The Future of Organ and Tissue Transplantation: Title and subTitle BreakCan T-Cell Costimulatory Pathway Modifiers Revolutionize the Prevention of Graft Rejection?

Curative transplantation has become a widely accepted and utilized treatment modality. In 1997 alone, nearly 20,000 Americans received therapeutic and often life-saving organ transplants.¹ Unfortunately, curative transplantation is made available to only a fraction of those who need it. United Network for Organ Sharing data indicate that patients in need of organ transplants outnumber available organs 3 to 1. Even that statistic minimizes the health benefit that could be achieved were organ and tissue transplantation more widely available. An estimated 75% of annual US mortality is caused by chronic illnesses (eg, cardiovascular and cerebrovascular diseases, chronic obstructive pulmonary disease, chronic liver disease and cirrhosis, malignant neoplasms, and diabetes) that would benefit from more widely available transplant therapies.²

Several factors, aside from the inadequate organ supply, limit more widespread transplantation. Despite great improvements in immunosuppressive therapies to prevent acute rejection, chronic graft rejection remains a significant problem. For example, half of renal allografts fail within 10 years of transplantation.^{3 - 4} Indeed, the majority of kidney grafts that fail after the first year are lost to chronic graft rejection.⁵ Unfortunately, while immunologists understand the mechanisms underlying acute graft rejection in some detail, chronic rejection is less well understood and probably represents several different processes (immunological and nonimmunological) that converge to negatively affect allograft function.⁶ Because the primary histological feature of chronic graft rejection is a proliferative vasculopathy, several investigators have explored the role growth factors may play in the process.⁷ Important factors also shown to influence the incidence of chronic graft rejection include the number of acute rejection episodes that have occurred, donor organ quality, the use of certain immunosuppressive drugs, and more classic risk factors for vascular disease, such as elevated blood pressure and serum lipid levels.⁵

In any case, the shortcomings of current therapies have prevented life-saving skin allografts for serious burn victims, curative pancreatic islet cell transplants for patients with type 1 diabetes,^{8 - 10} and small bowel transplants for patients with short-bowel syndromes.¹¹ Even when current immunosuppressive therapy is effective, as it is in most patients receiving solid organ transplants today, the therapy is costly,¹² usually must be taken daily for life, and is associated with a significantly increased risk for (depending on the specific agent used) infection, cancer, and other morbidity, including osteoporosis, hyperglycemia, cataracts, and renal dysfunction.^{13 - 18} Transplant recipients also must comply with a fairly complex daily drug-dosing schedule, and they must be followed up closely to minimize the risk of known complications while they strike a balance between too much or too little immunosuppression. There is a real need for improved ways to maintain allograft function by preventing immune system–mediated rejection.

All multicellular organisms have systems that defend against invasion by other biological organisms. Even a sponge will destroy tissues grafted from an unrelated sponge.^{19 - 20} Thus, it is safe to surmise that multicellular organisms without immune systems do not survive long under evolutionary pressure, since none are known to exist today. Transplantation of tissues or organs from one member of a species to another is not a likely natural event, however, so the immune response to an allograft is a by-product of the response to more immediate threats to the survival of a species.

Evolutionary immunologists have subdivided the immune system into 2 broad and interrelated systems, the innate and adaptive immune systems.²¹ The innate system, the more primitive of the 2, consists of phagocytic cells that use a variety of clues to recognize, engulf, and digest invading pathogens. In mammalian species, the innate immune system consists of polymorphonuclear cells, monocytes, macrophages, and dendritic cells. The adaptive immune system first appeared some 400 million to 600 million years ago along with vertebrate evolution, and a more advanced system, with definable T and B lymphocyte subsets, evolved with the common predecessor for all mammalian and avian species some 200 million to 300 million years ago. The defining characteristic of this advanced system is the expression, on the surface of the adaptive immune system cells (T and B lymphocytes), of antigen-specific receptors designed to recognize only 1 target. B lymphocytes make antigen-specific antibodies, and T lymphocytes are responsible for what is called the cell-mediated immune response. It is this cellular immune response that is primarily responsible for the rejection of allografted tissues and organs.

