Increased Proliferation of FoxP3 Regulatory T Cells after Allogeneic Hematopoietic Stem Cell Transplantation Is Counterbalanced by Increased Susceptibility to Apoptosis

CD4+FoxP3+ Regulatory T cells (Treg) play a critical role in the maintenance of tolerance after allogeneic hematopoietic stem cell transplantation (HSCT). We previously demonstrated that patients with active chronic graft-versus-host disease (cGVHD) have a reduced frequency of Treg. However, the mechanisms responsible for inadequate Treg reconstitution in patients with cGVHD have not been characterized. We therefore examined phenotypic and functional characteristics of Treg in 16 patients 2–41 months (median 10 months) post-HSCT to elucidate these mechanisms. Treg were compared to conventional CD4+FoxP3-T cells (Tcon) within individual patient samples and to healthy donors. All patients received TBI-based myeloablative conditioning, peripheral blood stem cells from HLA-matched donors (12 MRD; 4 URD) and acute GVHD prophylaxis (11 tacrolimus and sirolimus; 5 tacrolimus and methotrexate). At the time of analysis, 9 patients had no chronic GVHD, 5 had active chronic GVHD (1 limited disease; 4 extensive disease) and 2 had inactive chronic GVHD. Total CD4 counts were relatively low after HSCT compared to healthy donors (median CD4 273/ul vs 756/ul). After HSCT, patient Treg exhibited a predominant CD45RA(−)CCR7(−) effector/memory phenotype. Expression of CD31 on CD45RA+ Tcon and Treg was used to identify cells within these subsets that were recent thymic emigrants (RTE). In patient samples, 16.5% of Tcon and 2.8% of Treg expressed CD31+CD45RA+. In healthy donors, 22.9% of Tcon and 5.4% of Treg were CD31+CD45RA+. The lower fraction of RTE within the Treg population after transplant suggests that Treg primarily reconstitute through peripheral proliferation rather than through thymic generation. The proliferative capacity of both Tcon and Treg was examined by evaluating expression of Ki-67 in these subsets. After transplant, Ki-67 expression was significantly higher in Treg (5.2%) than in Tcon (1.5%) (p<0.001). This was significantly higher in both populations compared to healthy donors where 2.5% of Treg (p<0.05) and 0.2% of Tcon (p<0.01) expressed Ki-67. In both patients and healthy donors, Ki-67 expression was found almost entirely in cells that were CD45RA-indicating that proliferation was primarily occurring within the memory subsets of Tcon and Treg. Increased expression of Ki-67 on Treg was associated with low CD4 T cell counts (p<0.001), but not with time after HSCT (p=0.21) and chronic GVHD status (p=0.35). Treg Ki-67 expression after HSCT showed a strong positive correlation with CD95 (FAS) expression (p<0.01), but this association was not present in Tcon post-HSCT or in Treg from healthy donors. To determine whether increased expression of CD95 results in apoptosis of Treg, we purified 4 different CD4+ T cell subsets by cell sorting (CD45RA+ Tcon, CD45RA− Tcon, CD45RA+ Treg and CD45RA− Treg) from healthy donors and HSCT patients. Purified cells were cultured with or without agonistic FAS antibody (anti-FAS) and apoptosis was measured using Annexin-V staining. Anti-FAS rapidly induced apoptosis of CD45RA− memory-like Treg from HSCT patients while all other Treg and Tcon subsets were relatively resistant to apoptosis. In summary, these results indicate that Treg reconstitution post-HSCT is characterized by high levels of peripheral proliferation, which appear to be driven primarily by persistent CD4 T lymphopenia. However, post-HSCT Treg are also highly sensitive to FAS-mediated apoptosis. This process does not affect the survival of other CD4 T cell subsets. In the absence of thymic generation of Treg from hematopoietic precursors, this dynamic process results in a relative deficiency of Treg post-HSCT. Our findings provide important information for developing strategies aimed at monitoring and modulating Treg to promote immune tolerance following HSCT.