Chimerism in organ transplantation: conflicting experiments and clinical observations

The concept of chimerism has influenced our thinking about tolerance and rejection of organs and tissues since the beginning of modern transplantation. In macrochimerism, persisting donor-specific cells are easily detectable by flow cytometry at levels of several to 100%, usually after transient lymphoablation and bone marrow (or other cell) transplantation. Microchimerism refers to a state in which donor cells persist at low levels (1 cell per 104 or 105 or less), frequently detectable by molecular techniques and usually consisting of class II dendritic cells. Although macrochimerism is frequently associated with donor-specific tolerance in many experimental animals and people, instances occur in which macrochimerism can be produced but tolerance is not achieved. Also, in large animal models, macrochimerism and associated tolerance can be produced but macrochimerism can then disappear, and yet tolerance persists. Clinically, states of microchimerism can exist, but rejection still occurs. Also, persisting microchimerism does not necessarily correlate with clinical tolerance or the ability to wean from or reduce immunosuppressive drugs. Recent experiments in several rodent models using bone marrow to induce macrochimerism and tolerance have shown that establishment of the macrochimeric state does not necessarily produce tolerance. The presence of class II positive cells in the donor bone marrow inoculum is essential for tolerance induction in these models. The concept of chimerism (i.e., the presence of replicating living cells of two distinct genetic backgrounds in one organism) has influenced our thinking about tolerance and rejection of organs and tissues since the beginning of modern transplantation. The idea of possible immunologic consequences of chimerism in relation to tissue transplantation essentially started with Owen’s (1) observation that cattle twins, which commonly share a single placenta, are erythrocyte chimeras at birth, with this chimeric state usually persisting into adulthood. In addition, most dizygotic cattle twins were later found to accept skin grafts from each other, regardless of differences in sex and color, and this mutual tolerance was specific, because third-party grafts were quickly rejected (2). Billingham and colleagues (3) recognized the significance of the chimeric state and reproduced this experiment of nature in the laboratory by injecting replicating lymphoid cells of one donor strain into allogeneic naturally incompetent neonatal recipients that could not reject the donor cell inoculum. Injected neonates were chimeras as adults and were specifically tolerant of donor strain skin grafts applied in adulthood. Furthermore, when chimerism was abrogated by experimental means, tolerance was lost. Thus, the association of chimerism and tolerance became a fundamental concept in modern transplantation immunobiology. Billingham and colleagues (3) never thought the idea of inducing chimerism had any practical significance or value. Rather, they thought their experiment had moral or inspirational value to transplant surgeons because it clearly demonstrated that the vigorous rejection reaction could be circumvented. Nevertheless, the modern observation that patients treated for malignancy with myeloablative therapy and allogeneic bone marrow (BM) who became chimeras accepted organ allografts from their BM donors without immunosuppression (4) definitely suggested that clinically produced chimerism could possibly induce clinical tolerance. A second clinical observation that long-term hepatic and renal allograft recipients on minimal or no immunosuppressive therapy frequently have persisting donor cells in peripheral tissues (so-called microchimerism) also suggested that donor cell chimerism might be related to donor-specific tolerance or hyporesponsiveness (5). MACROCHIMERISM AND MICROCHIMERISM Concepts of chimerism generally focus on two types of chimerism. Macrochimerism usually is produced after BM (or a related cell type) transplantation; to achieve macrochimerism invariably requires some type of lymphoablative preconditioning. Hematologic stem-cell (HSC) engraftment occurs, giving rise usually to multilineage chimerism. Donor-specific cells are easily detectable by flow cytometry at levels from 2% or 3% to 100%. Theoretically, macrochimeric states can be produced by syngeneic, allogeneic, or xenogeneic HSC transplants. In contrast, microchimerism refers to the state in which donor-specific cell levels are low (1×104 or 105 cells or less), frequently detectable only by molecular techniques and usually consisting of class II dendritic cells. The potential role or significance of microchimerism has been debated. Microchimerism could reflect just effective immunosuppression; also, such cells could mediate tolerance by functioning as immature antigen presenting cells (in the absence of co-stimulation) or