Pathways of allorecognition: implications for transplantation tolerance

Transplant Immunology 10 (2002) 101–108

Review

Pathways of allorecognition: implications for transplantation tolerance David S. Game, Robert I. Lechler* Department of Immunology, Imperial College School of Medicine, Hammersmith Campus, Du Cane Road, London W12 0NN, UK

Abstract Allorecognition occurs when the host immune system detects same-species, non-self antigens and this is the trigger for allograft rejection. Host T cells detect these ‘foreign’ antigens which are mostly derived from a highly polymorphic region of the genome called the major histocompatibility complex. Allorecognition can occur by two distinct, but not mutually exclusive pathways: direct and indirect. The direct pathway results from the recognition of foreign major histocompatibility molecules, intact, on the surface of donor cells. Indirect allorecognition occurs when donor histocompatibility molecules are internalised, processed, and presented as peptides by host antigen presenting cells—this is the manner in which the immune system normally sees antigen. However, in addition to antigen recognition, T cell activation requires the provision of costimulatory signals, the prerogative of bone marrow-derived, specialised antigen-presenting cells (APC). Once these have been depleted from a transplanted organ, as occurs within weeks of transplantation, the parenchymal cells of the transplant are incapable of driving direct pathway activation of recipient T cells. Alloantigen recognition on these non-professional APCs may have a tolerising effect and indeed, the frequency of T cells reactive to the direct pathway diminishes with time irrespective of whether or not chronic transplant rejection occurs. This implies that while the direct pathway plays a dominant role in acute rejection, it is unlikely to contribute to chronic rejection. Assays of T cell responses have, however, found an association between the indirect pathway and chronic rejection and animal models support a role for the indirect pathway in both acute and chronic rejection. The indirect pathway is likely to be permanently active due to traffic of recipient APCs through the graft. The challenge that this poses in the pursuit of clinical tolerance is how to induce tolerance in T cells with indirect allospecificity. The answer may lie in manipulation of the environment of the interaction between the T cell and APC. Apart from recognition without costimulation, there are other circumstances when recognition without activation can occur although the in vivo relevance is uncertain. The presence of regulatory cytokines or inhibitory surface molecules either from a distinct regulatory cell, or as a negative feedback loop may prevent activation; this could also happen without sufficient stimulatory support: the final outcome is likely to be decided by the overall balance. Furthermore, some peptides may act as antagonists to T cell activation, usually when the agonist peptide is structurally very similar. It is hoped that the careful study of these mechanisms will reveal ways of ensuring allorecognition without activation and thus donor-specific tolerance. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Allorecognition; Tolerance; Costimulation; Regulation

1. Introduction The afferent limb of the alloresponse, that is allorecognition, refers to T cell recognition of genetically encoded polymorphisms between members of the same species. The main targets of the immune response to allogeneic tissues are the major histocompatibility complex (MHC) molecules, which are present on the donor cells. Indeed, the term MHC highlights the fact that MHC molecules were discovered in the context of tissue transplantation between incompatible individuals w1x. *Corresponding author. Tel.: q44-208-3833255; fax: q44-2083832788. E-mail address: [email protected] (R.I. Lechler).

This recognition of allograft MHC antigen is the primary event that ultimately leads to graft rejection. The T cell response to allogeneic MHC molecules is uniquely strong, as reflected in vitro in the mixed leukocyte reaction (MLR) and in vivo by the vigour of transplant rejection w2x. Although MHC incompatibility can provoke strong immune responses, rapid graft rejection can also occur when MHC-matched tissues are transplanted. This is due to the recognition of minor histocompatibility (mH) antigens, peptides derived from allelically polymorphic host proteins, other than MHC molecules, presented in the groove of MHC class I and II molecules. These antigens have been identified in mouse and man encoded

0966-3274/02/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 6 6 - 3 2 7 4 Ž 0 2 . 0 0 0 5 5 - 2

