|Home | About | Journals | Submit | Contact Us | Français|
Immune regulation or suppression has long been proposed to play a role in transplantation. Various types of regulatory T cells have been documented since the early 1970’s1. However, the absence of specific and reliable marker(s) to discriminate regulatory T cells (Treg) from effector T (Teff) cells hampered their characterization. Seminal work by Sakaguchi et al revealed that CD4+CD25+ T cells were potent suppressors of autoimmunity2. These CD4+CD25+ T cells were subsequently termed natural regulatory T cells, or nTreg cells. Later, a transcription factor, forkhead box protein 3 (Foxp3) was identified as a master regulator for Treg generation and differentiation, and as a more specific marker than CD25 (i.e., the α-chain of interleukin [IL]-2 receptor) for Treg cells because the latter is also expressed on activated Teff cells3–5. Nonetheless, CD25 remains a reliable cell surface marker for isolation of nTreg cells as staining for Foxp3 requires cell permeation leading to cell death, making Foxp3 unhelpful for cell isolation.
Foxp3+CD4+CD25+ nTreg cells have been increasingly documented to suppress allograft rejection even though rejection is often an inevitable outcome in non-immunosuppressed recipients due to vigorous activity of alloreactive Teff cells. In vitro, naïve CD4+CD25− Teff cells were more reactive than whole CD4+ Teff cells against alloantigen (alloAg) stimulation6. Both freshly isolated and ex vivo-activated nTreg cells potently suppressed alloAg-driven Teff cell activation in mixed lymphocyte reaction (MLR) assays6,7. In vivo, pre-transplant depletion of Foxp3+CD4+CD25+ nTreg cells by anti-CD25 mAbs resulted in subversion of tolerance into chronic rejection in otherwise spontaneously tolerant mouse models or conversion of chronic rejection into acute rejection in otherwise chronically rejecting rodents8–12.
Foxp3+CD4+CD25+ Treg cells have also been shown to mediate in vivo allograft tolerance. Interestingly, CD4+CD25+ cells have been shown to be the primary mediator of allograft tolerance in rats even before CD4+CD25+ cells were recognized as a population of Treg cells13. Recently, CD4+CD25+ Treg cells have been repeatedly implicated in initiation and/or maintenance of allograft tolerance induced by a variety of tolerizing treatments14, 15. Under most conditions, the induced tolerance is transferable to naïve hosts through the adoptive transfer of the Foxp3+CD4+CD25+ population composed of expanded Foxp3+CD4+CD25+ nTreg cells and/or induced Treg (iTreg) cells from Foxp3−CD4+CD25− precursors14, 15 from the tolerized hosts. Antigen-specific Foxp3+CD4+CD25+ iTreg cells could also be generated ex vivo under special culture conditions from Foxp3−CD4+CD25− precursors in the presence or absence of Foxp3+CD4+CD25+ nTreg cells16–19. In humans, accumulating evidence suggests that a favorable allograft outcome is frequently associated with a robust population of FOXP3+CD4+CD25+ Treg cells despite conflicting evidence emerging in part due to the confounding expression of FOXP3 by activated Teff cells20–23.
Here, we will review current approaches for amplification of allo-specific Foxp3+CD4+CD25+ Treg cells for prevention of allograft rejection. Available strategies are: (i) immunotherapy with ex vivo-expanded Foxp3+CD4+CD25+ nTreg cells; (ii) immunotherapy with ex vivo-induced Foxp3+CD4+CD25+ iTreg cells; and/or (iii) in vivo expansion and/or induction of Foxp3+CD4+CD25+ Treg cells by tolerizing treatments.
