Deleterious Roles of the KIR Ligand Mismatch on Acute Graft Versus Host Disease in Unmanipulated HLA-Mismatched/Haploidentical Blood and Marrow Transplantation.

Abstract

The beneficial effect of KIR ligand mismatch with a very low incidence of rejection, acute graft versus host disease (aGVHD), and leukemia relapse had been demonstrated by Perugia's group in the Haploidentical mismatched Hematopoitic cell transplantation (HCT) with extensive T cells depletion in vitro. However, many analytical results of the haploidentical and unrelated mismatched transplantation appeared to be inconsistent with Perugia's work. The disputes in the inconsistent roles of KIR ligand mismatch seem to be caused by the different transplant protocols with different extent of T cells depletion in vitro or in vivo. In recent years, we successfully established a novel protocol—conditioning including antithymocyte globulin followed by un-manipulated HLA-mismatched/haploidentical blood and marrow transplantation, which can achieve comparable outcomes to HLA-identical sibling transplantation. Following the contradictory results about the KIR ligand mismatch in the haploidentical related and mismatched unrelated HCT, we have evaluated the roles of the KIR ligand mismatch in 94 leukemia patients undergoing unmanipulated HLA-mismatched/haploidentical blood and marrow transplantation. Multivariate analysis showed that both KIR ligand mismatch (HR 2.833, CI, 1.286–6.241, p=0.01) and doses of T cells (HR 3.059, CI, 1.292–7.246, p=0.011) were independent risk factors causing the acute graft versus host disease (aGVHD). In addition, compared to patients without KIR ligand mismatch, KIR ligand mismatch worsened the adverse effect of ‘high' dose T cells (>1.48×108/kg) on aGVHD (100% vs 63.3%, p=0.036), and increased the incidence of aGVHD in patients with HLA-C mismatch (80% vs 57.4, p=0.056). Since multivariate analysis demonstrated that high risk leukemia is the only predictor for TRM, relapse and OS, we further analyzed the effect of KIR ligand mismatch on prognosis in standard and high risk patients. The differences in TRM (50% vs 7.6%, p=0.005) and OS (50% vs 88.4%, p=0.014) between patients with and without KIR ligand mismatch were most striking for standard risk. Therefore we conclude that due to the presence of large dose T cells in the allograft, the alloreactivity of NK cells had been inhibited and KIR ligand mismatch directed alloreactive T cells played crucial roles in our model.