| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
PHYSIOLOGICAL REVIEWS Vol. 79 No. 1 January 1999, pp. 99-141
Copyright ©1999 by the American Physiological Society
Department of Medicine, University of Massachusetts Medical School, Worcester, Massachusetts
I. INTRODUCTION
A. Overview: From the Miraculous to Medawar and Merrill
B. The Modern Era of Transplantation
C. Beyond Immunosuppression
II. TRANSPLANTATION IMMUNOLOGY
A. Transplantation Terminology
B. Immune Response to Foreign Tissues
C. Host Immune Responses to Allografts
D. Host Immune Responses to Xenografts
E. Mechanisms of Graft Destruction
III. IMMUNOLOGIC TOLERANCE
A. New Approaches to Transplantation
B. Definition of Tolerance
C. Genesis of the Concept of Transplantation Tolerance
D. Graft-Based Tolerance Induction
E. Host Response-Based Tolerance Induction
IV. METHODS TO INDUCE TRANSPLANTATION TOLERANCE IN VIVO
A. Methods Successful in Rodents
B. Methods Successful in Large Animals and Primates
C. Evidence of Tolerance Induction in Humans
V. CONTEMPORARY CLINICAL ORGAN TRANSPLANTATION
A. Criteria for Successful Transplantation Program
B. Issues of Graft Placement
C. Immunosuppression
VI. TRANSPLANTATION AFTER TOLERANCE
A. Demand for Transplantation Services
B. Potential Pitfalls
C. Issues of Donor Tissue and Organ Availability
D. Special Case of Xenografts
E. Conclusion
REFERENCES
 |
ABSTRACT |
|---|
 Top References
|
|---|
Rossini, Aldo A., Dale L. Greiner, and John P. Mordes. Induction of Immunologic Tolerance for Transplantation. Physiol. Rev. 79: 99-141, 1999. 
In the second half of the 20th century, the transplantation of replacement organs and tissues to cure disease has become a clinical reality. Success has been achieved as a direct result of progress in understanding the cellular and molecular biology of the immune system. This understanding has led to the development of immunosuppressive pharmaceuticals that are part of nearly every transplantation procedure. All such drugs are toxic to some degree, however, and their chronic use, mandatory in transplantation, predisposes the patient to the development of infection and cancer. In addition, many of them may have deleterious long-term effects on the function of grafts. New immunosuppressive agents are constantly under development, but organ transplantation remains a therapy that requires patients to choose between the risks of their primary illness and its treatment on the one hand, and the risks of life-long systemic immunosuppression on the other. Alternatives to immunosuppression include modulation of donor grafts to reduce immunogenicity, removal of passenger leukocytes, transplantation into immunologically privileged sites like the testis or thymus, encapsulation of tissue, and the induction of a state of immunologic tolerance. It is the last of these alternatives that has, perhaps, the most promise and most generic applicability as a future therapy. Recent reports documenting long-term graft survival in the absence of immunosuppression suggest that tolerance-based therapies may soon become a clinical reality. Of particular interest to our laboratory are transplantation strategies that focus on the induction of donor-specific T-cell unresponsiveness. The basic biology, protocols, experimental outcomes, and clinical implications of tolerance-based transplantation are the focus of this review.
A. Overview: From the Miraculous to Medawar and Merrill
1. Cosmas and Damian
I. INTRODUCTION
Top
Next
References
2. The first successful transplant
There is remarkable evidence that successful transplantation surgery was actually performed by ancient Hindu vaidya, perhaps 2,000 years ago (236a). The Ayurvedic physicians, among them the author Sushruta, reconstructed noses using pedicle flap grafts from the patient's own forehead (236a). By 1597, the physician Gaspare Tagliacozzi was performing similar reconstructive rhinoplasty using skin flaps from the patient's arm (371). There were no reports of success using skin from another person, however (236a, 371). Transplantation progressed little until the 20th century.3. Laws of transplantation
By the beginning of the 20th century, advances in surgery, asepsis, and anesthesia had revivified the ancient dreams of Cosmas and Damian. Animal and some human transplantation experiments began to be performed. These suggested that the serious barriers to successful transplantation would not be technical, but biological. George Schöne summarized these insights in 1912. He defined classic "laws of transplantation" that, in the absence of immune intervention, still apply today (371). These are as follows. 1) Transplantation into a foreign species invariably fails. 2) Transplantation into unrelated members of the same species usually fails. 3) Autografts almost invariably succeed. 4) There is a primary take and then delayed rejection of the first graft into an unrelated member of the same species. 5) There is accelerated rejection of a second graft in a recipient that had previously rejected a graft from the same donor, or of a first graft in a recipient that had been preimmunized with material from the same donor. 6) The closer the "blood relationship" between donor and recipient, the more likely is graft success.4. Transplantation genetics and immunosuppression
It was decades more before the basis for these empirical laws was found in studies of the small but prolific mouse. Clarence C. Little and George Snell (who later won a Noble prize for this work) had the inspiration to develop and analyze "congenic" strains of mice. Congenic strains are genetically identical except at a single chromosomal locus. By constant breeding and testing for allograft acceptance, Little and Snell pioneered the analysis of what came to be called the major histocompatibility complex (MHC), a genetic locus designated H-2 in mice (371). The MHC is the molecular embodiment of "blood relationship." Armed with genetic insight, Sir Peter Medawar attempted to apply this information to aid World War II burn victims who could be helped by skin grafts. Like Schöne, he observed that 1) autografts succeed, 2) allografts fail after an initial take, and 3) a second allograft from the same donor undergoes accelerated rejection. He suggested that "destruction of the foreign epidermis was brought about by a mechanism of active immunization" (371). As Medawar continued his analyses using rabbits, his insights led to the recognition that immunosuppression might overcome the laws of transplantation. The development of immunosuppressive agents like nitrogen mustard and corticosteroids and their evaluation in animal models soon led to the practical application of transplantation as a medical therapy. In 1954, Merrill, Murray, and Harrison performed the first successful human vascular organ graft, a kidney transplant (253). The donor and recipient were monozygotic twins in this "proof-of-principle" intervention, with genetic identity obviating the need for immunosuppression. The graft survived until the death of the patient 7 yr later of heart disease. With the advent of an effective immunosuppressive regimen, the same group went on in 1959 to perform the first kidney graft between unrelated individuals; that graft survived for 20 yr (252).B. The Modern Era of Transplantation
The modern era of transplantation has also benefited from improved therapeutics for those awaiting transplantation, making them better surgical candidates, improved methods of donor organ preservation, and broadening societal awareness of the need for organ donation.
Above all, however, the key to successful transplantation has been the discovery and improvement of strategies for immunosuppression. Radiation of the host came first, but was rapidly supplanted by immunosuppressive drugs including glucocorticoids, azathioprine, and antilymphocyte serum. Newer immunosuppressive drugs with greater potency and wider margins of safety have improved the outcome of renal allografts and generated the first consistent success with cardiac, liver, lung, and pancreas grafts (138). The first of these second-generation drugs was cyclosporine. Newer agents include tacrolimus (formerly called FK-506), monoclonal antibodies like anti-CD3, and many other biologic (393) and nonbiologic agents (262).
C. Beyond Immunosuppression
The success achieved by solid organ transplant surgeons since the pioneering work of the 1950s has been extraordinary. More than 90% of living related donor and 80-85% of cadaveric kidney grafts are now functional after 1 yr (279, 294). Comparable success has been achieved in the areas of heart, lung, liver, and to a lesser extent pancreas transplantation. This success, however, belies some important residual problems.