Foxp3+CD4+CD25+ nTreg cells have a broad T cell receptor (TCR) repertoire including clones recognizing alloAgs24, 25. Presence of such allo-suppressive clones within the fresh repertoire has been confirmed in vitro and in vivo. In vitro, fresh Foxp3+CD4+CD25+ nTreg cells from naïve rodents or humans effectively suppress alloactivation of recipient T cells against donor alloAg in MLR assays6, 26. In vivo, sufficient quantities of freshly isolated Foxp3+CD4+CD25+ nTreg cells from naïve animals have been demonstrated to suppress simultaneously co-transferred Teff cell-mediated allograft rejection in immunodeficient rodents27. Due to the low precursor frequency, efficient ex vivo expansion of fresh naïve Foxp3+CD4+CD25+ nTreg cells is likely essential for generating sufficient numbers of cells to therapeutically target a large repertoire of alloreactive Teff clones in the recipients. Additionally, in vitro-activated nTreg cells exert stronger suppression than fresh cells against alloAgs6,28. Ex vivo expansion could be achieved through alloAg-nonspecific and alloAg-specific manners. Regardless of the methods, preserving certain cell characteristics would be necessary for the effectiveness of the ex vivo-expanded nTreg cells in targeting allograft rejection. These characteristics include: sufficient cell number, preservation of Foxp3, survival in vivo after infusion for some unknown period of time if not permanently, in vivo migration to sites of alloreactivity, and the ability to suppress alloreaction in an alloAg-specific manner.
The basic requirements to expand anergic fresh Foxp3+CD4+CD25+ nTreg cells include triggering of TCR (signal 1), activation of costimulation pathway(s) (signal 2), and the presence of IL-2. It is logical to achieve alloAg-specific expansion by co-culture of fresh naïve nTreg cells in the presence of donor-type dendritic cells (DC). Indeed, DC-expanded nTreg cells seem to be highly suppressive against same donor alloAgs in vitro and in vivo29. However, it is not an efficient strategy in terms of cell yield6, 29. Thus far, alloAg-nonspecific expansion using anti-CD3/CD28 mAb-coated beads plus IL-2 appears to be the most efficient method for ex vivo expansion of both mouse and human nTreg cells30–32. These alloAg-nonspecifically expanded nTreg cells have been shown to suppress alloactivation in vitro as well as to prevent allograft rejection and graft-versus-host disease in vivo, although they are less efficacious than alloAg-specific nTreg cells 6,32. We have developed a protocol achieving a reasonable level of expansion while achieving alloAg specificity by re-stimulation of nTreg cells with donor BM-derived DC (BM-DC) after initial expansion using anti-CD3/CD28 mAb-coated beads and IL-233. Cells expanded by this method were potent suppressors against donor but not 3rd-party alloAg in vitro and in vivo33.
Unexpectedly, a rapid decline of the percentage of intracellular Foxp3+ cells during alloAg-nonspecific and alloAg-specific expansion was observed in our study, but overall both Foxp3+ and Foxp3− fractions were effectively expanded33,34. By using flow-sorted Foxp3+ cells from Foxp3-green fluorescence protein (GFP) knock-in mice, we have recently confirmed activation-induced loss of Foxp3 in our culture system with fewer than 30% cells remaining Foxp3+ at day 18 of IL-2 containing culture (Xia G, et al. manuscript submitted). Decrease of Foxp3 has also been observed in other culture conditions. An example is a recent study using Foxp3+ Treg cells isolated from Foxp3-transgenic mice which showed that Foxp3+ Treg cells lost Foxp3 expression and became Teff cells during in vitro Th1 and Th2 differentiation35. Another recent study using Foxp3+ nTreg cells purified from Foxp3-GFP knock-in mice has shown that 12–17% Foxp3+ nTreg cells became Foxp3− cells at 24 hours after in vitro anti-CD3 stimulation36. Decrease of Foxp3 at the gene transcript and/or protein levels has also been reported after in vitro activation of human CD4+CD25+ nTreg cells 37,38.
