Secondary Logo

Share this article on:

AMPLIFICATION OF NATURAL REGULATORY IMMUNE MECHANISMS FOR TRANSPLANTATION TOLERANCE1

Chen, Zhonghua K.2; Cobbold, Stephen P.3; Waldmann, Herman3; Metcalfe, Su2,4

Rapid Communication

There is a need to derive donor-specific tolerance in clinical organ transplantation, where potential benefits remain overshadowed by chronic rejection and side effects of continual immunosuppressive therapy. It is known that the mature immune system in mice can be reprogrammed to accept a foreign graft as if it were “self.” Here we show that, once generated, this state of operational tolerance becomes self-sustaining, imposing itself on new cohorts of lymphocytes as they arise. These new cohorts retain specificity for the tolerizing antigen and can be selectively amplified to tolerate new antigens that have linked expression with the original tolerogen. Regulation is critically dependent upon the continuous presence of tolerizing antigen and is mediated by the CD4 + lymphocyte population. We propose that such natural mechanisms of immune regulation may eventually be exploited for transplantation tolerance, even in fully immune-competent recipients.

Department of Surgery, Addenbrooke's Hospital, University of Cambridge, Cambridge CB2 2QQ, and Dunn School of Pathology, University of Oxford, Oxford OX1 3RE, United Kingdom

1We are grateful to the British Heart Foundation for funding this work (BHF grant PG/93102 to S.M. and H.W.). S.M. and S.P.C. are funded by the Medical Research Council.

2Department of Surgery, Addenbrooke's Hospital, University of Cambridge.

3Dunn School of Pathology, University of Oxford.

4Address correspondence to: Su Metcalfe, Department of Surgery, Addenbrooke's Hospital, University of Cambridge, Cambridge CB2 2QQ, UK.

Received 4 June 1996.

Accepted 16 August 1996.

Immunoregulation within the peripheral immune system has been reported to control potentially autoreactive lymphocytes to maintain self-tolerance(1, 2). In rodents, new populations of suppressive(regulatory) cells can be generated for a foreign graft if first exposure of the T cell receptor to graft antigen, derived from an organ or skin graft, or from a donor-type blood transfusion, coincides with blockade of the CD4 and CD8 co-receptors of the T cell receptor(3-6). Moreover, when tolerant recipients are infused with naive cells (7, 8), or when T-cell-depleted recipients are infused with donor-specific regulatory cellsplus naive cells (9), then the naive donor-reactive cells can also become regulatory and suppressive(“infectious tolerance”). The conditioning of the recipients in the above infusion experiments was designed to expose any regulatory properties of the transferred cells. Having found that such regulation occurs, it became important to determine whether similar infectious tolerance could be demonstrated in fully immune-competent animals.

In preliminary experiments, we have demonstrated that“suppressive” or “regulatory” cells can be generated to MHC and minor antigen mismatched heart allografts (BALB/c to CBA) if these grafts are introduced under cover of nonlytic CD4 and CD8 antibodies. We also showed that these “regulatory” cells are able to transfer tolerance to naive, fully immune-competent recipients(10). We now extend this work to examine mechanisms underlying the maintenance of tolerance in unmodified recipients with particular emphasis on amplification, antigen specificity, spread to bystander antigens, and “suppressor” memory.

Back to Top | Article Outline

MATERIALS AND METHODS

Mouse heart allografts. Donor-recipient combinations were mismatched for both minor and major histocompatibility antigens. Untreated CBA/Ca (H2k) recipients of a BALB/c (H2d) heart rejected their graft at 7 days, whereas untreated CBA/Ca recipients of C57/Bl10 (H2b) heart rejected the graft at between 9 and 20 days, with a median rejection time of 10 days. The techniques for abdominal and cervical heart transplantation have been described previously (11, 12). Graft function was monitored by daily palpation of the heart graft.