The T lymphocyte is primarily responsible for directing the immune response that recognizes and then destroys cells damaged by toxins or viruses. The T lymphocyte also directs the immune response against cells, tissues, or organs transplanted from another member of the same species. Each adult human has approximately 10¹⁰ to 10¹² T lymphocytes, any 1 of which can recognize only 1 target. But since most specific lymphocytes exist in multiple copies, it is estimated that a normal human's T lymphocyte repertoire can recognize approximately 10⁹ to 10¹⁰ different antigens.²²

T lymphocytes, however, do not recognize whole antigens. Rather, antigen-specific T-cell receptors (TCRs) recognize small peptide fragments of whole antigens presented by the major histocompatibility complex (MHC) proteins present on cell surfaces (Figure 1). Class I MHC molecules, present on nearly all tissue cells, present fragments of proteins made within the cell (eg, both self-proteins and proteins that may be encoded by infecting virus genes). Class II MHC molecules, found only on specialized antigen-presenting cells, present fragments of antigens made outside the cells, which are then phagocytosed, digested, and finally presented by the specialized, antigen-presenting cell. The MHC molecules themselves are polymorphic, so it is highly unlikely that 2 unrelated individuals will have identical MHC types. Indeed, this donor-host MHC disparity is primarily responsible for triggering the antigraft immune response and is the reason that measures are taken to scan the list of potential recipients for an MHC match whenever a donor organ becomes available. Finding a complete match greatly decreases the likelihood of a vigorous immune response against the transplanted tissue.⁴

T lymphocytes are often categorized into 2 subgroups based upon the presence of other protein complexes on their surface, the so-called CD4⁺ and CD8⁺ T-cell subsets (Figure 1). CD4⁺ T cells recognize antigens presented by the MHC class II complexes only on professional antigen-presenting cells (cells of the innate immune system), while CD8⁺ T cells recognize antigens presented by MHC class I complexes. Both CD4⁺ and CD8⁺ T cells play important roles in the immune response that results in graft rejection. Moreover, both the innate and the adaptive immune systems discussed above have important roles in the antiallograft immune response.

Approximately 1% of an individual's T cells are cross-reactive for another (unmatched) individual's MHC complex proteins. Thus, approximately 1% of a recipient's T cells might directly recognize MHC molecules present on transplanted tissue cells. In addition, when the host's innate immune system cells encounter damaged tissue or cells within the transplant (from surgical or ischemic injury), those damaged donor cells are ingested and antigens from them are presented to the host T cells (indirect recognition), further promoting the antigraft immune response.

While TCRs must recognize MHC-presented antigen peptides to activate an antigen-specific T cell, Lafferty and Cunningham²³ suggested that recognition alone was not sufficient to activate a T cell. Their study expanded on earlier work by Bretscher and Cohn^{24 - 25} and resulted in the development of the 2-signal model for T-cell activation (Figure 2). The model proposed that TCR recognition of an appropriately presented antigen would deliver a signal 1 to the T cell, but that a simultaneously delivered signal 2 (the costimulatory signal) was required to activate the T cell. Two important corollaries resulted from this hypothesis. First, a signal 2 delivered without antigen recognition was a neutral event for the T cell. Second, if a T cell encountered its cognate MHC antigen but did not receive a signal 2 at the time of the antigen recognition, that T cell would either die or be rendered resistant to activation in future encounters with that antigen.

The second corollary suggested a novel way to prevent an antiallograft immune response after therapeutic transplantation. The other typical antirejection regimens used therapeutically to date have resulted in nonspecifically blunted T-cell responses. For instance, glucocorticoids and calcineurin phosphatase inhibitors (cyclosporine and tacrolimus) are now known to impair TCR-mediated signaling (ie, signal 1) to the T cell. Other agents in wide use, such as azathioprine and mycophenolate mofetil, interfere with purine synthesis and therefore impede all rapidly proliferating cells, such as T cells responding to an antigen challenge. The most nonselective tools were those that indiscriminately depleted T cells, eg, hemi-body irradiation, thoracic duct drainage, and agents such as antithymocyte globulin.⁴ The 2-signal model suggested that an agent that specifically interrupted the proposed T-cell costimulatory receptor when the T cells first encountered the transplanted organ could inactivate only those antigen-specific T cells, leaving unimpaired other T cells that did not encounter their cognate antigens.