102

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

on autosomes, sex-chromosomes and mitochondrial DNA. It is estimated that approximately 40 mH antigen differences exist between strains of inbred mice and several mHAgs have been identified in humans, the first being HA-2, restricted by HLA-A2 and probably originating from the class I myosin family. Immunity against mH antigens is a clinically significant problem as evidenced by the need to give systemic immunosuppression to recipients of HLA-identical organ grafts, and by the incidence of graft vs. host disease following HLA-identical stem cell transplantation. Most of the work in allorecognition utilises differences in major histocompatibility antigens and therefore, for brevity, minor antigens will not be further discussed: for a review see Hernandez-Fuentes w3x. Alloreactive T cells recognise alloantigens via two distinct, but not mutually exclusive, pathways: direct and indirect. Direct recognition occurs when recipient T cells recognise intact donor MHC molecules complexed with peptide on donor stimulator cells. In contrast, indirect recognition occurs when the recipient APC process the donor-MHC molecules prior to presentation to recipient T cells in a self-restricted manner (Fig. 1). Whilst two pathways of allorecognition have been proposed, the relative contributions of these pathways have not been clearly defined although some patterns are emerging. Historically, evidence favours a predominant role of the direct pathway especially during acute rejection and initiation of the allograft response. The majority of data published prior to the mid-1980s focused on this pathway. Recent findings, however, emphasise the role of the indirect route of allorecognition w4x. With the development of knockout and transgenic mice it has been possible to study the different pathways of rejection in isolation. The direct and indirect pathways are not necessarily mutually exclusive: both may be involved in mediating allograft rejection either simultaneously or at different times. 2. The molecular basis of direct allorecognition It was noted several decades ago that, underlying the strength of proliferative response of the MLR, was a uniquely high frequency of T cells with direct allospecificity. Evidence from a variety of sources, including a recent structural analysis of an alloreactive TCR w5x, indicates that this mode of allorecognition results from cross-reactivity by T cells specific for a self MHC molecule ‘A’ with peptide ‘x’ on an allogeneic MHC molecule ‘B’ with peptide ‘y’. The generality of the cross-reactive hypothesis was shown by the observation that a large fraction of the direct alloresponse was derived from T cells with a memory phenotype, which suggests that these cells had been primed against foreign antigens in the context of self MHC molecules w6,7x. Two hypotheses have been proposed to account for the

high precursor frequency of alloreactive T cells within the direct pathway of allorecognition. The high determinant density model proposes that the specificity of the alloreactive T cells’ receptor is for exposed amino acid polymorphisms on the foreign MHC molecule: the nature of the peptide in the binding groove is not relevant w8x. By this reasoning every foreign MHC molecule of a given type could act as a ligand for the alloreactive T cell, thereby creating a very high ligand density. As a consequence the affinity of the alloreactive T cell’s receptor could afford to be substantially lower that that required for a peptide–self-MHC complex, thereby calling into the alloreactive T cell repertoire T cells with low, medium and high affinity, and creating a high precursor frequency w9x. The multiple binary complex model proposes that specificity of the alloreactive T cell receptor is for a particular peptide derived from a normal cellular or serum protein. Differences in the allo-MHC peptide-binding groove will cause binding of a substantially different set of peptides from those of the self-MHC homologue causing allorecognition. Therefore any one MHC mismatch will be able to stimulate a large number of different T cells responsive to different antigens. At face value, both of these hypotheses concerning the recognition of allogeneic MHC break the rules of thymic positive selection for self-MHC restriction. However, the structural similarity between the TcR contact surfaces of many MHC alleles allows a substantial fraction of direct alloresponses to be accommodated within the framework of positive selection. In responder–stimulator combinations in which such structural similarities apply, allorecognition can be regarded as mimicking self-MHC restriction and directed against novel peptides that are bound by the allogeneic but not the self MHC molecules due to differences in the peptide-binding groove, i.e. the multiple binary complex hypothesis w10x. In combinations in which there are multiple amino acid differences in the TcR-contact surfaces, the alloresponse is likely to result from a chance higher affinity cross-reaction with the exposed residues of the foreign MHC structure, i.e. the high determinant density hypothesis. Given the bias that appears to exist in TcR genes for MHC recognition w7x, this could occur with sufficient frequency to account for the numbers of alloreactive T cells identified by limiting dilution analysis assays. 3. Direct allorecognition and graft rejection It has long been assumed that acute transplant rejection represents the in vivo correlate of the in vitro MLR. However, relatively little evidence has been produced in support of this contention. Early studies of the effects of depletion of donor bone marrow-derived ‘passenger’ leukocytes were taken to

