that the stimulatory, or T-cell replacing, factor could be obtained only after 24 hours in culture, and prolonged stimulation beyond that time led to the production of inhibitors (Armerding and Katz, 1974). An additional finding of considerable significance in the paper, is that, while the allogeneic effect factor could replace T-cell helper activity in B-cell populations allogeneic to the strain in which the factor was produced, it was most effective in purified form tested on B cells with the same histocompatibility type as the T-cell population of origin. Subsequently, it was shown that the allogeneic effect factor may share specificities in common with the Ia determinants of the H-2 complex ( Armerding et al., 1974). In a different system, evidence had been presented that such a factor may have only limited strain specificity, since both parental and F, spleen cells coiild be positively influenced by supernatants from an in vitro allogeneic effect in their ability to

GRAFT-VERSUS-HOST

REACTIONS

169

respond to SRBCs (Moller and Coutinho, 1973). Whether or not the allogeneic effect factor acts with any significant specificity, it is clear that it is only one of the two prerequisite signals leading to antibody production, the other cardinal element being the presence of specific antigens, engaged by the B-cell receptor (Britton, 1972; Armerding and Katz, 1974; Trenkner, 1974).

2. Host Immunocompetence during GVHRs From the discussion just concluded, it is apparent that profound and powerful factors are set in motion by GVHRs. The intimate and extensive involvement of the host’s immunologic apparatus, the widespread destruction of lymphoid tissue, and the elaboration of mediators via the allogeneic effect or the action of suppressor T lymphocytes, all conspire to modulate the ability of the host to display normal immunocompetence. Although certain effects would be expected to enhance the immunologic reactivity of the host as GVH disease progresses, there are increasing indications of suppressed immunocompetence. The most severe forms of the disease are accompanied by profound leukopenia, which would fully account for such immunoincompetence; nevertheless, significant depression occurs even before lymphocytes disappear. Blaese et al. (1964) have cited examples of this defect in immunologically runted animals. Prolonged survivals of allografts have been shown in adult F, hybrid mice grafted from 5 to 19 days after the injection of parental strain lymphocytes (Howard and Woodruff, 1961; Lapp and Moller, 1969). Moller ( 1971) has provided evidence that suppression may occur even when the donor lymphocytes have been presensitized to the test antigen (in this case, SRBCs), but that the suppression is dependent upon the genotype of the donor cells. He concluded that this suppression was promoted by antigenic competition, in which the antigens displayed by the host assumed immediate priority. Evidence advanced in another system suggests that the antigenic competition in GVH disease may be due to a nonspecific inhibitor produced by the allogeneic donors (Sjoberg, 1971). Zaleski and Milgrom (1973) have gone so far as to use the suppression of an anti-theta response in an F, host as a measure of the severity of the GVHR. The data now available do not allow the construction of a concise model for GVH-induced host immunosuppression. It does clearly involve host lymphoid cells and seems to be more complex than the simple destruction of host lymphocytes or lymphopoietic tissues. Lapp ( 1973) has postulated that GVH disease results in the depletion of some “nonantigen specific thymic mediator,” but the role of such a factor (o r

170


the effect of its absence) is not clear at this time. It is tempting to speculate that the production of an antibody or some immunoregulatory protein during the course of the disease might be implicated in generalized immunoincompetence, despite the fact that there is no apparent effect of antibody in the pathogenesis of the disease. In GVHRs, the two extremes of immunologic expression are represented by the “allogeneic effect”and “immunologic incompetence.” While the allogeneic effect (Katz and Benacerraf, 1972) is a transient phenomenon and not a true manifestation of immunocompetence per se, but rather a heightened form of existent immunological capability, there are nevertheless some important clinical features of the GVHR that can be explained by it. As noted above, the accumulation of large numbers of host lymphoid cells in the lymph nodes and spleens would provide the cellular elements necessaiy to allow an allogeneic effect to occur, since donor T cells trigger the synthesis of antibody by host B cells that have been previously immunized to experimental or environmental antigens. The appearance of autoantibodies in the course of GVH disease in F, hamsters may be a reflection of an allogeneic effect (Streilein and Stone, 1973; Streilein et al., 197513). Similarly, both the appearance of Coombs antibodies of host origin and a marked stimulation of 7 S mouse anti-SRBC antibody in mice with GVHRs induced neonatally (Lindholm et al., 1973) suggest strongly the provocation of host B-cell activity by allogeneic donor lymphoid cells. At the other extreme, animals suffering from GVHRs become severely immunosuppressed (Howard and Woodruff, 1961). In most instances, this suppression occurs even in animals that are not lymphopenic. This depressed responsiveness may involve humoral immunity ( Moller, 1971) , cellular immunity (Lapp and Moller, 1969), or both (Streilein, 1972a; Treiber and Lapp, 1973). There is an apparent dichotomy in the ability of the hosts to recover some competence, since multiple immunization of the immunosuppressed animals evokes cell-mediated immunity, but does not produce humoral immunity to the antigens tested (Treiber and Lapp, 1973). One explanation of this discrepancy is that some aspect of the GVH process interferes with interactions of T cells with B cells through production of either an inhibitory cell or cell product, but does not modify T-cell functions. It is also possible that host B cells are destroyed preferentially when compared to host T cells in the GVH process, and recovery of T-cell function occurs more readily. Regardless of the mechanism of GVH-associated immunosuppression, it may provide an important, although not yet apparent, link to the understanding of immunoregulation. As many of the studies of altered host immunocompetence have shown,

GRAFT-VERSUS-HOST

REACTIONS

171

any given experiment may provide documentation of both heightened and suppressed immunoreactivities depending on cell dosage, time aftcr GVHR initiation (Lapp et aE., 1974), or time span of in uitru exposure of T cells to allogencic antigens ( Armerding and Katz, 1974). More penetrating experiments may detcrmine that many factors are involved; it is also possible that the allogeneic effect (an early event) and immunosuppression ( a later occurrence) may be mediated by a single factor, the activity of which is primarily concentration dependent. VI. Experimental Modification of Graft-versus-Host Reactions

A. GENERAL CONSIDERATIONS From the previous sections it is clear that GVHRs are dictated by donor lymphocytes as well as by the host. Consequmtly, alteration of a GVHR can be considered at several levels. In the first place, an experimental maneuver may bc undertaken to modify the donor of the attacking cells prior to their removal for injection; moreover, donor cells may be subjected to experimental manipulation in oitro that materially influences their further reactivity. In the second place, modification of the GVHR may be applied to the host either while the reaction is developing or when it is clinically manifest. The kinds of experimental manipulations that can be brought to bear may be classified as immunologic (such as exposure to ALS) or nonimmunologic (treatment with a pharmacologic agent such as mitomycin C ) . In addition, immunologic maneuvers may attempt to influence the GVHR in an antigen specific way, or they may be more broad-based in action such that an entire component of the immunologic apparatus is disrupted. For the purposes of this review, modification of GVHRs will be considered in light of altered donors or donor cells and altered hosts after initiation of the GVHR.

