CD62L and CD31 Matching Are Independent Risk Factors for Survival after Bone Marrow Transplantation.

In hematopoietic stem cell transplantation between HLA identical siblings (HLA-id sib), graft versus-host disease (GVHD) and possibly graft anti-tumor activity have been attributed to minor histocompatibility antigen (mHA) disparity between patients and their donors. Because of the lack of human mHA alloantbodies, putative differences between patient and donor mHAs have been identified by genetic analysis of single nucleotide polymorphisms (SNPs) in coding regions of expressed proteins. For example, we and others have previously shown that mismatch at the polymorphic codons in CD31, a cell surface adhesion molecule expressed on endothelial cells and important in leukocyte diapedesis, correlates with an increased incidence of acute GVHD and with poorer survival following HLA-id sib bone marrow transplantation (BMT). Identification of other immunodominant mHAs would provide clinical benefit, as well as insight into the biology of GVHD and graft anti-tumor activity. Because of suggestions in the literature regarding a possible role for CD62L and CD49b as acute GVHD (aGVHD) risk factors, we investigated the polymorphic codons in those adhesion molecules as candidate mHAs. Paired samples of recipient and donor DNA were collected pre-transplantation in 120 consecutive patients receiving unmanipulated bone marrow transplants (BMT) from HLA-id sibs at our institution. Patient age ranged from 1 year to 60 years old, with 40 patients under 18 years old. Fifty-five patients had acute leukemia, 29 chronic myelogenous leukemia, 25 primary bone marrow failure disorders, and 11 lymphoid malignancies. The ablative transplant regimen was total body irradiation (TBI) based in 68 patients and non-TBI based in 52 patients. After an 18 month period of observation, the overall incidence of severe aGVHD (grade III–IV) was 8%, and the actuarial overall survival was 66%. Identity at the CD62L locus (codon 213, Pro versus Ser) was associated with decreased overall survival (60% versus 90%, P = 0.007). CD62L identity also correlated with an increased incidence of severe aGVHD and disease relapse with p values approaching significance. Identity at the CD49b locus (codon 505, Glu versus Lys) had no effect on survival, aGVHD or relapse. Identity between donor and recipient at the CD31 locus (codon 563, Ser versus Asn) was associated with a decreased incidence of severe aGVHD (5% versus 15%, P = 0.003) and increased overall survival (75% versus 55%, P = 0.018) when compared with nonidentity. CD62L identity and CD31 nonidentity remained independent risk factors for improved overall survival when tested in multivariate analysis with the covariates of age, sex, ethnicity, disease, pre-transplantation disease risk assessment, TBI versus non-TBI based preparative regimens, and post-transplantation immunosuppression. Comparison of the best case donor:recipient allelic combination (CD62L nonidentity and CD31 identity) with the worst case combination (CD62L identity and CD31 nonidentity) yielded the highest level of significance for improved actuarial overall survival (90% versus 50%, P = 0.003). These results suggest that the mHA’s CD62L and CD31 but not CD49b may be important determinants of clinical outcomes following BMT. One may speculate that CD62L donor recipient mismatch combined with CD31 matching confers protection against GVHD, and promotes graft anti-tumor activity, thereby improving on overall survival.