At this time, all successful treatment of human disease by transplantation (other than between monozygotic siblings) requires the use of general immunosuppressive agents (138). All such drugs are toxic to some degree, and toxicity often leads to patient noncompliance. The drugs are also known to predispose to both infection and neoplasia. Some 10-45% of persons who are chronically immunosuppressed after transplantation develop a neoplasm after 10 yr, and 40-75% do so after 20 yr (75, 272). In addition, many immunosuppressive drugs have deleterious long-term effects on the function of transplanted organs (138), e.g., 
-cells in the case of transplantation for diabetes and cyclosporine-induced nephrotoxicity in cardiac and liver transplant recipients (410). New immunosuppressive agents are constantly under development (262, 393), but transplantation of any organ or tissue into an unrelated human currently requires patients to accept the risks of life-long systemic immunosuppression.
Another critical problem is that of "chronic rejection." Although the majority of renal grafts are functional at 1 yr, only 20% will be functional at 10 yr (294). The pathology and mechanisms of this indolent form of graft failure are only beginning to be understood (15, 145, 146, 298).
A final issue is recurrence of disease in a successfully transplanted organ. Recurrence may take the form of atherosclerosis in a cardiac allograft or recurrence of tissue-specific disease processes, for example, recurrent autoimmune destruction of pancreatic islets or recurrent lupus nephritis.
Finding alternatives to immunosuppression that will prevent both immediate and long-term graft rejection will require further understanding of the mechanisms by which Schöne's six laws operate. Immunosuppression is based on interference with the immune system. Understanding immunologic processes better may allow us to redirect the immune response to favor graft acceptance. We next review our current understanding of transplantation immunobiology and newer approaches to transplantation. These include modulation of transplant immunogenicity (sects. IIID1 and IIID2), removal of passenger leukocytes (sect. IVA7), transplantation into immunologically privileged sites like the testis or thymus (sect. IIID3), mechanical encapsulation of tissues (199), and the induction of a state of immunologic tolerance. It is the last of these alternatives that has, perhaps, the most generic promise of revolutionizing tissue transplantation. Induction of transplantation tolerance is the focus of work in our laboratory and the primary focus of this review.
| |
II. TRANSPLANTATION IMMUNOLOGY |
|---|
|
TopPreviousNextReferences
|
|---|
A. Transplantation Terminology
Transplauten is a Middle English agricultural word used literally to describe the removal of a plant from one place to another. Its modern surgical adaptation refers to the transfer of an organ or tissue from one part of the body to another or from one individual, the donor, to another, the recipient. If donor and recipient are the same individual, a graft is autologous. If donor and recipient are monozygotic, a graft is syngeneic. If donor and recipient are any other same-species individuals, a graft is allogeneic. If the donor and recipient are of different species, a graft is xenogeneic (174).
Orthotopic grafts are placed in the normal anatomic location of that organ; heterotopic grafts are located in other sites of convenience, e.g., intrahepatic placement of pancreatic islets. If a graft heals and functions, it is accepted; if destroyed by the immune system, it is rejected. Rejection may be hyperacute (within minutes to hours) or acute (usually within the first month after grafting). Rejection can also be chronic, with slow, gradual destruction over months to years. If the immune system has not previously been sensitized to the donor's tissues, the rejection is termed a first-set rejection. In presensitized recipients, the rejection is a second-set rejection.
B. Immune Response to Foreign Tissues
Many processes participate in the response to all foreign tissue grafts. These include the local inflammatory response to surgery, the processes that initiate wound repair and vascular endothelialization, and the immune response to the recognition of nonself antigen. The basic elements of this response are schematized in Figure 1.
|
The response to nonself antigen involves both cellular and humoral immunity. The goal of the response is to reject that antigen, and its nature and intensity are determined by two factors. The first is the biology of the foreign tissue, whether it is an allograft or a xenograft, a vascularized organ or dispersed tissue, a fresh tissue or one that has been pretreated to reduce antigenicity. The second factor is the host response to the encounter with that specific foreign tissue. These responses fall into three categories: hyperacute, acute, and chronic rejection.
1. Hyperacute rejection of discordant xenografts
"Natural antibodies" in human recipients cause hyperacute rejection of "discordant" vascularized xenografts (67, 109, 328, 426). These predominantly IgM antibodies recognize
-galactosyl residues that are absent on human (and all old-world primate) endothelial cells but present on endothelium in many other species. They bind to xenogeneic endothelium and fix complement, leading to vascular leakage and thrombosis. The process can cause xenograft rejection in minutes to hours. Preformed anti--galactosyl natural antibodies are a major obstacle to the transplantation, for example, of porcine xenografts into humans (128, 149). In contrast, transplantation of allografts never encounters the obstacle of hyperacute rejection due to anti--galactosyl natural antibodies. Hyperacute rejection is, however, a well-recognized risk in the case of allografts transplanted into sensitized recipients who have preformed antibodies against allogeneic human leukocyte antigens (HLA). Anti-HLA antibodies occur in individuals exposed to non-self HLA antigens by blood transfusion, pregnancy, prior transplantation, or bacterial infections that induce cross-reacting antibodies to HLA epitopes (170).
Based on the presence or absence of -galactosyl residues or anti--galactosyl natural antibodies, xenografts can be classified into two groups (11, 343). Concordant xenografts do not elicit hyperacute humoral rejection responses. This occurs most often when donor and recipient are members of closely related species (e.g., rat and mouse). Discordant xenografts do elicit hyperacute rejection responses and commonly occur when donor and recipient are members of less-related species.
Strategies for overcoming hyperacute rejection include genetic engineering of xenograft donors to express human complement inhibitors like CD46, CD55, and CD59 (67, 280, 338) or to knock out genes required for the expression of -galactosyl residues (228, 441). Another approach is prophylactic pretreatment of the human xenograft recipient, for example, by plasmapheresis to remove natural antibodies or by injection of soluble inhibitors of complement such as CD35 (110, 228). Other proposed therapies address events downstream of natural antibody binding and include the administration of antioxidant drugs to maintain the platelet-inhibiting enzyme ecto-ATPase in its active form (67, 280, 338) and depletion of complement by cobra venom. Because none of these promising strategies for overcoming hyperacute rejection involves the induction of tolerance, they are not discussed further.
