Translating transplantation tolerance in the clinic: where are we, where do we go?

Abstract

Research networks dedicated to translation of immune tolerance in the clinic currently support pilot trials aiming at immunosuppression withdrawal in kidney or liver allograft recipients. Although results obtained so far indicate that significant hurdles still need to be overcome before organ transplant recipients can be weaned off drugs safely and routinely, recent advances suggest that immunosuppression minimization on the basis of validated biomarkers might become standard practice in a near future.

Introduction

Following the seminal observations of Medawar and colleagues in newborn mice [1], a number of experimental studies demonstrated that it is possible to take advantage of a transient period of immunosuppression to achieve engraftment of long-lived allogeneic haematopoietic cells, resulting in tolerance to tissue or organ allografts bearing the same alloantigens as those expressed by the tolerogenic cells [2]. The best proof of principle for the induction of tolerance to solid allografts in humans using this strategy is provided by the observation that patients having received an allogeneic bone marrow graft under a myeloablative regimen subsequently accepted an organ transplant from the same donor without immunosuppression [3,4]. In these settings, all the recipients' haematopoietic cells are eliminated and substituted by cells of donor origin, a situation referred to as ‘full chimerism’. The toxicity of the treatments required to achieve full chimerism clearly prevents its application in non-malignant conditions. Milder, non-myeloablative conditioning regimens were found to result in a state of ‘mixed chimerism’ where significant numbers of donor-type and recipient-type haematopoietic cells co-exist, at least in the short term. There are several advantages of mixed chimerism over full chimerism, including the reduced risk of graft-versus-host disease because of deletion of donor-type host-reactive T lymphocytes [2]. On this basis, several clinical protocols aiming at complete immunosuppression withdrawal after combined haematopoietic cell and solid organ transplantation have been launched recently. Herein, we first review the major lessons learned from these pilot trials and discuss the hurdles which currently limit their further development. We then underline the progress made in the field of biomarkers which might allow minimization of immunosuppression in selected individuals according to novel stratifications of organ transplant recipients.

Split tolerance after combined transplantation of haematopoietic cells and solid organs in humans

Following the pioneering report by Monaco et al.[5] and supported by the microchimerism theory of Starzl [6], several investigators attempted to facilitate acceptance of solid allografts by infusing donor bone marrow cells in conjunction with standard immunosuppression, which usually included an induction course with polyclonal or monoclonal anti-T lymphocyte antibodies [7–10]. Although favourable clinical outcomes were reported, and molecular techniques eventually demonstrated low levels of chimerism (so-called ‘microchimerism’), there was no evidence for tolerance as long-term immunosuppression was maintained in most patients.

Sachs, Sykes and Cosimi reported the first attempts of deliberate induction of transplantation tolerance through mixed lymphohaematopoietic chimerism in patients with multiple myeloma and end-stage renal failure [11–13]. The preparative regimen consisted of treatment with cyclophosphamide, anti-thymocyte globulin and thymic irradiation, and a human leucocyte antigen (HLA)-matched sibling served as donor of bone marrow and kidney. Donor bone marrow infusion was performed on the day of renal transplantation and was followed by infusions of donor leucocytes. Cyclosporine A was given for about 2 months post-transplantation and immunosuppression was then withdrawn. Despite the repeated administration of donor cells, lymphohaematopoietic chimerism declined gradually and eventually disappeared in four of six patients. However, kidney grafts were tolerated, except in one patient in whom reversible acute kidney graft rejection occurred. In parallel, Millan et al. attempted to facilitate acceptance of HLA-mismatched living related renal transplants using total lymphoid irradiation plus anti-thymocyte globulin as the conditioning regimen and granulocyte-colony-stimulating factor mobilized CD34⁺ stem cells as the source of donor haematopoietic cells [14]. As in the previous study, chimerism was lost 2–3 months post-transplantation. Only one patient was reported to have been weaned off immunosuppressive drugs, indicating that this protocol could not induce allograft tolerance reliably.

