Dendritic and T Cell Subsets Correlate with the Development and Site of Acute GVHD Following Unrelated Cord Blood Transplantation. A Multivariate Analysis of Host, Graft, and Day+50 Immune Profile.

HLA-mismatched unrelated umbilical cord blood transplantation (uUCBT) is a suitable option for those who lack HLA-matched sibling donors. Although UCB T cells mount significantly less cytotoxicity and cytokine response than adult T cells uUCBT can lead to significant aGVHD. We hypothesized that select features of T cell and dendritic cell (DC) reconstitution will correlate with the development of aGVHD. Here we report on 106 prospectively studied pediatric patients, all full donor chimera. Methods: On day +50 we determined by 4-color surface and ic. FACS the reconstitution of T cell subsets, lymphoid, and myeloid DC, and measured their homeostatic proliferation, activation state, and inducible cytokine secretion to correlate with the risk for developing aGVHD. Logistic regression models were created on variables with a univariate P value <0.05 by Wilcoxon rank-sum test. Results: Mean age of patients was 6.4 years, 65 patients are alive with a mean time of 36 months follow-up (61%). Survival is lower (26%) in patients who developed aGVHD at grade III–IV (n=21), versus those with grade 0–II (73%, n=85, p<0.0001), while it is not different (p=0.5) for those with grade II–IV (n=52) versus 0–I (n=54). More patients with aGVHD grade III–IV experienced de novo opportunistic infection (OI) impacting their survival (p=0.0001), while death due to OI was similar for those with grII–IV GVHD vs 0–I. Non-infectious death was similar regardless of GVHD grade. Among potential confounding factors (Table I) younger age/smaller body size were associated with grade II–IV aGVHD while gender, CMV status, graft content did not. Younger children were less likely to have malignancy/receive TBI. HLA mismatch at 2–3 HLA alleles was the only confounder significantly related to developing gr III–IV aGVHD (p=0.008). Table I presents immune parameters correlating with grade II–IV aGVHD. Several variables correlating with grIII–IV GVHD in univariate analysis (e.g activated/DR+CD4 T cells) lost significance in the multivariable model when OI was added. Still higher % and absolute numbers of CD4+ T cells secreting IL-2 and TNF remained significant. Additionally, fewer CD123+ DC in the circulation (0.7/μl) correlated with gr III–IV aGVHD in contrast with grade 0–II (2.5/μl, p=0.009). More T cells were in cell cycle of those with grIII–IV GVHD (27% Ki-67+ cells vs 16%, p=0.04). In patients with grII–IV skin GVHD T cells expressing the skin homing molecule CLA appeared (2% vs 0%, p=0.05). Conclusion; HLA-mismatch at >1Ag and select patterns of T cell subset and DC reconstitution may identify those at risk for GVHD.

UNIVARIATE AND MULTIVARIATE ANALYSIS OF DAY+50 IMMUNE PROFILE AND THE RISK FOR DEVELOPING GRADE II–IV ACUTE GVHD