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

103

Fig. 1. Pathways of allorecognition.

reinforce this assumption. Permanent survival of some rat renal allografts was achieved by ‘parking’ the kidney in an intermediate recipient, thereby depleting the allograft of donor leukocytes w11x. Subsequent experiments, injecting the recipients of the retransplanted kidneys with small numbers of donor DC led to brisk rejection w12x. These data clarified previous studies by Lafferty et al. w13x, who achieved prolonged survival of cultured thyroid allografts—presumably due to the loss of the passenger leukocytes. These results established the fundamental distinction between antigenicity (the capacity to be recognised) and immunogenicity (the capacity to initiate a destructive immune response). It was assumed that the donor ‘immunogenic’ DCs were activating the direct pathway and precipitating graft rejection. However, it is also possible that the donor DCs, due to their propensity for trafficking to lymphoid tissue, were providing a source of antigen for priming the indirect pathway w14x.

The first clear evidence that T cells with exclusively direct allospecificity can effect transplant rejection was provided by a recent study. Reconstitution of SCID or Rag1yyy mice with syngeneic CD4q T cells led to rejection of MHC class II-expressing cardiac allografts but not MHC class II-deficient grafts w15x. Furthermore, Rag1yyy mice that were also MHC class II-deficient rejected allogeneic cardiac transplants when reconstituted with CD4q T cells. Since these mice have no CD8q cells and lack the capacity for MHC class IIrestricted indirect allorecognition, these results indicate that direct pathway CD4q T cells were both necessary and sufficient to mediate allograft rejection. ¨ Anti-donor alloreactive T cells derived from the naıve fraction of the recipients’ T cell repertoire must be primed in lymphoid tissue. As a consequence, the ¨ direct pathway alloreactive T cells is priming of naıve likely to only occur during the first few weeks after transplantation, while donor-derived dendritic cells are

104

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

¨ available. Once those dendritic cells have died, the naıve T cell repertoire of the recipient is likely to be irrelevant as far as direct anti-donor responses are concerned. As far as antigen-experienced memory CD4q T cells are concerned, it is likely that resting memory cells also are activated in lymphoid tissue. Following activation, such cells are capable of crossing the vascular endothelium and entering the graft. Once again, the crucial period for activation of recipient direct pathway memory T cells is the first few weeks following transplantation in response to donor dendritic cells. The fate of activated CD4q T cells in the graft is unclear. Recognition of allogeneic MHC molecules is confined to the parenchymal cells of the transplanted tissue, a situation that has no parallel in conventional antigen-specific responses. Under normal circumstances, inflammation would lead to the recruitment of monocytes and macrophages that would be capable of in situ antigen presentation. The requirement of previously activated CD4q T cells for B7-mediated costimulation is controversial. In some in vitro systems, it appears that the requirement of such cells for costimulation is substantially less stringent. However, other data suggests that full activation does not occur unless costimulation is provided. Indeed, culture of CD4q CD4ROq T cells with HLA-mismatched, g-interferon-treated primary epithelial cell cultures from human thyroid or kidney induced allospecific hyporesponsiveness w16x. Others, and we have studied the evolution of the antidonor direct pathway alloresponse in renal and cardiac transplant recipients. In the majority of patients, donorspecific hyporesponsiveness occurs as judged by a striking fall in the frequency of anti-donor T cells w17,18x w19x. In a recent study, we compared the anti-donor CD4 T cell frequency before and four months after renal ¨ transplantation, analysing the CD45RA (naıve) and CD45RO (memory) fractions of recipient T cells independently. A significant fall in anti-donor frequency was seen in the CD45RO but not the CD45RA T cell fractions, consistent with the postulate that induced hyporesponsiveness was indeed due to encounter with the graft parenchyma w20x. Collectively, these observations suggest that the direct pathway is of diminishing importance with time after transplantation. The picture may be less clear with regard to CD8q T cells. There is considerable evidence from rodent models that CD8q T cells are less costimulation dependent and may be capable of effecting graft rejection in the absence of CD4q T cell help. In a recent study of 19 recipients of living related renal transplants, the frequency of anti-donor cytotoxic T cells was substantially diminished in 17 out of 19 patients. However, these patients were studied many years after transplantation and it is possible that the direct CD8q T cell response is a threat to the graft for some time after