In a recent study by the Massachusetts General Hospital group, Kawai et al. have rejuvenated interest in tolerance induction in HLA-mismatched kidney transplantation. These investigators reported graft acceptance after discontinuation of immunosuppression in four patients conditioned with anti-CD2 monoclonal antibody, cyclosporine A and thymic irradiation in association with anti-CD20 monoclonal antibody in two of them, after the observation of irreversible acute humoral rejection in one patient [15]. Again, chimerism was only transient in this series of patients. A similar observation was made in two patients who received a living donor liver transplant for advanced intra-hepatic malignancy 6–8 weeks after conditioning consisting of cyclophosphamide, anti-thymocyte globulin and purified donor CD34⁺ stem cells [16]. Both patients remained rejection-free for several months after immunosuppression withdrawal despite detectable chimerism, but eventually died of tumour recurrence. Taken together, these studies suggest that functional solid allografts can be maintained for prolonged periods of time despite disappearance of the donor cells used to reprogramme the recipient's immune system. However, the robustness of these situations of ‘split tolerance’ is not established. In order to achieve long-term mixed chimerism, more aggressive conditioning regimens, including total lymphoid irradiation, might indeed be needed, as suggested by a recent report by Strober et al.[17]. Whether long-term mixed chimerism and hence more aggressive pretransplant conditioning is necessary for immunosuppression reduction/withdrawal in kidney or liver transplant recipients who receive donor bone marrow remains to be determined.

Beside haematopoietic stem cells, other cell types are currently considered as potential therapies for the induction of transplantation tolerance, including tolerogenic dendritic cells, regulatory T cells, myeloid suppressor cells and mesenchymal stem cells. The latter cells arouse special interest, as a recent study showed that they are clinically efficient in controlling graft-versus-host disease [18]. While the ability of those cell types to harness tolerance to solid organ allografts in the clinic remains to be investigated, in the setting of human kidney transplantation Fändrich et al. recently assessed a conditioning regimen based on ‘transplant acceptance-inducing cells’ obtained by culturing donor peripheral blood mononuclear cells with recombinant macrophage-colony-stimulating factor before pulsing with interferon (IFN)-γ[19]. After infusion of such cells at the time of living-donor renal transplantation, one patient enjoyed normal graft function for 8 consecutive months in the absence of immunosuppression, whereas drug minimization was successful in four other patients [19].

Why do human beings resist mixed chimerism and transplantation tolerance?

Results of the pilot trials reviewed in the previous section establish clearly that stable mixed chimerism and associated transplantation tolerance are much more difficult to achieve in human beings than in animal models. Future advances in the field will depend upon the design of strategies to overcome the multiple hurdles we are facing in the clinic. Indeed, it is now clear that even when very large numbers of donor cells are infused (the so-called ‘megadoses’), robust engraftment is difficult to achieve when the haematopoietic graft is depleted of T cells to prevent graft-versus-host disease and when the recipient is conditioned with a non-myeloblative regimen with reduced toxicity, which are prerequisites in solid organ transplantation (reviewed in [20]).

