Chronic graft-versus-host disease (GVHD) is the most common complication in the late stage after allogenic hematopoietic-stem-cell-transplantation (SCT), but the pathophysiology and treatment strategy of chronic GVHD remain poorly defined. Prolonged administration of cyclosporine (CSA) did not decrease the risk of chronic GVHD. Recent studies using a mouse model have shown that regulatory T cells (Tregs) can influence immune responses, and Tregs in the grafts can prevent acute GVHD when injected together with donor T cells. However, it is not known whether Tregs remain in the grafts in the late stage of SCT and play a role in preventing chronic GVHD.First, we examined the origin of Tregs using a major histocompatibility complex (MHC) mismatched mouse SCT model. Lethally irradiated C3H/HeN(H-2k) recipient mice received 10x106 T-cell-depleted bone marrow (BM) cells from B6.Ly-5a(H-2b, CD45.1) mice and 1x106 spleen cells from C57BL/6(B6, H-2b, CD45.2) mice. Spleen cells were collected from SCT recipient mice at serial time points and subjected to fluorescence-activated cell sorting (FACS) analysis. Transplanted mice displayed complete donor hematopoietic chimerism and mild acute GVHD at day14. On day 21 (early stage) after SCT, host type Tregs (CD4+FoxP3+ H-2k) were no longer detectable, and most of the Tregs (83±3%) were derived from donor spleen Tregs (H-2b, CD45.2). However, the homeostatic expansion of spleen Tregs gradually contracted and newly arising donor BM-derived Tregs (H-2b CD45.1) became dominant (93.8±0.5%) in the late stage of SCT (day 120). As in the spleen, BM-derived Tregs reconstitution in the late stage was seen in the thymus and mesenteric lymph nodes. Moreover, in a minor MHC-mismatched SCT model (B6 into C3H.SW), Tregs in the late stage were derived from donor BM cells (97.0±0.2%). These BM-derived Tregs suppress alloreactivity in the same manner as naturally occurring Tregs isolated from naïve mice in the MLR.Next, we compared the effects of CSA and the mTOR inhibitor rapamycin (RAPA) on Tregs reconstitution. Mice receiving CSA or RAPA showed the same Tregs reconstitution pattern: in the early and late stages, Tregs were derived from donor spleen and BM cells, respectively. However, the number of Tregs in the spleen was reduced significantly in mice receiving CSA, as compared to control mice receiving phosphate-buffered saline (PBS; 1.3±0.2x106 vs. 2.4±0.6x106) at day 110. In particular, the number of Tregs in the thymus was reduced dramatically in mice receiving CSA (0.7±0.2 x105 vs. 2.6±0.5x105 , P<0.02). By contrast, the numbers of Tregs in both the thymus and spleen from RAPA-treated mice were the same as those from PBS-treated mice. Mice treated with everolimus, another mTOR inhibitor, also showed no reduction in the numbers of Tregs. Histologic examination revealed that CSA-treated mice showed pathogenic features of chronic GVHD, including sclerodermatous skin changes, bile duct loss, fibrosis in the portal area of the liver and fibrosis and atrophy of acinar tissue in the salivary glands, while RAPA-treated mice showed no sign of chronic GVHD.Our findings indicate that a) Tregs cannot remain in grafts in the late stage, and newly arising donor BM-derived Tregs became dominant; b) CSA hampers BM-derived Tregs reconstitution and may be associated with the development of chronic GVHD; and c) mTOR inhibitors do not hamper Tregs reconstitution and might prove beneficial for the treatment of both acute and chronic GVHD. [Display omitted]