Neoantigen-Specific T Cells in a Novel Cutaneous Squamous Cell Carcinoma Model

Abstract

Non-melanoma skin cancer, which includes cutaneous squamous cell carcinoma (cSCC), is the most common cancer. In the US, there are more than a million cases of cSCC diagnosed annually. Although most early stage cSCC are successfully treated with excision, cSCC results in significant morbidity, and approximately 4% of cSCC patients develop metastases and 2% die. Immune checkpoint inhibitors and other immunotherapies are becoming increasingly important in treating multiple cancer types, including cSCC, but only a fraction of patients respond. Yet, there is a paucity of clinically relevant cSCC murine models that can be used to study neoantigen-specific T cell responses. Outbred SKH-1 mice are highly susceptible to solar UV light-induced cSCC. However, an outbred strain limits the ability to evaluate MHC-restricted, antigen-specific T cell responses and perform studies with genetically engineered mice. To address this need, solar UV light was used to induce tumors in inbred FVB/N, BALB/c, and C57BL/6 mice. Solar UV light reliably induced tumors in FVB/N and BALB/c mice. In contrast, C57BL/6 mice were relatively resistant to tumor formation. Most tumors were histologically diagnosed as actinic keratosis, cSCC in situ, or invasive cSCC. Histologically confirmed solar UV-induced invasive cSCC tumors were used to create clonal cSCC cell lines on the BALB/c background. The cSCC cell lines recapitulate the high mutational burden, mutational profile, and driver mutations observed in human disease. These cSCC cell lines constitutively express MHC class I, and thus, can be targeted for destruction by CD8 T cells. T cells constrain the cSCC cell lines, as cSCC tumors grew faster in athymic mice, which lack mature T cells, compared to wild-type mice. In vivo depletion of CD8 and CD4 T cells demonstrated a major role for CD8 T cells and a supportive role for CD4 T cells in controlling cSCC growth. Vaccination with irradiated cSCC cells completely protected mice from tumor challenge, and this response was dependent on CD8 T cells. This supports that MHC class I neoantigens in the cSCC model can mediate tumor destruction. To identify tumor-rejecting MHC class I neoantigens, mutations were prioritized based on the predicted neoantigen: MHC interaction and mRNA expression. ELISPOT was used to evaluate the immunogenicity of the prioritized neoantigens, and IFN-γ-secreting CD8 T cells that recognize mutant Kars and mutant Picalm.2 were identified. Prophylactic vaccination with mutant Kars or mutant Picalm.2 constrained cSCC tumor growth, demonstrating that both mutant Kars and mutant Picalm.2 are tumor-rejecting neoantigens. To identify characteristics that distinguish tumor-rejecting neoantigens, we evaluated the non-immunogenic neoantigens from in the cSCC model and tumor-rejecting neoantigens from the cSCC model and other murine cancer models. Among neoantigens with a low differential agretopic index (DAI), tumor-rejecting neoantigens had a greater solvent accessible surface area (SASA) of the mutated residue compared to non-immunogenic neoantigens. Thus, SASA of the mutated residue may be an important characteristic to consider when prioritizing neoantigens that do not have a large difference in MHC binding compared to the wild-type sequence. As this cSCC model recapitulates human disease, future work using this model may provide insights into the neoantigens to target in neoantigen-based immunotherapy that can improve therapeutic outcomes for patients with cSCC and other high mutational burden cancers.