Early CD4+ T-Cell Effector Alloreactivity Towards Multiple Mismatched HLA Class II Alleles Is Associated with Graft Predominance after Double Umbilical Cord Blood Transplantation (dUCBT)

While dUCBT may be associated with less graft failure as compared to single UCBT, hematopoietic recovery following dUCBT generally originates from only a single graft, designated as the ‘winning' graft. Graft predominance was suggested to be T-cell mediated, but is still incompletely understood. We recently showed that CD4+ T-cells rapidly expand after dUCBT and early CD4+ T-cell chimerism predicts for graft predominance (Somers et al BBMT 2012; Haematologica 2014). Given the frequent HLA class II allele mismatches between the 2 UCB units in dUCBT, we hypothesized that HLA class II-specific CD4+ T-cells from the ‘winning' CBU may be responsible for rejection of the 'loser' CBU. In order to test that hypothesis, we evaluated whether ‘wining' CD4+ T-cells specifically recognize individual HLA class II allele mismatches, expressed by the rejected graft. Patient T-cells were propagated in-vitro by 1. HLA unbiased polyclonal expansion (by K562 clone 2D11+CD3mAb and sequential addition of cytokines IL-7, Il-15 and IL-12), and 2. specific activation by HLA class II allele transduced HELA stimulator cells. In-vitro reactivity of propagated T cells was assessed in a co-culture with class II allele transduced HELA cells, and measured by analysis of T cell activation or effector markers by flow cytometry (FCM). 11 patients with poor-risk hematological malignancies were included, receiving 22 UCB units matched at HLA A, B, and DRB1 for 5/6 (n=7) or 4/6 (n=15) with the recipient. The median number (range) of class II allele mismatches between the 2 UCB units per transplant was 2 (1-6). In total, 33 different class II allele mismatches were tested, including 16 at HLA DRB1, 7 at DQB, and 10 at DP. Conditioning with TBI (4 Gy) combined with fludarabine and cyclophosphamide was applied and mycophenolate and ciclosporin were given for GVH-prophylaxis. All patients engrafted, at a median number of 26 days (range: 20-50) after transplantation (without G-CSF), and all developed complete single unit chimerism. Peripheral blood CD3+ T-cell numbers of samples taken at 1-6 months post UCBT were low (median: 0.207; range 0.030-0.699 x 10-9/L) with 74% (range, 8-96%) consisting of CD4+ T-cells. In all 11 patients, alloreactive CD4+ T-cells towards one or more mismatched class II alleles were detectable. In total, CD4+ alloreactivity towards 29 out of 33 (88%) mismatches was detected, including 15/16 for DRB1 (94%), 7/7 for DQ (100%), and 7/10 (70%) for DP alleles. All mismatched alleles but one (DPB1*04:01) elicited a CD4+ T-cell response. Stronger CD4+ T-cell reactivity was observed towards DRB1 and DQ. Analysis of activity towards matched, control alleles showed positivity in 2/11 (18%) combinations. Reactivity towards irrelevant third party alleles showed positivity in 3/17 (18%). The class II alloresponse was significantly higher for mismatched versus matched alleles (median fold increase over control: 7.4 (range 4.4-21.5) versus 1.4 (1.1-2.2), respectively).  The highest alloreactive responses were observed in samples (n=14) taken at 1 month post UCBT (Figure). Alloreactive CD4+ T-cells upregulated CD137 and CD134, PD1 and the effector markers CD107 and Interferon-gamma. Collectively, these results demonstrate that specific effector alloreactivity by ‘winner' CD4+ T-cells directed to multiple class II mismatched alleles was present in all patients, already at 1 month post UCBT. These results suggest that immediate cytotoxicity exerted by CD4+ T-cells from the ‘winning' cord represent a novel mechanism of rapid rejection of the ‘losing' unit after dUCBT following a non-ATG conditioning regimen. Furthermore, these observations suggest that the lower incidence of graft failure observed after dUCBT may, in part, result from an immunological, graft-potentiating effect evoked by mismatched class II alleles expressed by the rejected graft. Therefore, allele matching at HLA class II between the 2 units can be ignored in dUCBT. Class II mismatches between de UCB units might, alternatively, even be aimed for, but this should be confirmed in a prospective study. The potential graft versus leukemia effect of these alloreactive CD4+ T-cells is subject of ongoing investigation.

Figure. Kinetics of anti-HLA class II allele CD4+ T-cell alloreactivity.