HLA-DR and DQ matching by DNA restriction fragment length polymorphism methods and the outcome of mixed lymphocyte reaction tests in unrelated bone marrow donor searches. The IMUST Study.

The use of DNA restriction fragment length polymorphism (DNA-RFLP) typing for HLA-DR and DQ genes was assessed in 96 patients who were HLA-A,B and DR matched by serology with one or more potential unrelated marrow donors (UD). Two hundred recipient-donor pairs from 10 transplant centres in the UK were studied. DNA-RFLP revealed serological errors in HLA-DR typing and identified additional recipient-donor mismatches. Of the 200 allegedly serologically matched pairs, 55 (28%) were mismatched for HLA-DR and/or DQ by DNA-RFLP typing. Of the 200 pairs, 68 (34%) mixed lymphocyte reactions (MLR) were negative and 132 (66%) positive. There was a significant correlation between DNA-RFLP defined mismatch and MLR positivity. However quantitative studies revealed no trend towards increasing MLR relative response index (RRI) with cumulative RFLP defined mismatch (i.e. for DR plus DQ compared with DR or DQ alone). RFLP matching for HLA-DR and DQ failed to predict a negative MLR. The RRI in the graft-versus-host direction was greater in RFLP matched pairs who carried HLA-DR4 than in matched pairs lacking DR4, suggesting that failure of RFLP to characterize DR4 subtypes was one reason why routine prediction of a negative MLR was not possible. Despite these limitations we have shown that DNA-RFLP is a reliable method for HLA-DR, DQ typing in routine UD searches in diverse clinical centres. By reducing the number of UD tested in MLR, RFLP typing can substantially reduce the work involved in donor selection.