Antigen recognition by CD8+T cells is governed largely by the pool of peptide antigens presented on the cell surface in the context of human leukocyte antigen-class I complexes (HLA-I). We, and others, have recently shown that a high proportion of presented peptides are generated through proteasome-mediated splicing of non-contiguous regions of proteins to form novel peptide antigens.
We have developed a novel workflow for the identification of spliced peptide antigens and applied it to HLA-I peptides derived from multiple cancer cell lines including melanoma, triple-negative breast cancer, mesothelioma, DIPG and colon cancer, as well as mouse models of breast cancer and melanoma. Surprisingly we found 20-30% of peptides presented by HLA-I are generated through post-translational peptide splicing. Of note, in each sample 100-300 spliced peptides were derived from cancer-associated antigens (CAA), thus are cancer-specific spliced peptides. Moreover, a subset of these CAA spliced peptides demonstrated immunogenicity in unrelated patients. Vaccination of mice with cancer-specific spliced peptides significantly reduced the size of tumours, whilst clinical studies for vaccination of melanoma patients are ongoing.
These observations highlight the breadth and complexity of the repertoire of immunogenic peptides available for exploitation therapeutically and suggest that spliced peptides may be a major class of tumour antigens. Indeed, we found spliced peptides may yield more immunogenic epitopes than genome template viral/cancer peptides and, in the absence of high-affinity CAA-derived HLA ligands, can generate higher affinity neoepitopes for interaction with host HLA allomorphs. Understanding the nature and abundance of spliced peptides has high relevance for the discovery of novel targets of T cell immunity and will have significant implications for further immunotherapeutic approaches in cancer and other diseases.