Graft-versus-host disease is locally maintained in target tissues by resident progenitor-like T cells.

In allogeneic hematopoietic stem cell transplantation, donor αß T cells attack recipient tissues, causing graft-versus-host disease (GVHD), a major cause of morbidity and mortality. A central question has been how GVHD is sustained despite T cell exhaustion from chronic antigen stimulation. The current model for GVHD holds that disease is maintained through the continued recruitment of alloreactive effectors from blood into affected tissues. Here, we show, using multiple approaches including parabiosis of mice with GVHD, that GVHD is instead primarily maintained locally within diseased tissues. By tracking 1,203 alloreactive T cell clones, we fitted a mathematical model predicting that within each tissue a small number of progenitor T cells maintain a larger effector pool. Consistent with this, we identified a tissue-resident TCF-1+ subpopulation that preferentially engrafted, expanded, and differentiated into effectors upon adoptive transfer. These results suggest that therapies targeting affected tissues and progenitor T cells within them would be effective.

Stromal cells control the epithelial residence of DCs and memory T cells by regulated activation of TGF-ß.

Cells of the immune system that reside in barrier epithelia provide a first line of defense against pathogens. Langerhans cells (LCs) and CD8(+) tissue-resident memory T cells (TRM cells) require active transforming growth factor-ß1 (TGF-ß) for epidermal residence. Here we found that integrins αvß6 and αvß8 were expressed in non-overlapping patterns by keratinocytes (KCs) and maintained the epidermal residence of LCs and TRM cells by activating latent TGF-ß. Similarly, the residence of dendritic cells and TRM cells in the small intestine epithelium also required αvß6. Treatment of the skin with ultraviolet irradiation decreased integrin expression on KCs and reduced the availability of active TGF-ß, which resulted in LC migration. Our data demonstrated that regulated activation of TGF-ß by stromal cells was able to directly control epithelial residence of cells of the immune system through a novel mechanism of intercellular communication.

Bim regulates the survival and suppressive capability of CD8+ FOXP3+ regulatory T cells during murine GVHD.

CD8+ Foxp3+ T cells (Tregs) are a potent regulatory population whose functional and ontological similarities to CD4+ Fox3+ T cells have not been well delineated. Using an experimental model of graft-versus-host disease (GVHD), we observed that CD8+ Tregs were significantly less potent than CD4+ Tregs for the suppression of GVHD. To define the mechanistic basis for this observation, we examined the T-cell repertoire and the transcriptional profile of in vivo-derived CD4+ and CD8+ Tregs that emerged early during this disease. Polyclonal and alloantigen-induced CD8+ Tregs had repertoire diversity that was similar to that of conventional CD8+ T cells, indicating that a restricted repertoire was not the proximate cause of decreased suppression. Transcriptional profiling revealed that CD8+ Tregs possessed a canonical Treg transcriptional signature that was similar to that observed in CD4+ Tregs, yet distinct from conventional CD8+ T cells. Pathway analysis, however, demonstrated that CD8+ Tregs had differential gene expression in pathways involved in cell death and survival. This was further confirmed by detailed mRNA sequence analysis and protein expression studies, which demonstrated that CD8+ Tregs had increased expression of Bim and reduced expression of Mcl-1. Transplantation with CD8+ Foxp3+ Bim-/- Tregs resulted in prolonged Treg survival and reduced GVHD lethality compared with wild-type CD8+ Tregs, providing functional confirmation that increased expression of Bim was responsible for reduced in vivo efficacy. Thus, Bim regulates the survival and suppressive capability of CD8+ Tregs, which may have implications for their use in regulatory T-cell therapy.

PD-L1 Prevents the Development of Autoimmune Heart Disease in Graft-versus-Host Disease.

