A comparison of cytotoxic T lymphocyte precursor frequencies in responder/stimulator pairs with increasing degrees of mismatch

Limiting dilution assays of helper or cytotoxic T cell precursor (HTL-p and CTL-p) frequencies have been developed to measure, with high sensitivity and specificity, donor/recipient alloreactivity before bone marrow transplantation (BMT). Most studies demonstrated a significant correlation between the frequency of donor host-specific CTL-p before transplant and the incidence of severe (≥ grade II) acute graft-versus-host disease (aGvHD) in unrelated donor BMT. Recently we demonstrated that the CTL-p assay was sensitive enough to provide similar predictive information also in HLA-identical sibling BMT. In the present study, the CTL-p frequencies were compared in four groups of responder/stimulator pairs with different degree of mismatch: (1) monozygotic twins (MZ) (n=7), (2) HLA-identical sibling pairs (SIB) (n=67), (3) HLA-matched unrelated pairs (HLA-A, B serologically and HLA-DRBI allele-matched) (MUP) (n=59), (4) 2-4 antigen HLA-mismatched pairs (MM) (n=27). Logistic regression analysis showed that as mismatch increased, the probability of a high (>1:100.000) CTL-p frequency (HF) significantly increased. Then, MUP group was divided in two subgroups: (3a) well-matched pairs (WM) (n=20): matched for HLA-DRB3, B4, B5 alleles and, by DNA heteroduplex analysis, for HLA-A, B genes; (3b) matched pairs (M) (n=39): with either a single HLA-A, B mismatch as revealed by DNA heteroduplex analysis, or a single mismatch in HLA-DRB3, B4, B5 alleles. The CTL-p frequencies were analysed again in the five groups, re-using a logistic regression model. The probability of a HF response was significantly higher in the M than in SIB group. Otherwise, WM and SIB groups had the same probability of a HF response. The probability of a HF response in the M group was significantly higher than in WM group, and lower, though not significantly, than in MM group. The results confirmed not only the power of CTL-p assay in the identification of functionally significant polymorphisms, but also in measuring the degree of mismatch in responder/stimulator pairs. They also showed that when unrelated pairs were more closely matched, CTL-p frequencies approached those of SIB. In conclusion, the CTL-p assay may be an important "in vitro" tool in BMT helping also to predict alloreactivity due to minor histocompatibility antigen differences, which still remain difficult to define despite the improvements in unrelated donor matching. © 2001 Blackwell Science Ltd,.