HLA anchor optimization of the melan-A-HLA-A2 epitope within a long peptide is required for efficient cross-priming of human tumor-reactive T cells.

The uptake and long-term cross-presentation of tumor Ag long peptides (LP) by dendritic cells (DC) make them attractive cancer vaccine candidates. However, it remains to be established whether LP can prime long-lived tumor-reactive CTL and whether other cell types are able to cross-present them. Using HLA-A2 healthy donor and melanoma patient-derived PBMC, we studied the in vitro cross-priming potential of Melan-A 16-40 LP bearing the HLA-A2-restricted epitope 26-35 or its analog 26-35(A27L) and compared it to the priming capacity of the short analog. We then addressed LP priming capacity in vivo using HLA-A2 mice. We also studied LP cross-presentation by monocyte-derived DC, plasmacytoid DC, monocytes, and B cells. We showed that the modified LP gave rise to high and sustained cross-presentation by monocyte-derived DC. This led to cross priming in vitro and in vivo and to expansion of long-lived tumor-reactive cytotoxic T cells. In contrast, the LP containing the natural 26-35 epitope primed specific T cells poorly, despite its long-lived cross-presentation, and T cells primed against the short analog were short-lived. We further showed that LP cross-presentation is restricted to monocytes and conventional DC. These results document for the first time, to our knowledge, the strong immunogenicity of a human tumor Ag LP. Of note, they underscore that this property is critically dependent on sufficient HLA binding affinity and/or TCR ligand potency of the cross-presented epitope. We conclude that LP fulfilling this requirement should be used as tumor vaccines, together with DC maturating agents, especially the Melan-A 16-40(A27L) LP, for the treatment of HLA-A2(+) melanoma patients.