The exact mechanism(s) underlying the loss of Foxp3 during in vitro expansion are not clear. Utilizing IL-2 alone in our current protocol may be insufficient and may partially be responsible for the loss of Foxp3. Besides the crucial and dominant role of IL-2, other members of the ©-chain cytokines including IL-7 and IL-15 have also been shown to play a role in the survival of CD4+CD25+ nTreg and in the maintenance of Foxp3 expression, although it needs to be further studied whether they synergize with IL-2 in vitro and/or in vivo39,40. A series of studies have shown that TGF-β is critical in generating, maintaining, and expanding the in vivo pool of CD4+CD25+ Treg cells via Foxp3 modulation and/or induction, and is likely also important in mediating in vivo suppressive function41–44. Indeed, adding TGF-β to our cultures has partially rescued the loss of Foxp3 and the cells retained stronger suppressive function, while similar levels of expansion were achieved33. This is consistent with a recent study showing that addition of TGF-β to cultures helped to maintain Foxp3 in nTreg cells isolated from Foxp3-GFP knock-in mice36. Interestingly, in the presence of TGF-β, nTreg cells not only retained higher levels of Foxp3 during ex vivo expansion (from 10% to 25%) but expressed much higher levels of Foxp3 in vivo (from 15% to 55%) after adoptive transfer33. This is fundamentally different from de novo in vitro-generated Foxp3+CD4+CD25+ iTreg cells, which generally tended to lose their Foxp3 expression in vivo (fewer than 10% remained Foxp3 expression in 2-wks post transfer) after adoptive transfer into allograft-bearing Rag−/− mice (Xia G, et al., manuscript submitted and Luo, unpublished observations).
Addition of retinoic acid and histone deacetylase inhibitor trichostatin A into our aforementioned culture system could further improve expression of Foxp3 in vitro (>45% remaining Foxp3+) and after adoptive transfer in vivo (85~90% remaining Foxp3+) to prevent allograft rejection (Xia G, et al. manuscript submitted). Thus in our culture system, retinoic acid and trichostatin A synergize with IL-2 and TGF-β to preserve expression of Foxp3 in vitro. Maintaining Foxp3 promoter CpG demethylation and protein acetylation by these agents appear to be important45,46,47. Lastly, the role of chemokines in the function of Treg cells has been mainly focused on chemokine-medicated recruitment of Treg cells to site of inflammation48–51. Interestingly, there is growing evidence to suggest that expression (or the lack of) certain chemokines is associated with altered expression levels of Foxp352, 53, adding another potential avenue for modification during Treg expansion. At present time, optimizing ex vivo expansion protocols remains a challenge for maximally preserving Foxp3 while achieving an efficient level of expansion.
After adoptive transfer, ex vivo-expanded nTreg cells survived, expanded long-term and trafficked through the peripheral blood and lymphoid tissues in allograft-bearing mice, suggesting that similar to fresh Treg cells, their maintenance could be driven by self- or allo-Ags and that their homing capability was not apparently impaired after ex vivo-expansion6,32,33. Moreover, nTreg cells migrated to and resided within allografts after adoptive transfer suggesting their intra-graft presence may be essential for exerting graft protection14,32,33. We have observed that various degrees of loss of Foxp3 expression occurs after adoptive transfer of the in vitro expanded or generated Treg cells (ref.32 and Luo, unpublished observations), although others have shown stable maintenance of such Foxp3+ cells in vivo54. Emerging data now suggest that cross-differentiation of Foxp3+ cells to other Th-type cells such as IFN-γ-secreting Th1 cells or IL-17-secreting Th17 cells with or without concurrent Foxp3 expression is possible in vitro and in vivo under appropriate stimulation or inflammation micro-environment35,55,56. Single cell cloning of FOXP3+ cells in humans further revealed significant heterogenecity of this cell population57. These data suggest that further identification of better-defined starting populations as well as optimal control of in vivo cytokine milieu will assist optimal survival and sustained expression of Foxp3 of the adoptively transferred Treg cells generated ex vivo.
