|Home | About | Journals | Submit | Contact Us | Français|
Regulatory T cells (Tregs), in particular CD4+ Foxp3+ T cells, have been shown to play an important role in the maintenance of tolerance after allogeneic stem cell transplantation. In the current study, we have identified a population of CD8+ Foxp3+ T cells that are induced early during GVHD, constitute a significant percentage of the entire Treg population, and are present in all major GVHD target organs. These cells expressed many of the same cell surface molecules as found on CD4+ Tregs and potently suppressed in vitro alloreactive T cell responses. Induction of these cells correlated positively with the degree of MHC disparity between donor and recipient and was significantly greater than that observed for CD4+ induced Tregs (iTregs) in nearly all tissue sites. Mice that lacked the ability to make both CD8+ and CD4+ iTregs had accelerated GVHD mortality compared to animals that were competent to make both iTreg populations. The absence of both iTreg populations was associated with significantly greater expansion of activated donor T cells and increased numbers of CD4+ and CD8+ T cells that secreted IFN-γ and IL-17. The presence of CD8+ iTregs, however, was sufficient to prevent increased GVHD mortality in the complete absence of CD4+ Tregs, indicating at least one functional iTreg population was sufficient to prevent an exacerbation in GVHD severity, and that CD8+ iTregs could compensate for CD4+ iTregs. These studies define a novel population of CD8+ Tregs that play a role in mitigating the severity of GVHD after allogeneic stem cell transplantation.
Graft versus host disease (GVHD) is the major complication associated with allogeneic stem cell transplantation and is attributable, in large part, to an imbalance between the effector and regulatory arms of the immune system (1). A preponderance of evidence in experimental murine models and humans indicates that there is a progressive loss of regulatory T cells (Tregs) during GVHD (2–5). This decline in Treg numbers unleashes cytotoxic T cells and proinflammatory cytokine pathways that subsequently mediate pathological damage. Conversely, the adoptive transfer of Tregs at the time of transplantation can enhance overall survival and abrogate GVHD lethality (6–10), providing confirmation that these cells play a central role in the maintenance of transplantation tolerance. The most well characterized population of Tregs in GVHD biology has been CD4+ T cells which express the forkhead box P3 (Foxp3) transcription factor (11). This population is comprised of two major subsets which have been termed natural (nTregs) and induced (iTregs), based on the unique ontological and developmental characteristics that are specific for each cell population (12). The majority of experimental murine BMT studies have focused on the role of nTregs, whereas the contribution of iTregs to the prevention of GVHD lethality is still largely unclear. CD4+ iTregs that are in vivo-derived have been identified in GVHD recipients (13,14), but their ability to mitigate GVH reactivity has not been critically examined. Analysis of this population has also been confounded by the presence of nTregs in most experimental models of GVHD which has limited the ability to isolate the effects of these cells. Studies in other inflammatory disease models, however, have provided strong evidence that these two populations have nonredundant, complementary roles in maintaining immunological tolerance (15,16), indicating that Tregs are not a monolithic population, but constitute a heterogeneous population of cells with differing specificities and functions.
The premise that Tregs constitute a heterogeneous population has been bolstered by the identification of a population of CD8+ Foxp3+ T cells in autoimmune disorders and after allergen exposure (17–20). These cells which express many of the cell surface molecules such as GITR, CD103, and CTLA-4 commonly found on classical CD4+ Tregs have also been shown to suppress immune responses in vitro (21). The potential importance of this cell population is highlighted by their more recent identification in humans who received stem cell transplants for autoimmune disorders and diabetes where they were found to correlate in an inverse manner with the level of ongoing inflammation (22,23). Furthermore, these cells have been detected in tumor-bearing animals along with biopsies from patients with cancer where they have been implicated in suppressing the host immune response against the underlying malignancy (24,25). Whether these cells are present or have any functional role in allogeneic stem cell transplantation or, more specifically, GVHD biology is not known. In the current study, we demonstrate that CD8+ Foxp3+ Tregs are induced early during the course of GVHD and constitute a significant percentage of the entire Treg population. Moreover, these cells play a role in preventing GVHD-mediated lethality and are able to complement CD4+ iTregs, establishing them as a novel regulatory T cell population in GVHD biology.