June and colleagues²⁶ found that stimulating T cells with a mixture of 2 antibodies, 1 that stimulated the TCR and another directed against a previously uncharacterized 44-kd cell surface receptor, induced rapid T-cell proliferation that, unlike all other ways of activating T cells via surface receptor antibodies, was not inhibited with cyclosporine.²⁷ The cell surface receptor identified by that antibody, originally called Tp44 and later named CD28, was the first of several costimulatory T-cell receptors to be reported and extensively studied. While costimulatory receptorlike function has since been reported for several receptor–counter-receptor pairs, the 2 shown in Figure 3, A and B, have been of particular interest to transplant immunologists because of their ability to prevent allograft rejection in experimental animal models.

The CD28-B7 counter-receptor group consists of 4 unique receptors. Two B7 receptors, B7-1 (or CD80) and B7-2 (or CD86), have been cloned^{28 - 30} and are known to be expressed by activated antigen-presenting cells, albeit with slightly different kinetics and levels of expression. The interaction of either B7 with a T cell's CD28 can costimulate T-cell activation if a TCR-generated signal is delivered simultaneously.³¹ T cells also express another receptor called CTLA4 (or CD152) that also serves as a counter-receptor for both B7 receptors.^{32 - 33} While the functional relevance of CD152 was debated for several years after its identification, 2 simultaneous reports^{34 - 35} describing the phenotype of CD152 knock-out mice cleared up much of the confusion. Both articles reported that mice with disrupted CD152 genes died prematurely, showing evidence of profound lymphocyte activation manifested by massive splenomegaly, lymphadenopathy, and lymphocytic infiltration of the heart, pancreas, and other parenchymal tissues. Thus, while many of the details remain obscure, CD28-B7 interactions stimulate T-cell activation, while CD152-B7 interactions seem to restrain T-cell responses.

Linsley et al³³ sought to develop a reagent that would block the CD28-B7 interaction. Using molecular biological techniques, they created chimeric molecules, composed of the extracellular domain of either CD28 or CD152 coupled with the IgG heavy chain (CD28-Ig and CTLA4-Ig, respectively). CD28-Ig was subsequently found to ineffectively interfere with CD28-B7 signaling, but CTLA4-Ig³⁶ appeared to compete effectively with CD28 for binding to both B7 ligands and therefore to prevent T-cell activation.

The CD40-CD154 receptor pair (Figure 3, B) has an even more recent history in transplantation immunology. B lymphocytes express CD40, a cell surface receptor known to be important in mediating immunoglobulin class switching from the early IgM response to the later IgG response following antigen stimulation. The study of hyper-IgM syndrome determined that such patients had a defect in the T-cell counter-receptor for the B cell's CD40.³⁷ This T-cell counter-receptor was originally called gp39, then CD40 ligand (or CD40L), and, most recently, CD154. Thus, while CD154 was identified because B-lymphocyte function is defective when it is genetically absent, dramatic anti-CD154 antibody effects on T-cell function were subsequently identified.

Interest in T-cell costimulatory receptor signaling greatly increased when Lenschow et al³⁸ reported a landmark study demonstrating that costimulatory receptor blockade could specifically prevent graft rejection in mice. Diabetes was induced in mice by administering a pancreatic β-cell toxin (streptozotocin), and the mice were then transplanted with human pancreatic islets injected under 1 kidney capsule. The mice were temporarily cured of diabetes, but the human islets were rejected a few days later, and disease recurred. Lenshow et al demonstrated that CTLA4-Ig administered during the 2 weeks immediately after the transplantation allowed for long-term graft function and diabetes-free survival. They also reported that if the kidney containing the human islets was excised several weeks after the transplant, diabetes promptly recurred, demonstrating that the transplanted human islets were responsible for maintaining euglycemia. Islets injected under the animal's contralateral kidney capsule following the surgical removal of the initially transplanted islets could permanently cure diabetes only if the islets were from the original human donor; islets harvested from another human donor were rejected. It appeared that the mouse immune system had learned to recognize tissues from the original human donor while responding normally to similar cells (islets) from a different human donor. Other reports followed, demonstrating that allograft survival in rodents could be induced reproducibly by agents interfering with CD28-B7 signaling.^{39 - 42} More recent reports have suggested that such agents could even abrogate the vasculopathy characteristic of chronic rejection.^{43 - 46}

