Abstract
In unrelated donor hematopoietic stem cell transplantation (URD-SCT) patients are preferably transplanted with stem cells from a fully HLA matched donor, usually defined as identical for HLA-class I, -DR and -DQ. Since HLA-DPB1 is often not taken into consideration in donor selection, 80–90% of URD-SCTs are mismatched for HLA-DPB1. The role of HLA-DPB1 as transplantation antigen has been unclear, since clinical reports on the impact of matching for HLA-DPB1 on transplant outcome showed conflicting results. HLA-DPB1 mismatching has been associated with an increased risk of graft versus host disease (GVHD). However, we recently demonstrated that HLA-DPB1 specific T cells can mediate a potent graft versus leukemia effect without inducing GVHD. It has been suggested that the controversial effects of matching for HLA-DPB1 in URD-SCT could partly be explained by the assumption that not all HLA-DPB1 differences are immunogenic. This theory was based on the cross-reactive recognition of two HLA-DPB1* 09 specific T cell clones that recognized other HLA-DPB1 alleles sharing amino acids (aa) in position 8–11 of HLA-DPB1 (Zino et al, blood 2004). It was hypothesized that there would be no induction of T cell responses between individuals expressing HLA-DPB1 molecules sharing this aa sequence. This was translated into a classification of permissive and non-permissive HLA-DPB1 mismatches in order to allow a broader donor selection. To investigate whether cross-reactive recognition of other HLA-DPB1 molecules by our previously generated HLA-DPB1*02 or *03 specific CD4+ T cell clones depended on the presence of specific aa sequences we tested recognition of a panel of 14 EBV-LCL expressing 9 different HLA-DPB1 molecules. All HLA-DPB1*02 as well as all *03 specific T cell clones showed cross-reactivity with other HLA-DPB1 alleles and each T cell clone exhibited its own pattern of cross-reactivity. Two HLA-DPB1*0201 specific T cell clones with different TCR-Vβ showed also recognition of EBV-LCL expressing HLA-DPB1*1001 and *1701 or HLA-DPB1*1001, *0901 and *1601 respectively. Five HLA-DPB1*03 reactive T cells clones with different TCR-Vβ showed differential cross-recognition of EBV-LCL expressing HLA-DPB1*0101, *0601, *1101, *1301 and *1401. To identify immunogenic differences the aa sequences of the HLA-DPB1 molecules recognized by the various T cell clones were compared. The HLA-DPB1 molecules recognized by the HLA-DPB1*02 specific T cell clones shared an aa substitution at position 69 compared to the responder cell. However, HLA-DPB1*0601,*0901 and *1901 with the same substitution were not recognized by both T cell clones. This phenomenon was also observed for the HLA-DPB1*03 specific T cell clones, indicating that the cross-reactive recognition of HLA-DPB1 could not be predicted by aa sequences. Next, we analyzed the immunogenicity of various HLA-DPB1 alleles in different stimulator/responder combinations to verify the classification of permissive and non-permissive mismatches. We developed a model to generate allo-HLA-DP responses by transducing HLA-class II negative HELA cells with various HLA-DP molecules and used these cells to stimulate purified CD4+ T cells from HLA-DPB1 homozygous donors. HELA cells transduced with HLA-DPB1*0101, *0201, *0301, *0401, *0402, *0501, *0601, *0901, *1101, *1301, *1401 or *1701 were used as stimulator cells. Responder CD4+ T cells were typed HLA-DPB1* 0201, *0301, *0401 or *0402. 14 days after stimulation, CD4+ T cells were tested for recognition of the stimulator cells and of HELA cells transduced with the responder HLA-DPB1 molecule as a negative control. For these 4 responders, stimulation with 12 different HLA-DP transduced HELA cell lines resulted in specific IFN-γ production in response to the stimulator cells in 47 out of 48 stimulations. 28 CD4+ T cell lines also showed cross-reactive recognition of HELA cells transduced with at least one other HLA-DPB1 molecule. In conclusion, we showed that cross-reactive recognition of various HLA-DPB1 molecules by HLA-DPB1 specific T cells is a common observation. However, we demonstrated that cross-reactivity between HLA-DPB1 molecules by allo-HLA-DPB1 specific T cells does not exclude the generation of immune response between individuals expressing these HLA-DPB1 molecules. By generating multiple allo-HLA-DP specific T cell lines, we showed that all HLA-DPB1 mismatch combinations are immunogenic.