they could act as surrogate targets of rejection, protecting the graft. OBSERVATIONS OF CHIMERISM IN LARGE-ANIMAL MODELS Although there is substantial evidence that induced macrochimerism correlates with tolerance in rodents (see below), such is not always the case in larger species. Gottlieb et al. (6) established canine allogeneic chimeras (20%–50% donor cell chimerism) using total lymphoid irradiation and BM infusion without development of graft-versus-host disease. However, Koretz (7) from the same group showed that after chimerism was established with a similar protocol, four of four dogs that were macrochimeras (>25%) rejected donor-specific heart transplants. Also, macrochimerism to achieve tolerance to renal allografts has been induced in nonhuman primates after nonmyeloablative conditioning and simultaneous donor BM and kidney transplantation; in these studies (8–10), macrochimerism became undetectable after several months, although tolerance persisted. Furthermore, Spitzer et al. (11) reported induction of tolerance to renal allografts after lymphoablative conditioning and simultaneous donor-specific BM and kidney transplants in multiple myeloma patients with renal disease. As in the nonhuman primate studies, these recipients remained tolerant but chimerism did not persist. CLINICAL OBSERVATIONS WITH MICROCHIMERISM Clinical analyses have not substantiated an association between microchimerism and tolerance or donor-specific hyporesponsiveness. Indeed, allograft rejection can occur in the presence of microchimerism, long-term excellent graft survival can occur in the absence of microchimerism and, most importantly, microchimerism cannot be used as a criterion for immunosuppressive withdrawal. In one study (12), only one third of long-term renal allograft recipients (20–30 years) and only 14% of short-term successful patients (>2 years) were microchimeric. Furthermore, microchimerism (1 in 104 cells were detectable) did not correlate with human leukocyte antigen matching, rejection, or various clinical and immunologic parameters. Using more sensitive methodology (1 in 105 cells detectable), Hisanaga et al. (13) showed that 75% of heart and 72% of liver allograft recipients were microchimeras at some time in the posttransplant period; they noted that the microchimerism was dynamic and variable. Microchimerism did not correlate with rejection and showed no diagnostic or predictive value for the individual patient. Devlin et al. (14) subjected 18 microchimeric liver allograft recipients with stable function (>5 years) to a staged program of immunosuppressive drug withdrawal. When the frequencies of achieving complete or partial drug withdrawal were compared, there was no statistical difference between the number of patients who were chimeric versus the number who were nonchimeric. Schlitt et al. (15) reported the extraordinary case of the occurrence of extensive donor-type microchimerism associated with graft rejection 8 years after liver transplantation. The only significant data for an association of microchimerism and donor-specific hyporesponsiveness are reported by Ciancio et al. (16), who treated 63 cadaveric renal allograft recipients with one or two donor-specific BM infusions and compared their course with 219 noninfused controls given identical immunosuppression. Actuarial graft survival at 5 years was superior in the BM-infused patients, especially when graft survival was censored for recipient death. Acute rejections were significantly reduced in patients given two BM infusions, and chronic rejection was dramatically reduced in all BM-treated recipients. The most interesting finding was that through 4 years, the degree of microchimerism increased in the group of patients given BM that had exhibited no rejection episodes. Nevertheless, with the exception of this report in which additional donor BM was given with organ grafting, a significant association of microchimerism and donor-specific hyporesponsiveness after solid organ grafting has not been substantially identified. VARIABLE EFFECTS OF CHIMERISM ON TOLERANCE IN RODENT MODELS Chimerism and Tolerance in the Non–Radiation-Based Model of Tolerance Using Antilymphocyte, Rapamycin, and Donor-Specific Bone Marrow. Monaco and Wood (17, 18) first demonstrated that donor-specific BM infusion into mice transiently lymphoablated with polyclonal antilymphocyte serum induced donor-specific unresponsiveness to skin allografts in class I disparate combinations. Although the degree and duration of tolerance could be augmented with a finite course of cyclosporine A, the tolerance achieved continued to be limited to class I differences. It was suggested by the authors that BM contained a unique, tolerogenic stem cell; they also proposed the possibility that some tissues and organs, such as the liver, might contain such stem cells that in turn could facilitate tolerance induction. Subsequently, Hale et al. (19,20) showed that addition of rapamycin to