Effector memory T cells (TEM) are less capable of inducing graft-versus-host disease (GVHD) compared with naive T cells (TN). Previously, in the TS1 TCR transgenic model of GVHD, wherein TS1 CD4 cells specific for a model minor histocompatibility Ag (miHA) induce GVHD in miHA-positive recipients, we found that cell-intrinsic properties of TS1 TEM reduced their GVHD potency relative to TS1 TN Posttransplant, TS1 TEM progeny expressed higher levels of PD-1 than did TS1 TN progeny, leading us to test the hypothesis that TEM induce less GVHD because of increased sensitivity to PD-ligands. In this study, we tested this hypothesis and found that indeed TS1 TEM induced more severe skin and liver GVHD in the absence of PD-ligands. However, lack of PD-ligands did not result in early weight loss and colon GVHD comparable to that induced by TS1 TN, indicating that additional pathways restrain alloreactive TEM TS1 TN also caused more severe GVHD without PD-ligands. The absence of PD-ligands on donor bone marrow was sufficient to augment GVHD caused by either TEM or TN, indicating that donor PD-ligand-expressing APCs critically regulate GVHD. In the absence of PD-ligands, both TS1 TEM and TN induced late-onset myocarditis. Surprisingly, this was an autoimmune manifestation, because its development required non-TS1 polyclonal CD8+ T cells. Myocarditis development also required donor bone marrow to be PD-ligand deficient, demonstrating the importance of donor APC regulatory function. In summary, PD-ligands suppress both miHA-directed GVHD and the development of alloimmunity-induced autoimmunity after allogeneic hematopoietic transplantation.

Graft-versus-host disease.

Allogeneic haematopoietic stem-cell transplantation (SCT) is a curative therapy for haematological malignancies and inherited disorders of blood cells, such as sickle-cell anaemia. Mature alphabeta T cells that are contained in the allografts reconstitute T-cell immunity and can eradicate malignant cells in the recipient. Unfortunately, these T cells recognize the recipient as 'non-self' and employ a wide range of immune mechanisms to attack recipient tissues in a process known as graft-versus-host disease (GVHD). The full therapeutic potential of allogeneic haematopoietic SCT will not be realized until approaches to minimize GVHD, while maintaining the positive contributions of donor T cells, are developed. This Review focuses on research in mouse models pursued to achieve this goal.

Engineering human peripheral blood stem cell grafts that are depleted of naïve T cells and retain functional pathogen-specific memory T cells.

Graft-versus-host disease (GVHD) is a frequent major complication of allogeneic hematopoietic cell transplantation (HCT). Approaches that selectively deplete T cells that cause GVHD from allogeneic stem cell grafts and preserve T cells specific for pathogens may improve HCT outcomes. It has been hypothesized that the majority of T cells that can cause GVHD reside within the naïve T cell (TN) subset, and previous studies performed in mouse models and with human cells in vitro support this hypothesis. As a prelude to translating these findings to the clinic, we developed and evaluated a novel 2-step clinically compliant procedure for manipulating peripheral blood stem cells (PBSC) to remove TN, preserve CD34(+) hematopoietic stem cells, and provide for a fixed dose of memory T cells (TM) that includes T cells with specificity for common opportunistic pathogens encountered after HCT. Our studies demonstrate effective and reproducible performance of the immunomagnetic cell selection procedure for depleting TN. Moreover, after cell processing, the CD45RA-depleted PBSC products are enriched for CD4(+) and CD8(+) TM with a central memory phenotype and contain TM cells that are capable of proliferating and producing effector cytokines in response to opportunistic pathogens.

Profound depletion of host conventional dendritic cells, plasmacytoid dendritic cells, and B cells does not prevent graft-versus-host disease induction.

The efficacy of allogeneic hematopoietic stem cell transplantation is limited by graft-versus-host disease (GVHD). Host hematopoietic APCs are important initiators of GVHD, making them logical targets for GVHD prevention. Conventional dendritic cells (DCs) are key APCs for T cell responses in other models of T cell immunity, and they are sufficient for GVHD induction. However, we report in this article that in two polyclonal GVHD models in which host hematopoietic APCs are essential, GVHD was not decreased when recipient conventional DCs were inducibly or constitutively deleted. Additional profound depletion of plasmacytoid DCs and B cells, with or without partial depletion of CD11b(+) cells, also did not ameliorate GVHD. These data indicate that, in contrast with pathogen models, there is a surprising redundancy as to which host cells can initiate GVHD. Alternatively, very low numbers of targeted APCs were sufficient. We hypothesize the difference in APC requirements in pathogen and GVHD models relates to the availability of target Ags. In antipathogen responses, specialized APCs are uniquely equipped to acquire and present exogenous Ags, whereas in GVHD, all host cells directly present alloantigens. These studies make it unlikely that reagent-based host APC depletion will prevent GVHD in the clinic.

Chemokine receptors are required for effector T cell trafficking to GVHD tissues but not to bone marrow.