donor specialised antigen presenting cells have been eliminated w19x. 4. Indirect allorecognition The original proposal that an alternative pathway of T cell allosensitisation exists arose from the same passenger cell depletion model that was described above w11x. One of the strain combinations in which retransplanted kidney grafts were accepted without exogenous immunosuppression was (AS=AUG) F1 into AS. However, if the strain combination was changed, and fully allogeneic AUG donors were used, the retransplanted grafts were invariably rejected, albeit at a slower tempo. Based on the assumption that the parenchymal tissues of the kidney were incapable of activating direct pathway anti-donor T cells (antigenic but not immunogenic), it was proposed that a second, ‘indirect’, pathway of allorecognition was responsible for T cell sensitisation. This is the way that the immune system normally sees antigen. Thus, via this pathway allogeneic MHC molecules are recognised in a truly self MHC-restricted manner. In recent years the indirect pathway of recognition has become the focus of much research. The results of peptide elution studies using human and murine systems demonstrated that peptides of MHC molecules represented a substantial fraction of the naturally processed peptides occupying cell surface MHC molecules w21x. The most striking demonstration of the propensity of MHC molecules to be presented in peptide form by other MHC molecules involved the first monoclonal antibody that was shown to be specific for a MHC– peptide complex, called Y-Ae. This antibody sees a peptide fragment from the H2-Ea chain presented on H2-Ab products. The corresponding MHC–peptide complex is expressed at high levels on DCs and B cells from strains co-expressing H2-Ab and H2-E but not from strains expressing H2-Ab or H2-E only w22–24x. The high level of staining obtained with this antibody illustrates the capacity of MHC molecules to present MHC-derived peptides. This has also been demonstrated in the DCs of the thymic medulla, an important site for the development of thymic or central tolerance by negative selection w25x. Y-Ae has also been used to monitor the processing and presentation of H2-Ea in vivo. When H2-E-bearing DCs were injected into H2Ab recipients, within 2 days most of the recipient DCs in the draining lymph node became reactive with Y-Ae. The number of donor cells in the lymph node was much smaller than the number of recipient DCs that had processed the donor H2-E w26x. This result implies that when migratory donor DCs die upon reaching the lymph node, they are phagocytosed and processed by resident recipient DCs.

