Genome-wide minor histocompatibility matching as related to the risk of graft-versus-host disease

Abstract

The risk of acute graft-versus-host disease (GVHD) is higher after allogeneic hematopoietic cell transplantation (HCT) from unrelated donors as compared with related donors. This difference has been explained by increased recipient mismatching for major histocompatibility antigens or minor histocompatibility antigens. In the current study, we used genome-wide arrays to enumerate single nucleotide polymorphisms (SNPs) that produce graft-versus-host (GVH) amino acid coding differences between recipients and donors. We then tested the hypothesis that higher degrees of genome-wide recipient GVH mismatching correlate with higher risks of GVHD after allogeneic HCT. In HLA-genotypically matched sibling recipients, the average recipient mismatching of coding SNPs was 9.35%. Each 1% increase in genome-wide recipient mismatching was associated with an estimated 20% increase in the hazard of grades III-IV GVHD (hazard ratio [HR], 1.20; 95% confidence interval [CI], 1.05-1.37; P = .007) and an estimated 22% increase in the hazard of stage 2-4 acute gut GVHD (HR, 1.22; 95% CI, 1.02-1.45; P = .03). In HLA-A, B, C, DRB1, DQA1, DQB1, DPA1, DPB1-phenotypically matched unrelated recipients, the average recipient mismatching of coding SNPs was 17.3%. The estimated risks of GVHD-related outcomes in HLA-phenotypically matched unrelated recipients were low, relative to the large difference in genome-wide mismatching between the 2 groups. In contrast, the risks of GVHD-related outcomes were higher in HLA-DP GVH-mismatched unrelated recipients than in HLA-matched sibling recipients. Taken together, these results suggest that the increased GVHD risk after unrelated HCT is predominantly an effect of HLA-mismatching.

Figure 1.

Unrelated donor-recipient pairs have more genome-wide recipient mismatching than sibling pairs. For each coding SNP, recipient mismatching was counted as 0 or 1, where a value of 0 indicates no recipient mismatching and a value of 1 indicates mismatching for 1 or both alleles. The extent of genome-wide mismatching for each recipient was calculated as the proportion of SNPs with mismatching. Histograms show the numbers of patients for each 0.001 increment of genome-wide recipient mismatching for coding SNPs among full-sibling recipients (red, left-side peak) (N = 1840) and unrelated recipients (blue, right-side peak) (N = 1217). Light green dots show the day-100 cumulative incidence of grades III-IV acute GVHD within deciles across the respective ranges of genome-wide recipient mismatching in sibling and unrelated recipients.

Figure 2.

Parental chromosomal segregation and meiotic recombination account for variation in genome-wide recipient mismatching between siblings. (A) Differential inheritance of paternal and maternal chromosomes causes mismatching between sibling pairs. If a pair of paternal chromosomes is designated “a” and “b” and the respective pair of maternal chromosomes is designated “c” and “d,” the chromosomes inherited by the offspring in the absence of meiotic recombination will be “ac,” “bc,” “ad,” or “bd.” Inheritance of the same chromosome pairs by a sibling pair (eg, “ac”) as shown for siblings 1 and 2 is described as IBD with an identity by state (IBS) value of 2, indicating identity for both chromosomes. Inheritance of 1 identical chromosome and 1 nonidentical chromosome by a sibling pair (eg, “ac” and “ad”) is described as partial IBD with an IBS value of 1. Inheritance of 2 nonidentical chromosomes by a sibling pair (eg, “ac” and “bd”) is described as having an IBS value of 0. (B) Meiotic recombination introduces different IBS values across segments of chromosomes inherited by sibling pairs. The figure shows results for 3 sibling pairs with high (top), medium (middle), and low (bottom) genome-wide mismatching across chromosome 6, analyzed as described by Roberson and Pevsner. HLA-matching of these pairs is indicated by IBS values of 2 (green) across the MHC region. Genome-wide mismatching increases as the length of segments with IBS values of 2 decreases. Segments with IBS values of 0 (red) and 1 (blue) have different probabilities of recipient mismatching (see supplemental Appendix for details).

Figure 3.

Genome-wide recipient mismatching appears to have limited effects on the risks of GVHD-related outcomes. HRs (diamonds) and 95% CIs (lines) show GVHD-related outcomes per 0.01 increment of genome-wide recipient mismatching for coding SNPs among (A) sibling recipients and (B) unrelated recipients, adjusted for HLA-DPB1 T-cell epitope matching. Recipient mismatching was counted as described in Figure 1. Results were similar when recipient mismatching was counted as 0, 1, or 2 for each SNP, corresponding to no mismatching, mismatching for a single allele, or mismatching for both alleles, respectively.

Figure 4.

The increased GVHD risk after unrelated HCT is predominantly an effect of HLA-mismatching. Panels compare the cumulative incidence frequencies of GVHD-related outcomes for sibling recipients, HLA 16/16–matched unrelated recipients, and HLA-DP GVH-mismatched unrelated recipients. Results are shown for (A) grades II-IV GVHD, (B) grades IIb-IV GVHD (see “Methods”), (C) grades III-IV GVHD, (D) stages 2-4 gut GVHD, (E) extensive chronic GVHD, and (F) the competing risks of nonrelapse mortality or recurrent or progressive malignancy. The higher incidence of competing risks in the HLA 16/16–matched unrelated recipients is entirely attributable to a higher incidence of recurrent or progressive malignancy (see Figure 5A). DP, HLA-DP; URD, unrelated donor.

Figure 5.

HLA-DP-mismatching has a large effect on the risks of GVHD-related outcomes. HRs (diamonds) and 95% CIs (lines) show GVHD-related outcomes for (A) HLA 16/16–matched unrelated recipients and (B) HLA-DP GVH-mismatched unrelated recipients as compared with sibling recipients. HRs for GVHD-related endpoints were adjusted for the risk factors listed in Table 1, and the HRs for recurrent or progressive malignancy were adjusted for low-risk disease (chronic myeloid leukemia in chronic phase and refractory anemia), high-risk disease (malignancy not in remission), and intermediate-risk disease (all others).