HLA-DRB1 Hed Is Associated with Improved Survival and Decreased Relapse in Patients with Hematologic Malignancies Following Allogeneic Hematopoietic Stem Cell Transplant

Abstract

Introduction: Despite advances in disease monitoring and maintenance therapies, the risk of relapse following allogeneic hematopoietic cell transplant (HCT) remains a major concern. Attention has turned to understanding how diversity across the immune system can influence tumor surveillance. Due to their protein's integral role in binding and presenting antigens to immune cells, HLA genes have been a major focus because a high number of polymorphisms alter their peptide-binding grooves and impact their repertoire of antigens. Based on the heterozygote advantage hypothesis, heterozygous HLA genotypes were thought to better elicit immune responses due to a higher probability of binding antigens. Further refinement of HLA diversity centers on the immunopeptidome and variations in peptide-binding specificities. HLA evolutionary divergence (HED) represents a quantitative measure of physiochemical sequence divergence between peptide-binding grooves encoded by HLA alleles. Greater sequence divergence across HLA alleles leads to a more expansive repertoire of antigens and helps to explain that high HED was associated with improved overall survival (OS) in patients with solid tumors treated with immune checkpoint inhibitors (Chowell et al. Nat Med, 2019). With a more diversified immune response, we hypothesize that high HED will improve tumor surveillance and impact survival and relapse following allogeneic HCT. Methods: Patients who underwent 8/8 HLA-A, -B, -C and -DRB1 matched first allogeneic HCT for AML, MDS, ALL, CML, or lymphoma between 2008-2018 were identified from the CIBMTR database. All patients received myeloablative conditioning, except for those patients with lymphoma. Using donor HLA typing to the allele level and reference amino acid sequences for peptide binding domains, HLA HED was calculated for each HLA class I allele (-A,-B,-C) and HLA class II allele (-DRB1) using the Grantham Distance metric (Pierini, et al. Mol Biol Evol, 2018). An overall mean HED score was calculated for HLA class I and HLA-DRB1 alleles and was initially divided at the median to classify high versus low HED. HED was then evaluated based on quartiles with high HED defined as overall mean HED greater or equal to the upper quartile. The primary outcomes were OS and relapse. Secondary outcomes included transplant-related mortality (TRM), disease-free survival (DFS), and graft-versus-host disease (GVHD). Multivariate analysis using Cox proportional hazards model was performed with factors violating the proportional hazards assumptions adjusted through stratification. P-value ≤ 0.01 was significant for this study. Results: A total of 9,231 patients received an 8/8 HLA-matched allogeneic HCT from 2008-2018 according to selection criteria. With high HED defined by the median, there was no association with OS and HLA class I HED (HR 1.00; 95% CI 0.93 - 1.07, P=0.935) or with HLA-DRB1 HED (HR 0.96; 95% CI 0.91 -1.02, P=0.225). Relapse was not associated with HLA class I HED (HR 1.04, 95% CI 0.96-1.12, P=0.354) or with HLA-DRB1 HED (HR 0.99, 95% CI 0.93 - 1.06, P=0.857). There were no significant associations with HED and the secondary outcomes of TRM, DFS, or GVHD. HED was then characterized by quartiles. High HLA-DRB1 HED (upper quartile) was associated with improved OS (HR 0.91; 95% CI 0.85 - 0.96, P=0.0019) and decreased relapse (HR 0.87; 95% CI 0.80 - 0.95, P=0.0014) when compared to low HED (Q1-Q3). Improved DFS was observed with high HLA-DRB1 HED but failed to achieve statistical significance (HR 0.92; 95% CI 0.86 - 0.98, P=0.011). There was no association between HLA class I HED and OS, relapse, or DFS. Conclusions: High HLA-DRB1 HED, defined as upper quartile, was associated with improved OS and decreased relapse in patients with hematologic malignancies following allogeneic HCT. High HED promotes a diverse repertoire of antigens, which may heighten tumor surveillance and enhance graft versus leukemic effect. Since similar associations were not replicated when high HED was defined by the median, the effects of HLA-DRB1 HED are most likely observed in the most divergent donor-recipient HLA pairings. Previous work has implicated the role of HLA class II alleles in affecting immune function and contributing to relapse following HCT. Therefore, HLA-DRB1 may play a more integral role in tumor surveillance post-HCT, which can have important implications in donor selection, especially in HLA mismatched settings.