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

5. Indirect allorecognition and graft rejection It was the results of Auchincloss et al. w27x, working with MHC deficient mice that established the role of the indirect pathway in transplant rejection. Their most compelling evidence that the indirect pathway is sufficient to mediate transplant rejection was the observation that MHC class I knock out recipient mice could reject skin grafts from MHC class II knock out donor mice. The recipient mice lacked CD8q cytotoxic T cells capable of recognising donor MHC class I molecules directly, and the CD4q T cells in the recipient animals could only be stimulated by recognising donor MHC class I molecules indirectly, presented in the context of recipient MHC class II molecules. These findings were supported using a murine skin allograft model, which demonstrated that the indirect pathway alone was sufficient to elicit allograft destruction in the absence of direct allorecognition w28x. Primed cells from mice previously immunised with allogeneic spleen cells or skin cells were shown to proliferate in response to peptides derived from donor-MHC in the context of self-MHC. Indeed, Fabre’s group had shown similar results several years earlier in rats. Immunisation with certain peptides corresponding to MHC class I a-helices of DA strain rats could accelerate rejection of a DA kidney graft in a LEW strain recipient w29x. In addition, CD4q T cells from non-immunised LEW recipients of DA skin or kidney grafts proliferated specifically in the presence of these peptides and recipient APCs w30x. Because not all peptides elicit this response, the implication is that the nature of the peptide presented (as a result of processing) is crucial to determine the alloresponse. ELISPOT analysis on lymph node T cells demonstrated that whilst more than 90% of the allospecific T cell repertoire was directed towards intact donor MHC, selfrestricted T cells recognising indirectly presented donorderived peptides accounted for a small but detectable population (1–5%) of the T cell repertoire as early as day 11 w31x. Returning to the original experiments with donor DCdepleted kidney grafts, in the strain combinations in which rejection occurred, the tempo was invariably slower, with a mean rejection time of 21 vs. 10 days for DC-replete grafts. This was the basis for suggesting that the indirect pathway might be most prominent in later, more chronic forms of rejection. The hypothesis that T cells with indirect anti-donor allospecificity are important drivers of chronic transplant rejection has received support from several clinical studies. First, as mentioned above, the frequency of T cells specific for the direct pathway declines with time in the majority of patients. This was equally pronounced in patients with classical features of chronic rejection as in

105

those with stable good function w19,32x. This finding suggests that the direct pathway of anti-donor alloreactivity is unlikely to be an important driver of chronic rejection. The second series of clinical observations comes from attempts to measure the indirect pathway in patients with established chronic rejection. Two approaches have been taken to reveal T cells with indirect alloreactivity. One involves offering donor MHC molecules as synthetic peptides, thereby by-passing the need for antigen processing. Several groups have reported raised frequencies of T cells with indirect anti-donor specificity in patients with chronic heart, kidney, and lung transplant rejection w33–37x. The other approach that has been used involves offering donor antigens in the form of lysed whole cells, thereby making no assumption as to which peptides will be recognised. Using this approach we have detected raised frequencies of indirect pathway T cells in patients with, but not without, chronic heart transplant rejection w38x. Despite this, all groups have found lower frequencies of indirectly alloreactive T cells than might be expected if this pathway was the main player in chronic rejection; there are several possible explanations: First, there could be a potent APC efficiently presenting to few cells, the B cell being a good candidate w39x; second, regulatory cells could reduce activity of alloreactive cells; and third, the assays might be insufficiently sensitive; finally, other risk factors for chronic rejection might allow a small number of T cells to suffice w40x. Improved assays for indirect pathway T cells may address some of these issues. If the indirect pathway is critical in mediating transplant rejection, strategies that promote tolerance in the indirect pathway should enhance allograft survival. Compelling evidence that induction of tolerance in the indirect pathway favours graft survival came from experiments in which recipient animals were pre-treated with donor MHC peptides. The peptides were administered either intrathymically w41x, orally w42x, or as donor peptide-pulsed recipient APCs w43x. All of these treatments induced graft prolongation. Although the mechanisms of these strategies have not been defined, the use of donor peptides (for presentation by host APCs) means that the enhancement of graft survival can only be mediated through the indirect pathway. 6. Interactions between the direct and indirect pathway As mentioned above, allorecognition per se is not sufficient to invoke an alloresponse: there must, in addition, be an appropriate environment of cell surface molecules and cytokines. It is becoming increasingly clear that the picture of an APC displaying a MHC– peptide complex and costimulatory molecules to a spe-

106

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

Fig. 2. Co-stimulation in trans can be seen as part direct, and part indirect pathway depending on the nature of the APC presenting peptide– MHC.