The first reports suggesting that agents modulating the CD40-CD154 interaction could similarly induce long-term graft survival first appeared in 1995. Parker et al⁴⁷ found that donor-specific small lymphocytes and anti-CD154 given before and for 7 weeks after the transplant of allogeneic islets under a kidney capsule permanently prevented graft rejection and relapse of diabetes. Moreover, they proposed that the mechanism of graft acceptance was similar to that induced by CTLA4-Ig,⁴⁸ that is, small lymphocytes are known to express little or no B7. Furthermore, the interaction of CD40 on small B lymphocytes with CD154 present on activated T cells is known to lead to the up-regulated expression of B7 on B lymphocytes. Rossini and colleagues⁴⁸ proposed that the anti-CD154 had prevented B7 expression from being up-regulated, allowing the MHC molecules present on the donor small lymphocytes to interact with the host's anti–donor-specific T cells without receiving the required B7-CD28 costimulatory signal, thus making the host tolerant to the donor tissues. Several other reports documenting the graft-sparing effect of anti-CD154 have since appeared.^{48 - 53}

While these rodent studies generated considerable interest, transplant immunology literature is filled with promising therapies developed in mouse transplant models that have failed in other species. Indeed, using conditions similar to those that resulted in great success in rodents, we⁵⁴ and others⁵⁵ found that CTLA4-Ig did not reliably prevent allograft rejection in nonhuman primates. Regardless of the relative ease with which graft acceptance can be induced in mice, no simple and nontoxic regimen had ever been reported to prevent the rejection of full-thickness allogeneic skin grafts, even in mice, prior to a study reported by Larsen et al.⁵⁶ Reasoning that anti-CD154 prevented the expression of B7 from being up-regulated, and CTLA4-Ig bound to B7 and prevented B7-CD28 signaling (Figure 4), Larsen and colleagues tested the effect of both agents in combination using a mouse full-thickness skin allograft model. Remarkably, while neither agent alone was effective, a short course of both agents combined allowed long-term skin allograft survival in nearly all experimental animals tested. Equally surprising was their observation that the graft-promoting effect of CTLA4-Ig plus anti-CD154 was abrogated if the agents were coadministered with cyclosporine. Because cyclosporine interferes with the signal transduced by the antigen-engaged TCR (signal 1 in the 2-signal model), these data strongly suggested that the costimulatory pathway–modifying reagents were not simply another class of immunosuppressive drugs. Indeed, the agents appeared to require TCR-mediated antigen recognition for their full therapeutic effect.

Upon learning of the efficacy achieved using combination CTLA4-Ig plus anti-CD154 therapy in the murine full-thickness skin allograft model, we tested a similar regimen in a rhesus monkey renal allograft model.⁵⁴ Adolescent rhesus monkeys underwent bilateral nephrectomy and then were given purposely mismatched renal allografts. All the control allografts were rejected within 8 days, as expected. The animals treated with a combination of CTLA4-Ig plus anti-CD154 for 28 days posttransplant all maintained normal graft function for at least 6 months. In contrast to the results reported in the mouse transplant models, however, anti-CD154 alone was effective at preventing acute renal allograft rejection. Consistent with the rodent studies, the costimulatory pathway–modifying reagents were well tolerated. The animals' hematologic parameters were not affected, they did not display any evidence of susceptibility to infectious agents, their wounds healed normally, and they gained weight consistent with colony mates. Since those studies were published, we have conducted others to determine the minimally effective dose of the costimulatory pathway–modifying reagents, whether the agents can maintain their effectiveness when administered with traditional graft-sparing agents, and tests to further delineate the mechanism underlying the agents' graft-sparing effect. Anti-CD154 monotherapy is highly effective at preventing acute renal allograft rejection.⁵⁷ For instance, 2 animals given anti-CD154 in 6 doses over the first month after their kidney transplantations demonstrated impressive graft survival; 1 animal rejected the kidney 120 days posttransplantation, and the other continues to display normal graft function more than 750 days after surgery (ie, no antirejection therapy for more than 700 days). Nine additional animals that received a kidney allograft, 6 anti-CD154 doses over the first month, and monthly doses thereafter for 5 subsequent months did not reject their grafts. However, consistent with the mouse data that preceded our primate studies, traditional immunosuppressive agents appeared to abrogate anti-CD154 effectiveness. The outcome in animals treated with tacrolimus, steroids, or mycophenolate mofetil was not uniformly successful.⁵⁷ Most intriguing, allograft biopsies revealed that T cells that appeared to be activated infiltrated the transplanted tissue, yet those cells did not invade the kidney blood vessels, tubules, or glomeruli.³⁷ Moreover, we have collaborated with other investigators to test whether the reagents can support the survival of islet allografts transplanted into rhesus monkeys that underwent pancreatectomy. Those experiments have yielded similar results.⁵⁸