In allogeneic hematopoietic stem cell transplantation (alloSCT), alloreactive donor T cells mediate the graft-versus-leukemia (GVL) effect but also attack nonhematopoietic tissues causing graft-versus-host disease (GVHD). Reducing alloreactive T cell trafficking to GVHD target tissues while allowing their access to bone marrow (BM) and spleen, major sites of malignant hematopoiesis, is a rational strategy for reducing the GVHD risk when using alloreactive T cells as a therapeutic. Here we show that effector T cell (Teff) entry into BM and spleen in unmanipulated mice and in mice transplanted without alloreactive T cells is augmented by pertussis toxin-sensitive (PTX) chemokine receptor signaling. However, unexpectedly, in the presence of a GVH response, chemokines no longer draw T cells into BM and spleen but remain critical for their recruitment to GVHD target tissues. Consistent with this, PTX-treated Teff cells were as efficacious as untreated T cells in killing leukemia cells in BM and spleen in mice with a concurrent GVHD response. These results suggest a strategy to improve the safety of alloreactive T cell therapeutics in treating leukemias in the context of an alloSCT.

Mechanisms of antigen presentation to T cells in murine graft-versus-host disease: cross-presentation and the appearance of cross-presentation.

Recipient antigen-presenting cells (APCs) initiate GVHD by directly presenting host minor histocompatibility antigens (miHAs) to donor CD8 cells. However, later after transplantation, host APCs are replaced by donor APCs, and if pathogenic CD8 cells continue to require APC stimulation, then donor APCs must cross-present host miHAs. Consistent with this, CD8-mediated GVHD is reduced when donor APCs are MHC class I(-). To study cross-presentation, we used hosts that express defined MHC class I K(b)-restricted miHAs, crossed to K(b)-deficient backgrounds, such that these antigens cannot be directly presented. Cross-priming was surprisingly efficient, whether antigen was restricted to the hematopoietic or nonhematopoietic compartments. Cross-primed CD8 cells were cytolytic and produced IFN-γ. CD8 cells were exclusively primed by donor CD11c(+) cells, and optimal cross-priming required that they are stimulated by both type I IFNs and CD40L. In studying which donor APCs acquire host miHAs, we made the surprising discovery that there was a large-scale transfer of transmembrane proteins from irradiated hosts, including MHC class I-peptide complexes, to donor cells, including dendritic cells. Donor dendritic cells that acquired host MHC class I-peptide complexes were potent stimulators of peptide-specific T cells. These studies identify new therapeutic targets for GVHD treatment and a novel mechanism whereby donor APCs prime host-reactive T cells.

Langerhans cells are not required for graft-versus-host disease.

Graft-versus-host disease (GVHD) is initiated and maintained by antigen-presenting cells (APCs) that prime alloreactive donor T cells. APCs are therefore attractive targets for GVHD prevention and treatment. APCs are diverse in phenotype and function, making understanding how APC subsets contribute to GVHD necessary for the development of APC-targeted therapies. Langerhans cells (LCs) have been shown to be sufficient to initiate skin GVHD in a major histocompatibility complex-mismatched model; however, their role when other host APC subsets are intact is unknown. To address this question, we used mice genetically engineered to be deficient in LCs by virtue of expression of diphtheria toxin A under the control of a BAC (bacterial artificial chromosome) transgenic hu-man Langerin locus. Neither CD8- nor CD4-mediated GVHD was diminished in recipients lacking LCs. Similarly, CD8- and CD4-mediated GVHD, including that in the skin, was unaffected if bone marrow came from donors that could not generate LCs, even though donor LCs engrafted in control mice. Engraftment of donor LCs after irradiation in wild-type hosts required donor T cells, with immunofluorescence revealing patches of donor and residual host LCs. Surprisingly, donor LC engraftment in Langerin-diphtheria toxin A (DTA) transgenic hosts was independent of donor T cells, suggesting that a Langerin(+) cell regulates repopulation of the LC compartment.

Memory T cells from minor histocompatibility antigen-vaccinated and virus-immune donors improve GVL and immune reconstitution.

Donor T cells contribute to the success of allogeneic hematopoietic stem cell transplantation (alloSCT). Alloreactive donor T cells attack leukemia cells, mediating the GVL effect. Donor T cells, including the memory T cells (T(M)) that are generated after infection, also promote immune reconstitution. Nonetheless, leukemia relapse and infection are major sources of treatment failure. Efforts to augment GVL and immune reconstitution have been limited by GVHD, the attack by donor T cells on host tissues. One approach to augmenting GVL has been to infuse ex vivo-generated T cells with defined specificities; however, this requires expertise that is not widely available. In the present study, we tested an alternative approach, adoptive immunotherapy with CD8+ T(M) from donors vaccinated against a single minor histocompatibility antigen (miHA) expressed by leukemia cells. Vaccination against the miHA H60 greatly augmented T(M)-mediated GVL against mouse chronic-phase (CP-CML) and blast crisis chronic myeloid leukemia (BC-CML). T(M)-mediated GVL was antigen specific and was optimal when H60 expression was hematopoietically restricted. Even when H60 was ubiquitous, donor H60 vaccination had a minimal impact on GVHD. T(M) from lymphocytic choriomeningitis virus (LCMV)-immune and H60-vaccinated donors augmented GVL and protected recipients from LCMV. These data establish a strategy for augmenting GVL and immune reconstitution without elaborate T-cell manipulation.

