High HLA evolutionary divergence mismatch increases the risk of acute graft-versus-host disease and mortality after haploidentical hematopoietic stem cell transplantation

Abstract Background HLA evolutionary divergence (HED) quantifies physicochemical differences between HLA protein sequences, determining immunopeptidome diversity presented to T cells. In haploidentical HSCT (haplo-HSCT), mismatched HLA molecule/peptide complexes create structural disparities between recipient antigens and donor T-cell receptors (TCRs). While such immunopeptidome divergence may enhance graft-versus-leukemia (GVL) effects, it concurrently increases graft-versus-host disease (GVHD) risk. Beyond established factors like HLA mismatch load and specific high-risk mismatches, we investigated the impact of donor-recipient HED discrepancy (HED-mismatch, HED-mis) on transplantation outcomes in the context of haplo-HSCT utilizing ATG platform. Methods Between May 2012 and December 2023, a total of 797 patients with hematological malignancies were included in this study. These patients had undergone their first haplo-peripheral HSCT with ATG at our institution, and only those with complete donor-recipient HLA data and clinical information were eligible for inclusion. For each mismatched HLA locus (A, B, C, DRB1, DQB1) per donor-recipient pair, HED-mis was quantified using Grantham distances (measuring amino acid physicochemical differences). For each donor allele at a locus, the minimum distance to the recipient's alleles was found. The HED-mis for the locus was the mean of the two donor allele minimum distances. To summarize class I and class II effects, two composite scores were derived: HED-misTotal1: Mean of HED-mis values for HLA-A, -B, and -C. HED-misTotal2: Mean of HED-mis values for HLA-DRB1 and -DQB1. HED-mis variables were dichotomized at median values (high vs. low) in both univariate and multivariable analyses. We analyzed the association of HED with acute GVHD (aGVHD), chronic GVHD (cGVHD), non-relapse mortality (NRM), cumulative incidence of relapse (CIR), and overall survival (OS). Results The diagnoses of patients were AML (n=425), ALL (n=302), MDS (n=20), others (n=25). The median age of recipients was 40 years (range, 8–70), with 46% being female; all received peripheral blood stem cell (PBSC) grafts. Among donors, the median age was 33 years (range, 8–64), and 34% were female. The median follow-up were 24.1 months after HSCT [range: 0.6-115.5]. Median HED-mis values [range] were 2.41 [0–10.95] for HLA-A, 3.36 [0–10.49] for HLA-B, 1.76 [0–4.19] for HLA-C, 4.95 [0–15.66] for DRB1, and 3.72 [0–9.71] for DQB1, respectively. For HLA class I, the HED-misTotal1 was 2.40 [0–6.43], whereas for HLA class II, the HED-misTotal2 was 4.15 [0–9.43]. In univariate analysis stratified by median cutoffs, higher HED-misB predicted increased risk of grade II-IV aGVHD (HR 1.36, 95%CI 0.99-1.87, p=0.05), higher NRM (HR 1.75, 95%CI 1.01-3.01, p=0.04), and inferior OS (HR 1.33, 95%CI 1.00-1.77, p=0.05). Similarly, high HED-misTotal2 was associated with greater incidence of overall aGVHD (HR 1.37, 95%CI 1.11-1.70, p=0.004) and grade II-IV aGVHD (HR 1.50, 95%CI 1.09-2.06, p=0.013). Notably, no HED-mis metrics showed significant associations with cGVHD or relapse (all p>0.1). Multivariable analysis adjusted for sex, diagnosis, and age revealed disease-specific effects: in the AML cohort, patients with high HED-misB (above median) had a significantly higher risk of grade II-IV aGVHD (HR 1.23, 95%CI 1.06-1.43, p=0.007) compared to those with low HED-misB. Similarly, high HED-misDRB1 (above median) was associated with increased incidence of aGVHD (HR 1.07, 95%CI 1.01-1.13, p=0.02) and higher non-relapse mortality (HR 1.18, 95%CI 1.04-1.34, p=0.01). Among ALL patients, high HED-misDQB1 (above median) predicted increased risk of aGVHD (HR 1.09, 95%CI 1.02-1.16, p=0.01) and worse overall survival (HR 1.09, 95%CI 1.01-1.17, p=0.03) compared to low HED-misDQB1. Conclusion Donor-recipient HED discrepancy at HLA-B (HED-misB) and the DRB1/DQB1 composite (HED-misTotal2) independently predict increased aGVHD and mortality in haplo-HSCT. The locus-specific effects observed—particularly the association of HED-misB with aGVHD in AML—highlight HED-mis quantification as a functional biomarker for donor selection to mitigate post-transplant complications.