B. REDUCTION OF GVH POTENTIAL BY MANIPULATION OF DONOR AND/OR DONORLYMPHOCYTES 1 . General Measures Attempts to remove GVH reactivity from an inoculum of donor lymphocytes are either nonspecific or specific with regard to the immunologic character of the deficit; that is, at least those T lymphocytes capable of inciting GVHRs and probably many additional cells are removed or, alternatively, only the specifically involved H-ARCS and their progenitors are extracted. The latter situation would ultimately be the most desirable procedure in the case of human bone marrow transplants, pre-

172


venting any potential GVHRs but allowing the transfer of the widest possible spectrum of immunocompetence for the already immunosuppressed recipient. However, designers of the majority of experimental and clinical protocols have been forced to elect a more broad-ranging course, depleting putative cellular inocula of most or all T cells. Experimentally, such treatments as neonatal thymectomy or lethal whole-body irradiation of donors have been widely used (Miller and Osoba, 1967; Billingham, 1968; Elkins, 1971a), but they demonstrate only the mechanical basis of GVH induction and have no clinical value. While techniques to eliminate H-ARCS specifically are only now coming into vogue, the field of “immunosuppression” in GVHRs has been extant since the earliest studies of GVH disease (Simonsen, 1962a; Billingham, 1968). As will be apparent in the following sections, the majority of these techniques are distinguished by their singular lack of specificity, often imposing their effect more by “shot gun” overkill than by the incisiveness of a sharpshooter. 2. Alteration of GVH Reactivity by Treatment

with Pharmacologic Agents The recent literature on immunosuppression describes an abundancc of drugs that depress one or another important aspect of immune responsiveness. With regard to the extinction of GVHRs in experimental animals by pretreatment of donor animals or donor cells, there have been a number of important acquisitions to the pharmacopeia, as well as some new insights into the mode of action of more commonly used reagents ( Floersheim, 1972,1973; Levy, 1973). In this latter category, cyclophosphamide has come under increased scrutiny. Although used extensively to inimunosuppress recipient animals for successful allograft “takes,” to provide acceptable, unresponsive allogeneic recipients for GVHRs (Santos and Owens, 1966; Owens and Santos, 1971, Sandberg et al., 1971), or to treat F, recipients after administration of allogeneic cells (Owens and Santos, 1971), it has also been used to decrease the GVH reactivity of spleen cells by pretreatment of the donors (van Winkel, 1971; Fink and Cloud, 1974). In the latter report, maximal suppression of the GVHR (indicated by the longest survival of lethally irradiated rats injected with spleen and bone marrow cells from parental donors ) was procured when cyclophosphamide-treated donors were “sensitized with recipient histocompatibility antigens 1 day later, and 6 days before harvesting donor cells (Fink and Cloud. 1974). Presumably, the H-ARCS responding to the sensitized alloantigens were most susceptible to the cytotoxic effect of cyclophosphamide at this time.


173

However, it is now apparent that cyclophosphamide may act differentially on T and B lymphocytes, with important immunoregulatory consequences. After three intraperitoneal doses of cyclophasphamide on alternate days, spleen and lymph nodes of treated mice showed changes in the proportion of theta-positive lymphocytes, from 25%and SO%,respectively, to greater than 90% (Poulter and Turk, 1972). More important, when cyclophosphamide-treated animals are tested for delayed hypersensitivity and antibody to such antigens as ovalbumin (Turk and Parker, 1973; S. I. Katz, et al., 1974a,b), SRBC (Lagrange et al., 1974b) and histocompatibility antigens ( Kerkhaert et al., 1974), after sensitization either directly or in adoptive transfer experiments, their antibody responses are depressed and cell-mediated immunity is concomitantly increased. Thus, antibody appears to wield an immunosuppressive influence over certain forms of T-cell directed responses. Under these conditions, cyclophosphamide may prevent the formation of enhancing antibodies (Stockman et al., 1973). The implications of this finding will be discussed in Section VII. The administration of certain drugs to animals donating lymphoid cells that are placed in suspensions may either augment or suppress the GVH reactivity of the animals, depending on the dose of reagent, time and frequency of administration, and accompanying manipulations. For example, when similar numbers of hydrocortisone-treated and normal parental spleen or bone marrow cells were infused into adult F1 mice, the first were the most active in promoting splenomegaly (Cohen et al., 1970). On the other hand, steroid treatment of recipients with GVH disease successfully aborted the disease only when given at the same time as donor lymphocytes (Schwartz and Beldotti, 1965). Many studies have focused on the influence of a drug when given to the host after the inoculation of donor cells, so that any effect seen may be against the donor cells, the host, or both. In a number of systems, pretreatment of donor lymphoid cells with mitomycin C has effectively inhibited subsequent GVHRs ( Meuwissen and Good, 1967; Lemmel and Good, 1969, Grebe and Streilein, 1975a). In a recent report (Scollay et al., 1974) it was suggested that mitomycin C-treated parental lymphocytes may produce splenomegaly in appropriate F, recipients if the drug-treated donor inoculum is admixed with F, cells prior to injection. Thus cells that are unable to divide may nevertheless elaborate whatever factors are involved in splenomegaly, providing that the donor cells are exposed to the alloantigen at the time of or before injection. An alternative explanation is that mitomycin C aIters the traffic or homing pattern of donor lymphocytes, so that they do not reach the spleen in the same fashion as normal lymphocytes. This

174


explanation seems unlikely as niitomycin C-treated parental lymph node cells, injected into the hind foot pads of allogenic F, rats, were unable to elicit PLN hypertrophy (Grebe and Streilein, 1975a)-a system that would obviate the requirement that injected cells must travel through the peripheral blood. In clinical bone marrow transplant, however, mitomycin C treatment of bone marrow cells would almost certainly block cell division in hematopoietic stem cells as well as in the H ARCS. 3. Alteration of GVH Reactivity with Plant Lectins