A repertoire-independent and cell-intrinsic defect in murine GVHD induction by effector memory T cells.

Effector memory T cells (T(EM)) do not cause graft-versus-host disease (GVHD), though why this is has not been elucidated. To compare the fates of alloreactive naive (T(N)) or memory (T(M)) T cells, we developed a model of GVHD in which donor T cells express a transgene-encoded TCR specific for an antigenic peptide that is ubiquitously expressed in the recipient. Small numbers of naive TCR transgenic (Tg) T cells induced a robust syndrome of GVHD in transplanted recipients. We then used an established method to convert TCR Tg cells to T(M) and tested these for GVHD induction. This allowed us to control for the potentially different frequencies of alloreactive T cells among T(N) and T(M), and to track fates of alloreactive T cells after transplantation. T(EM) caused minimal, transient GVHD whereas central memory T cells (T(CM)) caused potent GVHD. Surprisingly, T(EM) were not inert: they, engrafted, homed to target tissues, and proliferated extensively, but they produced less IFN-γ and their expansion in target tissues was limited at later time points, and local proliferation was reduced. Thus, cell-intrinsic properties independent of repertoire explain the impairment of T(EM), which can initiate but cannot sustain expansion and tissue damage.

Graft-versus-host disease is independent of innate signaling pathways triggered by pathogens in host hematopoietic cells.

Graft-versus-host disease (GVHD) is initiated by APCs that prime alloreactive donor T cells. In antipathogen responses, Ag-bearing APCs receive signals through pattern-recognition receptors, including TLRs, which induce the expression of costimulatory molecules and production of inflammatory cytokines, which in turn mold the adaptive T cell response. However, in allogeneic hematopoietic stem cell transplantation (alloSCT), there is no specific pathogen, alloantigen is ubiquitous, and signals that induce APC maturation are undefined. To investigate APC activation in GVHD, we used recipient mice with hematopoietic cells genetically deficient in pathways critical for APC maturation in models in which host APCs are absolutely required. Strikingly, CD8-mediated and CD4-mediated GVHD were similar whether host APCs were wild-type or deficient in MyD88, TRIF, or MyD88 and TRIF, which excludes essential roles for TLRs and IL-1ß, the key product of inflammasome activation. Th1 differentiation was if anything augmented when APCs were MyD88/TRIF(-/-), and T cell production of IFN-γ did not require host IL-12. GVHD was also intact when APCs lacked the type I IFNR, which amplifies APC activation pathways that induce type I IFNs. Thus in GVHD, alloreactive T cells can be activated when pathways critical for antipathogen T cell responses are impaired.

Regulatory T cells require mammalian target of rapamycin signaling to maintain both homeostasis and alloantigen-driven proliferation in lymphocyte-replete mice.

Rapamycin (Rapa), an immunosuppressive drug that acts through mammalian target of Rapa inhibition, broadly synergizes with tolerogenic agents in animal models of transplantation and autoimmunity. Rapa preferentially inhibits conventional CD4(+) Foxp3(-) T cells (Tconv) and promotes outgrowth of CD4(+)Foxp3(+) regulatory T cells (Treg) during in vitro expansion. Moreover, Rapa is widely perceived as augmenting both expansion and conversion of Treg in vivo. However, most quantitative studies were performed in lymphopenic hosts or in graft-versus-host disease models. We show in this study that in replete wild-type mice, Rapa significantly inhibits both homeostatic and alloantigen-induced proliferation of Treg, and promotes their apoptosis. Together, these lead to significant Treg depletion. Tconv undergo depletion to a similar degree, resulting in no change in the percent of Treg among CD4 cells. Moreover, in this setting, there was no evidence of conversion of Tconv into Treg. However, after withdrawal of Rapa, Treg recover Ag-induced proliferation more quickly than Tconv, leading to recovery to baseline numbers and an increase in the percent of Treg compared with Tconv. These findings suggest that the effects of Rapa on Treg survival, homeostasis, and induction, depend heavily on the cellular milieu and degree of activation. In vivo, the resistance of Treg to mammalian target of Rapa inhibition is relative and results from lymphopenic and graft-versus-host disease models cannot be directly extrapolated to settings more typical of solid organ transplantation or autoimmunity. Moreover, these results have important implications for the timing of Rapa therapy with tolerogenic agents designed to increase the number of Treg in vivo.