The exact mechanism of allosuppression mediated by adoptively transferred nTreg cells is unclear, but multiple mechanisms may be involved. In vitro data suggest that once nTreg cells become activated, their suppression is Ag non-specific28. However, recent data have suggested that the repertoire of nTreg cells could be shaped in vivo to suppress T-cell reactivity in an Ag-specific manner58–60. In our system, alloAg-specific clones were first selected ex vivo by co-culture with donor BM-DC and subsequently maintained in vivo via cognate interactions with alloAg expressed on heart allografts, likely via both direct and indirect alloAg presentation pathways33, 58. Indeed, some recent studies have suggested that CD4+CD25+ nTreg are particularly potent at recognizing alloAg via the indirect pathway, and that ex vivo priming through the indirect pathway using recipient DCs is necessary for overcoming chronic rejection in vivo after adoptive transfer in mice61,62. Furthermore, adoptively transferred Treg cells may facilitate development and/or expansion of endogenous Treg cells specific for donor alloAg, thereby prompting a cascade of Treg cell-amplification resulting in a robust status of allograft tolerance, a process known as “infectious tolerance”63. This possibility could be mediated through direct T cell-T cell interaction or through induction of and interactions with “toleragenic” dendritic cells64–68. Thus, a potential state of pan-immunosuppression is avoided and alloAg-specific immunotherapy with Foxp3+CD4+CD25+ nTreg cells may be used as a non-toxic modality for allograft rejection.
In reality, tolerance does exist after liver and kidney transplantation in humans, albeit a rare event69. Amplification of host CD4+CD25+ Treg cells would be of critical value in promoting transplantation tolerance in humans. The basic principles and expansion strategy successful in rodents seem to be generally applicable to humans70–72. In the organ transplantation setting, nTreg cells would be of recipient origin. For application in clinical practices, all procedures including nTreg cell isolation and expansion, facility, equipments, reagents, and quality control measures should be in compliance with good manufacturing practice requirements73. Besides stringent regulatory, logistic, and financial challenges, there are unique technical challenges for the success of human Treg cell immunotherapy. First, purification of bona fide FOXP3+CD4+CD25+ nTreg cells is problematic due to contamination of activated Teff cells if purification is simply based on the expression of CD25high. Recent studies have progressed to using a combination of surface markers such as CD127, CD49d, CD45RA and CD25 to purify nTreg cells74–76. However, continuing search for better membrane markers is still needed to identify bona fide nTreg cells. Second, expression of FOXP3 is not necessarily an indicator of suppressive capacity, as activated human Teff cells also express FOXP3, albeit transient and/or at a lower level, without suppressive function77. Third, a likely expansion strategy is a combination of alloAg-nonspecific (e.g., by anti-CD3/CD28-coated artificial antigen presenting cells, or aAPC) and alloAg-specific (alloAg-bearing APC-mediated) expansion78. However, the optimal form of APCs has yet to be identified79. Fourth, in experimental studies using adoptive transfer model, we and others have shown that the ratio of Treg:Teff determines the efficacy of Treg cells in controlling allograft rejection; usually at least 1:1 Treg:Teff ratio is required to achieve significant graft-prolonging benefits27,32,80. The number of cells necessary for achieving clinical benefits in humans is yet to be defined. Fifth, pre- and/or peri-transplant induction therapy (to deplete or functionally inactivate Teff cells) is probably important for maximizing the benefits of nTreg cell immunotherapy33. Induction therapies that do not interfere with the survival and function of adoptively transferred ex vivo-expanded, recipient-origin nTreg cells will need to be designed. Likewise, choice of maintenance immunosuppressants is likely to be critical as well, as they may interfere with the adoptively transferred nTreg cells81. Lastly, further challenges exist in human nTreg cell therapy in tracking, monitoring and functional assessment of these cells in vivo once they are delivered to the recipients.
As discussed earlier, one of the main challenges for expanding existing nTreg cells is the lack of reliable cell surface markers that would allow sorting out nTreg cells of high purity, thus risking contamination with non-Treg CD4+ cells and subsequent expansion of the latter as an effector population. Therefore, if culture conditions exist that would convert non-Treg cells into functionally suppressive Foxp3-expressing Treg cells (iTreg), this would not only alleviate concerns over contamination of the expanding nTreg cell population with non-Treg cells, but would also allow exploration of the much more abundant Foxp3−CD4+CD25− T cell population for conversion to Foxp3+CD4+CD25+ suppressor cells.