The vast majority of pharmacologic agents used in experimental or clinical subjects as a means of attenuating GVHRs act primarily on dividing cells, thus placing at risk of destruction not only the undesirable GVH effectors and their progenitors, but also potentially beneficial hematopoietic stem cells. However, recently much attention has been drawn to the putative immunosuppressive properties of a well known lymphocyte mitogen, concanavalin A (Con A ) . The capacity of this plant lectin to stimulate T-cell proliferation in vitro has been widely documented (Anderson et al., 1972; Greaves and Janossy, 1972; Stobo, 1972; Stobo et al., 1973), but it is a novel notion that inmunoresponsiveness can be specifically depressed in animals treated in viuo with 100-1000 pg of Con A (Nirmul et al., 1972; Anaclerio et al., 1974; Davis and Boone, 1974; Egan et al., 1974). Doses over 500 pg may be lethal, but effective immunosuppression of skin graft rejection (Nirmul et al., 1972), delayed hypersensitivity to tuberculin (Davis and Boone, 1974) and antibody to SRBCs, but not to T independent lipopolysaccharide (Egan et al., 1974), could be demonstrated with little or no host mortality. Interestingly, while Con A was equally effective in suppressing antibody responses in normal and sensitized mice (Egan et al., 1974), it conferred a more pronounced suppression on the delayed hypersensitive response of sensitized mice than on normals (Nirmul et nl., 1972; Davis and Boone, 1974), suggesting that long-lived memory T cells are more susceptible than unprimed T-cells. Under some circumstances, suboptimal mitogenic doses of Con A may activate cells to become cytotoxic effectors, after either in vivo administration (Anaclerio et al., 1974) or in vitro cultivation. However, in vitro mitogenic concentrations of Con A (1-5 mglml) almost totally suppress CMC (Stavy et al., 1971; Peavy and Pierce, 1974). The addition of 10%Con A-activated, theta-positive cells to a mixed lymphocyte-sensitizing culture was sufficient to abrogate the CMC of the aggressor cells, and the inhibitory impact was effective only after the initiation of the response to alloantigens (Peavy and Pierce, 1974). Fortified by such intriguing possibilities, Ledney (197213) and Tyan


175

(1974) have shown that the CFU of hematopoiesis are less susceptible to the toxic effect of Con A in uivo or in uitro than the immunocompetent GVH-inducing cells. One hundred micrograms of Con A given intraperitoneally to allogeneic donors of spleen and bone marrow cells, or 500 pG/ml of the mitogen incubated with the donor spleen and bone marrow cells at room temperature for less than an hour, reduces the 100%mortality of lethally irradiated F, mice, which would otherwise succumb to GVH disease. While the mechanism by which Con A may selectively influence these cells is unknown, possibly they are stimulated into a kind of “sterile activation,” as was originally suggested by Medawar (1969) to explain the mechanism by which ALS might act in an immunosuppressive manner. Alternatively, it might act by some form of direct cytotoxicity (Tyan, 1974). 4. Treatment of Donors nncl Donor Cells with Lymphocyte Reactive Antibody In human clinical bone marrow transplants, many protocols utilize heterologous ALS or antihuman thymocyte globulin ( ATG ) , primarily administered to the patient bcfore and after transfusion of donor marrow (see Section II,D) (Congdon, 1971; Mathk et al., 1974; Thomas, 1974). These sera are important both for suppressing any residual HVG reactivity and for confining potential GVHRs. There is some lingering controversy as to the mechanism b y which ALS suppresses T-cell-mediated immunity, but it is unlikely that it acts by “blindfolding” T lymphocytes or simply stimulating them to proliferate uiiproductively by “sterile activation” ( Medawar, 1969). More probably, it may afford macrophages a better opportunity to phagocytose ALS-opsonized lymphocvtes, since the immunosuppressive index of the antiserum most closely correlates with its opsonic index in uitro and is less closely related to its mitogenic, agglutinating, or cytotoxic properties (Greaves et al., 1969; Saleh et al., 1969). The phagocytosis of ALS-coated lymphocytes is important to the immunosuppressive action; not only does the hepatic reticuloendothelial system take up large quantities of ALS-coated cells (Lydyard and Ivanyi, 1974), but treatment of skin grafts and recipients with antimacrophage serum or silica ( agents that subvert reticuloendothelial function) reverses the immunosuppressive effects of ALS ( Harris et al., 1971). These studies suggest that effective immunosuppression by ALS depends upon (1) the affinitv of lymphocytic antibody for target lymphocytes, and ( 2 ) the complete removal of opsonized cells by phagocytosis. In general, despite discrepancies among many in uitro assays of serum titer and in uiuo immunosuppressive potency, ALS that is effective in

176


prolonging skin or renal allograft survivals also effectively reduces GVH reactivity (Harris et al., 1971; Nouza et al., 1971; van Bekkum et al., 1972). A number of T-cell functions and characteristics have been dissected with the aid of heterologous ALS (Waksman, 1971), and it is now apparent that not all thymus-dependent lymphocytes are at equal risk from its suppressive or cytotoxic effects. For example, after chronic (4month) administration of ALS to mice ceases, the relative GVH reactivities of thymus, spleen, and bone marrow cells are depressed for 2 days. By 7 days afterward, the thymus returns to normal, control levels; however, the actvity of the spleen and bone marrow remains depressed even at this time (Burleson and Levey, 197213). ALS treatment of donor mice may effectively exhaust the capacity of their cells to cause GVH disease (Brent et al., 1967; Ledney and van Bekkum, 1968; Ledney, 1969). However, because in vivo admixture of parental spleen cells and ALS destroys not only lymphoid cells, but also colony-forming stem cells, papain-digested ALS, consisting of individual Fab fragments have been used with suppression of GVH, but retention of hematopoietic activity (Gallagher et al., 1972; Richie et at., 1973). Whether or not the ALS Fab-coated lymphocytes are phagocytized as are F( ah),-pepsin-digested, ALS coated cells ( Lydyard and Ivanyi, 1974) is not known. Nevertheless, 0.84 mg of Fab per lo6 parental spleen and bone marrow cells was capable of ablating acute GVH disease when treated cells were injected into lethally irradiated mice, which would otherwise all have been dead by 9 days after injection of cells (Richie et al., 1973). Similarly, ALS absorbed first with spleen cells substantially reduced GVH disease by suppressing cell-mediated immunity, but without cytotoxic effects on CFU of bone marrow (Trentin and Judd, 1973).

5 . Physical Means

Reducing GVH Reactivity of Donor Cells The use of density gradient cell centrifugation, which may ultimately be of enormous clinical value, is typical of the broad-stroke approach to immunosuppression. In many attempts to associate GVH reactivity with a specific cellular buoyant density, linear gradents have been used (Szenberg and Shortman, 1966; Shortman, 1968), but the majority of studies have depended on discrete bovine serum albumin (BSA) gradients (Dicke et al., 1968; Guyer et al., 1970; Phillips and Miller, 1970; Dicke and van Bekkum, 1971; El-Arini and Osoba, 1973; Argyris, 1974; Mandel and Asofsky, 1974). In general, it has been possible to eliminate GVH-reactive T-cells, which are of relatively higher density than CFU hematopoietic stem cells (Phillips and Miller, 1970) or rosette-forming of