C57BL/6 (B6) (H-2b), Balb/c (H-2d), FVB/N (H-2q), B6.SJL (CD45.1), B6.PL (Thy1.1+), B6.129S7-Rag-1 (B6 Rag-1), C.B10-H2b/LilMcd (Balb.B) (H-2b) and B6.C-H2bm1/Byj (bm1) (H-2Kbm1) mice were bred in the Animal Resource Center (ARC) at the Medical College of Wisconsin (MCW) or purchased from Jackson Laboratories (Bar Harbor, ME). Foxp3EGFP mice and Foxp3ΔEGFP in which there is mutation in the Foxp3 coding region which renders the Foxp3 gene nonfunctional were bred at MCW and have been previously described (26). The latter mice were reconstituted with 40–60 × 106 spleen cells from B6.SJL Foxp3EGFP (CD45.1) animals 1–2 days after birth to prevent the development of lethal autoimmunity. All animals were housed in the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC)-accredited Biomedical Resource Center of the Medical College of Wisconsin. Experiments were all carried out under protocols approved by the MCW Institutional Animal Care and Use Committee. Mice received regular mouse chow and acidified tap water ad libitum.
Bone marrow (BM) was flushed from donor femurs and tibias with Dulbecco’s modified media (DMEM) (Gibco-BRL, Carlsbad, CA) and passed through sterile mesh filters to obtain single cell suspensions. Host mice were conditioned with total body irradiation (TBI) administered as a single exposure at a dose rate of 900–1100 cGy using a Shepherd Mark I Cesium Irradiator (J.L. Shepherd and Associates, San Fernando, CA). Irradiated recipients received a single intravenous injection in the lateral tail vein of BM with or without added spleen or purified T cells. Mice were weighed two-three times per week and were euthanized when they attained pre-defined morbidity criteria. In experiments which utilized Foxp3ΔEGFP mice as donors, CD45.2+ CD4+ EGFP− or CD45.2+ CD8+ EGFP− T cells were sorted from the spleens of reconstituted animals prior to transplantation.
Spleen cells were collected from Foxp3EGFP and/or Foxp3ΔEGFP mice and sorted on a FACSAria II (Becton-Dickenson, Mountain View, CA). Sort purity was typically 98–99%. Spleen, liver, lung and colon cells from transplant recipients were labeled with monoclonal antibodies (mAb) conjugated to fluorescein isothiocyanate (FITC), phycoerythrin (PE), BD Horizon 450 (V450), BD Horizon 500 (V500), PE-Cy7, PE-Texas Red or allophycocyanin (APC) that were obtained from BD Biosciences (San Diego, CA), or Brilliant Violet 421 from Biolegend (San Diego, CA). Cells were analyzed on a FACSCalibur or LSRII flow cytometer with Cellquest or FACSDiva software (Becton-Dickenson). Data were analyzed using FlowJo software (Treestar, Ashland, Oregon).
To isolate lamina propria lymphocytes, pooled colons were incubated in Hank’s Balanced Salt Solution (HBSS) buffer (Gibco-BRL) supplemented with 2% fetal bovine serum (FBS), EDTA (0.05 mM) and 15 µg/ml dithiothreitol (Invitrogen, Carlsbad, CA) at 37°C for 30 min and subsequently digested in a solution of collagenase D (Roche Diagnostics, Mannheim, Germany, 1 mg/ml) in DMEM with 2% FBS for 75 min at 37°C. The resulting cell suspension was then layered on a 44%/67% Percoll gradient (Sigma, St. Louis, MO). Liver and lung lymphocytes were isolated by collagenase D digestion followed by layering on a Percoll gradient.