While lack of functional Foxp3 does not seem to affect the intrathymic development of Treg cells82, maintenance of Treg cell function has been shown to be critically dependent on the continued expression of functional Foxp383. Consequently, much effort has been focused on the induction of Foxp3 for generation of iTreg cells.
Chen et al demonstrated that when naïve Foxp3−CD4+CD25− T cells were stimulated by anti-CD3 and anti-CD28 in the presence of TGF-β1 and IL-2, both CD25 and Foxp3 were up-regulated, and consequently these cells phenotypically became Foxp3+CD4+CD25+ T cells16. Functionally, these cells were suppressive toward naïve cell proliferation stimulated by anti-CD3 and anti-CD28. This phenomenon was later reproduced by many. In some cases, IL-2 appeared to be obligatory84,85, whereas in others it appeared to be dispensable86,87, although the possibility remained that the activated T cells were the source of IL-2 in the process of conversion.
To delineate the role of APCs in the process of TGF-β– dependent induction of Treg cells, we have taken advantage of transgenic T cells with a TCR that recognizes a single diabetogenic epitope represented by the mimotope BDC peptide, and tested dendritic cell sub-populations in this process86. We found that compared to granulocyte macrophage colony stimulation factor (GM-CSF)-differentiated BM-DCs that express high levels of costimulation molecules, splenic CD11c+ DCs expressing much lower levels of costimulation molecules are far more conducive to TGF-β-dependent induction of Foxp3 expression. In a more recent study, splenic DCs were further fractionated and the CD8α+ DCs were shown to be the main subset within the splenic DCs that determined inducibility of Foxp3 by TGF-β in vitro88. While coinhibitory signals through programmed death ligand 1 was obligatory for conversion, costimulatory signals through glucocorticoid-induced TNFR-related protein (GITR) suppressed conversion, implicating a balance of signal strengths being an important determinant for de novo Treg induction. In vivo, this subset of DCs likely also plays a role in differentiation of peripheral Foxp3+ Treg cells, and this ability is at least in part due to their ability to produce TGF-β89. Another recent series of reports identified that DCs derived from gut-associated lymphoid tissue, particularly the lamina propria, are highly efficient at de novo generation of Foxp3+ Treg cells from Foxp3− precursors, and that this ability can be attributed to their capacity to produce retinoic acid (RA) and the gut-homing receptor CD10390,91. This is an important finding in iTreg biology, because TGF-β has also been found to promote the differentiation of naïve T cells into proinflammatory IL-17 producing cells in the presence of IL-692. It appears that TGF-β-dependent T cell differentiation is reciprocally controlled by RA and IL-6. In the presence of RA, IL-17 differentiation is inhibited and iTreg cell differentiation is enhanced; whereas in the presence of IL-6, the opposite occurs93.
At a molecular level, it has recently been shown that the promoter of the Foxp3 gene is highly de-methylated in nTreg cells, correlating with high levels of Foxp3 expression; whereas in non-Treg cells, this region is highly methylated correlating with minimal Foxp3 expression45, 94. When TGF-β is present in the process of naïve T cell activation, this region can be induced to become partially de-methylated, leading to induction of Foxp3 expression45,95. We have recently shown that the activation state of the extracellular-regulated kinase (ERK) is implicated in the TGF-β-induced demethylation of the Foxp3 gene promoter46. It appears that ERK phosphorylation during T cell activation can be significantly inhibited by the presence of TGF-β. Inhibition of ERK phosphorylation in turn leads to down-regulation of DNA methyl-transferases, which could account for the partial de-methylation of the Foxp3 promoter. This hypothesis was corroborated by demonstrating that Foxp3 expression could be induced by direct ERK inhibition using a small molecule inhibitor of ERK, UO126, during T cell activation. Another recent report also implicated other signaling network involving phosphatidyl inositol 3-kinase, protein kinase B, and the mammalian target of rapamycin in the epigenetic regulation of Foxp3 gene expression96.