GRAFT-VERSUS-HOST

REACTIONS

177

cells (Argyris et d., 1972; Argyris, 1974). This dissociation of GVH H-ARCS and rosette-forming cells would suggest that removal of the latter from bone marrow would not be a fruitful means of preventing GVHRs. Levey and Burleson ( 1972) cxteiisively characterized thymocytes with regard to their buoyant density and GVH-inflicting potential. Sedimentation of thymocytes on a 17-35% discontinuous BSA gradient revealed that cells of similar densities could both produce splenomegaly in neonatal F, mice and home to the peripheral lymph nodes of a normal syngeneic mouse. However, thymocytes which spontaneously incorporated tritiated thymidine in uitro, which responded to stimulation by PHA or were sensitive to the in vivo administration of ALS were not found in the GVH-inducing fraction. Therefore, based on cell density differences, there was an apparent dichotomy between the primitive T-cell characteristics ( PHA responsiveness and ALS sensitivity) and the manifestations of mature immunocompetence ( GVH-potential and homing capacity), Although the bone marrow of mice contains relatively few thta-positive cells (Raff, 1969, 1971), and bone marrow cells are relatively less effective for initiating GVHRs than spleen, lymph nodes, or peripheral blood lymphocytes ( Billingham et al., 1962; Billingham, 1968; Elkins, 1971a; Yoshida and Osmond, 1971), clearly the incidence and severity of clinical GVH disease in human bone marrow transplantation indicate the presence of immunocompetent H-ARCS or their precursors within the hematopoietic stem cell pool (see Section 11,D). Thus, it has been possible to identify a fraction of higher density cells in the bone marrow which are enriched with GVH-producing cells. Using a linear sucroseserum gradient ( 5 to 15%), Yoshida and Osmond (1971) demonstrated a 5-fold increase in the potency of one fraction over unfractionated cells. When mouse bone marrow was applied to a discontinuous BSA gradient, a high density fraction, which contained only 20% of the original cells, produced nearly twice the splenomegaly in newborn F, mice than was obtained using a lower density fraction ( Burleson and Levey, 1972a). Although there is little substantial rationale for using such broad spectrum physical agents as irradiation to reduce GVH reactivity, there may be some potential in developing newer and more sophisticated tools that interact with immunocompetent cells through some specific cell feature. For example, a recent report has cited the efficacy of ultraviolet light irradiation in decreasing such T-cell functions as GVH reactivity and PHA responses of human and murine lymphocytes (Horowitz et al., 1974). The process is geared to the differential susceptibility of T and B lymphocytes to a given dose of ultraviolet light. Similarly, it may be possible to utilize the fluorescent cell sorter (Jones et al.,

178


1973) to remove serologically detectable GVH-inducing cells from bone marrow preparations, although it would be necessary to expand considerably the capacity of the apparatus in order to handle large numbers of cells. Classically, the use of X-irradiation as an immunosuppressive agent has been widely accepted and used in a variety of immunologic contexts; GVHRs are no exception. The administration of whole-body irradiation to the recipients of allogeneic lymphocytes produces a somewhat complex picture of altered immunologic competence (see Section I1,D) ; however, the irradiation of donor lymphoid cells or of the donors themselves prior to the harvesting of lymphoid cells has much less equivocal results (Simonsen, 1962a; McBride, 1966; Brent and Medawar, 1966a; HaSkovL and GansovL, 1970). Sublethal doses of irradiation may not completely suppress the GVH reactivity of lymphocytes, but doses of 1000 rads or :moretare almost totally effective in suppressing the antihost reactivity. Occasionally, lymphocytes from specifically sensitized donors are somewhat more radioresistant; nevertheless, they are susceptible to higher doses of irradiation. 6. Antigen-Specific Reduction in GV H Reactivity of Donor Cells

In an attempt to find a practical way of separating H-ARCS from nonreactive cells by in vitro physical means, Bonavida and Kedar (1974) have shown that normal and immune BALB/c spleen cells adsorbed onto C57 spleen cell monolayers with aid of light centrifugation could be almost totally exhausted of CMC activity. Furthermore, when 40 to 60 X lo8 of the nonadhering BALB/c cells were injected into lethally irradiated C57 recipients, 70%survived for longer than 6 months without signs of GVH disease, whereas BALB/c cells adsorbed onto BALB or unrelated SLJ monolayers were fatal. In similar work by Lonai and co-workers (Lonai et al., 1972; Wekerle et al., 1972, 1974 ) , populations of spleen cells were depleted of antigen-reactive cells by adsorption onto monolayers of allogeneic cells. In other laboratories attempts to remove specific GVH reactivity by adsorbing cells onto allogeneic monolayers were unsuccessful ( Mage and McHugh, 1973). Using monolayers of fetal fibroblasts or adult periotoneal macrophages as the allogeneic experimental substrate and using syngeneic controls, these workers showed that immune splenic lymphocytes that did not adhere to these monolayers incited normal GVH reactivity but had a reduced cytotoxic capacity against the allogeneic target cells. As a method for subverting GVH reactivity by the specific depletion of reactive cells, their experiments were unrewarding. They speculated that unadsorbed nonreactive cells may have acquired reactivity after injection into the neonatal F, recipients used in their GVH assay. How-


179

ever, take at face value, their data reflect the diverse functional composition of H-ARCS demonstrated in many other systems. Ford and Atkins (1971) developed an interesting in uivo analog for removing specific GVH-inducing cells from a parental inoculum injected intravenously into F, hybrid rats. I n their hands, cells obtained by the thoracic duct cannulation of such F, hybrids 24-48 hours after injection of parental cells intravenously were deficient in their ability to produce a GVHR when injected subsequently into a second syngeneic F, recipient. However, these “filtered” cells, in a mixture of parental and F, lymphocytes, retained specific reactivity to third-party recipients. Furthermore, the cells that were retained within the spleens of these primary “filter F,” recipients were specific H-ARCS and possibly constituted as high as 4-54: of the original parental thoracic duct lymphocyte population (Atkins and Ford, 1975; Ford et al., 1975). Thus, there is a selective depletion of antigen-reactive cells in the recirculating pool (represented by the thoracic duct lymph), and there is a marked enrichment of the antigen-reactive cell population in the first way station of the peripheral blood-the spleen. Cell division following specific stimulation by alloantigens is another functional characteristic of H-ARCS which gives them a higher profile than nonreactive cells. Although there are relatively few data available on the destruction of these susceptible dividing cells after their stimulation in uitro, some investigators have been successful. For example, when specifically alloantigen-activated lymphocvtes are cultured with BUdR and then exposed to light, the dividing cells that have incorporated the pyrimidine base analog are destroyed (Zoschke and Bach, 1970, 1971: Salmon et al., 1971; Clark and Ncdrud, 1974; Lazda and Baram, 1974; Merino et al., 1974). The remaining cells are thus depleted of antiallogeneic reactivity. An adaptation of the BUdR and light treatment has been used to kill H-ARCS in a mixture of parental and F, spleen cells, which were subsequently unable to incite GVH in the appropriate Fl hvbrid rat ( Rich et at., 1972). Similarly, the “hot-pulse” technique of Dutton and Mishell (1967) might be used to incapacitate dividing cells by the incorporation of radiolabeled niicleic acid precursors with extremely high specific activity, while nondividing cells would bc spared ( Moorehead and Claman, 1974).

