Systematic evaluation of donor-KIR/recipient-HLA interactions in HLA-matched hematopoietic cell transplantation for AML

In acute myeloid leukemia (AML), donor natural killer cell killer immunoglobulin-like receptors (KIR) and recipient HLA interactions may contribute to the graft-versus-leukemia effect of allogeneic hematopoietic cell transplantation (HCT). Analyses of individual KIR/HLA interactions, however, have yielded conflicting findings, and their importance in the HLA-matched unrelated donor (MUD) setting remains controversial. We systematically studied outcomes of individual donor-KIR/recipient-HLA interactions for HCT outcomes and empirically evaluated prevalent KIR genotypes for clinical benefit. Adult patients with AML (n = 2025) who received HCT with MUD grafts in complete remission reported to the Center for International Blood and Marrow Transplantation were evaluated. Only the donor2DL2(+)/recipient-HLA-C1(+) pair was associated with reduced relapse (hazard ratio [HR], 0.79; 95% confidence interval [CI], 0.67-0.93; P = .006) compared with donor-2DL2-/recipient-HLA- C1(+) pair. However, no association was found when comparing HLA-C groups among KIR-2DL2(+)-graft recipients. We identified 9 prevalent donor KIR genotypes in our cohort and screened them for association with relapse risk. Genotype 5 (G5) in all recipients and G3 in Bw4+ recipients were associated with decreased relapse risk (HR, 0.52; 95% CI, 0.35-0.78; P = .002; and HR, 0.32; 95% CI, 0.14-0.72; P = .006; respectively) and G2 (HR 1.63, 95% CI, 1.15-2.29; P = .005) with increased relapse risk in C1-homozygous recipients, compared with other patients with the same ligand. However, we could not validate these findings in an external data set of 796 AML transplants from the German transplantation registry. Neither a systematic evaluation of known HLA-KIR interactions nor an empiric assessment of prevalent KIR genotypes demonstrated clinically actionable associations; therefore, these data do not support these KIR-driven strategies for MUD selection in AML.