Monoclonal antibodies. Primary recipients were treated with anti-CD4 and anti-CD8 monoclonal antibodies (mAbs*) YTS 177.9.6 and YTS 105.18.10, respectively: both of these mAbs are of the nondepleting rat IgG2a isotype (13). Two milligrams of monoclonal antibodies were administered every other day from day of transplant up to day 20 after transplant. To deplete CD4+ cells from tolerant recipients, three intraperitoneal doses of mAb YTS 191.1.2 (1 mg) plus YTA 3.1.2 (1 mg) were given on alternate days. Both YTS 191.1.2 and YTA 3.1.2 are depleting rat IgG2b anti-CD4 mAbs (13, 14). The same procedure was used to deplete CD8+ cells from tolerant recipients, except that the depleting mAbs were directed against CD8, namely YTS 169 and YTA156, both rat IgG2b isotypes (15). For depletion of Thy1.2+ cells, two intravenous doses of 1 mg each of 30-H12 mAb were used (30-H12 is a depleting rat mAb of IgG2b isotype) (16).

Interleukin (IL) 2. Supernatant from X6310, the IL-2-secreting murine cell line kindly donated by Dr. Fritz Melchers, was used as a source of IL-2. Levels of IL-2 were quantitated by comparison with commercial mIL-2(Genzyme) in a bioassay using the IL-2-dependent HT-2 cells, as described previously (17).

Flow cytometry. Numbers of CD4+ and CD8+ spleen cells adoptively transferred into naive recipients were calculated from aliquots labeled for CD4+ cells using biotin-conjugated YTA3.1.2, or biotin-conjugated YTS 156.7.7 for CD8+ cells. These were then stained with streptavidin-phycoerythrin (Serotech Star 4B). The total number of T cells was stained by labeling a cell aliquot with KT3, a rat IgG2a anti-CD3 mAb (18), followed by fluorescein isothiocyanate-labeled mouse anti-rat IgG2a (Serotech). Two-color flow cytometric analysis allowed discrimination between the different cell populations. In some experiments, Thy1.2+ cells were depleted in vivo: the levels of depletion were determined by staining for Thy1.2+ and Thy1.1+ cells, with their respective mAbs (16, 19) labeled with streptavidin-phycoerythrin, with CD3+ cells stained as described above.

Back to Top | Article Outline

RESULTS

Regulatory suppressive cells are CD4 +. We have previously shown that 3×107, but not 1×107, (BALB/c tolerant)CBA spleen cells were sufficient to establish operationally dominant tolerance in naive, fully immune-competent CBA recipients of a BALB/c heart graft (10). Here we have determined which subpopulation of cells within the spleen is required for the transfer of tolerance, using selective depletion in vivo. When the CD4+ cells were depleted from (BALB/c tolerant)CBA mice before spleen cell transfer, then the naive recipient rejected the donor-type graft at normal tempo, despite receiving on average 2×106 CD8+ (BALB/c tolerant)CBA spleen cells. In contrast, tolerance was successfully transferred by (BALB/c tolerant, CD8-depleted)CBA spleen cells in 8 out of 10 naive CBA recipients of a BALB/c heart. It became clear that a minimum number of CD4+ (BALB/c tolerant)CBA spleen cells was critical for transfer of tolerance, since the two naive recipients that rejected their BALB/c heart had received less than 3×105 adoptively transferred CD4+ cells: the other eight recipients that became BALB/c tolerant had received at least 3.5×105 adoptively transferred CD4+ (BALB/c tolerant) CBA spleen cells. This implies that the suppressive cell population generated under cover of the nondepleting mAbs to CD4 and CD8 has a CD4+ phenotype and that 3.5×105 of these cells is sufficient to adoptively transfer tolerance to a fully immune-competent recipient. Figure 1 summarizes these results.

Exogenous IL-2 does not prevent the tolerant cell population being dominant in naive recipients. Specific nonresponsiveness to antigen may result from anergy due to insufficient levels of IL-2(20). If so, provision of exogenous IL-2 might reverse tolerance, as has been found for tolerance induced by preoperative blood transfusion (21, 22). We tested the possibility that IL-2 insufficiency might underly tolerance generated by regulatory cells adoptively transferred to naive recipients. Naive CBA recipients of a BALB/c heart plus 5×107 (BALB/c tolerant)CBA spleen cells were treated with 1600 U of mIL-2 daily for 6 days from the time of grafting. The IL-2-treated recipients retained their grafts, suggesting that insufficient IL-2 is not central to the established regulatory mechanism operated by the CD4+ suppressive cell population. This result contrasted with our finding when we treated the primary recipient of a graft during the therapeutic period; here, exogenous IL-2 prevented tolerance from becoming established and grafts were rejected by day 20. In some experiments, we explored the possibility that the anti-IL-4 mAb, 11b11, might influence tolerance be reducing IL-4 levels. However, 11b11 mAb treatment of recipients at the time of tolerance induction, or of naive recipients at the time of adoptive transfer of (BALB/c tolerant)CBA spleen cells, had no effect on subsequent induction of tolerance. These results are summarized inTable 1.