7. Other Means of Reducing Donor Cell GVH Readiuity The literature contains a significant number of references to other agents that reduce GVH activity when donor lymphocytes are pretreated, either in liiuo or in uitro. The only feature common to them is that they defy easy codification, One of the most interesting of these is the so called “chalone” (Math6 et al., 1973), an alcohol extract of either

180


spleen (Garcia-Giralt et al., 1972) or thymus (Kiger et al., 1973). When parental bone marrow and lymph node cells are cultivated with one of these unrefined extracts, their capacity to incite fatal GVH disease in F, recipient mice is significantly suppressed. No good data are available as yet to account for this phenomenon, but its action apparently turns upon the preferential inactivation of immunocompetent cells over hematopoietic stem cells. When spleen cells in solutions containing dimethyl sulfoxide are carefully frozen, thawed, and administered to recipients, they are less active in their GVH potential than untreated controls (Bohm and RokosovA, 1972). This inhibition is apparently readily reversible, since normal GVH reactivity is restored when the cells are washed free of the reagent, signifying that this exposure of the cells to the chemical does not kill them. Neuraminidase-treated lymphocytes, when injected into susceptible hosts in relatively small numbers, are unable to mount a GVHR, although larger numbers of treated cells maintain their reactivity ( Adolphs, 1973). While the reason for this enzymatic change remains unknown, it has been shown that neuraminidase-treated lymphocytes do not home normally to syngeneic host lymphoid organs after intravenous injection ( Gesner and Woodruff, 1969). The increased phagocytic activity of the reticuloendothelial system and weight loss seen in “control” GVHRs, but not splenomegaly, were absent in F, recipients after the injection of donor mice with normal PHA, denatured PHA, or hemocyanin 24 hours before donor cells were harvested (Hunter et al., 1969). This subversion of GVH reacti-Jity may be related to the phenomenon of “antigenic competition” ( Liacopoulos et al., 1967); the effect of PHA was not related to its mitogenic activity. Bacterial endotoxin may have a variable effect on GVH reactivity, but is primarily a depressant ( Skopiiiska, 1972). Finally, when killed Corynebacterium pamum were given to CBA mice 7 days before cell harvest, the donor cells subsequently produced less splenomegaly in (CBA X C57) F, hybrid mice (Scott, 1972). The incomplete abrogation of GVH reactivity was only partly correlated with the total inhibition of PHA responsiveness, presumably because two somewhat different T-cell populations are sensitive to stimulation by mitogen-versus-allogeneic cells. C. ATTEMPTSTO ALTER GVHRs BY TREATING THE HOST Many of the maneuvers already described purporting to reduce the GVH potential of donor lymphocytes have been employed to abrogate an already extant GVHR in a susceptible host. From a theoretical standpoint, this type of experimenta1 design gives results that are difficult


181

to interpret because most agents that have any effect whatever exert that effect on donor lymphocytes as well as on the host. Thus, treatment of the host with GVH disease carries little hope of dissectirig components of the response that are donor in origin from those of the host. 1 . General Measures In clinical GVH disease, however, the situation is quite different. Theoretical considerations aside, the clinician is compelled to attempt whatever measures he can employ in order to rein in an ongoing, threatening CVHR. Thus, a variety of treatment protocols have been devised, generally on a strictly empiric basis, and although the spin-off in terms of new and enlightening information has not been overwhelming, nonetheless some interesting observations have been made. An attempt to review this aspect of GVH disease is beyond the scope of this article; excellent and comprehensive articles have appeared in the past few years (Math6 et al., 1974; Thomas, 1974). The data from these efforts can be briefly summarized as follows: whole-body X-irradiation, radiomimetic drugs, hydrocortisone, and antimetabolites ( which are used primarily in cancer chemotherapy) have all been employed in an effort to control an ongoing GVH disease. Almost without exception, control of the disease either is not achieved or the host dies of the consequences. It is fairly clear that the successful management of clinical GVH disease will depend upon the use of agents that have more specificity for the immunocompetent attacking donor lymphocytes, at the same time sparing the host’s own native specific and nonspecific defense apparatus. ALS and serum-containing antibodies directed at antigenic determinants on the plasma membrane of donor lymphocytes offer such an opportunity. Survival rates improved when animals with GVH disease were treated with ALS (Boak et al., 1968; Ledney, 1969; Corneci et al., 1971). However, in one notable exception, pretreatment of F, hybrid hamsters with heterologous ALS was completely ineffectual, regardless of the dose regimen, although in vivo or in vitro administration of the ALS to donor and/or donor cells could totally abort the systemic GVH disease (Shaffer et al., 1971).It is interesting that the best survival rates may depend upon relatively gennfree conditions, especially with regard to low intestinal flora, highlighting the important balance between immunosuppressing the GVHR and maintaining sufficient immunocompetence to thwart systemic infection ( Ledney, 1969) .

2. Host Treatment with ALS The origin of ALS may have a significant influence on the effectiveness of the serum, especially with regard to such adverse effects as toxicity

182


and infection. While some investigators have used sera raised in horses (Saleh et al., 1969; Gallaher et al., 1972; Trentin and Judd, 1973; Richie et aZ., 1973), ALS is most routinely raised in rabbits (Suvatte d al., 1968; Ledney, 1969, 1974a; Medawar, 1969; Harris et aZ., 1971; Shaffer et aZ., 1971; Burleson and Levey, 1972b; Kinnamon et al., 1973; Lydyard and Ivanyi, 1974). The results of one study, monitoring the benefits of rabbit ALS in renal transplantation, point out the lower toxicity and decreased incidence of infection when using rabbit instead of horse ALS (Davis et aZ., 1971). In a rhesus system, a comparison of the two types of ALS revealed prolonged skin graft survival in 55% of the animals receiving rabbit ALS, whereas only 12% of horse ALS recipients showed prolongation (Darrow et al., 1971). In summary, the composite experience with ALS suggests that it has inconstant advantages over other means to control GVHRs. In some instances, early and continuous treatment of recipient mice is beneficial in moderating runt disease (Suvatte et al., 1968) and secondary disease (Kinnamon et aE., 1973). Ultimately, its clinical applicability will depend heavily on its specificity in blunting a GVH attack while sparing hematopoietic stem cells. Even now there is no consistent agreement upon optimal source, dosage, and route of administration or period of application to achieve the most effective immunosuppression of the allograft reaction (Taub, 1970; Lance et al., 1973; Brendel et al., 1974). 3. Host Treatment with Alloantisera