Careful promoter studies of the Foxp3 gene revealed that in mouse T cells, TGF-β-induced promoter hypo-methylation involves only selected CpG islands, and that the temporarily hypo-methylated sites could be readily re-methylated upon retrieval of TGF-β, suggesting that the induced expression of Foxp3 may not be stable45. Furthermore, in human T cells, it appears that TGF-β-induced FOXP3 expression was not associated with FOXP3 promoter de-methylation at all97. A recent study showed that in human cells TGF-β-induced FOXP3+ T cells were neither anergic nor suppressive77. These studies raise further questions with respect to the utility of TGF-β-induced Treg cells in humans.
As described earlier, MLR has been traditionally used as an in vitro measurement for allogeneic responses in which recipient T cells are cultured with donor APCs. Interestingly, in combination with TGF-β and IL-2, naïve recipient T cells stimulated in MLRs have been shown to convert to CD4+CD25+ regulatory T cells66, 98. These early studies utilized irradiated donor whole splenocytes with or without T cell depletion as allogeneic stimulators. Such in vitro MLR-generated CD4+CD25+ regulatory T cells were capable of prolonging donor-specific cardiac graft survival. Furthermore, in such MLRs containing both CD4+ and CD8+ recipient T cells, TGF-β appeared to be able to induce Foxp3-expressing suppressive (regulatory) cells in both T cell subsets, a phenomenon described by several other groups as well99–101.
As with the induction of autoantigen-specific iTreg cells, the role of APCs in efficient de novo induction of alloantigen-specific iTreg cells remains to be defined. In our experience, immature spleen CD11c+ DCs isolated from the donor strain are poor inducers for conversion of recipient bulk CD4+ T cells to Foxp3+CD4+CD25+ iTreg cells in primary cultures, even in the presence of IL-2 (Luo, unpublished observation). On the other hand, BM-DCs that are “alternatively activated” have been shown to be “toleragenic” and are able to induce Treg cells in vivo and in vitro. BM-DCs differentiated in the presence of IL-10 and TGF-β have been shown to be poor stimulators of allogeneic T cell proliferation, but they strongly induce IL-10 production and expansion of Foxp3+CD4+CD25+ cells in cultures102. Likewise, BM-DCs differentiated in pharmacological agents such as the mTOR inhibitor sirolimus or LF 15-0915, a chemical analogue of 15-deoxyspergualin, have also been shown to selectively increase the Foxp3+CD4+CD25+ population103–105. In these systems, the fact that the starting population was comprised of naïve whole CD4+ cells made it difficult to discern whether an increase in the Foxp3+CD4+CD25+ T cell population came from expansion of existing nTreg cells or de novo conversion of previous non-Treg cells. In addition, whether the direct (by using donor APCs) or indirect (by using recipient APCs) presentation is more efficient in iTreg cell generation in vitro remains to be defined.
Several well-studied tolerance protocols in transplant models have shown in vivo expansion and/or induction of Foxp3+CD4+CD25+ Treg cells as one of the mechanisms for tolerance induction. These protocols utilized mitogenic or non-mitogenic anti-CD3 monoclonal antibodies, non-depleting anti-CD4 and anti-CD8 antibodies, anti-CD154 (anti-CD40L) antibody, CTLA4-Ig, and more recently toleragenic DCs. Anti-CD3 mAbs have been widely studied, particularly in autoimmunity (reviewed in ref.106). Contradicting the initial assumption that these were T cell depleting agents, it is now known that depletion of T cells by these antibodies was rather modest (20% to 50%, varying depending on FcR-binding capacity)107. Rather, T cell anergy, as well as a significant increase in the Foxp3+CD4+CD25+ Treg cell population, can be induced108,109. The capacity of anti-CD3 mAb to induce operational tolerance has been demonstrated in an allogeneic cardiac transplant model110, as well as in the spontaneous autoimmune diabetes NOD model108. The induced tolerance was exquisitely dependent on TGF-β in the NOD model. Subsequent studies demonstrated that treatment with anti-CD3 results in two sources of increased TGF-β production: first, by phagocytic macrophages and immature DCs triggered by the apoptotic T cells induced by anti-CD3111; and second, by the CD4+CD25+ T cells themselves108,112. Therefore, it is conceivable that a positive feedback loop may be established involving initial iTreg cell induction by phagocyte-produced TGF-β which is later augmented by TGF-β production by Treg cells themselves. Further expansion of nTreg and/or iTreg cells through such an “infectious tolerance” process leads to a durable state of immunoregulation by the anti-CD3 treatment.