2. Acute rejection
In the absence of a hyperacute rejection response, all transplanted tissues, except for those from identical siblings, engender an acute rejection response. This response is based on the recognition of the foreign tissue as nonself. The basis for recognition of self resides with a system of cell-surface glycoproteins. All living tissues express families of cell-surface proteins that comprise the MHC. For historical reasons, MHC proteins are given the prefix H2 in mice, RT1 in rats, and HLA in people. Alleles of MHC proteins are extremely polymorphic, and chance MHC identity between individuals other than monozygotic siblings is extremely rare. When an allograft is transplanted, the disparity in MHC, or "alloantigenicity," constitutes recognition of nonself and incites a number of reactions. The interaction of graft alloantigens with the host immune system leads to the activation of several classes of cells. In the first stages of transplantation, as the organ becomes vascularized, there is rapid infiltration of host immune cells into the graft. The infiltrate is composed mostly of mononuclear cells. Although the exact composition of the infiltrate over time is not precisely known, it is presumed that those host mononuclear cells with the capacity to present nonself antigens to the host predominate. Details of this process are presented in section IIC2.3. Chronic rejection
The third major form of graft rejection occurs months to years after transplantation, even in the presence of continued immunosuppression. The majority of kidney transplants, which now have an initial acceptance rate of >90%, inevitably fail due to the development of chronic rejection and loss of function (279, 294). The pathological hallmark of an organ undergoing chronic rejection is fibrosis, leading to distortion of normal architecture. The pathology is comparable to that which accompanies wound healing, and it has been suggested that chronic rejection is the end result of healing in response to recurring episodes of acute rejection. Other suggested mechanisms include delayed-type hypersensitivity with activation of T-helper cells and B cells (346) leading to activation of macrophages and the secretion of tissue growth factors. Another possibility is antibody-mediated humoral rejection (see sect. IIE10) or endothelial cell damage leading to ischemia (1). In addition, nonimmunologic factors, e.g., ischemia and vascular injury, may contribute to chronic rejection. These factors include maladaptive responses to prior acute injury. The basis of the development of chronic rejection is unknown, and currently available therapies generally fail to reverse chronic rejection, eventually leading to graft failure (15, 146, 298).4. Graft versus host disease
Immune responses to foreign tissue are not totally unidirectional. Essentially all organ and tissue grafts carry with them elements of the host immune system. In the special case of lymphohemopoietic grafts, the immune system components of the graft are able to mount an immune response to the host. This is termed graft versus host disease (GVHD). In the case of solid organ grafts, however, GVHD is not a typical complication. Donor immune cells, may, however, persist at low levels in the recipient, leading to a state of "microchimerism."5. Microchimerism
It has been suggested that passenger leukocytes present in at least some successful allografts migrate after organ transplantation and produce persistent chimerism (385). The level of engraftment is low, but it has been argued that a microchimeric state may be essential for sustained survival of allografts (385). The intentional augmentation of microchimerism by cotransplantation of bone marrow together with an organ graft to induce tolerance (46, 276, 446) is discussed in section III.C. Host Immune Responses to Allografts
1. Role of the major histocompatibility complex
As noted above, the host MHC plays the defining role in acceptance or rejection of a graft. The more closely the donor graft and the host are matched for MHC, the greater the likelihood of the graft acceptance by the host. The family of MHC cell-surface proteins is subdivided into two major subclasses that play pivotal yet different roles in the determination of self versus nonself by the immune system. These are designated as the MHC class I and the MHC class II antigens. Host T cells can recognize a foreign antigen and generate an immune response, only when foreign antigens are "presented" to its T-cell receptor (TCR) by an MHC class I or class II molecule. In most cases, host MHC class I does not readily present antigens derived from a donor graft, although a mechanism for this type of antigen presentation has been proposed (69, 331, 453). In general, MHC class I molecules present self-antigens derived from intracellular degradation of proteins. These processed antigens are expressed in a cleft present in the extracellular domain of the MHC molecule. Presentation of foreign donor antigens and activation of host-reactive T cells occurs through the MHC class II molecule. Antigens taken up by endocytosis and processed by host antigen presenting cells (APC) are readily presented to the host immune system by the host MHC class II molecule (69, 331, 453). In transplantation, this is primarily the presentation of alloantigen or xenoantigen. In addition to MHC antigens, the immune system can also respond to so-called minor histocompatibility antigens present on grafts. Response to these antigens is indolent, and graft rejection based solely on these antigens is slow (360, 432). An example of a minor histocompatibility antigen is the male Y antigen (360). Peptides noncovalently bound to MHC molecules provide the immune system with antigenic targets. In addition, the presence of peptide bound to MHC molecules lengthens the life span of those MHC molecules on the cell surface of the APC (270). Major histocompatibility complex without peptide has a very short life span and is rapidly lost from the cell surface. The predominant response of the host immune system to foreign tissue is not simply to alloantigen but rather to the allo-MHC plus peptide unit. Direct recognition of allo-MHC without bound peptide can occur, however, and T-cell clones reactive to allo-MHC in the absence of peptide can mediate skin allograft rejection (375). The strategies to induce tolerance described in section IV are directed at altering the acute rejection response. In molecular terms, they are directed at redefining the MHC plus peptide complexes present on the graft as "self" by reeducating the host immune response.2. Antigen presenting cells
The primary APC in the immune system are dendritic cells and macrophages. These cells are termed "professional antigen presenting cells" because of their ability to provide accessory cytokines and costimulatory molecules needed for initiation of maximal T- and B-lymphocyte immune reactivity. There are many other types of cells that can present antigen to the immune system, but in general, these do not provide all of the factors necessary for the initiation of an immune rejection response. These "nonprofessional" APC include endothelial cells and resting B lymphocytes. Macrophages are among the first APC to infiltrate a graft. They both initiate and mediate the inflammatory process. Macrophages are able to 1) process and present the antigen to the immune system, 2) release monokines that are both chemoattractant and immunostimulatory, 3) release chemokines that are also chemoattractant and proinflammatory, and 4) scavenge debris from necrotic graft tissue. Another class of professional APC, dendritic cells, can be distinguished from monocytes and macrophages by cytochemical, physical, and morphological characteristics. Dendritic cells reside in the interstitial space of many tissues, process and present antigen, interact with and stimulate T cells, and contribute to allograft rejection. It is now recognized that the "passenger leukocytes" responsible for the immunogenicity of many grafts are probably dendritic cells. Treatments that remove or block dendritic cells have been shown to prolong the survival of transplanted grafts (142) (see sect. IVA7).3. T lymphocytes
After presentation of graft antigens by donor or host APC, host lymphocytes become sensitized and activated, and then proliferate. These lymphocytes are found in lymph nodes draining the graft site as well as in peripheral blood and spleen. They are directed against graft antigens presented by either MHC class I or class II molecules on APC. The net effect is activation of the immune system and initiation of immune rejection. Graft destruction can be effected by direct cell-to-cell contact between activated effector T cells and the targeted graft resulting in delivery of a cytotoxic molecule. Alternatively, graft rejection may be mediated by indirect mechanisms including lymphokine-induced destruction (see sect. IIE).4. B lymphocytes
The role of antibodies in graft rejection is controversial. T lymphocytes alone are sufficient for rejection of a graft. The humoral response to the graft could, however, participate in this process either by synergizing with T cells or by antagonizing them so as to impede rejection. The humoral immune response to a graft generates antibodies that could synergize with the cellular immune system to enhance graft rejection in several ways. First, graft-reactive antibodies may provide an adherence signal that "marks" a graft and directs macrophages, mononuclear cells, and neutrophils to recognize and attack it. Second, bound antibodies could initiate antibody-dependent cell-mediated cytotoxicity (ADCC), a process in which cells like natural killer cells recognize the Fc portion of bound immunoglobulin and, again, attack the marked graft. Third, antibodies may bind to many target antigens, form immune complexes, and impair the function of the graft. Immune complexes are also recognized by macrophages, again enhancing cellular immunity to the graft. Finally, the antibody could bind complement. This process generates two effects. The bound complement binds macrophages through their C3 receptors. This again enhances macrophage targeting of the marked graft. Another effect is the activation of complement and direct complement-mediated lysis of the targeted graft. Alternatively, antibodies to the graft could act to impair rejection. It has been proposed that antibodies can provide a protective effect termed immunologic enhancement, or simply enhancement (see sect. IIIE3). What determines whether antibodies to grafts facilitate destructive mechanisms or instead act to protect grafts from cell-mediated destruction is not known. However, as detailed in section IVA6D, this phenomenon has been demonstrated experimentally in islet transplantation. In those studies, pretreatment of islets with anti-MHC class I antibody extended the survival of human islet xenografts in mice (94). The extent to which enhancing antibodies of host origin may contribute to graft survival is not known.5. Direct and indirect antigen presentation pathways
In transplantation, activation of the T-cell component of the immune response by MHC plus peptide can occur in two ways. These are termed direct and indirect antigen presentation (225). These processes are summarized in Figure 2. In direct antigen presentation, the MHC molecules on or within the graft present donor antigens directly to host T cells (23, 349). In indirect antigen presentation, antigens "shed" by the graft are thought to be presented to the host immune system by host APC (12, 24, 117, 208, 405).