Heterologus ALS may have relatively broad specificity and, when administered to recipients after the injection of donor cells, may react against both donor and recipient lymphocytes. In contrast, alloantibodies raised in recipients against donor T cells or T-cell antigens can drastically reduce GVH reactivity of donor cells when they are pretreated in uitro in the presence of complement. Successful abrogation of GVH reactivity has been achieved by using antibodies directed against the theta-antigen ( Cantor, 1972; Tyan, 1973) in mice, thymus-specific antigens in rats ( Lubaroff, 1973), and histocompeatibility alloantigens in both species (Kren et al., 1962; Russell, 1962; Silvers and Billingham, 1969; Clancy and Riecke, 1973). These last studies focused primarily on the efficiency with which alloantiscra against donor histocompatibility type raised in adult rodents syngeneic with GVH recipients could allow neonatal recipients to survive “runt disease,” even if they did not receive the serum until 13 days after birth and donor lymphoid cell inoculation (Silvers and Billingham, 1969). One of the most surprising findings in terms of surface antigens on the membranes of H-ARCS is that heteroIogous antisera with specificity

CHAFT-VEHSUS-HOST 1lEACTIONS

183

for whole-mousc serum (Tyan, 1971, 1973), mouse 7 S globulin (Colt and Maki, 1971), and nioiise light chain (Mason and Warner, 1970; Riethmuller et d.,1971) can specifically inactivate GVH-inducing cells when incubated with the donor cells prior to their inoculation. Similar abrogation of CAM pocks in chick embryos was shown when adult peripheral blood leukocytes werc pretreated with anti-light chain sera, and the activity of this antibody could not be absorbed with washed thymocytes (Rouse and Warner, 1972). Of special interest to clinicians is the finding that, although mortalitv from GVH disease could be reduced by the addition of anti-mousc- 7 S-y-globulin to parental lymph node cells (without complement), parallel studies with treated F, bone marrow injected into syngeneic recipicnts revealed that CFU were not eliminated by the anti-7-globulin antisera. Furthermore, such treatment of parental bone marrow cells given to lethally irradiated F, mice provided virtually complete protection against secondary disease ( Cole and Maki, 1971). Taken collectively, these corroborative data raise a fundamental question about the anatomic charactcristics of the T-cell surface. Although some laboratories have reportcd the presence of immunoglobulin on T-cell membranes (Marchalonis et al., 1972), others are unable to document this occurrence (Vitetta et al., 1972). If antibodies directed at immunoglobulins or immunoglobulin components such as light-chains can abrogate a widely acccpted T-cell function ( GVH reactivity), what role might these “surface” immunoglobulins play in promulgating that T-cell function? Are immunoglobulins indced a composite part of that intricate mosaic, the T-cell membrane, or arc they merely absorbed transiently? Whatever the origin of their association to the T lymphocyte, are immunoglobulins a fundamental component of the T-cell function, either in antigen recognition or iminunoregulation? These questions have not as yet been answered satistactorily . VII.

lmmunoregulation and the Graft-versus-Host Reaction

I n the past 10 vears, rescarch in transplantation and immunology have established bevond doubt the cellulur basis of the immune response. A great deal of information has been gathered concerning thc events by which the immunologic response is initintecl. Similarly, much is known about the phenomenon of allograft destruction and the fatal consequences of GVH disease wherein spccific CMC and inflammatory and nonspecific effector processes interact. A major deficit in understanding exists for the processing and differentiating events that proceed from the initiation of the response. It is obvious that physiologic mechanisms

184


have evolved which ensure that an immunologic response to a particular antigenic challenge is brisk, specific, of appropriate magnitude and duraton, and conducted to the host’s advantage. This finely orchestrated response is controlled by processes that we only vaguely perceive; it is assumed that these processes are operative in immunologic responses with the vigor of GVHRs. This section of the review deals with insights into the process of immunoregulation which have come or might be expected to come from the study of experimental GVHRs.

A. SPONTANEOUS ALTERATIONSOF GVHRs

AS

MANIFESTATIONS OF IMMUNOREGULATION

Based on immunogenetic considerations, one can make certain predictions about GVHRs that would follow from current immunologic dogma. In a host that is immunologically defenseless, a GVHR, once evoked, should continue until the host is consumed. As an example of cell-mediated immunity, it should be possible to passage GVH disease serially with lymphoid cells from affected to unaffected hosts for an indefinite number of transfers. Until the host is destroyed, it should be possible ever to increase the intensity of the disease by the further addition of donor lymphoid cells. These three predictions are particularly provocative because they have not been borne out experimentally. Many animals afflicted with GVH disease may survive the acute phase of the diseas’e and go on to apparent clinical recovery. Animals that have recovered from acute GVH disease are peculiarly resistant to a second GVH-inducing challenge. Although the GVHR can be adoptively transferred primarily from an affected animal to a fresh recipient, secondary transfers are much less successful, and further transfers are virtually without effect. Investigations into these paradoxes of GVH disease have been useful in highlighting potential regulatory influences brought into play during the course of response. 1. Natural Resolution of GVH Reactivity In the majority of local GVHRs whether measured as a palpable lesion in the skin, as the intensity of donor lymphocyte division in the kidney, or as the weight increase of an involved lymph node, inevitably the response subsides. There may be a residual lesion, as marked by the appearance of a necrotic nodule in the skin (Streilein and Billingham, 1967) or a histologically distinct “invasive-destructive” lesion in the kidney parenchyma (Elkins, 1964). Similarly, after their striking onset, PLN hypertrophy and the latent accompanying splenomegaly, subside within 4 weeks of initiation (Grebe and Streilein, 1975a). In all these instances,

GRAFT-VERSUS-HOST

REACTIONS

185

the reaction does abate, and frequently there is no further evidence of GVH reactivity. Likewise, there are numerous examples in which systemic GVH disease does not result in the death of the host but becomes attenuated. In hamsters, as already cited (see Section II,B), the inoculation of 200 X loGspecifically sensitized parental lymph node cells intravenously into F, hybrid recipients never incites lethal epidermolysis, although this syndrome regularly follows intracutaneous inoculation ( Streilein and Billingham, 1 9 7 0 ~ )The . experiments by Fox ( 1966) showed that several weeks after the onset of GVH processes in F, hybrid mice from the injection of parental spleen cells, there was severe lymphoid tissue depletion and subsequent repopulation; the repopulating cells might have been either donor or host in origin. Nevertheless, the GVHR, after running a violent course, eventually subsided. A singlular form of abrogated GVH reactivity, which could not be attributed to the permanent inactivation of elimination of donor cells, as demonstrated in some F, hybrid mice that had received several hundred million parental spleen cells but did not develop initially any signs of “homologous disease.” However, 13 months later, after receiving 400 R of whole-body X-irradiation, 6 out of 7 animals developed typical signs of GVH disease, 4 of them dying within 2 weeks of the reactivating dose of irradiation ( Schwartz and Beldotti, 1963). This quiescent form of GVHR could not be demonstrated in any of the control animals, suggesting that the original donor cells were directly responsible for the reawakening of GVH reactivity. As an extension of this work, Schwartz and Beldotti (1965) also showed that 6-mercaptopurine and Xirradiation could increase the severity of the GVH disease when administered either before or after parental lymphocyte injection. Thus, there are numerous examples of GVHRs that subside inexplicably and apparently without a concomitant “HVG’ reaction of sufficient magnitude to destroy donor H-ARCS. 2. Inability t o Passage GVHRs Serially