Another robust tolerance regimen for allogeneic transplantation in mice and in nonhuman primates is the i.v. infusion of donor cells (donor-specific transfusion (DST)) with concomitant administration of anti-CD154. Depletional studies demonstrated a critical role of CD4+CD25+ Treg cells in the tolerance induced by this regimen113–115. Interestingly, CD154 plays a direct role in dampening the suppressive function of CD4+CD25+ Treg cells. Blocking the effects of CD154 on CD4+CD25+ Treg cells by anti-CD154 or genetic deletion of CD154 enhanced the suppressive function of CD4+CD25+ Treg cells to an adequate level such that there was no additional need to block CD154 on effector CD25− T cells in order to establish graft protection116. More recently, it has been shown that in this regimen plasmacytoid DCs carrying alloantigens were preferentially directed towards the lymph nodes, rather than the spleen, where they efficiently induced the development of alloAg-specific Foxp3+CD4+CD25+ Treg cells from Foxp3−CD4+CD25− T cells65.
In recent years, direct manipulation of DCs has been exploited for generation of “tolerogenic DCs” for induction of peripheral tolerance to alloAg (reviewed in ref.67). The methods for DC manipulation can be categorized into: 1) in vitro generation of tolerogenic DCs from bone marrow-derived precursors; and 2) in vivo targeting of steady-state DCs. The principle of DC manipulation is to endow DCs with the ability to resist maturation and to subsequently deliver negative signals to the interacting T cells, down-regulate effector T cell responses and possibly expand Treg cells.
Among in vitro-generated DCs, BM-DCs cultured in the presence of rapamycin103, LF15-0195 (an analogue of deoxyspergualin)104, and vitamin D3 metabolites117, have all been shown to prolong allogeneic graft survival upon transfer in vivo, and do so at least in part through expansion and/or induction of CD4+CD25+Foxp3+ Treg cells. These agents have also been shown to inhibit DC maturation in vivo when directly administered to the host. Among in vivo existing DC subsets, both CD8α+ DCs and B220+ DCs have been implicated in tolerogenicity of specific tolerance protocols; it therefore appears that tolerogenicity is not restricted to any particular DC subsets, rather all DCs have the plasticity to promote either immunity or tolerance under the appropriate signal combinations.
Direct targeting of recipient DCs in vivo represents a more recent advance in DC-based tolerance therapy. The principal of in vivo DC targeting is to deliver the necessary pool of alloantigens to host DCs in such a way that the DCs remain un-activated while maintaining their ability to induce T cell tolerance. One such method is to deliver donor cells that are undergoing apoptosis118 or to deliver donor APC-derived exosomes119, both of which provide the entire spectrum of donor antigens to recipient DCs without altering the quiescent status of the recipient DCs. It is interesting to note that phagocytes of different classes, including macrophages, lymphoid DCs and myeloid DCs, may all be involved in recognizing and uptaking apoptotic cells/bodies, but may do so in a temporarily staggered fashion, and may subsequently influence each other. For example, macrophages engulfing apoptotic cells may direct subsequent apoptotic cells to be up-taken by the more toleragenic 8α+ DCs, rather than 8α− DCs120. In addition, 8α− DCs may acquire 8α positivity upon taking up apoptotic cells121. Interactions between apoptotic cells with host phagocytes may also induce robust generation of Treg cells, likely via a TGF-β-dependent mechanism111. Such in vivo DC targeting strategies in combination with transient immunosuppression have been shown to induce indefinite allograft survival118,119.
