Lineage-Specific Relapse Prediction After Haploidentical Transplantation With Post-Transplant Cyclophosphamide Based on Recipient HLA-B-Leader Genotype and HLA-C-Group KIR Ligand.

The role of NK cell alloreactivity on outcomes after T cell-replete haploidentical donor transplantation (HIDT) remains uncertain. After transplantation, newly formed NK cells are licensed through interactions of donor inhibitory KIR (iKIR) and NKG2A receptors with their cognate ligands on recipient cells. Donor NKG2A recognizes HLA-E bound by recipient HLA class I leader peptides, a process requiring methionine (M) at position -21 of the leader sequence. An rs1050458C/T dimorphism results in approximately 40% of individuals expressing at least one copy of -21M HLA-B (M/M or M/T [M+]), allowing ligand expression. We assessed the impact of recipient HLA-B-leader genotype (M+ versus M- [T/T]) and HLA-C-group iKIR missing ligand (ML, C1C1/C2C2 versus C1C2) on relapse and disease-free survival (DFS) in recipients of post-transplantation cyclophosphamide (PTCy)-based HIDT. Based on preclinical data, we hypothesized that the relative impact of each variable may depend on disease lineage (lymphoid versus myeloid). To this end, we analyzed outcomes of 322 consecutive PTCy-based HIDT recipients with hematologic malignancy who underwent transplantation at a single institution using standardized supportive care measures with mature follow-up (median 45 months). Primary endpoints were relapse and DFS of patients based on HLA-B-leader genotype and HLA-C-group iKIR ML. Planned subgroup analysis included patient with lymphoid versus myeloid malignancy. M+ HLA-B-leader genotype and HLA-C-group iKIR ML were seen in 42% and 49% of recipients, respectively. The presence of a recipient M+ B-leader (versus M-) improved overall survival (OS) and DFS and lowered cumulative incidence of relapse (CIR), an effect primarily seen in lymphoid malignancies (80% versus 51%, 72% versus 41%, 16% versus 42%, respectively). In contrast, myeloid malignancy patients benefited most from HLA-C-group iKIR ML with better OS and DFS and lower CIR (67% versus 51%, 64% versus 44%, 25% versus 45%, respectively). Multivariate analysis confirmed the disease-specific associations of improved relapse/DFS with M+ HLA-B-leader in lymphoid malignancy (hazard ratio [HR] 0.20, P < .001/HR 0.34, P <.001) and HLA-C-group iKIR ML in myeloid malignancy (HR 0.44, P = .004/HR 0.54, P = .009). Neither HLA-B-leader nor iKIR ML was associated with the incidence of non-relapse mortality or acute or chronic graft-versus-host disease. Two distinct NK cell education pathways predict relapse and DFS after HIDT-PTCy in a disease-specific manner: the presence of recipient M+ HLA-B-leader genotype improves outcome in patients with lymphoid malignancies, whereas HLA-C-group iKIR ML improves outcome in patients with myeloid malignancies. These findings strengthen the essential role of NK cells for optimal GVL in the context of HIDT-PTCy and may suggest different approaches to improving transplant outcome depending on disease type.