A multicenter kidney transplantation study identifies hierarchy and directionality of HLA-DQ alloimmune responses.

HLA-incompatibility between recipient and donor is conventionally assessed by counting serologic mismatches. More recently, tools quantifying differences at the amino acid level are advocated to prognosticate risk for generation of de-novo donor-specific antibodies (dnDSA). Using hierarchical clustering based on HLA evolutionary divergence and amino acid physiochemical unique qualities to evaluate HLA-DQ αβ-heterodimers, we demonstrate that HLA-DQ alleles segregate into two main evolutionary clades, and further subdivide into six functional sub-categories. Applying this biologically grounded immune risk-stratification approach to a cohort of 3003 kidney allograft recipients from five transplant centers shows that transplantation across the two evolutionary clades is associated with hierarchy and directionality of humoral immune responses, despite similar values when applying quantitative metrics. Our three-tier risk stratification model identifies patients at higher risk to develop dnDSA (HR:6.77; 95%CI:4.64-9.90), and antibody mediated rejection (HR:4.41; 95%CI:1.44-13.53). This work provides a thoughtful immunologic framework, that outperforms quantitative metrics, for alloimmune risk stratification.