We have developed a novel and robust tolerance regimen that utilizes infusions of donor cells pre-treated with a chemical cross-linker called 1-ethyl-3-(3’-dimethylaminopropyl)-carbodiimide, or ECDI, without further immunosuppressive therapy, for tolerance induction in allogeneic islet cell transplantation. Previous studies have shown that i.v. injection of antigen-pulsed splenic APCs chemically fixed with ECDI was a powerful and safe method for inducing antigen-specific T cell tolerance in vivo122,123. Specifically, myelin peptide-coupled, ECDI-fixed syngeneic APCs could effectively ablate induction and progression of experimental autoimmune encephalomyelitis (EAE), a murine Th1/17-mediated model of multiple sclerosis (MS)124. This protocol is also effective in preventing and treating autoimmune diabetes in the non-obese diabetic (NOD) mice125.
We found that i.v. infusions of ECDI-treated donor splenocytes induced indefinite donor-specific tolerance in allogeneic islet cell transplantation. Here, the antigens of interest are mainly donor MHC class I and II molecules that are an integral surface component of donor lymphocytes, and ECDI treatment presumably interferes with costimulatory signals, thereby inducing tolerance to the membrane-bound allogeneic MHC antigens126. Graft protection is donor-specific and is associated with markedly diminished donor-specific allo-responses measured by delayed type hypersensitivity, MLRs, IFN-γ production, as well as production of anti-donor antibodies, thus suggesting effective down-regulation of both donor-specific T and B cell responses127.
Interestingly, treatment with the ECDI-fixed donor splenocytes induced a significant increase in the number of CD4+CD25+Foxp3+ Treg cells in the spleen of tolerized mice vs. controls. The importance of the CD4+CD25+ Treg cells in this protocol was further supported by the fact that depleted/inactivated Treg cells by anti-CD25 mAb treatment around the time of initial tolerance induction completely blocked tolerance. Conversely, anti-CD25 treatment in long-term tolerized recipients (120 days post-transplant) did not break the established tolerance. These results indicate that CD4+CD25+ Treg cells are critical for the initial establishment of donor-specific tolerance, however once tolerance has been established, active regulation is less important. Thus, other mechanisms, such as anergy, may be required for long-term maintenance of donor-specific unresponsiveness in this model. The nature of the increased CD4+CD25+Foxp3+ population (are they expanded nTreg cells versus de novo induced iTreg cells?) is currently under intense investigation.
The critical difference between this tolerance protocol and others using infusions of donor cells, e.g., DST, is that efficient tolerance induction is achieved in the complete absence of immunosuppression, including transient cell-depletion, antibody-mediated blockade of costimulatory signals, and peri-transplant application of immunosuppressive drugs. The exact mechanism by which ECDI-treated cells induce donor-specific tolerance is not completely understood. Cell tracking indicates that ECDI-treated cells distribute widely, but intact cells disappear within 48 hours. Therefore, while direct presentation may play a role, this mechanism is likely transient. In contrast, indirect presentation of alloAg by host regulatory APCs is likely the predominant tolerance mechanism. Recent data in the EAE model suggest that host plasmacytoid DCs are crucial in the tolerogenic cross-presentation in this protocol128, and that the spleen is likely where the tolerogenic interactions occur.
Considerable insight has been gained into the biology of Treg cells and their role in mediating transplant tolerance. Both ex vivo expansion/induction of alloAg-specific Treg cells in tissue cultures and in vivo expansion/induction of alloAg-specific Treg cells via tolerizing regimens have shown promises in promoting durable allo-specific tolerance. It is possible that a favorable Treg:Teff ratio set forth by conditioning tolerance regimens through targeting of multiple mechanisms could be further enhanced by infusion of ex vivo-expanded or induced Treg cells. Therefore, combination cell-based therapies could be theoretically advantageous. Linked suppression may further allow effective in vivo suppression of allo-responses in the absence of specificities toward the entire spectrum of alloAg. A major challenge in this area of transplant immunobiology is translating the accumulated knowledge into human clinical application, with important limitations arising from heterogeneous immunity of the hosts, the need for reliable tools for measuring host tolerance, and regulatory issues.
This work was supported in part by National Institutes of Health (NIH) Career Award 1K08DK070029-01 and Type 1 Diabetes Pathfinder Award DP2 DK083099-01 to X.L.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.