Graft-versus-leukemia (GVL) against mouse blast-crisis chronic myelogenous leukemia (BC-CML) and chronic-phase chronic myelogenous leukemia (CP-CML): shared mechanisms of T cell killing, but programmed death ligands render CP-CML and not BC-CML GVL resistant.

Graft-versus-leukemia (GVL) against chronic-phase chronic myelogenous leukemia (CP-CML) is potent, but it is less efficacious against acute leukemias and blast-crisis chronic myelogenous leukemia (BC-CML). The mechanisms underlying GVL resistance are unknown. Previously, we found that alloreactive T cell targeting of GVL-sensitive bcr-abl-induced mouse CP-CML (mCP-CML) required TCR-MHC interactions and that multiple and redundant killing mechanisms were in play. To better understand why BC-CML is resistant to GVL, we performed a comprehensive analysis of GVL against mouse BC-CML (mBC-CML) induced by the retroviral transfer of the bcr-abl and NUP98/HOXA9 fusion cDNAs. Like human BC-CML, mBC-CML was GVL resistant, and this was not due to accelerated kinetics or a greater leukemia burden. To study T cell recognition and killing mechanisms, we generated a panel of gene-deficient leukemias by transducing bone marrow from gene-deficient mice. T cell target recognition absolutely required that mBC-CML cells express MHC molecules. GVL against both mCP-CML and mBC-CML required leukemia expression of ICAM-1. We hypothesized that mBC-CML would be resistant to some of the killing mechanisms sufficient to eliminate mCP-CML, but we found instead that the same mechanisms were effective against both types of leukemia, because GVL was similar against wild-type or mBC-CML genetically lacking Fas, TRAIL-R, Fas/TRAIL-R, or TNFR1/R2 or when donor T cells were perforin(-/-). However, mCP-CML, but not mBC-CML, relied on expression of programmed death-1 ligands 1 and 2 (PD-L1/L2) to resist T cell killing, because only GVL against mCP-CML was augmented when leukemias lacked PD-L1/L2. Thus, mBC-CML cells have cell-intrinsic mechanisms, distinct from mCP-CML cells, which protect them from T cell killing.

Tissue-resident memory T cell maintenance during antigen persistence requires both cognate antigen and interleukin-15.

Our understanding of tissue-resident memory T (TRM) cell biology has been largely developed from acute infection models in which antigen is cleared and sterilizing immunity is achieved. Less is known about TRM cells in the context of chronic antigen persistence and inflammation. We investigated factors that underlie TRM maintenance in a kidney transplantation model in which TRM cells drive rejection. In contrast to acute infection, we found that TRM cells declined markedly in the absence of cognate antigen, antigen presentation, or antigen sensing by the T cells. Depletion of graft-infiltrating dendritic cells or interruption of antigen presentation after TRM cells were established was sufficient to disrupt TRM maintenance and reduce allograft pathology. Likewise, removal of IL-15 transpresentation or of the IL-15 receptor on T cells during TRM maintenance led to a decline in TRM cells, and IL-15 receptor blockade prevented chronic rejection. Therefore, antigen and IL-15 presented by dendritic cells play nonredundant key roles in CD8 TRM cell maintenance in settings of antigen persistence and inflammation. These findings provide insights that could lead to improved treatment of chronic transplant rejection and autoimmunity.

Enhancing alloreactivity does not restore GVHD induction but augments skin graft rejection by CD4⁺ effector memory T cells.