cific T cell is a huge oversimplification. A recent study has shown that costimulatory support can be provided by a different APC than that which displays the antigen w44x. Sayegh’s group used B7-1yB7-2 knockout mice as cardiac donors and MHC II knockout recipients. In this model the donor APC displays MHC–peptide ‘directly’ to the host T cell but cannot provide costimulation because of the knockout. Conversely, the host APC cannot ‘indirectly’ display MHC–peptide but can provide costimulation. The cardiac grafts were rejected as quickly as those grafted into wild-type recipients suggesting that costimulation ‘in trans’ can be as potent as the conventional costimulation ‘in cis’ (Fig. 2). As mentioned above, based on in vitro models, a reasonable hypothesis is that donor parenchymal cells, by presenting antigen without costimulation, lead to graft acceptance. This is the main proposed mechanism of the reduction, with time, in the frequency of T cells specific for the direct pathway. If costimulation in trans is important in vivo, even when cis is functional, then the high antigenic load provided by the donor endothelium could become very immunogenic indeed: there is currently no evidence for this. Evidence for cross-talk between the direct and indirect pathway is the observation that CD4q T cells with indirect anti-direct anti-donor specificity can amplify direct pathway CD8q T cell responses (Auchincloss, per com). Similarly, tolerance in the indirect pathway appears to lead to regulation of direct pathway responses

in some models. Unless transfer of MHC–peptide complexes between donor and recipient APC occurs, these interactions break the rules of linked help and linked suppression in that they require communication between two T cells interacting with separate (donor for direct and recipient for indirect) APC. Indeed, evidence is emerging that MHC–peptide complexes can be acquired by both B cells and T cells which may both increase efficiency of presentation and possibly suggest novel pathways of antigen presentation (unpublished observations). 7. Allorecognition and tolerance Immunological tolerance, by which we mean a donorspecific absence of immune attack, can occur by a number of mechanisms, central and peripheral, deletional and non-deletional w45x. The intricacies are beyond the remit of this review, but some concepts will be discussed briefly as it is the precise context of allorecognition that determines whether an alloresponse, nonresponse or peripheral tolerance occurs. The main mechanisms of the latter are recognition with suboptimal stimulation, negative feedback and regulation. Suboptimal stimulation may result from recognition without costimulation as discussed above, but there are other mechanisms. Some peptides can act as antagonists (prevent binding) or partial agonists (cause inhibition) of peptide agonists causing T cell activation. These are

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

termed altered peptide ligands and are usually very similar in structure to the agonist. The in vivo relevance is uncertain but with further study, they could prove a useful tool. Suboptimal stimulation may also occur with insufficient cytokines such as IL2 although over expression of IL2 in an expanding immune response may lead to activation induced cell death as a negative feedback loop w46x. There is a large body of literature accumulating on the importance of CD4qCD25q T cells in immune regulation w47x. In mouse models, elimination of these cells leads to spontaneous development of autoimmune disease and in vitro studies of CD4qCD25q cells from humans confirm their suppressive properties. Their role, if any, in policing the pathways of allorecognition is not yet defined w48x. One of the problems inherent to clinical tolerance is in its measurement. If we are measuring the absence of a response, how do we know that this absence will be maintained if we, say, withdraw immunosuppressive therapy? Auchincloss w45x suggests that we should only trust the finding that something is present and therefore concludes that clinical tolerance induction will involve macrochimerism using donor stem cells ‘educated’ by residual host thymus: this would mean detecting the presence of donor cells. This is certainly beyond us at this time and is not helpful in guiding our management of patients who might be tolerant to their current graft on conventional immunosuppression. We may have to settle for a reliable validated assay of indirect pathway hyporesponsiveness. Alternatively, the answer may lie in the detection of regulatory cells that police T cells specific for the indirect pathway. The development of a robust tolerance assay is currently under intense investigation. 8. Summary The pathways of allorecognition have been discussed together with the implications for transplantation tolerance. First, the direct pathway anti-donor response may be predominant in triggering acute rejection but diminishes with time regardless of whether or not chronic rejection occurs. Second, graft acceptance is likely to require tolerance in T cells with indirect pathway allospecificity, which can mediate acute and chronic rejection. Third, there is emerging evidence of cross-talk between the direct and indirect pathways. Fourth, whether allorecognition results in activation or tolerance depends upon the precise context of peptides, cell surface molecules and cytokines at any given time: this may be influenced by regulatory cells. Finally, a major challenge to clinical tolerance induction is in its measurement.