Thy1.2+ cells (1 × 105) were purified from B6 spleens using the magnetic activated cell separation (MACS) system (Miltenyi Biotech, Auburn, CA) and co-cultured with 5 × 104 Balb/c dendritic cell-enriched stimulator cells in U-bottomed microwell plates (Becton Dickinson, Lincoln Park, NJ) at 37°C. Stimulator cells were obtained by digestion of spleens with collagenase D (1 mg/ml) followed by positive selection of CD11c+ cells using the MACS system. Nonirradiated stimulator cells were then seeded into microwell plates. Flow-sorted CD4+ or CD8+ Treg cells, or activated CD8+EGFP− T cells were added at varying ratios to wells containing T and dendritic cells. One microcurie of 3H-thymidine was added to triplicate wells for the final 12–18 hours prior to harvest. Thymidine incorporation was assessed using a Wallac 1450 Microbeta liquid scintillation counter (Perkin Elmer, Shelton, CT). Control wells consisted of responders, stimulators, and Tregs alone.
Lymphocytes isolated from spleen, liver, lung and colon were stimulated with 50 ng/ml phorbol 12-myristate 13-acetate (PMA) (Sigma) and 750 ng/ml ionomycin (Calbiochem, La Jolla, CA) for 1 hour and then incubated with GolgiStop (BD Pharmingen) for an additional 4 hours. Cells were surface stained with Brilliant Violet 421 anti-CD3, APC anti-CD8, and PE-Texas Red anti-CD4 and then intracellularly stained with PE-labeled antibody to IL-17 and Alexa Fluor 700-labeled antibody to IFN-γ. All antibodies were obtained from Becton Dickenson with the exception of Brilliant Violet 421 which was obtained from Biolegend.
Group comparisons of Tregs and T cell populations, intracellular cytokine staining, and thymidine incorporation were performed using the Mann Whitney U test. Survival curves were constructed using the Kaplan-Meier product limit estimator and compared using the Gehan Wilcoxon test. A p value ≤ 0.05 was deemed to be significant in all experiments.
During the course of studies designed to define the functional role of various CD4+ Treg populations in GVHD biology, we identified a population of CD8+ T cells that expressed Foxp3 and were present in recipients undergoing GVHD. To further examine this observation, lethally irradiated Balb/c mice were transplanted with MHC-incompatible marrow grafts from B6 Foxp3EGFP animals. A serial time course analysis revealed the emergence of CD8+ Foxp3+ T cells within five days post transplantation in GVHD target tissues (i.e. spleen, liver, colon and lung) (Figure 1A and Supplementary Figure 1). The percentage of these cells was no different from that of CD4+ Tregs for the first ten days post transplantation, with the exception of the spleen where an increased frequency of CD8+ Tregs was observed and the lung where the converse was present (Figures 1B). Overall absolute numbers of CD4+ and CD8+ Foxp3+ T cells, however, were similar in all tissue sites over this time frame (Figure 1C). This was noteworthy given that the CD4+ Treg pool that derives from the marrow graft is composed of both natural and induced Tregs, whereas the bulk of CD8+ Tregs (henceforth referred to CD8+ iTregs) are presumably induced due to the negligible expression of Foxp3 on CD8+ T cells in the spleen, thymus, and bone marrow (Supplementary Figure 2). Thereafter, the absolute number of both CD4+ and CD8+ Treg populations declined with time in most tissues, although the rate of decline was more precipitous for CD8+ iTregs. Consequently, the absolute number of CD8+ iTregs was significantly lower than CD4+ Tregs at days 14 and 21 post transplantation (Figure 1C). Notably, CD8+ iTregs were not detected beyond four weeks in any tissue site (data not shown), indicating that these Foxp3-expressing T cells had limited in vivo persistence in GVHD. CD8+ Tregs were induced in a different MHC-mismatched strain combination, indicating that the results were not strictly model-dependent (Supplementary Figure 3). The emergence of CD8+ iTregs in GVHD animals was not attributable to homeostatic expansion occurring in a post transplant lymphopenic environment, since mice transplanted with syngeneic marrow grafts did not have significant induction of these cells in any tissue sites (Figure 2).
