Association of Predicted HLA T-Cell Epitope Targets and T Cell–Mediated Rejection After Kidney Transplantation

Rationale & Objective: The relationship between human leukocyte antigen (HLA) molecular mismatches and T-cell-mediated rejection (TCMR) is unknown. We investigated the associations between the different donor HLA-derived T-cell targets and the occurrence of TCMR and borderline histologic changes suggestive of TCMR after kidney transplantation. Study Design: Retrospective cohort study. Setting & Participants: All kidney transplant recipients at a single center between 2004 and 2013 with available biopsy data and a DNA sample for high-resolution HLA donor/recipient typing (N = 893). Exposure: Scores calculated by the HLA matching algorithm PIRCHE-II and HLA eplet mismatches. Outcome: TCMR, borderline changes suggestive of TCMR, and allograft failure. Analytical Approach: Multivariable cause-specific hazards models were fit to characterize the association between HLA epitopes targets and study outcomes. Results: We found 277 patients developed TCMR, and 134 developed only borderline changes suggestive of TCMR on at least 1 biopsy. In multivariable analyses, only the PIRCHE-II scores for HLA-DRB1 and HLA-DQB1 were independently associated with the occurrence of TCMR and with allograft failure; this was not the case for HLA class I molecules. If restricted to rejection episodes within the first 3 months after transplantation, only the T-cell epitope targets originating from the donor's HLA-DRB1 and HLA-DQB1, but not class I molecules, were associated with the early acute TCMR. Also, the median PIRCHE-II score for HLA class II was statistically different between the patients with TCMR compared to the patients without TCMR (129 [IQR, 60-240] vs 201 [IQR, 96-298], respectively; P < 0.0001). These differences were not observed for class I PIRCHE-II scores. Limitations: Observational clinical data and residual confounding. Conclusions: In the absence of HLA-DSA, HLA class II but not class I mismatches are associated with early episodes of acute TCMR and allograft failure. This suggests that current immunosuppressive therapies are largely able to abort the most deleterious HLA class I-directed alloimmune processes; however, alloresponses against HLA-DRB1 and HLA-DQB1 molecular mismatches remain insufficiently suppressed.