HLAMMATCHMAKER: A Molecularly Based Donor Selection Algorithm for Highly Alloimmunized Patients

TRANSPLANT candidates are generally consideredhighly alloimmunized if their serum panel-reactive antibody (PRA) activity exceeds 85%. Such patients can be transplanted successfully provided that the donor has no mismatched HLA antigens that react with the patient’s alloantibodies. Serum screening against HLA-typed panels must be done to determine the alloantibody specificity spectrum so that unacceptable HLA antigen mismatches can be identified. This approach, however, does not work if the patient’s serum has a very high PR A. Often, many highly sensitized patients remain on the waiting list with little prospect of a transplant because insufficient informa-tion is available about HLA mismatch acceptability and the probability of finding a zero antigen mismatch is very low. This report shows an alternative strategy for identifying potential donors for highly sensitized patients. HLA-MATCHMAKER is a computer-based algorithm that addresses amino acid sequence polymorphisms as critical components of immunogenic epitopes that can elicit alloantibodies. Such amino acids reside in sequence positions accessible to alloantibodies, namely the a helices and b loops of the protein chain structure. The residues in the strands of the b-pleated sheets of the peptide-binding groove are ex-cluded from this matching algorithm because they cannot make direct contact with alloantibodies. Many investigators have observed the presence of multiple amino acid–en-coded antigenic determinants on the exposed parts of HLA class I molecules. This HLA matching algorithm is based on comparisons of linear sequences of amino acid triplets as motifs for potentially immunogenic epitopes