nTregs constitute the majority of the Treg pool in the periphery, as iTregs have been estimated to comprise only 4–15% of total CD4+ Tregs in various studies (16,27). Therefore, we reasoned that the majority of CD4+ Tregs that were present in the spleen and tissue sites of GVHD recipients early after transplantation were likely nTregs given that approximately 10% of CD4+ T cells in the marrow graft inoculum expressed Foxp3. Since the total number of CD4+ and CD8+ Tregs were equivalent early post transplantation, we hypothesized that there might be disparate induction of CD8+ and CD4+ Tregs during GVHD. To directly address this issue, we performed experiments in which mice were transplanted with B6 Rag-1 BM plus CD4+ Foxp3EGFP− and CD8+ Foxp3EGFP− T cells so that only induced Tregs could reconstitute in these animals. Examination of Treg reconstitution in the spleen and GVHD target organs 5, 10 and 14 days post transplantation revealed an increased percentage of CD8+ iTregs in the spleen, liver and lung, whereas there was no difference in the colon (Figures 3A and 3B). We also observed an increased absolute number of CD8+ iTregs in the spleen and liver at all time points, and in the lung on day 5 (Figure 3C). The increase in CD8+ iTregs, relative to CD4+ iTregs, was not due to an overall increase in the total number of CD8+ T cells as there was no significant difference in the absolute number of CD4+ and CD8+ T cells at any of these time points (Supplementary Figure 4). These data indicated that the induction of CD8+ Tregs from the conventional T cell compartment exceeded that of CD4+ Tregs during the early stages of GVHD.
The induction of Foxp3 expression in conventional T cells is dependent upon cell activation (28–30). T cell activation is also a prominent component of GVHD due to the recognition of alloantigens expressed by host antigen-presenting cells (31–33). We therefore examined whether the degree of MHC disparity between donor and recipient was a factor which modulated the magnitude of CD8+ Treg induction. To address this question, we performed transplant studies using donor/recipient strain combinations in which there were differing degrees of MHC disparity. Specifically, in addition to an MHC-disparate model, we employed murine models in which mice were MHC-matched and differed only at multiple minor histocompatibility antigens (B6→Balb.B), or had an isolated MHC class I-mismatch (B6→bm1). The dose of T cells administered to recipients was sufficient to induce lethal GVHD in all strain combinations (data not shown). These results demonstrated that CD8+ iTregs were detectable in recipients of MHC-matched, minor antigen mismatched grafts (Figure 4A). The percent of total CD8+ T cells that expressed Foxp3 and the overall number of CD8+ iTregs were significantly less than observed in MHC-mismatched recipients at days 5 and 10 post transplantation (Figures 4B and 4C). Moreover, the percentage of CD8+ iTregs was much lower than CD4+ Tregs (Figure 4A), in contrast to what was observed in Balb/c recipients of B6 marrow grafts where percentages tended to be equivalent (Figure 1). We also observed that there was only a negligible CD8+ iTreg population in recipients of class I-mismatched grafts where a three amino acid difference distinguishes donor from host (Figures 4A–4C). To confirm that the absence of CD8+ Treg induction in bm1 recipients could not be ascribed to lack of CD4+ T cells in the inoculum, we repeated studies with the inclusion of mature donor CD4+ T cells in the graft. Under these conditions where CD4+ T cells do not contribute to an anti-host response, CD8+ iTregs were still not detectable (data not shown). Collectively, these results indicated that the induction of CD8+ Tregs during GVHD directly correlated with the degree of MHC disparity between donor and recipient.
