Peptide-Dependency of Alloreactive T Cell Clones after HLA-Mismatch Hematopoietic Stem Cell Transplantation.

Donor-patient disparities at single or multiple HLA alleles increase the risks of graft rejection or severe graft-versus-host disease, which are crucial determinant for the outcome of hematopoietic stem cell transplant (HSCT). Cytotoxic T lymphocytes (CTLs) recognizing the mismatched major histocompatibility antigens, defined as alloreactive CTLs (allo-CTLs), are predominant mediators of the vigorous allo-response. In mice, allo-CTLs most probably recognize the peptide/MHC complex with a certain peptide dependency. However, in humans, the peptide dependency of those allo-CTLs is still unknown. To assess the peptide dependency of allo-CTLs after HLA-mismatch HSCT, we isolated CTL clones from a post-transplant patient who underwent HLA-mismatch HSCT. The donor was HLA-B One locus mismatch in GVH-vector (Recipient: B*5101, Donor: B*5201). Nine of 10 CTL clones recognized the COS cells transfected with HLA-B*5101 but not HLA-B*5201, and lysed donor-derived lymphoblastoid cells transfected with HLA-B*5101, suggesting these 9 CTLs were alloreactive. These CTLs showed different TCR-Vβ gene usage and variable region sequence, respectively. To address a possible contribution of peptides on the mismatched HLA molecule for the recognition by allo-CTLs, we generated the TAP-deficient cell lines, T2 and .174 cell lines, stably transfected with HLA-B*5101 molecule and tested as a target for a cytotoxicity assay by the allo-CTLs. All of the 9 allo-CTLs did not lyse T2-B*5101 and .174-B*5101 either, although the expression of HLA-B*5101 was confirmed >95% by flow cytometry analysis. This suggests that some particular peptides are likely to have importance for the recognition of allo-CTLs. The difference of amino acid sequence between HLA-B*5101 and B*5201 are only two amino acids at position 63 (B*5101:Asn, B*5201:Glu) and at position 67 (B*5101: Phe, B*5201: Ser). To examine which amino acid substitution makes the different recognition by these CTLs, we generated two chimeric genes termed Asn63Glu and Phe63Ser, transfected into COS cells, co-cultured with the CTLs and examined the IFN-γ release. All of the CTLs did not recognize COS cells with Phe67Ser, and 5 out of 9 CTLs recognized COS cells transfected with Asn63Glu. Because both the amino acid position 63 and 67 were localize on the α1-helix of peptide binding groove, the one amino acid substitution likely affects more local conformation around peptide binding groove. Substitutions localized on the α3-domain did not alter IFN-γ production of 9 allo-CTLs at all. Thus, we convinced that the peptide on HLA-B*5101 is essential for the recognition by the allo-CTLs, and allo-CTLs recognize certain peptide/HLA-B*5101 complex in the peptide-dependent manner. It is hypothesized that the strong allo-response after HLA-mismatch HSCT might associate with considerably low peptide specificity of allo-CTLs. To test the possible low peptide specificity of allo-CTLs, we synthesized 10 known peptides already shown to be presented on the HLA-B*5101 as candidate target pepitdes, pulsed onto T2-B*5101 and .174-B*5101, and tested the cytotoxicity by these allo-CTLs. However, all of the allo-CTLs did not lyse any of the peptides/T2- or .174-B*5101. This suggests allo-CTLs recognize the target with a sufficient specificity. In conclusion, a majority of CTLs after HLA-mismatch HSCT were alloreactive, and exhibit peptide-dependent recognition pattern with an extensive peptide-specific manner.