Increased Relapse Associated with Mixed Donor-Donor Chimerism Following Double Umbilical Cord Blood Transplantation (dUCBT)

Abstract ³⁵⁸

We previously showed lower rates of relapse in recipients of dUCBT, compared to patients treated with single UCBT (Verneris, Blood, 2009). One unit predominates by D+100 in ∼90% of patients and recent studies show that single unit dominance is associated with T cell recognition of the rejected unit (Gutman, Blood, 2010). In addition, higher rates of acute graft-vs-host disease (aGVHD) have been observed after dUCBT compared to single UCBT, regardless of age (MacMillian, Blood, 2009). Thus, we hypothesized that graft vs. graft interactions that occur during dUCBT may drive both GVL and GVHD reactions. To test this hypothesis, we reviewed the records of patients with hematological malignancies transplanted with a dUCBT from 2000 to 2011 at the University of Minnesota. Myeloablative (MA) conditioning was with cyclophosphamide (Cy 120 mg/kg), fludarabine (Flu 75 mg/m2) and total body irradiation (TBI 13.2 Gy) (n=178) and non-MA conditioning was with Cy (50 mg/kg), Flu (200 mg/m2) and TBI (2 Gy) (n=282). Acute GVHD prophylaxis consisted of cyclosporine and mycophenolic acid. At D+100 316 patients were evaluable and had either BM (n=273) and/or PB (n=178) donor engraftment data. Engraftment was assessed using short tandem repeat analysis. “Mixed donor-recipient chimerism” was defined as >5% host cells and “dual chimerism” was defined as complete donor chimerism (donor 1 + donor 2), with a minimum contribution of 5% by each unit (5% is the lower limit of detection for this clinical test). As expected, patients with evidence of mixed donor-recipient chimerism in the blood or marrow at D+100 (n=35) had high rates of relapse (69% [95% CI, 49–89%] at 2 years) and were excluded from further analysis. Of the patients with evaluable peripheral blood chimerism data, 90.4% (n=161) had hematopoiesis derived from a single unit, while 9.6% (n=17) showed dual chimerism after dUCBT. In multivariate analysis, dual chimerism at D+100 was associated with higher relapse (HR=1.97 [95% CI, 1.08–3.59], p=0.01), less grade II-IV aGVHD (HR=0.36 [95% CI, 0.17–0.78], p=0.01) and a trend toward worse DFS (HR=1.68 [95% CI, 0.9–3.13], p=0.1). Dual chimerism was not associated with age (p=0.32), nor non-MA conditioning (p=0.15). Recipients of two HLA 6/6 matched units (with each other and the recipient), were more likely to have dual chimerism (p<0.01). Nearly identical results were obtained in a separate model examining BM engraftment (single (n=247) and dual chimerism (n=26)) and transplant outcomes. While these findings should be confirmed in a separate dataset, they suggest that an unexpected consequence of selecting well matched units for dUCBT is higher mixed donor-donor chimerim, less aGVHD and more relapse. Further study of dUCBT and GVL is needed.