The significant increase in CD8+ Treg numbers in GVHD recipients led us to examine whether these cells were functionally suppressive. To address this question, we first sorted CD8+ Foxp3EGFP+ T cells from the spleen and liver of GVHD mice 5 days post-BMT. Phenotypic characterization of in vivo-derived CD8+ CD3+ Tregs demonstrated that they expressed many of the same surface antigens expressed on CD4+ Tregs (e.g. GITR, CD44, CTLA-4, CD25) with the exception that there was higher expression of CD103 (Figure 5A). Notably, CD8+ Tregs induced during GVHD expressed the CD8 beta chain indicating that they were not CD8αα cells which have been shown to have suppressive properties in other murine model systems (34–36). We then tested whether these cells were able to inhibit alloreactive T cell proliferation in an in vitro suppression assay. These studies demonstrated that CD8+ Foxp3+ T cells derived from GVHD mice suppressed alloreactive T cell proliferation in a dose-dependent manner and that suppression was no different from that observed with sorted CD4+ Foxp3+ T cells derived from the same tissue sites (Figure 5B). CD8+ Foxp3EGFP− T cells that did not undergo conversion also suppressed T cell responses, but this was significantly less than CD8+ Foxp3+ cells at nearly all effector: Treg ratios.
To determine whether CD8+ iTregs had a suppressive role in preventing GVHD mortality in vivo, we employed donor mice that have a mutation in the Foxp3 coding region (Foxp3ΔEGFP mice) which renders the protein nonfunctional such that these cells cannot become suppressive Tregs in vivo (26). CD4+ and CD8+ Foxp3ΔEGFP− T cells (CD45.2+) were isolated from the spleens of these mice and employed in transplantation experiments to determine the relative role of CD4+ and CD8+ iTregs in GVHD. Of note, iTregs constitute the only Treg populations in these studies since there are no transferred nTregs in the grafts. Lethally irradiated Balb/c mice were transplanted with B6 Rag-1 BM alone or together with CD4+EGFP− and CD8+EGFP− T cells from Foxp3EGFP mice (EGFP), CD4+EGFP− T cells from Foxp3ΔEGFP and CD8+EGFP− T cells from Foxp3EGFP animals (CD4Δ), CD4+EGFP− from Foxp3EGFP mice and CD8+EGFP− T cells from Foxp3ΔEGFP mice (CD8Δ), or CD4+EGFP− and CD8+EGFP− T cells from Foxp3ΔEGFP (CD4/8Δ). We observed that there was no difference in overall survival between mice that were transplanted with CD4Δ or CD8Δ marrow grafts compared to animals transplanted with T cells that were fully competent to differentiate into both CD4+ and CD8+ iTregs (Figure 5C). Mice that were co-transplanted with both CD4+ Foxp3ΔEGFP− and CD8+ Foxp3ΔEGFP− T cells (CD4/8Δ grafts), however, had significantly worse survival (p=0.01 compared to GVHD control). These data indicated that, in the absence of nTregs, at least one functional iTreg population was necessary to prevent an acceleration of GVHD-associated mortality. Furthermore, CD8+ iTregs could compensate for the absence of CD4+ iTregs (i.e. there was no difference in survival between GVHD and CD4Δ groups, but worse survival in the CD4/8Δ group), since otherwise one would have expected survival in animals transplanted with both CD4Δ and CD4/8Δ marrow grafts to be equivalent, and inferior to GVHD control mice.