Graft-versus-host disease (GVHD) caused by donor T cells attacking recipient tissues is a major cause of morbidity and mortality following allogeneic hematopoietic stem cell transplantation (alloSCT). Studies have shown that effector memory T (T(EM) ) cells do not cause GVHD but are capable of immune functions post-transplant, including graft-versus-leukemia (GVL) effects, but the reasons for this are unclear. In mice, the T(EM) pool may have a less diverse T-cell receptor (TCR) repertoire than naive T (T(N) ) cells with fewer alloreactive clones. We therefore tested whether enhancing the alloreactivity of T(EM) cells would restore their ability to cause GVHD. In an MHC-matched system, alloreactive T(EM) cells were created by transferring GVHD effector cells into syngeneic recipients and allowing conversion to T(EM) cells. Upon retransfer to freshly transplanted recipients, these cells caused only mild GVHD. Similarly, in an MHC-mismatched system, T(EM) cells with a proven increased precursor frequency of alloreactive clones only caused limited GVHD. Nonetheless, these same cells mounted strong in vitro alloresponses and caused rapid skin graft rejection. T(EM) cells created from CD4(+) T cells that had undergone lymphopenia-induced proliferation (LIP) also caused only mild GVHD. Our findings establish that conversion to T(EM) cells significantly reduces GVHD potency, even in cells with a substantially enhanced alloreactive repertoire.

NK cells delay allograft rejection in lymphopenic hosts by downregulating the homeostatic proliferation of CD8+ T cells.

T cells present in lymphopenic environments undergo spontaneous (homeostatic) proliferation resulting in expansion of the memory T cell pool. Homeostatically generated memory T cells protect the host against infection but can cause autoimmunity and allograft rejection. Therefore, understanding the mechanisms that regulate homeostatic T cell proliferation is germane to clinical settings in which lymphodepletion is used. In this study, we asked whether NK cells, which regulate immune responses in lymphocyte-replete hosts, also regulate homeostatic T cell proliferation under lymphopenic conditions. We found that T cells transferred into genetically lymphocyte-deficient RAG-/- mice proliferate faster and generate more CD8+ memory T cells if NK cells were absent. CD8+ T cells that underwent homeostatic proliferation in the presence of NK cells generated mostly effector memory (CD44highCD62Llow) lymphocytes, whereas those that divided in the absence of NK cells were skewed toward central memory (CD44highCD62Lhigh). The latter originated predominantly from proliferation of the "natural" central memory CD8+ T cell pool. Regulation of homeostatic proliferation by NK cells occurred independent of perforin but was reversed by excess IL-15. Importantly, NK depletion enhanced CD8+ T cell recovery in T cell-depleted wild-type mice and accelerated rejection of skin allografts, indicating that regulation of homeostatic proliferation by NK cells is not restricted to genetically lymphocyte-deficient animals. These results demonstrate that NK cells downregulate homeostatic CD8+ T cell proliferation in lymphopenic environments by competing for IL-15. Concomitant NK and T cell depletion may be undesirable in transplant recipients because of enhanced expansion of memory CD8+ T cells that increase the risk of rejection.

A crucial role for host APCs in the induction of donor CD4+CD25+ regulatory T cell-mediated suppression of experimental graft-versus-host disease.

Allogeneic bone marrow transplantation is an effective treatment for a number of malignant and nonmalignant diseases (Applebaum. 2001. Nature. 411: 385-389 and Copelan. 2006. N Engl J Med. 354: 1813-1826). However, the application of this therapeutic modality has been impeded by a number of confounding side effects, the most frequent and severe of which is the development of graft-versus-host disease (GVHD) (Copelan. 2006. N Engl J Med. 354: 1813-1826 and Blazar and Murphy. 2005. Philos Trans R Soc Lond B Biol Sci. 360: 1747-1767). Alloreactive donor T cells are critical for causing GVHD (Fowler. 2006. Crit Rev Oncol Hematol. 57: 225-244 and Ferrara and Reddy. 2006. Semin Hematol. 43: 3-10), whereas recent data demonstrated a significant role for the naturally occurring thymic-derived donor CD4(+)CD25(+)Foxp3(+) regulatory T cells (Tregs) (Bluestone and Abbas. 2003. Nat Rev Immunol. 3: 253-257 and Shevach. 2006. Immunity. 25: 195-201) in suppressing experimental GVHD after bone marrow transplantation (Blazar and Taylor. 2005. Biol Blood Marrow Transpl. 11: 46-49 and Joffe and van Meerwijk. 2006. Semin Immunol. 18: 128-135) . Host APCs are required for induction of GVHD by the conventional donor T cells. However, it is not known whether they are also obligatory for donor Treg-mediated suppression of GVHD. Using multiple clinically relevant MHC-matched and -mismatched murine models of GVHD, we investigated the role of host APCs in the suppression of GVHD by donor Tregs. We found that alloantigen expression by the host APCs is necessary and sufficient for induction of GVHD protection by donor Tregs. This requirement was independent of their effect on the maintenance of Treg numbers and the production of IL-10 or IDO by the host APCs.