The regulatory “tolerant” cell population is self-sustaining. We found that suppressive effects of the adoptively transferred (BALB/c tolerant)CBA spleen cells remained even upon sequential transfer to further naive recipients of BALB/c hearts, over nine serial transfers (see Fig. 2A). Here, even if all the 5×107 transferred cells home to the spleen (whole spleen≈2×108 cells) at the time of the next adoptive transfer), the calculated minimal dilution of the original (BALB/c tolerant)CBA cells at the ninth transfer would be greater than 1 in 106. Since we know that 107 tolerant cells is insufficient for transfer of tolerance per se, then the regulatory cell population must amplify during the serial adoptive transfers. This amplification may occur either by direct clonal expansion, or by recruiting recipient cells as suppressors by“infectious tolerance,” or by both mechanisms. We were able to test for infectious tolerance by adoptive transfer of Thy1.2 (BALB/c tolerant)CBA spleen cells into a Thy1.1 congenic CBA recipient of a BALB/c graft, followed by anti-Thy1.2 mAb-induced (anti-donor) depletion of the original tolerant cells. The congenic recipient's own Thy1.1 cells were untouched by the depleting mAb. If the adoptively transferred cells were removed within the first week of coexistence, then no tolerance was established. In contrast, removal of the adoptively transferred cells at 14 days had no impact on the tolerance occurring in the naive recipient. This shows that infectious tolerance required some 14 days to become established, as shown in Figure 2B. Although in accord with infectious tolerance described for minor histoincompatible skin grafts where naive cells had been transferred to tolerant recipients (8),this is the first demonstration of infectious tolerance to major MHC antigens in the highly stringent model where suppressor cells are transfused into a fully immune-competent recipient. The fact that a whole mouse is made tolerant by infusion of suppressive cells emphasizes the potency of this elicited, natural regulatory mechanism for antigen-specific tolerance.

The “tolerant” cell population is specific for donor antigen but tolerance may spread to “linked” antigens. To determine the specificity of suppression mediated by tolerant spleen cells, we adoptively transferred the (BALB/c tolerant)CBA spleen cells to a naive CBA recipient of a C57/Bl10 heart: this third-party heart was rejected within 8 days (Fig. 3A). Thus the regulatory cells appeared to be specific for the original tolerizing alloantigen. Importantly, their specificity was maintained even when the adoptively transferred tolerant cells were challenged simultaneously with heart grafts from both donor (accepted) and third party (rejected) into the same naive recipient(Fig. 3B), emphasizing the robust nature of this specific tolerance. We next asked whether the tolerant cell population was able to protect against rejection of third-party, C57/Bl10 antigensco-expressed with donor-type antigens on a (BALB/c×C57/Bl10)F1 heart graft. Some 50% of naive recipients of an (BALB/c×C57/Bl10)F1 heart and (BALB/c tolerant)CBA spleen cells accepted the F1 heart. Furthermore, when (F1 BALB/c×C57/Bl10 tolerant) spleen cells were adoptively transferred, to a naive recipient of a C57/Bl10 heart, some 50% of the recipients accepted the full third-party graft, demonstrating the spread of tolerance through linked antigen expression to a second generation of linked antigen (Fig. 4). Importantly, in the adoptive transfer experiments it was routine to prepare three or four naive recipients using a common pool of tolerant spleen cells for infusion: some of these recipients became tolerant and others rejected their graft when linked third-party antigens were expressed, even though the source and number of tolerant cells infused were the same.

As a control to the above experiments, we confirmed that tolerance to BALB/c is not “special” by performing a reciprocal experiment. Tolerance generated to C57/Bl10 was specific for C57/Bl10 upon adoptive transfer of (C57/Bl10 tolerant)CBA spleen cells to a naive CBA mouse in that a third-party heart (here, BALB/c) was rejected at normal tempo, whereas the donor-type, C57/Bl10 heart was accepted. Furthermore, tolerance could be generated to linked antigens on (BALB/c×C57/Bl10)F1 using (C57/Bl10 tolerant)CBA spleen cells, indicating that our observations are likely to reflect general mechanisms of immune regulation.