Another expectation that would follow from the basic premises about GVH reactivity is that it should be possible to transfer the reaction from one affected recipient to another syngeneic animal by the injection of the secondary host with lymphocytes from involved lymphoid organs of the Grst host. In virtually all experiments of this type, serial passage of GVHR has been limited to one or two transfers at most. Although in an early experiment (Simonsen, 1957) a GVHR was passaged through multiple hosts, it was later shown to be due to different populations of immunocompetent allogeneic lymphocytes accumulated along the

186

STEPHEN C . GREBE AND J. WAYNE STREILEIN

way, because the successive recipients were not members of an inbred strain, and the allogeneic interaction accelerated immunologic maturation ( Simonsen, 1965; McBride et ul., 1966). Some early attempts to passage GVH reactivity provided evidence that within 24 hours after injection into newborn F,s, parental lymphocytes lost their capacity to incite GVH reactivity in a secondary, isologous F,, but could still mount a response in a third-party host (Simonsen and Jensen, 1959). In another experiment, although GVH reactivity was not tested beyond a single passage, when spleen cells from 8-day-old mice, which had received 4 to 10 x lofi adult allogeneic spleen cells at birth, were injected into normal newborn mice syngcneic with the initial recipients, 74% of them succumbed to GVH disease (Siskind et al., 1960). Using reciprocal combinations of C3H and the notoriously GVH-resistant C57 newborns, Dineen (1961) was able to passagc GVH reactivity with spleen cells from intravcnously inoculated neonatal hosts only once at the most, and in some cases even the primary passage wasunsuccessful. By using multiple ( two or three) spleen donor-equivalents, 15%runting has been shown in tertiary hosts, but only when spleens from less severely affected runts were used in that second transfer; no runting was found in newborns given a third transfer (Jutila and Weiser, 1962). In a heroic attempt to transfer GVH disease from newborn rats which had received either allogeneic lymph node or buffy-coat lymphocytes, only massive numbers of spleen cell (SO x loG)from the primary recipient could produce a runting syndrome in secondary hosts; this single passage was the limit of transfer (Billingham et al., 1962). Using the C57B1/6 mouse strain as primary and secondary neonatal recipients, Russell (1962) was unable to incite runting with DBA donor cells. Finally, in adult F, hybrid hamsters, Streilein and Billingham (19701)) produced epidermolysis in about two-thirds of tertiary recipients, when the majority of secondary recipient lymph nodes were injected in two donor-equivalent doses into the skin of recipient hamsters; even three donor equivalents of lymph node cells were insufficient to cause epidermolysis in a “quaternary” recipient. In adult F, hybrid rats after injection of the hind foot pads with 20 x loF parental lymph nodc cells, the lymphoid cells (parental plus F, hybrid) recovered from the draining PLN up to 7 days after injection were capable of provoking additional PLN hypertrophy when inoculated into “secondary” F, recipients. The reactivity was not transferable to “tertiary” F, hybrid recipients when lymphoid cells were harvested from the PLNs of the “secondary” hosts (Grebe and Streilein, 1975b). There is, thus, considerable and imprcssive evidence that GVHRs siniply cannot be propagated ad infiniturn, although the studies cited have not elucidated the mechanism by which the transfers fail ( Blomgren and Anclersson, 1973).


187

,3. Refractoriness to Additional GVH Challenge In mice (Fou and Howard, 1963), rats (Field and Gibbs, 1966; Field et al., 1968, Grebe and Streilein, 1975b), and hamsters ( Streilein, 1972a) that have suffered and recovered froin initial GVH encounters, attempts to evoke GVHRs have frequently been unsuccessful. Field and Gibbs ( 1966) postulated the existence of a circulating antibody that inteferred with the recognition of host antigens bv the GVH-inducing cells; several reconstitution experiments in rcfractoiy hamsters suggestcd that the unresponsiveness to rechallengc could be attributed to a deficiency of host antigens, since the addition of F, antigens in the form of lymphocytes was sufficient to reinitiate the disease. It should be mentioned that, although hamsters are not known to make isoantibodies in response to transplantatioii antigens, they might be induced to do so under the influencc of a systemic GVHR. There is no evidence in the hamster study to rule out, uncquivocally, the presence of such an “enhancing” antibody. During the recovery phase of an initial PLN-GVHR (30 clays after injection with DA lymph node cells), (Fischer x DA) F, rats were profoundly refractorv to the induction of PLN hypertrophy by the addition of DA lymph node cell inocula ( “specific refractoriness”), and also werc hyporesponsive to GVH induction by ahgeneic parental or Fischer lymph node cclls (“nonspecific refractoriness”) ( Grebe and Streilein, 1975b).

B. CELLULAR AND MOLECULAR ASPECTSOF IMMUNOREGULATION IN GVHRs Each of the three phenomena just described represents a paradox that does not fit the cvpectations of the conventional GVH construct. In each instance, a predoniinant cell-mediated immune responsiveness has been curtailed or blunted in its expression. The niodulation of a GVHR, or of any response that is prcdominantly cell mediated, ultimately impinges on T-cell function-for example, the inhibition of cell-mediated killing by interference with the histaniinc receptors on T cells (Shearer et al., 1972, 1974; Plaut et nl., 1973, 1974). Accordiilgly, o m can postulate a variety of control relationships that are both specific and nonspecific, i.e., factors that regulate the responsc to a single antigen versus mechanisms that generally modify ininiunoreactivity to all antigens. In the latter category, one might include lymphokines (Section IV,C ) alpharegulatory proteins ( Section VIIB,4 ), and anatomically variable traffic patterns of lymphoid cells, i.e., whether an antigen-reactive cell is in the spleen, lymph node, peripheral blood, or thoracic duct. The specific expression of T-cell function may bc directed in any one of at least three ways: (1) by means of a “regulatory T cell”; ( 2 ) bv nicans of