|
6. Direct recognition pathway
The direct pathway is believed to be of primary importance in the immune response to allografts (42, 207, 349). All allografts contain passenger APC that are transplanted with the tissue. Their removal in most cases is difficult. These donor-origin APC, like most nucleated cells, constitutively express MHC class I molecules that continuously present self-peptides. In addition, they constitutively express coactivation molecules like CD40 and B7-1/2 (142). Alloantigenic peptides presented by donor class I MHC molecules appear to be key targets of T-cell recognition in allorejection (23, 349). T cells of the CD8+ MHC class I-restricted subset directly engage these peptide-MHC class I complexes on APC, receive costimulatory signals, become activated, and undertake to destroy the target (405). This process was first demonstrated in thyroid transplantation (192, 193) and appears to play an important role in islet allograft rejection (68, 95, 190, 372). The case of the islet graft is particularly interesting. The passenger donor-origin leukocytes present in grafted islets are APC that express both MHC class I and class II antigens. Islet cells express only class I MHC and no costimulatory molecules on their surface. Removal of passenger leukocytes from the graft removes APC that express MHC class II and costimulatory molecules. This leaves islets with only MHC class I antigen and no costimulatory molecules. This in turn inhibits the host's ability to respond to foreign alloantigen on the donor islets by the direct presentation pathway, and the graft survives. Several studies support this concept. In some studies, APC were depleted by culturing islet grafts at 24°C or with 95% O2 (189). In others, grafts were treated with anti-MHC class II antibody to remove APC. Still others have used a diet low in essential fatty acids to induce emigration of passenger leukocytes from donor islets (209). Each of these methods was reportedly effective in preventing allograft rejection, demonstrating clearly the important role of donor-origin APC present in a graft. Although direct antigen presentation appears to play a key role in acute rejection, it is less likely to be important in chronic rejection (42).7. Indirect recognition pathways
Indirect antigen recognition is based on the theory that donor graft antigen can be shed, subsequently taken up by endocytosis by host-origin APC, processed, and then presented to the host immune system as a self-MHC-plus-foreign peptide unit. In the indirect pathway, host T cells recognize graft alloantigens presented by a host APC. This form of antigen presentation by the MHC class II molecule targets class II-restricted host CD4+ T cells, resulting in their activation in response to the graft. Definitive demonstration of indirect antigen presentation and its role in transplantation rejection has been difficult, but convincing evidence has emerged. Indirect allorecognition appears to have a role in chronic allograft rejection (see sect. IIB3), a phenomenon that is mediated in part by the activation of T-helper cells and of alloantibody-producing B cells (42, 346). The activation of T cells following antigen presentation by the indirect pathway may be less efficient than that of T cells primed by the direct pathway (346). Furthermore, the T-cell repertoire appears to be restricted following indirect antigen presentation, suggesting that a more limited number of antigens may be present by this pathway (224). The classical mechanism of antigen presentation by the indirect pathway involves the uptake of exogenous antigen by the host MHC class II molecules on APC and presentation to CD4+ T cells. It has, however, been suggested that indirect antigen presentation to CD8+ T cells can occur, possibly permitting both CD4+ and CD8+ host T cells to be activated by donor grafts in the absence of direct antigen presentation (25, 181). Indirect antigen presentation is thought to be the major pathway involved in the host response to xenografts, the subject to which we turn next.D. Host Immune Responses to Xenografts
1. Immune rejection of xenografts
The cellular immune response to xenografts is distinct from and less well understood than the cellular immune response to allografts (405). Helper and cytotoxic T-cell responses to xenoantigens are generally slower and less intense than are responses to alloantigens (149). The weaker xenoresponse may result from a relative inability of T cells from one species to interact with APC from another (117, 445). Interactions requiring the adhesion of host immune cells to an organ graft may similarly be weaker across species (149).2. Direct presentation of xenoantigens
Studies of xenograft rejection suggest that MHC class II is the predominant antigenic target of the host immune response and that MHC class II-restricted CD4+ T cells are the mediators of graft destruction (180, 238, 421). It has been suggested, in fact, that xenograft rejection may use a CD4+, cytotoxic T lymphocyte (CTL)-independent pathway (117, 149). Lafferty, Gill, and co-workers (48, 117, 191) have proposed the concept of "indirect" antigen recognition leading to immune activation and rejection based on the interaction of recipient CD4+ T cells with recipient APC presenting graft xenoantigens. The basis of this preference for indirect recognition in the response to xenografts may stem from the inability of the host CD8+ T-cell population to recognize and interact with the xenogeneic class I MHC molecules present on the graft itself. The generation of cytotoxic T cells and the subsequent release of cytokines and free radicals in the vicinity of grafted tissue are thought then to lead to cytotoxicity or apoptosis (see sect.IIE). Studies of cytokine gene expression are consistent with this view. For example, Morris et al. (261), studying xenogeneic porcine proislets in CBA/H mice, have observed that Th2-like CD4+ T cells appear to be preferentially activated in the course of rejection.E. Mechanisms of Graft Destruction
There are several mechanisms available to the immune system of a mammalian host to mediate rejection of a foreign tissue. This redundancy presents additional hurdles that must be overcome in the area of transplantation. Grafts can be destroyed either directly by delivery of a "lethal hit" from cytotoxic cells or indirectly by molecules such as cytokines. Both mechanisms involve a broad diversity of elements within the cellular arm of the immune response (333).
