Matching Status at Amino Acid Positions 57 and 65 of the HLA-DPB1 Beta Chain Determines Outcome in Recipients of Unrelated Donor Haematopoietic Stem Cell Transplants.

There is now a bulk of evidence supporting a role for HLA-DPB1 matching status on the outcome in Haematopoietic Stem Cell Transplants (HSCT) from Unrelated Donors (UD). More recently analysis has focussed on matching status at particular epitopes or amino acid (aa) positions within the DPB1 molecule. Mismatches at this level may predict for transplant complications and this analysis may contribute towards better understanding of the function of DPB1. Within exon two of the DPB1 molecule (which forms the peptide binding groove, PBG) there are six Hypervariable Regions (HVR), A to F, which are present in patterns characteristic to antigen groups. Within each of these HVR there are one or more amino acids which are polymorphic and this polymorphism is important in determining the peptides which will bind, the recognition by T cell receptors or both. We analysed the outcome in 282 mixed transplant pairs who have received an HSCT from an unrelated donor. The majority of transplant protocols included T cell depletion (TCD) with CAMPATH (>75%). All pairs had high resolution tissue typing for six HLA loci and, in order to remove confounding effects, were matched for HLA-A, -B, -C, -DRB1, -DQB1. Transplant pairs were assessed as matched or mismatched for HVRs A to F, as well as the individual amino acid positions in each of these regions, in a graft versus host direction. Mismatches at both amino acid position 57 (within HVR C) and 65 (within HVR D) were associated with transplant complications. Position 57 was matched in 233 pairs (83%) and mismatched in 49 (17%). A match at this position resulted in a significant survival advantage compared to those that were mismatched (3 years survival probability: 50% compared to 33%; log rank p=0.014). Position 65 was matched in 231 (82%) pairs and mismatched in 51 (18%). Matched pairs had a significantly improved overall survival (OS) compared to mismatched pairs (49% versus 35%; log rank p=0.039). The reason for the inferior OS was investigated. There was no significant association between matching at either position and graft versus host disease or disease relapse. In contrast, mismatched pairs had a significantly increased rate of transplant related mortality (TRM). The one year TRM was significantly increased in those pairs mismatched at positions 57 (19/42, 45% versus 52/192, 27%; log rank p=0.013) and 65 (19/45, 42% versus 52/189, 28%; log rank p=0.042). In order to rule out that the effects of mismatching for these positions were influenced by the number of allele incompatibilities present, the analysis was repeated only in those pairs with two DPB1 allele mismatches (n=82). The results mirrored those seen in the whole cohort. Although the crystal structure of DPB1 has yet to be solved, comparisons can be made with the structure and function of DRB1. Position 65 is analogous to position 67 in DRB1, and is known to be involved in peptide interactions in that molecule. Position 57 points away from the PBG, however may still interact with peptide as it is well recognised that Class II peptides “hang out” of the ends of the groove. In addition, these amino acids may determine recognition by T cells or other molecules. In conclusion, we have shown a significantly worse OS, largely mediated by an increased TRM, in recipients of HSCT from UD which are mismatched at aa positions 57 and 65 of the DPB1 molecule. We suggest this supports pre-transplant tissue typing for DPB1, both to aid in donor selection and in the prediction of outcome.