antilymphocyte serum (ALS) treatment followed by low-dose (25×106 BM cells per mouse) BM infusion dramatically augmented tolerance so that it was possible to easily induce stable, long-lasting tolerance in class I and class II disparate combinations but not in complete mismatches. This tolerance was associated with persisting low levels of chimerism (3%–5%). Other immunologic mechanisms associated with tolerance such as nonspecific suppressor cells and clonal reduction or deletion (limiting dilution assay) were identified at various times during the tolerant state. Subsequently, Hale et al. (21) used progressive increases in the BM dose (from 25×106 to 200×106 cells per mouse) to produce stable, robust, indefinite tolerance to skin allografts across complete major histocompatibility complex (MHC) mismatches. The degree and duration of chimerism and the induction and persistence of tolerance were dependent on BM dose infused, with the higher doses producing indefinite tolerance and chimerism. Thus, the dose of BM was a critical factor influencing the degree of tolerance and chimerism achieved in this non–radiation-based tolerance model. The tolerance achieved with high doses of BM (200×106 BM cells per mouse) was robust and indefinite. There was elimination of donor-specific cell-mediated lymphocytotoxicity (CML) at 200 days, whereas third-party CML reactions were completely preserved. Second donor-specific skin grafts survived indefinitely, whereas third-party grafts were rejected. The tolerance could not be broken by exogenous interleukin-2, suggesting the absence of anergy as a significant underlying mechanism. In recipients given high-dose BM, the chimerism was maintained at a level of 5% to 10% and was also multilineage (lymphocytes, macrophages, monocytes, and granulocytes). The lymphocyte phenotype was that of B cells, that is, cells were MHC class II-positive, Fc receptor-positive, and CD45/B220-positive, but negative for CD4, CD8, and Thy 1.2, consistent with a B-cell lineage. No T-cell phenotype was identified. Indeed, the donor cell chimerism over time was consistently maintained as B lymphocytes and macrophage-monocyte populations. Flow cytometric analysis after 100 days of tolerance showed that 10% to 15% of B cells were of donor origin, whereas nearly two thirds of the macrophage-monocyte population was of donor origin. Thus, in this model of tolerance, which does not use radiation-based immunoablation, the chimerism produced with bone marrow involves essentially only B cells and macrophage-monocytes. Although these data do not clarify whether chimerism causes tolerance or whether an induced state of tolerance permits chimerism, they do show that the level of chimerism obtained was directly related to the probability of tolerance being induced in this model. It had been shown that tolerance in closely related donor and recipient combinations (20) could be achieved with lower doses of BM and sustained with lower levels of chimerism than tolerance in a completely mismatched combination (21). One possible implication of this finding is that the chimeric cells themselves are contributing to the state of tolerance (i.e., through clonal deletion or suppression). If this is the case, it could be postulated that more chimeric cells would be necessary to overcome the more vigorous rejection reaction directed against a more disparate target. In this regard, it is notable that in these studies with the completely mismatched combination, these mice that achieved a 10% level of donor hematopoietic chimerism invariably demonstrated robust tolerance, whereas only one half of the animals with approximately 5% chimerism showed long-lasting tolerance. The absence of T cells in donor chimerism associated with tolerance induction after ALS-rapamycin was surprising because it had been suggested that a correlation exists between high levels of initial T-cell reconstitution and stable chimerism; such T-cell engraftment was considered a prerequisite for induction of allograft tolerance in the radiation-based mixed allogeneic chimera model (22). This prompted a detailed study of the role of T and B cells in tolerance induction in our non–radiation-based ALS-rapamycin model. Umemura et al. (23) used various gene knockout mice to examine the role of donor T cells in tolerance induced to skin allografts in ALS-treated, rapamycin-treated, donor BM-infused recipients in fully MHC-incompatible (complete mismatch) mouse strain combinations. BM from CD4KO, CD8KO, and CD4/CD8 double KO mice was equally as effective as their wild-type BM counterparts in inducing tolerance in this model. Also, BM from CD3+ T-cell–deficient mice, CD3c-tg mice, all of which lack T cells and natural killer cells, was also as effective as the appropriate wild-type BM. Furthermore, B-cell–deficient BM from BM KO donors also was as tolerogenic as the wild-type BM. Phenotypic analysis of chimeric cells showed the absence of all donor T cells in tolerant recipients, as