Continuous stimulation by tolerizing antigen is required to maintain the “tolerant” cell population in an active state. We investigated whether antigen was necessary to maintain activity in the regulatory cell population. We had already established that when the primary heart graft is removed, the “graft-empty” recipient will accept a donor-type heart after a delay as long as 5 months, after which operational tolerance decayed with time (log rank gave P<0.01 significance for graft acceptance in 4-month graft-empty recipients compared with 6-month graft-empty recipients where hearts were rejected, data not shown). These kinetics of tolerance decay after graft removal are in accord with the findings of Hamano et al (23). Since the graft-empty mouse may retain donor-type antigens in the form of migrated donor-type cells and processed donor-derived peptides, we addressed the question of“memory” by use of adoptive transfer of (BALB/c tolerant)CBA spleen cells into a naive CBA mouse that was completely free of any donor-type antigen at the time of transfer: a donor-type heart was then engrafted at defined intervals into this naive mouse. In the absence of donor-type antigen, the (BALB/c tolerant)CBA spleen cells rapidly lost their capacity to suppress and by 14 days were unable to prevent rejection of a donor-type heart at normal tempo (Fig. 5). This was in marked contrast to the stable tolerance achieved when tolerant spleen cells were infused on the same day as allografting, and implies that persistence of peripheral tolerance mediated by spleen cells is dependent upon continuous antigenic stimulation.

Back to Top | Article Outline

DISCUSSION

We have previously shown that the peripheral immune system may be specifically manipulated to tolerate a fully mismatched (major MHC plus minor antigens) allograft by monoclonal antibody-mediated blockade of CD4 and CD8 co-receptors. We here report for the first time that the antigen-specific, CD4+ suppressor cells are able to maintain this tolerance when placed in a fully immune-competent mouse, and that the mechanism involved is a dominant process similar to that described by Qin et al.(8) for minor mismatched skin on tolerant recipients: naive graft-reactive cells are driven toward a dominant, tolerant state(infectious tolerance). Moreover, adoptive transfer of tolerant spleen cells into unmanipulated recipients was successful in inducing tolerance over multiple generations. Thus, the maintenance of tolerance is unlikely to be a special function of therapy with monoclonal antibodies, but resulted from a natural regulatory mechanism that was allowed to amplify and become dominant under cover of the initial therapy. Once established, this natural mechanism of peripheral tolerance is very robust.

In these transfer experiments, CD4+ regulatory cells are active in maintaining transplantation tolerance, even in the presence of large amounts of exogenous IL-2. Since peripheral tolerance due to clonal deletion should not be transferable or dominant, our results are indicative of an alternative mechanism wherein naive donor-reactive cells are actively guided toward the tolerant state. Cells reactive to third-party antigens retain aggressive immune responses unless they have “seen” third-party antigen in the context of (or linked to) the original tolerizing antigen after, for example, exposure to an F1 graft. In this case, reactivity against third-party antigen also became suppressive rather than aggressive. The true suppressive nature of these cells was demonstrated in these stringent experiments where cells were infused into fully immune-competent recipients. This adds to the observation by Davies et al (9), who found spread of tolerance to third-party, linked antigens within mAb-induced tolerant mice. Thus, once tolerance has been generated, the tolerant CD4+ cell population may force tolerance to spread to coexpressed novel antigens by an expansive mechanism that is specifically regulated.

A simple hypothesis to explain these observations would be a numerical dominance of suppressive cells associated with the graft, diluting out naive cells to a critical threshold below which potential aggression is switched to active suppression. Interestingly, not all recipients became tolerant to F1 grafts, and it is possible that the numbers of naive cells reactive to the third-party (F1) antigens are sufficient to destabilize the numerical dominance of the transferred tolerant cell population, the outcome depending upon which population first amplifies to impose Th1, or Th2, responses. This critical period may be influenced by the local cytokine milieu if suppressive mediators associated with tolerance to parental antigen impinge on cells specific for the F1 antigens.