Given that we observed inferior survival only in transplant recipients that were incapable of making both CD4+ and CD8+ iTregs, we examined Treg reconstitution to determine whether the absence of either iTreg population during GVHD resulted in a compensatory increase in the corresponding regulatory T cells. Animals that lacked the ability to make CD8+ iTregs had an increased percentage of CD4+ iTregs in the spleen and lung compared to GVHD control animals (EGFP) (Figure 6A). Also, mice that were unable to reconstitute CD4+ iTregs had increased frequencies of CD8+ iTregs in the spleen, liver, and lung. No differences were observed in the colon for either iTreg population. To determine whether increased frequencies resulted in an absolute increase in iTregs, we also quantified absolute numbers of these cells in all tissue sites. We observed that mice that had the inability to make CD8+ iTregs had a significantly greater absolute number of CD4+ iTregs than animals which were functionally competent to make both Treg populations in the spleen and lung (Figure 6B). Similarly, the absence of CD4+ iTregs was accompanied by an increased number of CD8+ iTregs in spleen, liver, and lung. Thus, frequencies and absolute numbers were concordant in all tissue sites. We observed small percentages/numbers of CD4+ and CD8+ iTregs in mice transplanted with CD4/8Δ grafts which we attributed to minor contamination of the sort population with B6.SJL Foxp3EGFP cells that were used to rescue B6 Foxp3ΔEGFP mice at birth from lethal autoimmunity. Overall, these results indicated that there was a compensatory increase in iTregs in several GVHD target organs when mice were unable to reconstitute the corresponding iTreg population. This was a potential explanation for why the absence of a single iTreg population alone did not lead to an increase in GVHD mortality (Figure 5C).
To determine why mice that lacked both CD4+ and CD8+ iTregs had accelerated mortality whereas mice that were deficient in only one iTreg population did not, we examined what effect the absence of these populations had on donor T cell expansion and inflammatory cytokine production. To address this issue, animals were transplanted as in Figure 5C and then euthanized two weeks post transplantation for analysis. Examination of spleen and GVHD target organ tissues revealed that there were significantly greater absolute numbers of activated donor CD4+ CD44hi CD62Llo and CD8+ CD44hi CD62Llo T cells within the spleen, lung, and liver of mice that were functionally deficient in both iTreg populations compared to animals which were competent to reconstitute both populations (Figures 7A and 7B). On average, this amounted to a two to four-fold increase in these tissue sites. An absence of CD8+ iTregs alone had more modest effects, but still resulted in increased numbers of activated CD4+ T cells in the spleen and lung, and donor CD8+ T cells in the spleen. The absence of CD4+ iTregs had no deleterious effect on donor T cell expansion in any tissue site when compared to GVHD controls.
The absolute number of CD4+ and CD8+ T cells that secreted IFN-γ was significantly higher (~2–4-fold) in the spleen, liver, and lung in animals that were unable to reconstitute both iTreg populations (Figures 8A and 8B). These animals also had increased numbers of CD4+ IL-17+ T cells in the liver and lung, and CD8+ IL-17+ T cells in the lung (Figures 8C and 8D). The absence of CD8+ iTregs alone resulted in a significant increase in the number of CD4+ IFN-γ+ T cells in the spleen and lung, CD8+ IFN-γ+ T cells in the spleen, and CD4+ IL-17+ T cells in the lung. Conversely, we observed an increase only in CD4+ IFN-γ+ T cells in the lung of mice that were functionally incapable of making CD4+ iTregs. Collectively, these data demonstrated that the absence of both CD4+ and CD8+ iTregs resulted in increased expansion of activated donor T cells, as well as a significant increase in the absolute number of CD4+ and CD8+ T cells that were capable of secreting the proinflammatory cytokines, IFN-γ and IL-17. The absence of CD8+ iTregs alone also had more pronounced effects on T cell expansion and cytokine production than the isolated absence of CD4+ iTregs.
Understanding the role of specific Treg populations for the prevention of GVHD is of increasing importance given that regulatory T cell therapy is now in the early stages of implementation in bone marrow transplant recipients. In initial studies, unselected CD4+ CD25+ T cells have been administered to allogeneic BMT recipients with the ultimate goal that this approach will attenuate GVHD severity (37,38). Thus far, while cells have been able to be administered safely it is not clear how potent they are in abrogating GVHD severity. Moreover, whether this is the optimal Treg population and to what extent various Treg subpopulations cooperate to maintain immune tolerance is undefined (39). It is against this backdrop that we examined whether other Treg populations might be important in the regulation of GVHD. The studies described herein now define CD8+ Foxp3+ T cells as another regulatory T cell population that plays a role in the maintenance of transplantation tolerance in allogeneic stem cell transplant recipients.