1. Initiation of the effector response
The cell population that appears to be most important in initiating rejection of allografts is the CD4+ T cell (182). Mice lacking CD4+ cells fail to reject allografts, whereas mice lacking CD8+ retain their ability to reject allografts (182). The ability of CD4-deficient mice to retain grafts may, however, be strain dependent (241). To a first approximation, it can probably be concluded that CD4+ T cells can both initiate and mediate allograft rejection, whereas CD8+ T cells are primarily mediators of graft destruction. The mechanisms available to the T-cell immune system to mediate rejection are described next.2. Cytotoxic T lymphocytes
Cytotoxic T lymphocytes are major mediators of allograft rejection. The CTL are classically CD8+ cells that recognize antigen in the context of MHC class I molecules and, therefore, are important in allograft rejection where the graft itself can present alloantigen to the immune system. In addition, however, subsets of CD4+ T cells can also act as CTL (139). There are at least two major mechanisms by which CTL can deliver a lethal hit through direct cell-cell contact. One is by interaction of Fas ligand (FasL, CD95L) on the activated CTL with Fas (CD95) expressed on the target cell (see sect. IIE3). The other is by delivery of cytotoxic molecules termed granzymes (perforin and other molecules like granzyme B; see sect. IIE6). There are several key features of CTL-mediated cytotoxicity. These include 1) antigen specificity, 2) a requirement for cell-cell contact for killing, and 3) the ability of CTL to destroy multiple target cells without injury to themselves. Cytotoxic T lymphocyte-mediated cytotoxicity occurs in several steps. The first is recognition of the target and conjugate formation, whereby the CTL binds to the target cell. This binding event leads to activation signals in the CTL. The activated CTL then delivers a lethal hit mediated either by granzymes or through the Fas pathway. After these steps, the CTL is released from its conjugation with the target cell, and the target then undergoes programmed cell death. Efforts in transplantation to prevent many of these steps have been attempted. The most recent experiments have targeted the Fas-FasL pathway.3. Fas and Fas ligand
Fas and FasL (CD95 and CD95L) are members of the tumor necrosis factor (TNF) family of molecules, and defects in this pathway have been implicated in defective apoptosis and autoimmunity in animals. Mice homozygous for the lpr or gld mutations develop lymphadenopathy and suffer from a systemic lupus erythematosus-like autoimmune disease (268). The lpr defect is due to a mutation in the gene encoding Fas, whereas the gld defect is due to a mutation in the gene encoding FasL (232, 408, 433). Mice with lpr and gld mutations cannot mediate cytotoxic activity through the Fas-FasL pathway, and as a result, they do not reject allografts well. The Fas-FasL pathway has been implicated in clonal selection and control of lymphocyte activation (257, 267, 374) as well as in killing mediated by cytotoxic T cells (339).4. Immunologically privileged sites
It has been known for many years that certain anatomic sites are exceptionally permissive to graft acceptance. Islet allografts, for example, are not rejected even in the absence of immunosuppression when they are placed heterotopically in the testis (20, 371). It is now recognized that several of these "privileged sites" are associated with high levels of FasL expression on cells within the site (127). The interaction between FasL on tissues in the site and Fas on responding host T cells leads to destruction of those responding cells and preservation of the graft.5. FasL-mediated bystander killing
An additional role of cytotoxic cells in graft rejection may stem from their ability to self-regulate. T cells activated in response to allogeneic stimuli are able to lyse bystander host alloantigen-activated T cells in an MHC-restricted manner (377). This process is not dependent on the perforin pathway but rather on the Fas-FasL pathway of cytotoxicity. Cytotoxic T lymphocyte responses to foreign tissues may therefore cause collateral damage to local host Fas-expressing cells (377). The role of bystander-cell killing in graft destruction or in the generation of collateral damage to host tissues is not yet known.6. Perforin
Perforin is a cytotoxic molecule released into target cells by CTL during conjugate formation. The role of perforin in cytotoxicity may be direct, forming holes in the target cell membrane similar to those produced by the complement cascade. Alternatively, the role of perforin may be indirect, by increasing the porosity of the target cell membrane and thereby enhancing the entry of granzymes. The importance of perforin in transplantation has been highlighted by the observation that mice genetically deficient in perforin are deficient in their ability to lyse allo-specific targets in vitro (162). Cytotoxic CD4+ T cells can also mediate cytotoxicity via the perforin pathway (444).7. Natural killer cells
Natural killer cells (NK cells, sometimes designated as large granular lymphocytes), like CTL, kill target cells using perforin, granzymes, and proteoglycans. The NK cells lyse target cells in an MHC-unrestricted manner using receptors that are not antigen specific. The role of NK cells in allograft rejection is uncertain. The NK cells by themselves cannot reject allografts, but they are in the graft-infiltrating cell population and may contribute to graft damage. The NK cells have been implicated in the rejection of xenografts (219). In the absence of T cells and antigraft antibody, athymic rats reject hamster heart xenografts. Depletion of NK cells by injection of anti-asialo GM-1 antiserum delays rejection, suggesting that NK cell-mediated cytotoxicity can be important in xenograft rejection (219).8. Macrophages
As outlined in section IIC2, macrophages may play multiple roles in initiating and propagating the immune response to grafts. Indirect evidence suggests that they can also mediate graft rejection, at least for xenografts (219). This conclusion is based on several observations. 1) Spleens from animals that are rejecting xenografts harbor a high percentage of macrophages. 2) Rejected xenografts are infiltrated predominantly by macrophages. Furthermore, macrophages adoptively transferred to athymic recipients produce accelerated rejection of newly transplanted xenografts.9. Dendritic cells
Dendritic cells are among the most potent APC and play an important role in both direct and indirect antigen presentation. Dendritic cells, however, are also known to express FasL and can kill activated Fas-expressing CD4+ T cells (400). This activity may contribute to the resistance of a graft to host immune-mediated killing. In addition, dendritic cells can both activate and suppress the host immune system as a function of their expression of FasL or costimulatory molecules (230). The presence or absence of dendritic cells in a graft may be a primary determinant of its acceptance (see sect. IVA7).10. B lymphocytes
As discussed in section IIC4, B lymphocytes and antibody responses have the potential to play many roles in graft rejection. This has become a matter of increasing concern with the recognition that
50% of patients on transplantation waiting lists exhibit a high "panel-reactive antibody" status (340). These high levels of antibody to multiple MHC antigens usually result from the blood transfusions patients receive in the course of treatment for their primary disease. They can also arise from pregnancy, prior transplantation, or bacterial infections that induce cross-reacting antibodies to HLA epitopes (170). As noted in section IIB1, this high level of antibody reactivity can lead to hyperacute rejection and therefore poses an additional obstacle to transplantation that must be overcome. High panel-reactive antibody status can be overcome either by selecting the donor graft on the basis of nonreactivity to recipient preformed antibodies or possibly by immunoabsorption of the host's preformed antibodies before transplantation.
Recent studies in animals suggest that presensitization by transfusion can also be prevented. Mice transfused with donor cells in the presence of CTLA4-Ig, a molecule that blocks costimulation, failed to generate antidonor antibodies capable of mediating accelerated rejection (340). To prevent sensitization, current clinical protocols for the care of transplant candidates call for removal of white blood cells before blood transfusion or concurrent treatment with immunosuppressive agents.
11. Natural antibodies and complement
Natural antibodies appear to have a role only in xenograft, not allograft rejection. The predominant natural antibodies are those directed against the-galactosyl residues present on many nonprimate mammalian endothelial cells. They are responsible for hyperacute rejection as discussed in section IIB1.
Complement appears to have a role in both allo- and xenograft rejection. It can perforate the target cell membrane and create a lethal electrolyte imbalance. Alternatively, it can complex with bound antibody and form a potent adhesion complex for the binding of macrophages and neutrophils, thus targeting these cells to the graft. These functions of complement depend on the generation of an immune response to the graft that leads to antibody formation and antibody-complement complex formation.
12. Cytokines
Cytokines can play both destructive and immunomodulatory roles in graft rejection (264). Cytokines that participate in rejection include TNF-, interferon-
(IFN-), and interleukin (IL)-1 (8, 273). They contribute to graft destruction either directly or by activating effector cells. Cytokines thought to be capable of impairing graft rejection include IL-4, IL-10, and transforming growth factor- (TGF-).
A contemporary controversy in immunology and transplantation concerns the roles of specific T-cell subsets, termed Th1- and Th2-type T cells, in graft rejection (4). The Th1-type CD4+ T cells are producers of IL-2, IFN-, and TNF- (also known as lymphotoxin). These cytokines can activate both T cells and macrophages. They can promote cellular immune responses that serve as terminal effector mechanisms in allograft rejection. It has also been suggested that expression of the IL-12 receptor 2-subunit is specifically restricted to the Th1-type of cells (332, 406). Interleukin-12 is a proinflammatory cytokine produced predominantly by activated macrophages. The Th1-type immune responses are proinflammatory, promoting CTL development, delayed hypersensitivity responses, and the production of the IgG2a antibody that is involved in ADCC.
In contrast, Th2-type cells produce IL-4, IL-5, and IL-10, cytokines thought to have immunosuppressive or downregulatory effects on the immune system. It has been suggested, however, that in the absence of CD8+ T cells, CD4+ T-cell production of Th2-type cytokines can also mediate graft rejection (56). Whereas Th1-type immune responses are proinflammatory, Th2-type immune responses are biased toward humoral, IgE-mediated allergic, and mucosal immune responses (264).
