New molecular insights into HLA immunogenicity

Alloreactivity is the major barrier to solid organ and hematopoietic stem cell transplantation (HSCT), determining important clinical events such as graft rejection and graft versus host disease. This biological phenomenon is due to the immunogenicity of allogeneic Human Leukocyte Antigens (HLA) expressed in transplanted tissues and recognized by T cell receptor (TCR) on the surface of alloreactive T cells. In the past, the hosting laboratory identified an immunogenic T cell epitope (TCE) encoded by a subset of HLA-DPB1 alleles, thereby defining permissive and non-permissive DPB1 mismatches between donor and recipient associated with different clinical outcomes after HSCT. Aim of this thesis is to elucidate the molecular basis underlying the different clinical impact of DPB1 mismatches in order to improve our understanding of HLA immunogenicity in the context of transplantation. To achieve this aim, we investigated if the clinical effect of DPB1 mismatches was reflected by the strength of their in vitro alloreactive T cell response and, subsequently, we characterized the alloreactive response to one of the most immunogenic DPB1 alleles, HLA-DPB1*09:01 (HLA-DP9), at the molecular level by homology modeling driven site directed mutagenesis. The strength of the alloreactive T cell response to HLA-DP was assessed in Mixed Lymphocyte Reactions between healthy responder/stimulator pairs mismatched only at the DPB1 locus, followed by quantitative assessment of responder T cells upregulating the cell surface activation marker CD137. We observed significantly higher frequencies of activated T cells against non-permissive mismatches (n=9; mean 10.13% ± 7.51%) compared to permissive mismatches (n=15; mean 2.34% ± 2.82% p<0.05). SDM of HLA-DP9 was performed on 6 polymorphic amino acid residues predicted to be crucial for interaction with bound peptide (positions 9, 35, 55, 69, 76 and 84), and on 2 amino acid residues putatively involved in direct interaction with the TCR (positions 56 and 57). These residues were mutated into amino acids naturally occurring in other HLA-DP variants. A lentiviral vector expression platform was used to express the wild type or mutant HLA-DP in 2 HLA homozygous reporter B lymphoblastoid Cell lines. A panel of 8 alloreactive T cell effectors, specific for wild type HLA-DPB1*09:01 (n=6) or for DPB1*10:01 and DPB1*1701, respectively, but crossreactive to DPB1*09:01 (n=2), was used to evaluate the impact of mutagenesis on T cell recognition by gIFN ELISpot or CD107a degranulation assay. For T cells specifically alloreactive to HLA-DPB1*09:01, recognition was influenced by a complex pattern of residues, which was different for each of the 6 effectors studied. In contrast, for the 2 T cells cross-reactive to HLA-DPB1*09:01, only 2 amino acids (positions 69 and 76) had an influence on allorecognition, suggesting the presence of a more restricted set of T cell epitopes in the context of cross-reactivity rather than nominal specificity. These results are consistent with the peptide involvement in alloreactivity to HLA-DPB1*09:01, and underline the complexity of allorecognition which may be one of the mechanisms underlying the immunogenicity of this molecule. Finally, an additional collaborative study with the North Italian Program for Solid Organ Transplantation was aimed at predicting the functional role of amino acid point mutations in newly described naturally occurring variants of HLA-A3 and A32, respectively, by homology modeling. These studies confirmed data from the literature suggesting a relevant putative role of amino acid substitutions at positions 114 and 116, present in a variant of HLA-A3, on the molecular shape and charge of the peptide binding groove. In contrast, point mutations at position 151 and 258 found in two variants of HLA-A32 and A3, respectively, were not predicted to have important functional impacts, due to limited structural or biochemical changes of the groove. In conclusion, our results shed new light on the molecular basis of T cell alloreactivity. In particular, we showed that the stronger immunogenicity of non-permissive DPB1 mismatches is reflected by a divergent and complex T cell epitope repertoire with a putative involvement of the peptide repertoire presented to the alloreactive TCRs. Identification of specific allopeptides associated with these responses will allow to further dissect their molecular basis.