Expansion of Th2 cells is associated with a shift in regulatory cytokine networks toward IL-4 in neonatal tolerance induction(24, 25). Although in our experiments six doses of the anti-IL-4 (mAb 11b11) over 21 days failed to prevent adoptive transfer of tolerance when tolerant cells were placed in a fully immune-competent naive recipient, there is evidence that IL-4 does play a role in maintaining tolerance in that: (1) anti-IL-4 may prevent tolerant cells imposing tolerance on naive graft-reactive cells in T-cell-depleted recipients where there is a close balance between rejection and tolerance (9); and (2) intragraft infusion of IL-4, but not IL-10, prolongs rat heart allograft survival (26). It is also relevant that lack of IL-2 may bias the immune system to Th2-type responses(20, 21, 22, 27). This may occur in our experiments since CD4 and CD8 co-receptor functions are required for IL-2 secretion (28), and exogenous IL-2 given at the time of CD4 and CD8 mAb therapy prevented tolerance induction. Taken together, these observations are in accord with the concept of amplification of graft-specific Th2-like cells under cover of CD4 and CD8 mAbs to result in a numerically dominant tolerant cell population that perhaps responds to IL-4 rather than IL-2 (29). These tolerant cells would be able to induce naive graft-reactive cells to become regulatory suppressive cells, so that the balance always favors tolerance.

Since the capacity to suppress was lost rapidly in the absence of antigen, we propose that the regulatory cells require continuous, or closely repetitive, antigenic stimulation to maintain intracellular signals associated with their dominant tolerant state. Where tolerant cells were removed from donor antigen, the rapid loss of tolerance may result from (1) reversion to aggressive responses after altered intracellular signaling; (2) inactivation of the regulatory cell population; or (3) failure of infectious tolerance to divert naive cells toward tolerance.

In the whole animal, removal of the graft resulted in eventual loss of tolerance, implying that residual microchimerism via passenger donor-type cells or donor-derived peptides is insufficient to maintain indefinite tolerance, probably due to levels gradually falling below that critical for maintaining tolerance and diverting new thymic emigrants from attacking the graft. It has been suggested that donor antigen-presenting cell microchimerism is required for maintenance of tolerance (30). Clearly, in these experiments donor-derived antigen-presenting cells in the absence of graft are insufficient to maintain tolerance.

In summary, we have demonstrated that it is possible to exploit natural regulatory mechanisms for specific transplantation tolerance. Use of murine models to understand mechanisms could eventually lead to identification of surrogate markers for tolerance, applicable to monitoring patients in clinical trials or on new therapeutic agents. Therapy may then be guided by identifying when a natural tolerant state has been reached so that immunosuppressive drugs may be reduced or withdrawn without danger to the patient or graft.

