Impact Of HLA Allele and Haplotype On Acute Graft-Versus-Host Disease and Survival After Hematopoietic Stem Cell Transplantation From Unrelated Donor

Although the effect of allele matching of each HLA locus on the clinical outcome of unrelated hematopoietic stem cell transplantation (UR-HSCT) has been characterized, the effect of HLA allele or haplotype (HP) itself has not been well elucidated. The HLA region is recognized as one of the most important genetic regions associated with human disease, especially autoimmune and infectious diseases. We therefore hypothesized that the immunological response and the clinical outcome following UR-HSCT depend not only on HLA allele matching but also on the HLA allele itself or HLA-linked genetic background of the patient and donor.

We analyzed 5237 patients who received T-cell-replete bone marrow transplants from serologically HLA-A, -B, and -DR antigen-matched unrelated donors facilitated by the Japan Marrow Donor Program between 1993 and 2008. HLA-A, -B, -C, -DRB1, -DQB1, and -DPB1 alleles were retrospectively genotyped. HLA allele frequencies were calculated by direct counting, and multi-locus HLA HP frequencies were estimated using the maximum-likelihood method with EM algorithm of PyPop software. Patients were stratified by HLA-matching status into a full match (FM) group (12/12-matched, n=733) and a mismatch (MM) group (≤11/12-matched, n=4504). The effect of HLA alleles or HPs with a frequency greater than 5% on acute graft-versus-host disease (aGVHD) and overall survival (OS) was analyzed using a multivariate competing risk regression model. The results are expressed as hazard ratios (HRs) comparing specific allele/haplotype-positive group to -negative group.

For each allele, the number of HLA alleles significantly associated with aGVHD (p <.01) in the MM group, were as follows: HLA-A (1 of 10), HLA-B (2 of 17), HLA-C (3 of 15), HLA-DRB1 (1 of 17), HLA-DQB1 (1 of 11) and HLA-DPB1 (0 of 10). In contrast, only one HLA-DPB1 allele was significantly associated with aGVHD in the FM group. The following patient and donor HLA alleles were significantly associated with a reduced risk of aGVHD in the MM group: HLA-A*33:03, C*14:03, B*44:03, DRB1*13:02, and DQB1*06:04. These alleles are located on a common HP (HP-P2) in the Japanese population, which showed a similar effect on grade II-IV (n=534; HR 0.79; p=.001) and III-IV (HR 0.70; p=.004) aGVHD. Strong linkage disequilibrium (LD) hampered determination of the allele responsible for the reducing risk of aGVHD. A significant association with an increased risk of grade III-IV aGVHD and a poor OS was observed in patient HLA-B*51:01 (n=756; aGVHD: HR 1.51, p<.001; OS: HR 1.19, p=.003, respectively) and donor HLA-B*51:01 (n=773; HR 1.46, p<.001; HR 1.15, p=.015) , patient HLA-C*14:02 (n=599; HR 1.55, p<.001; HR 1.19, p=.007), and donor HLA-C*15:02 (n=226; HR 1.62, p<.001; HR 1.38, p=.001) in the MM group. HLA-B*51:01 demonstrated strong positive LD with HLA-C*14:02 and -C*15:02. A significant association with an increased risk of grade III-IV aGVHD and a poor OS was also observed in patient HLA-C*14:02-B*51:01 (n=586; HR 1.52, p<.001; HR 1.19; p=.007) and donor HLA-C*15:02-B*51:01 (n=106; HR 1.98, p<.001; HR 1.53, p=.001). HLA-DPB1*04:02 was the only allele associated with an increased risk of grade II-IV aGVHD in the FM group (n=173; HR 1.64; p=.001). HLA-DPB1*04:02 was linked to two distinctive extended HPs, and the effect of these HPs on aGVHD was stronger in the patients with HLA-DRB1*04:05-DQB1*04:01-DPB1*04:02 (n=60; HR 2.15; p<.001) than in those with HLA-DRB1*01:01-DQB1*05:01-DPB1*-DPB1*04:02 (n=125; HR 1.40; p=.035). HLA-DRB1*04:05-DQB1*04:01-DPB1*04:02 was also significantly associated with poor OS in the FM group (HR 1.65; p=.01). HP-P2 showed a tendency to reduce the risk of grade II-IV aGVHD in the FM group (n=119; HR 0.70; p=.075).

Patient- and donor-specific HLA alleles and HPs themselves contribute to the risk of aGVHD and survival after UR-HSCT. In addition to HLA-B*51:01 being strongly associated with Bechet’s disease, we found this allele to be associated with an increased risk of aGVHD in UR-HSCT. Given that different HLA alleles and HPs were identified in the FM and MM groups, multiple mechanisms, including HLA-mismatch induced alloreactivity, might be involved in the development or exacerbation of aGVHD. These findings suggest that, in addition to HLA-matching status, consideration of patient and donor HLA alleles and haplotypes will provide predictive risk factors for UR-HSCT.

No relevant conflicts of interest to declare.