CD8+ Tregs were induced early during GVHD and constituted a sizable percentage of the total Treg pool within the first ten days post transplantation. This was surprising given that generation of these cells derives from the conventional T cell compartment since, in contrast to CD4+ Tregs, constitutive Foxp3 expression on CD8+ T cells in the thymus and in the peripheral T cell compartment of healthy mice is negligible (Supplementary Figure 2). In fact, the absolute number of CD8+ iTregs was significantly greater than CD4+ iTregs when the contributions of both were comparatively analyzed in the absence of CD4+ nTregs. Induction was not due to homeostatic expansion occurring in a lymphopenic environment, but rather was the consequence of alloreactivity (Figure 2). We observed that there was a correlation between the degree of MHC disparity between donor and host, and the magnitude of CD8+ Treg induction. A component of this was attributable to the increased percentage of total CD8+ T cells early post transplantation in recipients of completely MHC-mismatched grafts (i.e. B6→Balb/c). However, even after correcting for that factor, there was still a significantly higher percentage of CD8+ Tregs in most GVHD target tissues when compared to animals that were reconstituted with MHC-matched, minor antigen mismatched, or class I-mismatched grafts (Figure 4). The explanation for this observation is not entirely clear, but may be due, in part, to a higher precursor frequency for alloantigens in the setting of a complete MHC-mismatch since Foxp3 expression requires activation to be induced (20). Alternatively, it is possible that the strength of the T cell activation signal is more robust in this setting which serves to enhance CD8+ Treg induction. It should be noted, however, that there have been a number of reports in which CD4+ iTreg generation unexpectedly did not occur despite strong immunological stimuli (11, 40–42). Consequently, the conditions that lead to the generation of iTregs in specific contexts are still incompletely understood.
CD8+ iTregs were found to be equally suppressive to CD4+ Tregs when tested in vitro and expressed many of the same cell surface markers that defined the CD4+ Treg compartment. To determine whether these cells had functional activity to suppress GVHD in vivo, we employed a murine BMT model in which the only Treg populations that can be generated are iTregs, as there were no transferred nTregs in the marrow graft. This allowed us to make valid comparisons between CD4+ and CD8+ iTregs. Under these conditions, we observed that mice that lacked both iTreg populations had accelerated mortality that was attributable to GVHD, whereas animals that could make either iTreg population had survival commensurate with that of GVHD control mice. Hence, at least one functional iTreg population was required to mitigate GVHD severity. CD8+ iTregs could also compensate for the complete absence of CD4+ iTregs, as mice transplanted with grafts deficient in CD4+ iTregs had survival that was comparable to that of animals which could reconstitute both iTreg populations. This was evidence that CD8+ iTregs had suppressive capability in vivo. This was further substantiated by the fact that mice that were selectively deficient in CD8+ iTregs had significant increases in donor T cell expansion and proinflammatory cytokine production in some GVHD target organs (Figures 7 and and8).8). For the most part, enhanced immune reactivity in the absence of CD8+ iTregs was observed in the spleen and the lung, raising the question as to whether these cells may have an important and preferential regulatory role within these tissue sites.