A major barrier is represented by memory T cells, which are present in large numbers in humans as a consequence of repeated exposure to microbes, especially viruses. Indeed, it is now clear that protocols allowing establishment of mixed chimerism in naive mice are much less efficient in the presence of virally induced memory T cells. This has been well studied in mice treated with co-stimulation blockade agents (CTLA4-immunoglobulin, anti-CD40 ligand antibody) and exposed to viruses causing acute infections [21,22], but also upon latent virus infection [23]. Strikingly, T cell depletion regimens expand and activate memory T cells [24,25], and homeostatic proliferation of memory-like T cells in the context of lymphopenia has been shown clearly to prevent the induction of transplantation tolerance in mice [26]. In solid organ transplant recipients treated with T cell-depleting antibodies (alemtuzumab or polyclonal anti-T cell antibodies), activated memory CD4⁺ T cells appear to expand preferentially during the first 3 weeks post-lymphocyte depletion [24] while CD8⁺ T cells appear dominant at later time-points [27–29]. Whereas a fraction of the CD8⁺ T cells which emerge after T cell-depletion are CD28⁻CD8⁺ cells with an immunosenescent phenotype [29], other CD8⁺ T cells display effector activity. Indeed, in a recent trial of early immunosuppression withdrawal after high-dose anti-lymphocyte globulins in liver transplant recipients, acute rejection episodes were related temporally to the emergence of IFN-γ-producing activated CD8⁺ T cells [30]. Furthermore, autoreactive memory CD8⁺ T cells expanded preferentially and were possibly responsible for diabetes recurrence in islet transplant recipients [31]. In the latter study, the process of homeostatic proliferation was found to be inhibited by mycophenolate mofetil but not by either tacrolimus or rapamycin [31]. Because the infusion of regulatory T cells is considered currently as an adjunct cell therapy to facilitate mixed chimerism and transplantation tolerance, it is important to underline here that both CD4⁺ and CD8⁺ alloreactive memory T cells were found to be resistant to suppression by regulatory T cells upon transfer in lymphopenic hosts [32]. Moreover, there is evidence that regulatory T cells even enhance the activity of memory cells of the T helper 17 type [33], a cell population which might be responsible for alternative modes of rejection [34].

The quest for biomarkers to tailor immunosuppression

In view of the major barriers to overcome before large-scale clinical trials aiming at the induction of transplantation of tolerance can be launched safely, how can we translate our current knowledge of transplantation immunology into an improved standard of care for solid organ transplant recipients? As well as the development of new pharmaceutical products with reduced toxicity and more comfortable mode of administration [35], tailoring immunosuppression according to the immune status of each patient would clearly represent a significant advance. Although a number of immunological methods have been proposed to monitor alloreactivity in humans [36], the only tests which are accepted as robust biomarkers are those measuring humoral parameters. It is well established that the presence before transplantation of antibodies directed against donor HLA molecules is a major risk factor for early severe renal graft rejection, and more recently pretransplant antibodies against major histocompatibility complex class I-chain A (MICA) antigens were also found to be associated with an increased frequency of renal allograft loss [37]. Whereas tests assessing allospecific T cell-mediated immune responses will probably be difficult to standardize and introduce into routine clinical practice, several recent studies suggest that gene profiling might become a clinically relevant monitoring tool. Indeed, microarray analysis of peripheral blood mononuclear cells samples from cohorts of drug-free patients with a functional graft allowed the proposal of molecular fingerprints or signatures for operational tolerance in kidney and liver transplantation [38,39]. In cardiac transplantation, a similar approach has been suggested to differentiate graft quiescence from rejection, although additional studies will be necessary for definitive qualification of this assay for the detection of cardiac transplant recipients with low risk of acute cellular rejection [40]. As it is likely that the most informative biomarkers should be sought at the graft level, one can predict that molecular imaging of transplanted organs will become an area of intense investigation in the near future [41].

Patients' stratification: towards a new paradigm in transplantation medicine

As in other areas of modern medicine, future advances will depend upon tailoring the therapy according to the individual patient's needs [42]. In the context of transplantation, this would mean to reduce immunosuppression to the minimal level required to prevent rejection. Stratification of transplant recipients will be based upon established risk factors (e.g. cold ischaemia time, preformed anti-HLA or anti-MICA antibodies) and on post-transplant monitoring of non-invasive biomarkers. Although several candidate biomarkers are available, the challenge now is to validate these biomarkers in order to make them acceptable by the medical community and regulatory agencies. Large clinical trials will need to be designed specifically for that purpose [43]. Hopefully, this will be facilitated by research networks which have been assembled in the United States [44] and the European Union [45].