188


a “regulatory B cell” or B-cell product, and ( 3 ) by deflection of T-effector cells activity into T-helper function, thus abandoning cell-mediated immunity in favor of antibody production. While these functions can be permuted into donorlhost relationships in GVHRs, the majority of relevant studies to date have not involved GVHRs. However, the implications of these studies are clear to those interested in GVHRs. The regulation of T-cell mediated immunity by T cells has already been discussed (see Section II1,C). Extensive work has been published on the manner in which antibody may influence the progress of immune responses, most notably of antibody production (Uhr and Moller, 1968; Bystryn et al., 1971; Schwartz, 1971). In more comprehensive terms, antibody may also regulate delayed hypersensitivity ( Takasugi and Hildemann, 1969), and some important insights into the more universal mechanisms of immunoregulation may be gained by examining the relationship between delayed hypersensitivity and antibody production. 1. Cell-Mediated Immunity, Antibody Production, and lmmunoregulation As a prelude to the consideration of the immunoregulatory interplay between B cells, T cells and their products in a GVH context, several pertinent allusions to the control of antibody synthesis by antibody may be drawn from the literature. a. Specific immunosuppression by passive antibody ( SIPA) has an extremely low threshold. For example, the response to the Rh-positive antigen may be suppressed if only 10%of the antigenic determinants are covered (Mollison and Hughes-Jones, 1967); if only 1%of the sites on SRBCs necessary to cause cytolysis are covered, the response to SRBCs is suppressed ( Haughton and Nash, 1969). b. The SIPA of mice to SRBCs can be reversed by mixtures of purified peritoneal macrophages from SRBC-immune mice and lymph node cells from normal mice when given to the suppressed animals (Haughton, 1974). Thus, the maintenance or abolition of SIPA depends on the type of cells involved; conditions that modify the population of SIPA affected cells may tip the balance between available antigens and antibody and disrupt the equilibirum of unresponsiveness in favor of sensitization. This study also reconfirms the essential role of T cells, B cells, and macrophages, not only in the induction of immunity, but also with regard to the regulatory aspects. Passively immunized mice phagocytose SRBCs normally. However, when macrophages from passively immunized mice were “ f e d SRBCs, they failed to induce an immune response in normal recipients (Ryder and Schwartz, 1969). It is apparent that some effect of SIPA is exerted at the level of the macrophage, and that the macrophage must be educated.


189

c. B cells, which may also be described as “complement-receptor lymphocytes,” may bind to antigen-antibody complexes that have fixed complement-a step essential to “immune adherence” ( reviewed by Nussenzweig, 1974). While immune adherence may be an initial step in the activation of B-cell immunity, it may also be involved in the arrest of antibody production. This inactivation of B cells may be a control mechanism in the response against polymerized flagellin ( Diener and Feldman, 1970), among others. The amount of complement present determines the extent of binding to antigen-antibody complexes, and this conglomerate in turn binds to thc complement-receptors available on B cells ( Miller and Nussenzweig, 1974). Thus, complement levels are an important parameter in the control of immune responsiveness. d. Anti-idiotypic antibodies may suppress specific antibody production (Hart et nZ., 1972; Cosenza and Kohler. 1972). More recent work has suggested that idiotype-positive antibodies are removed from the animal by the obliteration of idiotype-forming cell precursors, under the influence of exogenous anti-idiotypic antibody ( Eichmann, 1974). Multiple successive immunizations led to the appearance of specific antibodies followed by the acquisition of anti-iodiotypic antibodies in mice repeatedly immunized with Pneumococcus R36A vaccine and producing antibodies to phosphorylcholine (Kluskens and Kohler, 1974) or in rats hyperimmunized to alloantigens ( McKearn, 1974; McKearn et d., 1974a,b). In the latter studies, the titers of the specific antibody and anti-idiotypic antibody displayed roughly reciprocal and inverse oscillations over a 12-week period, suggesting a potentially regulatory interaction between the two species of antibody ( McKearn et al., 1974). e. Finally, despite certain controversies, there appear to be important counterbalancing systems of anti-idiotypic antibody within the two major subclasses of IgGs. In the modulation of the response of mice to group A streptococci, heterologous anti-idiotypic antibodies of the IgG2 subclass, (raised in guinea pigs against mouse antibodies) were able to suppress the specific idiotypic response, yet guinea pig IgG, apparently had an amplifying effect on the production of the idiotype (Eichmann, 1974). By using mouse serum that was hyperimmune to SRBCs ( b u t without regard specifically to the presence or the absence of anti-idiotypic antibodies), it was discovered that IgG, suppressed the anti-SRBC response, although IgG, augmented it (Murgita and Vas, 1972; Gordon and Murgita, 1975). Whatever the details of this “on-off system may ultimately be, there are clearly antagonistic-protagonistic regulatory antibodies that could serve to fine-tune antibody responses during and after the inductive phase (Vuagnat et al., 1973; Hoffman et al., 1974). This eclectic array of immunorcgulatory “vignettes” has been chosen

190

STEPIIEN C.

Advances in Immunology Volume 22

Advances in Immunology Volume 45

Advances in Immunology Volume 34

Advances in Immunology Volume 1

Advances in Immunology Volume 10

Advances in Immunology Volume 30

Advances in Immunology Volume 82

Advances in Immunology Volume 31

Advances in Immunology Volume 52

Advances in Immunology Volume 48

Advances in Immunology Volume 56

Advances in Immunology Volume 12

Advances in Immunology Volume 21

Advances in Immunology Volume 41

Advances in Immunology Volume 71

Advances in Immunology Volume 84

Advances in Immunology Volume 43

Advances in Immunology Volume 73

Advances in Immunology, Volume 61

Advances in Immunology Volume 14

Advances in Immunology Volume 46

Advances in Immunology, Volume 106

Advances in Immunology Volume 19

Advances in Immunology Volume 99

Advances in Immunology Volume 81

Advances in Immunology Volume 4

Advances in Immunology Volume 80

Advances in Immunology Volume 64

Advances in Immunology Volume 62

Advances in Immunology Volume 28

Advances in Immunology Volume 22

Advances in Immunology Volume 45

Advances in Immunology Volume 34

Advances in Immunology Volume 1

Advances in Immunology Volume 10

Advances in Immunology Volume 30

Advances in Immunology Volume 82

Advances in Immunology Volume 31

Advances in Immunology Volume 52

Advances in Immunology Volume 48

Advances in Immunology Volume 56

Advances in Immunology Volume 12

Advances in Immunology Volume 21

Advances in Immunology Volume 41

Advances in Immunology Volume 71

Advances in Immunology Volume 84

Advances in Immunology Volume 43

Advances in Immunology Volume 73

Advances in Immunology, Volume 61

Advances in Immunology Volume 14

Advances in Immunology Volume 46

Advances in Immunology, Volume 106

Advances in Immunology Volume 19

Advances in Immunology Volume 99

Advances in Immunology Volume 81

Advances in Immunology Volume 4

Advances in Immunology Volume 80

Advances in Immunology Volume 64

Advances in Immunology Volume 62

Advances in Immunology Volume 28

Recommend Documents