ARE THE FUNCTIONAL ASSAYS USEFUL IN PREDICTING GRAFT-VERSUS-HOST DISEASE IN HLA-MATCHED SIBLING BONE MARROW TRANSPLANTATION?

Abstract

A reply to the letter from Affaticati et al. There is no doubt that the frequency analysis of helper T lymphocyte precursors (HTLp) and cytotoxic T lymphocyte precursors (CTLp) has great impact on the detection of alloreactivity. However, the debate on the clinical values of these functional assays for the prediction of graft-versus-host disease (GVHD) occurrence after bone marrow transplantation (BMT) has carried on for the last decade. One can easily find several reports in the literature either for (1–4) or against (5–9) the usefulness of these functional assays for predicting GVHD in HLA-matched sibling BMT. Although our observation (9) showed no significant correlation between the HTLp/CTLp frequencies and the incidence of acute GVHD after HLA-matched sibling BMT, there are other reports showing the contrary (5,7). Due to the lack of standard protocols for these functional assays, there are considerable variations in the methodology used by different investigators, which makes it difficult, sometimes impossible, to compare or correlate data from different research groups. For a rigorous examination on the association between the frequencies of HTLp/CTLp and the incidence of GVHD, it is crucial to establish standard protocols and to study a sufficient number of transplant recipients with homogeneity of disease and uniformity of conditioning and GVHD prophylaxis. The effectiveness of the immunosuppressive regimen after transplantation and the precision of the HLA matching also need to be taken into consideration. The discordance in one or more of these factors could be attributable to the controversy regarding the usefulness of HTLp/CTLp frequency estimation for GVHD prediction. During the last few years, the sensitivity of these assays has been improved by eliminating antigen-presenting cells from the responder population and/or T cells from the stimulator population (10,11). However, the complexity of the assay procedure makes these protocols impracticable for routine clinical use, and researchers are still searching for a better method. The combined limiting dilution assay (LDA) used in our study was established by Bouma (12) et al. and has proven to be as sensitive as single LDA to detect alloreactive HTLp and CTLp frequencies in HLA-mismatched settings. Our observation (9) has demonstrated that this combined LDA may not be ideal in HLA-identical sibling settings. The possible reasons have been discussed in detail in our previous articles (8,9). Although Affaticati et al. (13), using a single LDA, showed a positive correlation of CTLp frequency with clinical acute GVHD grades II–IV, the statistical significance remained marginal (P =0.04), with a third of patients who did not develop severe GVHD demonstrating high frequencies. It may well be that the positive CTLp frequencies detected in both publications by Affaticati et al. and ourselves were against minor histocompatibility antigens. Some of these responses detected in cohorts of patients developing no or low-grade GVHD may, however, not be relevant or strong enough in the in vivo clinical setting (in the presence of prophylaxis regimens for example) to give rise to GVHD. However, in our view, CTLp frequency estimation remains an inconclusive predictive indicator of GVHD occurrence and severity in the HLA-matched sibling setting, mainly due to overlapping results of high frequency within cohorts of patients demonstrating either no or low-grade GVHD (grades 0–1). Anne Dickinson Xiao-Nong Wang