The Th1- and Th2-type cytokines are not simply the end products of an immune response; they can also, depending on the sequence and intensity of their production, determine (or "polarize") the nature of an immune response. T cells exposed to antigen (e.g., grafts) in the presence of IL-4 are driven toward Th2-type immune responses. In contrast, exposure to antigen in the presence of IFN- directs T cells toward a Th1-type response. The IFN- accomplishes this polarization by preventing the differentiation of naive cells to Th2-type cells despite the presence of IL-4; IL-12, when available, then drives the cells toward a Th1-type response (406). Interleukin-12 appears to be a potent determinant of Th1- versus Th2-type polarization of the immune response to foreign tissue.
13. Current status of the Th1-Th2 paradigm
The Th1-Th2 paradigm as formulated in mice, however, has begun to break down in its translation to the human immune response. For example, the expression of IL-10 seems to be restricted to Th2 clones in mice, whereas in both humans and rats, IL-10 is expressed by both Th1 and Th2 clones. Moreover, cells other than CD4+ T cells can express IL-10 (79). Mosmann and Kelso suggest that the attribution of graft acceptance versus graft rejection to predominance of a Th1- over a Th2-type immune response may represent an oversimplification of in vivo events (167a, 264). They point to the redundancy of cytokines able to mediate a particular effect and the pleiotropic nature of individual cytokines. The Th1-Th2 paradigm nonetheless remains useful in the interpretation of many aspects of the immune response to foreign tissues.| |
III. IMMUNOLOGIC TOLERANCE |
|---|
|
TopPreviousNextReferences
|
|---|
A. New Approaches to Transplantation
The immune response to allo- and xenografts reviewed above is remarkably complex and redundant. Indeed, even drugs that induce a state of general immunosuppression by interfering with many elements of the immune system are still incapable of preventing rejection when MHC mismatches are extensive. It is no wonder that precisely targeted, limited interventions capable of inducing graft acceptance have proven elusive. There remains no question, however, that alternative approaches to transplantation that circumvent the morbidity and mortality inherent in generalized immunosuppression are required.
Based on the recent discoveries in immune system recognition and activation, an appealing alternative approach to transplantation is the induction of tolerance to the foreign tissue. Tolerance induction is viewed by many as the most promising alternative to immunosuppression (93, 149, 314, 342, 373, 420).
B. Definition of Tolerance
Transplantation tolerance may be defined in two complementary ways. It is defined "clinically" or "functionally" as the survival of foreign (allogeneic or xenogeneic) tissue in normal recipients in the absence of immunosuppression. It is defined "immunologically" as the absence of a detectable immune response to a functional graft in the absence of immunosuppression.
The functional definition of tolerance is based on graft outcome. Tolerance in immunologic terms, however, may not be induced in all recipients of successful grafts. There are many reports of animals and humans bearing intact and functional grafts in the absence of immunosuppression while at the same time retaining demonstrable host versus graft immune reactivity (119, 385). In addition, a common feature of grafts in "tolerized" recipients is the presence of a mononuclear infiltrate (120, 380). It is unclear whether these infiltrates are destructive or protective, and it is possible that these cells play a role in the maintenance of the tolerant state.
The three functional hallmarks of immunologic transplantation tolerance are 1) the lack of demonstrable immune reactivity to graft alloantigens, 2) the presence of immune reactivity to other alloantigens, and 3) absence of generalized immunosuppression for graft maintenance. The immunologically tolerant graft recipient retains a functional graft, retains immune reactivity to all other foreign antigens, and avoids the risks associated with generalized immunosuppression.
C. Genesis of the Concept of Transplantation Tolerance
One of the first insights into the immunologic basis of transplantation tolerance was the description by Owen (294a) of freemartin cattle. These cattle are not monozygotic, but because of placental anatomy, they share a common circulation during gestation. Owen's observation was that pairs of freemartin cattle mutually accept allografts without the need for immunosuppression. Billingham et al. (27) then extended this observation by demonstrating that graft tolerance could be induced by exposure of neonatal animals to allogeneic cells. Based on these and other observations, Burnet (47) proposed the clonal selection theory to explain how the immune system distinguished self from nonself. Clonal deletion remains one of the major theories of immunologic tolerance today.
Burnet's theory proposes that clonal selection deletes all self-reactive immune cells in central lymphoid organs like the thymus before their release into the circulation. This theory predicts that all self-reactive cells are deleted before release into the periphery and, by definition, all activation of the normal peripheral immune system must be in response to the detection of nonself antigens. A corollary of the Burnet theory would be that elimination or reeducation of the subset of immune system cells that recognize a specific foreign graft should permit long-term tolerance of a graft in the absence of immunosuppression.
It is now recognized, of course, that clonal deletion alone cannot explain all aspects of self-tolerance. Self-reactive lymphocytes and autoantibodies have been detected in circulation of healthy individuals. These observations have required the development of new (but complementary) theories of tolerance based on suppression, lack of appropriate costimulation, and microenvironmental factors that prohibit self-reactive cells from mediating an immune response. These newer theories, in turn, have suggested additional potential mechanisms for inducing long-term tolerance of a graft in the absence of immunosuppression.
Most such theories have continued to be based on the Burnet self-nonself paradigm, but the concept of the immune response as recognition of "nonself" has recently been reinterpreted in terms of response to the recognition of "danger" (249). This theory holds that when tissue is damaged (e.g., by infection), molecular signals inform the immune system that danger is present and an appropriate local response is mounted. The theory is based largely on the evolutionary argument that tissue damage would indicate an infection and therefore require an immune response to eliminate the infectious agent. Resolution of the infection would halt tissue damage and ensuing danger signals, and the immune response would recede. This theory predicts that well-healed, quietly functional grafts may not generate danger signals and may be able to survive in the absence of immunosuppression.
The richness of the immune response and the theories that have developed to explain self-nonself discrimination suggest that a broad range of strategies may have the potential to induce long-term tolerance of a graft in the absence of immunosuppression. These theories and the strategies that they have inspired are discussed in section IIID.
D. Graft-Based Tolerance Induction
1. Fetal tissue and MHC knockouts
One obvious strategy for the induction of tolerance to a graft is to reduce its antigenicity, i.e., its ability to provoke a host immune response. Many approaches to the reduction of graft antigenicity have been investigated. The focus of most interventions has been the donor MHC. Much research has centered on the use of fetal tissues because they express lower levels of MHC antigens than do adult tissues (386). Because of the highly experimental nature of the concept and the unresolved ethical issues that surround it, few transplants of fetal human tissues have been attempted. One example is the use of fetal tissues to treat Parkinson's disease (3, 178, 179). The available animal data suggest that, although the antigenicity of fetal tissues may be less than that of corresponding adult tissues, the reduction in antigenicity is not enough by itself to ensure permanent graft survival (237, 239, 386). The concept remains intriguing, however. In animal research, proof-of-concept experiments have used knockout mice that do not express any MHC class I or MHC class II molecules (130), but graft survival was not permanent. Application of this technique to humans would require the development and implementation of xenogeneic knockouts and protocols for transplanting genetically manipulated animal organs.2. Masking
Another approach to modulating the antigenicity of the donor graft has been to "mask" the relevant immunologic recognition sites on transplanted tissues. Human pancreatic islet xenografts have been successfully transplanted into mice without immunosuppression simply by coating the donor islets with F(ab')2 fragments of anti-HLA class I antibody or antibody to tissue-specific epitopes (94). To date, however, no clinical reports of the successful use of this approach in humans have appeared.3. Privileged sites
Immunologically privileged sites in hosts that permit transplantation of grafts without the requirement for immunosuppression have been recognized for more than 50 yr (115, 275, 371). Immunologically privileged sites include the brain, testis, mammary and subcutaneous fat pads, thymus, anterior chamber of the eye, matrix of hair follicles, and the uterus during pregnancy. In Syrian hamsters, the cheek pouch has also been investigated as an immunologically privileged site. Many explanations of immunologic privilege have been offered. These have included a blood-tissue barrier that prevents access by the immune system to the site. This explanation is particularly relevant to privileged sites in the central nervous system. Other explanations have included 1) high concentrations of suppressor cells at the site, 2) soluble molecules secreted by the tissue that modulate local immune responses, and 3) imprisonment of graft antigens within the site, precluding a peripheral immune response. Immunologically privileged sites can also exhibit at least two characteristics that differentiate them from other sites. First, extracellular fluid within these sites generally does not drain directly via lymphatics into lymph nodes, the site where host immune responses are typically initiated. Second, the extracellular fluid in such sites can contain high concentrations of cytokines such as TGF-, a potent immunomodulatory cytokine that directs immune responses to nondestructive Th2-type rather than Th1-type responses.
