NOD2/CARD15 Genotype and HLA-DPB1 Matching Status Can Be Used To Predict Disease Relapse in Recipients of an Unrelated Donor Haematopoietic Stem Cell Transplant for Acute Leukaemia.

Haematopoietic Stem Cell Transplantation (HSCT) using volunteer Unrelated Donors (UD) has become an important and viable option in the treatment of Acute Leukaemia (AL). While matching donors and recipients usually refers to five of the classical HLA genes (HLA-A, -B, -C, -DRB1 and -DQB1), the impact of a sixth gene, HLA-DPB1, on the outcome of UD-HSCT is increasingly emerging. We have previously shown an increased risk of relapse with HLA-DPB1 matching and independently, with NOD2/CARD15 genotype. In light of these data, we have analysed a larger UD-HSCT cohort in order to establish the impact on transplant outcome when both HLA-DPB1 matching status and NOD2/CARD15 genotype are considered. HLA typing and NOD2/CARD15 genotyping was performed on 304 AL patients and their Anthony Nolan Trust volunteer unrelated donors. Transplants occurred between 1996 and 2005 at UK transplant centres. Diagnoses were ALL (47%) and AML (53%). 67% of the cohort were a 10/10 HLA match with 16% also being matched for HLA-DPB1. Myeloablative conditioning regimens were used in 74% of transplants. T-cell depletion was included in 82% of conditioning protocols. Bone marrow was used in 72% of transplants with the remaining 28% using peripheral blood stem cells. Two forms of post-transplant immunosuppression predominated, Cyclosporine A and Methotrexate (47%) and Cyclosporine A alone (38%). Previous studies on a subgroup of this cohort showed that HLA-DPB1 matching and NOD2/CARD15 SNPs independently caused an increase in disease relapse. Consequently, the cohort was grouped into three categories to reflect this risk, group 1 (DPB1 matched; NOD2/CARD15 SNP, n=24), group 2 (HLA-DPB1 matched; NOD2/CARD15 Wild-Type (WT) or HLA-DPB1 mismatched; NOD2/CARD15 SNP, n=112) and group 3 (HLA-DPB1 mismatched; NOD2/CARD15 WT, n=168). There was a significant difference in disease relapse between the three groups (1 year: group 1; 68%, group 2; 48%, group 3; 30%, p=0.0038). This finding persisted in multivariate analysis where being in either group 2 or 3 was protective towards relapse as compared to group 1 (RR 0.321; 95% CI 0.167–0.616; p=0.001 and RR 0.478; 95% CI 0.244–0.934; p=0.031 respectively). In the group with the highest relapse risk (group 1), this resulted in a decrease in Overall Survival (OS) (33% vs 54% in group 3, RR 0.617; 95% CI 0.359–1.060; p=0.080). The best OS was seen in the group with the lowest risk of relapse (group 3). Here, in addition to low relapse, there was increased acute and chronic Graft-versus-Host Disease (GvHD) (p=0.0019 and p=0.0058 respectively). In this cohort, cGvHD (in its limited form) was associated with a significantly lower incidence of relapse (p=0.0066) and better OS (p<0.0001). In concordance with our previous theories, it appears that being HLA-DPB1 matched and having NOD2/CARD15 SNPs predicts for the worst outcome with a significant increase in relapse and reduced OS. Conversely, the ideal pairing would be HLA-DPB1 mismatched and NOD2/CARD15 WT. These data suggest that prospectively typing AL patients for HLA-DPB1 and NOD2/CARD15 SNPs will allow the prediction of disease relapse, aGvHD and cGvHD and in addition will allow the effects of being independently HLA-DPB1 matched or having a NOD2/CARD15 SNP to be offset by intelligently selecting a suitable, less precarious donor.