107

References w1x Gorer PA. The genetic and antigenic basis of tumor transplantation. J Pathol Bacteriol 1937;44:691 –697. w2x Sherman LA, Chattopadhyay S. The molecular basis of allorecognition. Annu Rev Immunol 1993;11:385 –402. w3x Hernandez-Fuentes MP, Baker RJ, Lechler RI. The alloresponse. Rev Immunogenet 1999;1:282 –296. w4x Gould D, Auchincloss H. Direct and Indirect recognition: the role of MHC antigens in graft rejection. Immunol Today 1999;20(2):77 –82. w5x Daniel C, Horvath S, Allen P. A basis for alloreactivity: MHC helical residues broaden peptide recognition by the TCR. Immunity 1998;8:543 –552. w6x Lombardi G, et al. Are primary alloresponses truly primary? Int Immunol 1990;2(1):9 –13. w7x Merkenschlager M, Ikeda H, Wilkinson D. Allorecognition of HLA-DR and -DQ transfectants by human CD45RA and CD45RO CD4 T cells: repertoire analysis and activation requirements. Eur J Immunol 1991;21:79 –88. w8x Warrens A, Lombardi G, Lechler R. MHC and alloreactivity: presentation of major and minor histocompatibility antigens. Transplant Immunol 1994;2:103 –107. w9x Lechler RI, et al. The molecular basis of alloreactivity. Immunol Today 1990;11(3):83 –88. w10x Lombardi G, et al. Allorecognition of DR1 by T cells from a DR4yDRw13 responder mimics self-restricted recognition of endogenous peptides wpublished erratum appears in Proc Natl Acad Sci USA 1989 Dec;86(24):10074x. Proc Natl Acad Sci USA 1989;86(11):4190 –4194. w11x Lechler RI, et al. Immunogenicity of retransplanted rat kidney allografts. Effect of inducing chimerism in the first recipient and quantitative studies on immunosuppression of the second recipient. J Exp Med 1982;156(6):1835 –1841. w12x Lechler RI, Batchelor JR. Restoration of immunogenicity to passenger cell-depleted kidney allografts by the addition of donor strain dendritic cells. J Exp Med 1982;155(1):31 –41. w13x Lafferty KJ, et al. Thyroid allograft immunogenicity is reduced after a period in organ culture. Science 1975;188(4185):259 – 261. w14x Austyn J, Larsen C. Migration patterns of dendritic leucocytes: implications for transplantation. Transplantation 1990;49:1 –7. w15x Pietra BA, et al. CD4 T cell-mediated cardiac allograft rejection requires donor but not host MHC class II. J Clin Invest 2000;106(8):1003 –1010. w16x Frasca L, et al. Role of donor and recipient antigen presenting cells in priming and maintaining T cells with indirect allospecificity. Transplantation 1998;66(9):1238 –1243. w17x Hornick PI, et al. Assessment of the contribution that direct allorecognition makes to the progression of chronic cardiac transplant rejection in humans. Circulation 1998;97(13):1257 – 1263. w18x Mason PD, Robinson CM, Lechler RI. Detection of donorspecific hyporesponsiveness following late failure of human renal allografts. Kidney Int 1996;50(3):1019 –1025. w19x Baker RJ et al. Assessment of the immunological contribution to chronic allograft nephropathy. J Immunol 2001:in press. w20x Baker RJ, et al. The role of the allograft in the induction of donor-specific T cell hyporesponsiveness. Transplantation 2001;72(3):480 –485. w21x Shoskes DA, Wood KJ. Indirect presentation of MHC antigens in transplantation. Immunol Today 1994;15(1):32 –38. w22x Rudensky AY, et al. On the complexity of self. Nature 1991;353:660 –662.

