The predictive value of epitope analysis in highly sensitized patients awaiting renal transplantation.

The accumulation of highly sensitized patients (HSP) on renal transplant waiting lists is a problem faced by all transplant registries. We have studied the HLA class I serological reactivity of 20 random HSP and have related antibody specificity to primary amino acid sequence. In six patients we identified significant correlations (chi 2 test, r > or = 0.93) between panel reactivity and specific amino acid substitutions characteristic of HLA-A, -B, and -C public epitopes. Antibody reactivity was associated with up to three public epitopes in each patient. The 12 separate antibody specificities identified were associated with 10 residues. Seven correlated with HLA-A locus substitutions (Glu-62/Gly-65, Lys-66, Arg-114, His-114/Tyr-116/Lys-127, Thr-142/His-145 [x2], and Thr-149), two with HLA-B locus substitutions (Thr-24, Ser-24) and three with interlocus antibodies associated with either HLA-A and B (Leu-82/Arg-83 [x2]) or with HLA-B and -C substitutions (Leu-163). This information allowed us to predict HLA class I allelic products of known primary sequence that would react negatively with each HSP serum. Windows of acceptable mismatches (WAMMs) can thus be delineated with a view to crossmatch negative transplantation without the need for exhaustive serological analysis. Surprisingly we found that WAMMs for these patients included up to 80% of the 10 commonest HLA class I haplotypes in the British population with four patients being crossmatch compatible with A1,3; B7,8. These observations lead us to propose a more intelligent approach to transplanting HSP based on epitope analysis and definition of WAMMs.