Indirect presentation of mismatched human leukocyte antigen-B associates with outcomes of cord blood transplantation

Abstract

Cord blood transplantation (CBT) is a valuable donor source for patients without human leukocyte antigen (HLA)-matched donors. While CBT has a lower risk of graft-versus-host disease and requires less stringent histocompatibility, it is associated with a higher transplantation-related mortality (TRM) compared to other donor sources. We hypothesized that assessing the immunogenicity of mismatched HLA could reveal non-permissive mismatches contributing to increased TRM. We retrospectively analysed 1498 single-unit CBT cases from 2000 to 2018 across eight Japanese institutions, evaluating the immunogenicity of mismatched HLA using the PIRCHE algorithm to examine binding affinities of HLA-derived epitopes to donor or recipient HLA. Results indicated that Class I epitopes from mismatched recipient HLA-B were significantly associated with poor outcomes due to higher TRM and lower neutrophil engraftment, particularly when presented on matched HLA class I. Notably, epitopes from HLA-B exon 1 showed stronger prognostic significance, with HLA-B alleles carrying M-type leader peptides exhibiting higher affinity for these epitopes. Patients with a matched M-type HLA-B and Class I epitopes derived from mismatched HLA-B exon 1 had worse outcomes. These findings suggest that immunogenicity-informed donor selection could improve CBT outcomes.

Keywords: HLA; cord blood transplantation; donor selection; transplantation‐related mortality.

FIGURE 1

Workflow of the current study.

FIGURE 2

Univariate analysis of the PIRCHE derived from each HLA locus. (A) The forest plot shows the hazard ratio of each clinical event in patients with immunogenic epitopes derived from each HLA locus. The upper plots show the data of Class I epitopes and the lower plots indicate the results of HLA class II epitopes. (B) The survival and cumulative incidence curves for each event stratified by the HLA‐B allele mismatch status and Class I epitopes derived from recipient HLA‐B. (C) Cumulative incidence curve of severe PIR stratified by recipient HLA‐B‐derived Class I epitopes. The p‐values for ad hoc analysis to compare the outcomes of each group are adjusted using the Benjamini–Hochberg method. aGVHD, acute graft‐versus‐host disease; PFS, progression‐free survival; PIR, pre‐engraftment immune reaction; TRM, transplantation‐related mortality.

FIGURE 3

Multivariate analysis of Class I epitopes derived from HLA‐B with other prognostic factors. PFS, progression‐free survival; TRM, transplantation‐related mortality.

FIGURE 4

Differences in significance of epitopes derived from HLA‐B by the presenting HLA (matched donor—HLA class I versus mismatched donor—HLA class I). (A) A conceptual figure for the analysis of the prognostic significance of recipient HLA‐B‐derived epitopes presented on mismatched donor HLA or matched HLA. (B) Cumulative incidence curve of transplantation‐related mortality (TRM) and neutrophil engraftment stratified by the presence of Class I epitopes derived from recipient HLA‐B presented on matched or mismatched HLA. (C) Multivariate analysis of Class I epitopes derived from recipient HLA‐B presented on matched and mismatched HLA. (D) Cumulative incidence curve of severe PIR stratified by Class I epitopes derived from recipient HLA‐B presented on matched or mismatched HLA for patients who can be evaluated. PFS, progression‐free survival; PIR, pre‐engraftment immune reaction.

FIGURE 5

Differential prognostic significance of HLA‐B‐derived epitopes according to their exons of origin. (A) The forest plot of the univariate and multivariate analyses of TRM and neutrophil engraftment by the exon of origin for Class I epitopes derived from recipient HLA‐B. (B) The association between HLA‐B exon 1‐derived Class I epitopes presented on matched HLA‐B and leader‐type matching status in mismatched HLA‐B or the type of leader peptide (T or M) for matched HLA‐B. (C) Survival curve and cumulative incidence curve of TRM and neutrophil engraftment stratified by the presence of Class I epitopes derived from recipient HLA‐B presented on matched HLA‐B and the type of leader peptide (T or M) for matched HLA‐B. (D) Classification of the peptide sequence of exon 1 of the observed HLA‐B. (E) Distribution of Class I epitopes derived from recipient HLA‐B exon 1 according to the site of origin. (F) Multivariate analysis of TRM and neutrophil engraftment by Class I epitopes derived from recipient HLA‐B exon 1 from different sites of origin. (G) Cumulative incidence of TRM and neutrophil engraftment for patients with Class I epitopes derived from positions 7 to 15 in exon 1 of HLA‐B (black) and patients with (green) and without (red) Class I epitopes derived from the other portions of HLA‐B. HR, hazard ratio; PFS, progression‐free survival; TRM, transplantation‐related mortality.

FIGURE 6

Association between HLA‐B‐derived epitopes and the composition of the immune cells after transplantation. (A) Correlation coefficient between Class I epitopes derived from recipient HLA‐B and the percentage of immune cell fraction of bone marrow approximately 30 days after transplantation. (B) Box plot of the percentage of CD8+ central and effector memory T cells in patients with and without Class I epitopes derived from recipient HLA‐B.