108

D.S. Game, R.I. Lechler / Transplant Immunology 10 (2002) 101–108

w23x Inaba K, et al. High levels of a MHC II–self peptide complex on dendritic cells from the T cell areas of lymph nodes. J Exp Med 1997;186:665 –672. w24x Murphy DB, Lo D, Rath S. A novel MHC class II epitope expressed in thymic medulla but not cortex. Nature 1989;338:765 –768. w25x Kishimoto H, Sprent J. Negative selection in the thymus includes semimature T cells. J Exp Med 1997;185:263 –271. w26x Inaba K, Turley S, Yamaide F. Efficient presentation of phagocytosed cellular fragments on the MHC complex class II products of dendritic cells. J Exp Med 1998;188:2163 –2173. w27x Auchincloss H, et al. The role of ‘indirect’ recognition in initiating rejection of skin grafts from major histocompatibility complex class II-deficient mice. Proc Natl Acad Sci USA 1993;90(8):3373 –3377. w28x Fluck N, et al. Indirect allorecognition is involved in both acute and chronic allograft rejection. Transplant Proc 1999;31:842 –843. w29x Fangmann J, Dalchau R, Fabre JW. Rejection of skin allografts by indirect allorecognition of donor class I major histocompatibility complex peptides. J Exp Med 1992;175:1521 –1529. w30x Fangmann J, et al. T cell recognition of major histocompatibility complex class I peptides during allograft rejection. Eur J Immunol 1992;22:1525 –1530. w31x Benichou G. Direct and indirect recognition: the pathways to allograft immune rejection. Front Biosci 1999;4:D476 –D480. w32x Hornick P, et al. Significant frequencies of T cells with indirect specificity in heart graft recipients with chronic rejection. Circulation 2000;101:2405 –2416. w33x Hornick P et al. Significant frequencies of T cells with indirect anti-donor specificity in patients with chronic cardiac allograft rejection. Circulation 1999:in press. w34x Ciubotariu R, et al. Persistent allopeptide reactivity and epitope spreading in chronic rejection of organ allografts. J Clin Invest 1998;101(2):398 –405. w35x Vella JP, et al. Indirect allorecognition of major histocompatibility complex allopeptides in human renal transplant recipients with chronic graft dysfunction. Transplantation 1997;64(6):795 –800.

w36x Liu Z, et al. Indirect recognition of donor HLA-DR peptides in organ allograft rejection. J Clin Invest 1996;98(5):1150 – 1157. w37x Suciu-Foca N, et al. Indirect allorecognition of donor HLADR peptides in chronic rejection of heart allografts. Transplant Proc 1998;30:3999 –4000. w38x Womer KL. T cell recognition of allograft target antigens. Curr Opin Organ Transplant 2000;5:23 –28. w39x Lanzavecchia A. Antigen specific interaction between T and B cells. Nature 1985;314:537 –539. w40x Jain S, Furness PN, Nicholson ML. The role of transforming growth factor beta in chronic renal allograft nephropathy. Transplantation 2000;69:1759 –1766. w41x Sayegh MH, et al. Mechanisms of acquired thymic unresponsiveness to renal allografts. Thymic recognition of immunodominant allo-MHC peptides induces peripheral T cell energy. Transplantation 1994;58(2):125 –132. w42x Zavazava N, et al. Oral feeding of immunodominant MHC donor-derived synthetic class I peptide prolongs graft survival of heterotopic cardiac allografts in a high responder rat strain combination. J Leukoctye Biol 2000;67:793 –800. w43x Oluwole SF, et al. Mechanism of acquired thymic tolerance induced by a single MHC class I peptide with the dominant epitope: differential analysis of regulatory cytokines in the lymphoid and intragraft compartments. Transplantation 1999;68:418 –429. w44x Mandelbrot DA, et al. Rejection of mouse cardiac allografts by costimulation in trans. J Immunol 2001;167:1174 –1178. w45x Auchincloss H. In search of the elusive holy grail: the mechanisms and prospects for achieving clinical transplantation tolerance. Am J Transplant 2001;1:6 –12. w46x Lechler R et al. T cell energy and peripheral T cell tolerance. In press, 2000. w47x Sakaguchi S. Policing the regulators. Nat Immunol 2001;2(4):283 –284. w48x Ng WF et al. Human CD4qCD25q T cells: a naturally occuring population of regulatory T cells. Blood 2001;98(9):in press.