Additional factors must, however, be involved, because infection in the privileged site does induce effector cell infiltration and local immune system activation. In addition, many grafts in privileged sites will survive despite the presence of an infection and nonspecific inflammation in the site.
At least four additional immunologic mechanisms have been invoked to account for the increasingly complex transplantation biology associated with these sites (392). Depending on the site, these include clonal deletion, clonal anergy, immune deviation, and T-cell suppression, each of which is discussed in a broader context below. None of these mechanisms is mutually exclusive.
A) ANTERIOR CHAMBER-ASSOCIATED IMMUNE DEVIATION. Transplantation of human corneas does not require immunosuppression, and for this reason, successful grafts of this tissue actually antedate the first human kidney grafts. These transplants were successful, at least in part, due to the presence of anterior chamber immune privilege (392). This system has been studied extensively, and it has been determined that privilege exists in part because antigens introduced into the eye are captured by distinctive APC that migrate via the blood to the spleen. There they generate a systemic immune response that is biased against the Th1-type CD4+ T cells that mediate delayed hypersensitivity and that help B cells to secrete complement-fixing antibodies. These APC do, however, stimulate CD8+ T cells that function as regulatory cells.
The term anterior chamber-associated immune deviation has been coined to describe this immune response. It appears to be mediated by antigen-specific regulatory T cells that secrete TGF- in an autocrine fashion and suppress effector functions of proinflammatory Th-1-type CD4+ T cells. Recent data suggest, however, that FasL (CD95L) expression by host cells is also involved in the maintenance of immunologic privilege at this site (126, 394). It is possible that two or more mechanisms may simultaneously be involved.
B) ROLE OF FAS-FASL (CD95-CD95L) IN IMMUNOLOGIC PRIVILEGE. Many activated cells express Fas (CD95), a cell-surface molecule belonging to the TNF family (267). When cells expressing Fas encounter cells expressing the counterreceptor FasL (CD95L), the binding of Fas and FasL leads the cell expressing Fas to undergo programmed cell death by apoptosis. The process is thought to be important generally in the modulation and termination of inflammatory immune responses. Because activated graft-reactive T cells express Fas, it was hypothesized that immunologic privilege might involve the expression of FasL within the privileged site.
The role of Fas-FasL interaction in immunologic privilege was first documented by the discovery that testis allografts survive indefinitely upon transplantation to allogeneic hosts and that this survival is associated with the expression of FasL (22). Testes from mutant gld mice with defective FasL expression do not survive in allogeneic hosts. Further analyses showed that FasL mRNA is constitutively expressed by testicular Sertoli cells and that Sertoli cells from normal mice, but not gld mutant mice (deficient in FasL), are accepted when transplanted into allogeneic recipients. Expression of FasL in the testis is thought to provide immunologic privilege by inducing apoptotic cell death in Fas-expressing recipient T cells that become activated in response to graft antigens (22).
C) ARTIFICIAL IMMUNOLOGIC PRIVILEGE. The ability of Sertoli cells to express high levels of FasL (CD95L) has prompted attempts to use them to provide immunologic protection for other transplanted tissues in a kind of artificial immunologically privileged site. A mixture of Sertoli cells with islets before transplantation into the renal subcapsular space reportedly results in islet allograft survival in the absence of immunosuppression (364).
Further evidence consistent with a role for FasL (CD95L) expression on graft survival, and with the concept of artificial immunologic privilege, was provided by Lau et al. (204). These authors expressed FasL on myoblasts, mixed the myoblasts with islets, and observed indefinite islet allograft survival in the absence of immunosuppression in chemically diabetic mice. Graft survival was not prolonged by mixtures of islets and unmodified myoblasts or Fas-expressing myoblasts. Close proximity of FasL-expressing myoblasts and grafted islets was required; islet allografts were rejected if the FasL-expressing myoblasts were placed in the contralateral kidney. The data were interpreted to suggest that the FasL signal provided site- and signal-specific protection of islet allografts.
The generation and maintenance of FasL-based artificial immunologic privilege is unlikely to depend solely on the FasL signal, however. For example, the site selected for the generation of artificial immunologic privilege affects the ability of FasL-expressing cells to survive. In one study, transfected hamster kidney cells expressing FasL survived indefinitely in the subrenal space but not subcutaneously in xenogeneic nude mouse recipients. This outcome was hypothesized to result from a FasL-mediated inflammatory reaction induced when the graft was placed subcutaneously but not subrenally (450). Expression of FasL is known to be able to generate a severe inflammatory reaction (5, 450).
Involvement of factors other than the FasL signal in artificial immunologic privilege has also been suggested by Allison et al. (5), who were unable to confirm the report of prolonged survival of subrenal Fas-L+ testis allografts mentioned above (22). Failure of Fas-L+ testis allografts has also been reported by others (386, 387).
Other attempts to exploit graft expression of FasL to achieve transplantation tolerance in the absence of immunosuppression have also failed (64, 163). Transgenic mice were engineered to express FasL under the control of the -cell-specific rat insulin promoter. In three experiments, expression of FasL on islets not only failed to prevent rejection but actually appeared to accelerate graft destruction. Graft destruction was due to expression of Fas by -cells in the presence of activated T cells, resulting in -cell self-destruction (64). Furthermore, in the FasL-expressing transgenic mice, a pancreatic granulocyte infiltration developed, suggesting that FasL expression may have proinflammatory activity when expressed in specific tissues (5). These observations have prompted cautious reevaluation of the use of FasL-expressing cells to induce artificial immunologic privilege for transplantation tolerance induction.
E. Host Response-Based Tolerance Induction
Induction of transplantation tolerance by altering host responses to the graft fall into two general categories, central and peripheral. We discuss each in detail, but before doing so, we review the mechanisms of T-cell activation that are central to all of the strategies in each category.