previously shown by Hale et al. (21) with the presence only of donor B cells and macrophage-monocytes in all recipients except those rendered tolerant with B-cell KO BM. Recipients rendered tolerant with B and KO BM exhibited only donor macrophage-monocyte chimerism. These studies are important in defining the nature of cells in BM responsible for tolerance induction in this model. The results obtained would seem to definitely rule out the participation of T cells or B cells, or both, in tolerance induction. Because both T and B cells contained in BM were not shown to be essential for tolerance induction in the ALS-rapamycin-BM tolerance model, Umemura and colleagues (24) investigated the role of other donor MHC class II antigen-expressing cells in BM for tolerance induction using appropriate gene KO mice as BM donors. In ALS-rapamycin–treated recipients, BM from class I-deficient (class I-knockout) donors was equal or even better in tolerogenicity compared with the wild-type BM. In contrast, class II-deficient BM (class II KO donors) failed to induce tolerance. Furthermore, depletion of I-Ab–positive cells from normal wild-type BM dramatically reduced the tolerogenic effect of normal BM. This was in spite of the fact that CDII-deficient BM restored all lymphoid populations of lethally irradiated wild-type recipients with the exception of class II antigen expression. Interestingly, CDII-deficient BM-infused mice showed the same low level of chimerism at day 60 as wild-type BM recipients in spite of reduced tolerance. Additional studies in ALS-rapamycin–treated recipients have shown that BM from FasL knockout and C40 knockout donors was not tolerogenic. Taken together, these studies showed that expression of MHC class II antigen by donor BM cells is absolutely required for induction of allograft tolerance in the ALS-rapamycin-BM tolerance model, whereas FasL or CD40 expression was not essential. Chimerism and Tolerance in the Radiation-Based Model Using Donor Specific Bone Marrow Maki and colleagues (24) also studied the role of class II BM cells on production of allogeneic chimerism and tolerance in a radiation-based model. Sublethally irradiated (750 rad) B10.A recipients received T-cell–depleted wild-type B6- or CDII-deficient BM (complete mismatch). Flow cytometry confirmed a high degree of chimerism with H-2b–positive donor cells in both groups, although CDII recipients had negligible I-Ab–positive cells. Allogeneic chimeras reconstituted with wild-type B6 BM accepted B6- or CDII-deficient skin grafts. Allogeneic chimeras reconstituted with CDII-deficient BM rejected both B6 and CDII-deficient skin grafts in spite of a high degree of chimerism that persisted after graft rejection. These studies showed that in the radiation-based model of chimerism and tolerance, peripheral chimerism with cells deficient in class II antigen does not guarantee allograft acceptance. Additional studies by Maki and colleagues (25) also investigated the role of T and B cells in tolerance induction in the radiation-based model of chimerism and tolerance using BM from appropriate knockout donor mice. Irradiated (750 rad) recipients of wild-type B6, CD4-KO, CD8-KO, CD4/8 double KO, CD3e-tg, and B-cell KO BM cells had substantial degrees of chimerism when tested at 30 and 90 days (50%–90% donor cell range). Donor cells regenerated high percentages of all lymphoid cell classes (T and B cells, monocytes, dendritic cells) with the exception that no donor T cells were identified in recipients of CD4-CD8 double KO BM or CD3e-tg BM. Allogeneic radiation chimeras that contained donor-type T cells accepted wild-type skin grafts; graft tolerance correlated with the presence of chimerism. Allogeneic radiation chimeras that did not contain donor-type T cells (irradiated recipients of CD4-CD8 double KO and CD3e-tg BM) rejected donor-type skin grafts in spite of a high degree of chimerism. Umemura and colleagues (25) used an elegant in vivo mixed leukocyte reaction assay to show that lymphocytes from mice reconstituted with T-cell–deficient BM proliferated when injected into irradiated donor strain mice, indicating that they were not specifically tolerant, whereas lymphocytes from mice reconstituted with wild-type BM were unresponsive to donor alloantigens (i.e., did not proliferate). Donor-specific allograft tolerance was restored when donor-type T cells were adoptively transferred to recipient mice given T-cell–deficient BM. Presumably, these donor T cells contained immunoregulatory T cells that were capable of inhibiting the host T-cell response directed against their own (donor) alloantigens. These results show that in this radiation-based model of chimerism and tolerance, donor T-cell engraftment is required for induction of allograft tolerance but not for creation of continuous hematopoietic chimerism after allogeneic BM transplantation and that a high degree of chimerism is not necessarily associated with specific allograft tolerance.