Limited role of MHC class I chain–related gene A (MICA) typing in assessing graft-versus-host disease risk after fully human leukocyte antigen–matched unrelated donor transplantation

To the editor:

We read the manuscript by Parmar with interest. In an exploratory analysis, these investigators found that mismatching of major histocompatibility complex class I–related chain A (MICA) between donor and recipient was associated with gastrointestinal (GI) acute graft-versus-host disease (aGVHD).¹  Based on the polymorphism of this protein, the constitutive expression on the GI epithelium and the T- and natural killer (NK)–cell immune activating function,^2,3  we too have hypothesized a role for MICA in GI aGVHD. To address this we developed a high-resolution MICA typing method (see supplemental Table 1 and supplemental Figure 1, available on the Blood website; see the Supplemental Materials link at the top of the online article). MICA alleles were determined in 38 donor-recipient pairs after human leukocyte antigen (HLA) 12/12 matched unrelated donor transplantations. One-half of these recipients developed severe (grade II-IV) GI aGVHD, whereas the other half did not. All but one donor-recipient pair were matched at the MICA locus. Thus in this setting, MICA mismatching was rare. This is likely due to tight linkage disequilibrium with HLA B (D′0.99484, P < .001). Given that the single MICA mismatch occurred in a patient without aGVHD, this also suggests that MICA mismatching per se did not uniformly lead to severe GI aGVHD.

In reconsidering our hypothesis, we reasoned that the important interaction might be between the immune receptor NKG2D (on donor T and NK cells) and recipient MICA (on the GI tract). MICA amino acid substitutions at position 129 (methionine or valine) are associated with “weak” or “strong” NKG2D binding.⁴  We thus hypothesized that strong NKG2D:MICA binding might be associated with GI aGVHD. However no association between recipient MICA allele binding affinity and GI aGVHD was observed (not shown). An additional exploratory analysis examined whether certain recipient MICA alleles were associated with GI aGVHD protection or risk. There was a trend toward less aGVHD for recipients with MICA*008 (P = .07). Similarly, a possible increase in severe GI aGHVD was noted in recipients with an amino acid motif encoded by MICA*004, *006, *009, *044, or *049 (P = .075).

Given that MICA is tightly linked to HLA-B, we examined whether certain recipient HLA-B alleles (based on MICA linkage) are associated with GI aGVHD. Using a second cohort of 1676 recipients of myeloablative, HLA 8/8 matched unrelated donor transplantation, we tested whether recipient HLA-B alleles correlated with transplantation-associated outcomes. HLA-B alleles were divided into low and high risk based on linkage to MICA*008 or *004, *006, *009, *044, and *049, respectively (described in Table 1 footnotes). All other HLA-B alleles were considered to have intermediate aGVHD risk. As shown in Table 1, in multivariate analysis there were no differences in transplantation outcomes between the high-, intermediate-, and low-risk groups.

In contrast with the data presented by Parmar et al, our studies do not support the concept that donor-recipient MICA mismatching plays a role in GI aGVHD. In addition, these data also do not suggest that certain recipient MICA alleles are associated with GI aGVHD. While the reasons for the differences between our study and that of Parmar et al are not entirely clear, potential explanations may include registry versus single-center reporting, high- versus low-resolution MICA typing, HLA disparity in the patient cohorts (matched vs potentially mismatched), and differing experimental approaches. As called for by Parmar et al, confirmation of the role of MICA in GI aGVHD in a larger cohort of patients is necessary, particularly using high-resolution typing methods.