The immunological background of these developments in transplantation immunology have recently been reviewed in greater detail elsewhere.⁵⁹ The simple model shown in Figure 4, as first proposed by Lenschow et al³⁸ and Rossini et al⁴⁸ to explain the ability of costimulatory pathway–modifying reagents to prevent the rejection of transplanted tissues and organs, is now known to be oversimplified.⁶⁰ For instance, the grafts of animals given costimulatory pathway–modifying reagents to prevent transplant rejection express B7 molecules to a degree that is indistinguishable from that seen in organs undergoing rejection in untreated animals.⁵⁶

It is currently believed that the interaction between T cells (of both the CD4 and CD8 subsets) and cells of the innate immune system underlies the immune response that leads to the rejection of a transplanted organ (Figure 5). Donor tissue antigens are released as a result of the tissue damage that invariably results during harvest and/or transplantation, and the host's innate immune system cells, including dendritic cells, pick up those foreign antigens. In addition, since inflammatory cytokines are locally released as a result of simple surgical trauma, local dendritic cells in the area of the transplant are partially activated. Resting dendritic cells express low levels of MHC class II and B7 costimulatory ligands, and they do not express cytokines such as interleukin (IL) 12. But when dendritic cells are activated, they produce more IL-12 and more highly express MHC class II and B7 molecules. If 1 of these partially activated dendritic cells presenting a fragment of the foreign MHC encounters a CD4 T cell with a TCR specific for that MHC antigen, that antigen-specific T cell will become activated.

While the resting T cell expresses little or no CD154, the activated CD4 T cell rapidly increases its CD154 expression. At this point, a positively reinforcing cycle is established. The CD4 T cell's CD154 serves as a ligand for the dendritic cell's CD40, which, in turn, leads to further activation of the dendritic cell. Fully activated dendritic cells highly express several gene products (eg, MHC class I and II molecules, B7 costimulatory molecules, and adhesion molecules) and produce more inflammatory cytokines such as IL-12. Based on studies by Ridge et al,⁶¹ Schoenberger et al,⁶² and Bennett et al,⁶³ Lanzavecchia described this fully primed dendritic cell as one that can pass along a "license to kill,"⁶⁴ in that it can effectively stimulate the maturation of CD8 T cells called cytotoxic T lymphocytes, (CTLs) known to be required for full antigraft immune responses (Figure 5).

Given this proposed sequence of events, anti-CD154 could interrupt the escalating sequence of activation at the outset by blocking the CD4 T cell from further activating the partially activated dendritic cell, thus preventing the maturation of CTLs. Even this model is oversimplified, however. For example, anti-CD154 can appear to interrupt an ongoing rejection episode in primates.⁵⁴ Other studies have reported that anti-CD154 can prevent ongoing autoimmune syndromes in laboratory animals, even late in the disease process.^{65 - 66} Furthermore, the model suggests only an indirect role for anti-CD154, that of blocking CD4 T cells from fully activating dendritic cells. However, we have identified some evidence for a direct effect of anti-CD154 on inactive CD4 T cells (P. J. Blair, PhD, unpublished data, 1999).

Advances in immunosuppressive therapy have resulted in organ transplants that have protected thousands of patients from end-stage disease. However, the immunosuppression paradigm responsible for many of the gains made to date may have reached the limit of usefulness. Newer immunosuppression regimens have led to proportional increases in posttransplant morbidity, with little additional improvement in graft or patient survival.^{67 - 68} The goal of current clinical research in transplantation must shift toward developing antirejection therapies that have long-lasting efficacy and rely less on global immunosuppression. Research in T-cell costimulation has disclosed the previously underappreciated ability of the immune system to finely regulate its activity. By exploiting these regulatory mechanisms, durable rejection-free allograft survival might be feasible using intermittently dosed, relatively nontoxic agents. Clinical trials rigorously investigating these new agents are warranted, and their findings will be anticipated with great interest.