These results indicate that there appear to be compensatory regulatory mechanisms that are operative during GVHD. When CD8+ iTregs were absent in recipient animals, there was a corresponding increase in the absolute number of CD4+ iTregs in nearly all tissue sites (Figure 6). We observed similar findings when CD4+ iTregs were unable to be generated, although not to the same extent as for CD8+ iTregs. We would speculate therefore that these reciprocal increases are an explanation for the fact that only in the absence of both iTreg populations was there an increase in GVHD lethality. The question then arises as to whether CD4+ and CD8+ iTregs have redundant or unique roles in the maintenance of transplantation tolerance. Prior studies have demonstrated that CD4+ Tregs constitute a heterogeneous population which can be distinguished, in part by phenotypic differences which alter migration and function (43–45). Thus, it is not out of the question to postulate that CD8+ iTregs may also have a functional niche that distinguishes them from CD4+ Tregs. We observed that the absolute numbers, temporal appearance, and tissue localization differed between CD4+ and CD8+ iTregs which is circumstantial evidence that they may have divergent roles in suppressing the GVH response. In this regard, it will be of interest to determine whether CD8+ iTregs employ similar mechanistic pathways as CD4+ Tregs to suppress GVHD. The latter cells have been shown to suppress immune responses through a variety of mechanisms which include the production of immune suppressive cytokines such as IL-10, inhibition of dendritic cell function, modulation of effector T cells, and secretion of cytotoxic molecules such as granzymes (46). Consequently, there may not be only one mechanistic pathway that is employed by CD8+ iTregs.
A suppressive role for CD8+ iTregs was noteworthy given that these cells were not detected in mice beyond four weeks post transplantation. Why CD8+ iTregs appear to have limited persistence in vivo during GVHD is not clear. Since these cells were continuously exposed to recipient alloantigens, the absence of a persistent antigenic stimulus to enforce CD8+ Treg induction does not appear to be a plausible explanation for this observation. One possible explanation is that the proinflammatory milieu may adversely impact the continued induction and survival of CD8+ iTregs. In previous studies (13), we have shown that CD4+ iTreg reconstitution is modest during GVHD in comparison to nTreg regeneration. Similar results have been reported by Bucher and colleagues (14) who also noted a very low frequency of CD4+ iTregs in the spleen and colon of GVHD animals in which nTregs were not administered in the graft. However, in vivo-derived CD4+ iTreg reconstitution can be enhanced by blockade of proinflammatory cytokines such as IL-6 and IL-21 (13,14) with a corresponding reduction in GVHD severity. IL-6 exposure has also been shown to result in methylation of the upstream Foxp3 enhancer and in repression of Foxp3 expression, both of which inhibit Treg development and function (47). Alternatively, the inflammatory milieu could alter the availability of TGF-β which is necessary for continued expression of Foxp3 in CD4+ Tregs (48). In that regard, stability of Foxp3 expression is necessary for functional activity and loss of expression has been shown to be one explanation for why in vitro-differentiated CD4+ iTregs have reduced in vivo survival (49). Thus, instability of Foxp3 expression could be another potential explanation for why CD8+ iTregs are not detectable beyond four weeks. Finally, it should be noted that these explanations are not mutually exclusive and could all be operative to some extent in limiting the persistence of these cells in GVHD recipients.
In summary, these studies define CD8+ Foxp3+ Tregs as a novel population of regulatory T cells that emerge during the course of GVHD and serve to mitigate disease severity after allogeneic stem cell transplantation. These results also serve to highlight the heterogeneity that exists within the Treg population in GVHD biology. Understanding the complexity that is resident within these CD4+ and CD8+ Treg populations will be critical to the successful application of regulatory T cell therapy into the clinic.
We thank Hope Campbell and Kyle Upchurch in the Flow Cytometry Core for assistance with the sorting experiments.
This research was supported by grants from the National Institutes of Health (HL64603, HL081650, and DK083358) and by awards from the Midwest Athletes against Childhood Cancer (MACC) Fund.
Author Contributions: A.B. designed and performed research, analyzed data, and wrote the manuscript. D.H., A.C., and P.G. performed research. C.B.W. helped design experiments and interpret data. W.R.D. designed experiments, analyzed data and wrote the manuscript. The authors declare no competing financial interests.
Financial Disclosure: The authors have no financial disclosures.