Figure 1

Figure 1

Figure 2

Figure 2

Figure 3

Figure 3

Figure 4

Figure 4

Figure 5

Figure 5

Back to Top | Article Outline

REFERENCES

1. Fowell D, Mason D. Evidence that the T cell repertoire of normal rats contains cells with the potential to cause diabetes. Characterisation of the CD4+ T cell subset that inhibits this autoimmune potential. J Exp Med 1993; 177(3): 627.
2. Hutchings P, Parish N, O'Reilly L, Dawe K, Roitt IM, Cooke A. The regulation of autoimmunity through CD4+ T cells. Autoimmunity 1993; 15(suppl) 21.
3. Qin S, et al. Induction of tolerance in peripheral T cells with monoclonal antibodies. Eur J Immunol 1990; 20: 2737.
4. Cobbold SP, Martin G, Waldmann H. The induction of skin graft tolerance in MHC mismatched or primed recipients: primed T cells can be tolerised in the periphery by CD4 and CD8 antibodies. Eur J Immunol 1990; 20: 2747.
5. Chen Z, Cobbold S, Metcalfe S, Waldmann H. Tolerance in the mouse to major histocompatibility complex-mismatched heart allografts, using monoclonal antibodies to CD4 and CD8. Eur J Immunol 1992; 22: 805.
6. Saitovitch D, Bushell AR, Morris PJ, Wood KJ. Induction of transplantation tolerance using a nondepleting anti-CD4 Mab and donor-specific transfusion before transplantation: evidence that a critical period of time is required for the development of immunological unresponsiveness. Transplant Proc 1995; 27: 117.
7. Scully R, Qin S, Cobbold SP, Waldmann H. Mechanisms in CD4 antibody-mediated transplantation tolerance: kinetics of induction, antigen dependency and role of regulatory T cells. Eur J Immunol 1994; 24: 2383.
8. Qin S, Cobbold P, Pope H, et al. Infectious transplantation tolerance. Science 1993; 259: 974.
9. Davies J, Leong LY, Mellor A, Cobbold SP, Waldmann H. T cell suppression in transplantation tolerance through linked recognition. J Immunol 1996; 156: 3602.
10. Chen Z, Cobbold SP, Waldmann H, Metcalfe S. Stability of tolerance in mice generated by CD4 and CD8 monoclonal antibody treatment: cell transfer experiments. Transplant Proc 1993; 25: 790.
11. Corry RJ, Winn HJ, Russell PS. Primary vascularized allografts of heart in mice. Transplantation 1973; 16: 343.
12. Chen Z. A technique of cervical heterotopic heart transplantation in mice. Transplantation 1991; 52: 1099.
13. Qin S, Cobbold SP, Tighe H, Benjamin R, Waldmann H. CD4 monoclonal antibody pairs for immunosuppression and tolerance induction. Eur J Immunol 1987; 17(8): 1159.
14. Cobbold SP, Jayasuriya A, Nash A, Prospero TD, Waldmann H. Therapy with monoclonal antibodies by elimination of T cells subsets in vivo. Nature 1984; 312: 548.
15. Waldmann H, Cobbold SP, Qin S, Benjamin RJ, Wise M. Tolerance induction in the adult using monoclonal antibodies to CD4, CD8 and CD11a (LFA 1). Cold Spring Harb Symp Quant Biol 1989; 54: 885.
16. Ledbetter JA, Herzenberg LA. Xenogeneic monoclonal antibodies to mouse differentiation antigens. Immunol Rev 1979; 47: 63.
17. Metcalfe SM, Alexander D, Turner J. FK506 and cyclosporin A each inhibit antigen-specific signaling in the T cell line 171 in the absence of a calcium signal. Cell Immunol 1994; 158: 46.
18. Tomonari K. A rat antibody against a structure functionally related to the mouse T-cell receptor/T3 complex. Immunogenetics 1988; 28: 455.
19. Oliveira OL, Thomas DB. A new thymocyte membrane differentiation antigen for mouse and rat. Immunogenetics 1985; 21(6): 559.
20. Lombardi G, Sidhu S, Batchelor R, Lechler R. Anergic T cells as suppressor cells in vitro. Science 1994; 264: 1587.
21. Dallman MJ, Wood KJ, Morris PJ. Recombinant IL2 can reverse the blood transfusion effect. Transplant Proc 1989; 21: 1165.
22. Dallman MJ, Shiho O, Page TH, Wood KJ, Morris PJ. Peripheral tolerance to alloantigen results from altered recognition of the interleukin 2 pathway. J Exp Med 1991; 173: 79.
23. Hamano K, Rawsthorne M, Bushell A, Morris PJ, Wood KJ. Donor microchimerism is not responsible for the maintenance of tolerance to donor alloantigens in recipients tolerant of cardiac allografts. Transplant Proc 1995; 27: 151.
24. Chen N, Field EH. Enhanced type 2 and diminished type 1 cytokines in neonatal tolerance. Transplantation 1995; 59: 993.
25. Chen N, Gao Q, Field EH. Expansion of memory TH2 cells over TH1 cells in neonatal primed mice. Transplantation 1995; 60: 1187.
26. Levy AE, Alexander JW. Administration of intragraft interleukin-4 prolongs cardiac allograft survival in rats treated with donor-specific transfusion/cyclosporine. Transplantation 1995; 60: 405.
27. Jenkins MK, Chen CA, Jung G, Mueller DL, Schwartz RH. Inhibition of antigen specific proliferation of type 1 murine T cell clones after stimulation with immobilized anti-CD3 monoclonal antibody. J Immunol 1990; 144(1): 16.
28. Glaichenhaus N, Shastri N, Littman DR, Turner JM. Requirement for association of p56lck with CD4 in antigen-specific signal transduction in T cells. Cell 1991; 64: 511.
29. Cobbold SP, Adams E, Marshall S, Davies J, Waldmann H. Mechanisms of peripheral tolerance and suppression induced by monoclonal antibodies to CD4 and CD8. Immunol Rev 1996; 149: 1.
30. Starzl TE. Chimerism and donor specific non-reactivity 27 to 29 years after kidney allotransplantation. Transplantation 1993; 55: 1272.
© Williams & Wilkins 1996. All Rights Reserved.