Abstract 328Low CD4+ T cell counts early post-SCT are associated with a greater risk of developing chronic GVHD (cGVHD). However, the factors affecting post graft CD4+ T cell recovery are not known. Furthermore, the CD4 T cell subset correlating best with protection from cGvHD is not defined. We hypothesized that the donor immune repertoire determines CD4+ reconstitution after SCT. We studied 220 donor-recipient pairs undergoing a myeloablative matched-sibling T-cell depleted SCT for a variety of hematological malignancies including AML, ALL, CML, and MDS. Median donor and patient age was 36 years (range, 9 – 70). Median CD34+ cell dose was 5.56 ×106/kg (range, 2.3–14.5), with a fixed T cell dose per protocol of 1–5 ×104/kg. Ninety-four patients (43%) developed chronic GvHD (32 limited, 64 extensive). Prior to SCT a cryopreserved mononuclear fraction (PBMC) was obtained by leukopheresis on all donors. PBMC from 139 donors were stained with fluorescently conjugated antibodies against CD3, CD4, CD8, CD27, CD45RA, CD31, CD25, CD14, CD19, FOXP3, and Helios and acquired on a custom-built LSR Fortessa (BD) flow cytometer. The following T cell subsets were examined: naive, CD27+CD45RA+; central memory, CD27+CD45RA-; effector memory, CD27-CD45RA-; and effector T cells, CD27-CD45RA+. Regulatory T cells (Tregs) and Treg subsets were identified within the CD4+ T cell population using a combination of FOXP3 and Helios The combination of FOXP3 with Helios has recently been shown to identify natural, thymus-derived Tregs (nTregs), whereas CD4+FOXP3+ T cells lacking Helios represent a pool of induced Tregs (Thornton et al., 2010, J Immunol. 184:3433–3441). Combined expression of CD31 and CD45RA in the CD4+ and nTreg populations was used to identify recent thymic emigrants (RTE). Data were analyzed by FlowJo (Treestar, Ashland, OH, version 8.6.6), and both univariate and multivariate analysis analyses using the Cox regression models with right censored time-to-event variables were performed on the readouts (SPSS 15.0, S-Plus 8 and Prism 4 Software). All p-values were computed based on the log-rank test statistics. Absolute lymphocyte counts in 220 donors were 2.11/μl (range 0.89 – 4.12). A low absolute lymphocyte count was significantly correlated with a higher cumulative probability for time to incidence of extensive cGVHD (HR 2.38, p = 0.0008). This relationship between lymphocytes and cGVHD was retained in CD4+ T cell subsets: CD3+CD4+ (HR 2.48, p = 0.005); CD3+CD4+CD25+ < 30 cells/μL (HR 2.1, p = 0.02); nTregs < 8 cells/μL (HR 1.97, p = 0.043). RTE of total CD4 did not correlate with RTE nTregs, indicating that nTreg production varies among donors with similar levels of thymic activity. Furthermore recipients of donors with low RTE nTregs relative to overall thymic output had a significantly higher likelihood of developing extensive chronic GvHD (HR 2.78, p=0.0014). In multivariate analysis including patient age, acute GvHD (grade II – IV), graft direction (female donor to male recipient vs. other combinations), disease and status at transplant, and administration of post SCT donor lymphocyte infusion, only thymic output of nTregs (HR = 0.8, p = 0.037) and graft direction (HR = 2.1, p = 0.023) remained significant risk factors for extensive cGvHD. These data show that the thymic output of nTreg varies inherently among donors. Donors with low thymic output of nTreg may recreate a less tolerant immune system in the recipient leading to cGVHD of greater severity.(ZAM and JJM contributed equally to this work). [Display omitted] Disclosures:No relevant conflicts of interest to declare.