1. The two-signal hypothesis
How immune responses are initiated has been investigated intensively, and historically, many of the investigations have been tied closely to transplantation biology. Nearly 30 yr ago, Bretscher and Cohn (43) observed that presentation of antigen to the immune system was necessary but not in itself sufficient to initiate an immune response, in that case the production of antibody by B cells. Based on this finding, they proposed the concept of functional deletion or "paralysis." They introduced a model for B-cell activation that featured a requirement for two signals to achieve activation (43). Bretscher and Cohn (43) proposed that recognition of antigen by a B cell (signal 1) sent a signal that, in the absence of any further intervention, would paralyze the cell. To overcome this paralysis in appropriate circumstances, they proposed that simultaneous delivery of a second signal would induce B-cell responsiveness. They hypothesized that signal 2 came from a T cell. Signal 2 is now termed "costimulation." The "two-signal" concept was extended to T-cell activation some 5 yr later by Lafferty et al. (192), who proposed that T cells require not only the recognition of antigen in the context of MHC (signal 1) but an additional signal (signal 2) to become fully activated (192). They proposed that signal 2 came from hematopoietic cells, specifically APC. The proposal was based on the finding that survival of allogeneic thyroid grafts was prolonged if the tissue was first cultured in vitro for several days to remove passenger leukocytes, i.e., APC. Their observations were quickly extended to islets, and it was shown again that culture to remove passenger leukocytes prolongs allogeneic islet graft survival (21, 411). The two-signal concept was further developed in in vitro experiments, demonstrating that the delivery of signal 1 (antigen) to T cells in the absence of signal 2 (costimulation) shut down IL-2 production, downregulated TCR expression, and induced a state of nonresponsiveness termed anergy (159, 265, 330, 352, 355-357). The process of activating T cells by MHC plus peptide (signal 1) and costimulation (signal 2) is now understood at the molecular level and is known to be more complex than originally thought. To achieve full T-cell activation, three (not two) receptor-ligand interactions must occur. The first interaction is antigen-specific binding of the TCR to peptide-MHC complexes (signal 1). The second is binding of CD154 (CD40 ligand) on T cells to CD40 on APC (coactivation). The third is binding of B7-1/2 on APC to CD28 on T cells (signal 2, costimulation). With certain exceptions not germane to our discussion of tolerance, these three interactions characterize the T-cell response to foreign tissue. These three steps leading to T-cell activation are depicted schematically in Figure 3.
|
The MHC-peptide-TCR interaction (signal 1) was described in Figure 2 and section IIC5. The coactivation and costimulatory interactions are described below. A) CD40-CD154 (CD40L) COACTIVATION. Antigen-specific ligation of the TCR with a specific peptide-MHC complex (signal 1) induces the rapid but transient upregulation of CD154 (also called gp39 or CD40 ligand) on the surface of the T cell. The expression of CD154 on T cells then permits the T cell to activate APC (principally the one that is delivering signal 1). This occurs through the interaction of constitutively expressed CD40 on the APC with the newly expressed CD154 on the T cell (197, 277, 295, 381). CD154 is a glycoprotein expressed predominantly on the surface of activated CD4+ T cells (277, 295), but also on other types of activated cells (66, 103, 195). B) B7-1/2-CD28 (CD80/CD86-CD28) COSTIMULATION. The engagement of CD40 with its ligand, CD154, results in the rapid upregulation of the B7-1 and B7-2 costimulatory molecules on the surface of APC (66, 103, 168, 319). The upregulation of B7-1/2 (CD80/CD86) comprises the signal 2 postulated by Lafferty et al. (192) to be required for T-cell activation. This signal is actually transduced to T cells by the receptor-ligand binding of newly expressed B7-1/2 on APC with constitutively expressed CD28 on T cells. Encounter of T cells with antigen in the absence of costimulation by B7-1/2 through CD28 results in their failure to become activated (59, 159, 220, 221, 357, 409, 414). These processes are depicted schematically in Figure 3. C) CTLA4-B7-1/2 COSTIMULATION. A second receptor for B7-1/2 on T cells has been discovered and designated cytotoxic T lymphocyte antigen-4 (CTLA4) (44, 221, 357). This CTLA4 is upregulated on T cells 48-72 h after their activation by encounter with antigen and CD40-CD154 coactivation. Expression of CD28, in contrast, is constitutive. The CTLA4 binds to B7-1/2 with 10- to 100-fold greater affinity than does CD28 (165, 357). The role of CTLA4 in regulation of T-cell activation is a subject of considerable interest and controversy. CD28 binding to B7-1/2 is generally believed to deliver signal 2 to T cells and complete the process of T-cell activation. CTLA4 binds to B7-1/2 with higher affinity than does CD28 and competes with CD28 for this ligand. Binding of CTLA4 to B7-1/2 is thought by some to deliver additional costimulation and by others to interfere with costimulation and to deliver a downregulatory signal. Many lines of evidence are consistent with the hypothesis that CTLA4 downregulates activated T cells (165, 184, 413, 417). The most convincing evidence has been provided by CTLA4 knockout mice (415, 434). These animals develop a severe lymphoproliferative disease and die early in life. The disease can be prevented by administration of CTLA4-Ig fusion protein, a hybrid molecule that blocks B7-1/2 interaction with CD28. Cessation of CTLA4-Ig treatment leads to recrudescence of the lymphoproliferative disease and death (416). Another study has shown that anti-CTLA4 antibody treatment of young transgenic mice with a diabetogenic TCR is permissive to the expression of autoimmune diabetes (231). Treatment of mice with anti-CTLA4 monoclonal antibody (MAb) has also implicated CTLA4 as an important downmodulator of T-cell activation. Treatment with a blocking anti-CTLA4 MAb is permissive to increased T-cell proliferation (222); a different anti-CTLA4 MAb can induce apoptosis of T cells, the ultimate form of downregulation (125). Ligation of CTLA4 may also deliver a signal that is necessary for tolerance induction, another manifestation of the downregulation of T-cell activation (187, 243, 305). Although these data strongly suggest that CTLA4 acts primarily to downregulate immune responses, other data suggest that the opposite effect may occur in certain circumstances. For example, CD28-deficient mice can activate T cells and induce their proliferation through the CTLA4 molecule (448). Moreover, the antibodies used to define an immunomodulatory role for CTLA4 may in fact be blocking antibodies that prevent the enhancement of T-cell responsiveness by CTLA4 (223). D) THE B7 FAMILY OF COSTIMULATORY MOLECULES. The expression of B7-1 (CD80) and B7-2 (CD86) follow different kinetics. The molecules appear at different times during the development of an immune response (144, 449). Costimulation provided through these molecules may generate different cytokine profiles in stimulated T cells (107, 185). Ligation of CD28 by CD86 appears preferentially to induce expression of IL-4, a Th2-type cytokine, whereas ligation of CD28 by CD80 appears to induce a less biased cytokine profile. The CD80/CD86 system is growing in complexity. Expression of CD80 and CD86 may be differentially regulated by cytokines in the microenvironment in which they interact with CD28 (202). Antibodies to B7-1 and B7-2 have differential effects on immune responses, especially on autoimmune responses (185, 213). In addition, at least three forms of CD80 have been identified in mice (16, 40, 106, 133, 143, 154, 155). Molecules in this B7 family are designated B7-1, B7-1a, and B7-1cytII. These molecules exhibit differential and variable expression, diverse immune response effects, and a range of binding affinities to CD28 and CTLA4. The differential activity of each in relation to the provision of costimulation is yet to be determined (34, 300). For the purposes of this review on induction of transplantation tolerance, the key observation is that blockade of B7-CD28 interaction has potent immunosuppressive and perhaps tolerance-inducing properties in allograft and xenograft transplantation (see sect. IVA8). E) ACTIVATION OF NAIVE VERSUS MEMORY CELLS. We conclude this discussion of the two-signal model with the observation that costimulation (and cytokine) requirements for activation of naive versus memory T cells are very different. The former have never been activated; the latter have been activated by antigen exposure and retain the memory of that exposure. Much more stringent criteria must be met for the activation of naive T cells than for the reactivation of memory cells (41, 87,
), natural killer cells (NK), and B cells (B). CD8+ T cells interact with major histocompatibility complex (MHC) class I plus peptide present on graft, including its endothelial cells; CD4+ T cells interact with MHC class II plus peptide present on both graft and endothelial cells. Endothelial cells are depicted as activated in response to inflammatory cytokines, expressing surface molecules that include both MHC molecules and adhesion molecules. Bottom depicts mechanisms of graft destruction associated with various responding host immune cells. Finally, also depicted is possibility of deviation of immune response by Th2-type cytokines produced by CD4+ T cells. ADCC, antibody-dependent cell-mediated cytotoxicity.
