Quantification and Identity of Neoantigens in Cutaneous T-cell Lymphoma

Abstract

Background Mycosis fungoides (MF) is the most common cutaneous T-cell lymphoma (CTCL). Early stage disease, limited to skin patches & plaques, is indolent. However, progression to tumour stage disease marks a significant drop in survival to less than 4 years. Treatments for advanced disease are palliative and MF remains incurable. Given evidence that CTCL is immunogenic, immunotherapies are a promising avenue for newer treatments. However, trials of immune checkpoint inhibitors in MF have shown low to moderate response rates. The prevailing understanding of MF is that it is a highly mutated tumour demonstrating suboptimal response to immunotherapies. Neoantigens are predictors of immunotherapy response, but they have never been studied in CTCL. There is also no knowledge of neoantigen clonality in MF. Clonal tumours, comprised of genetically identical cells expressing the same neoantigen, are more susceptible to immunotherapies compared to subclonal tumours. There is an urgency in identifying neoantigens in MF given the mortality associated with advanced disease and the lack of significant advances in treatment. Objective The objective of this thesis was to gain insight into the immunogenicity of MF by characterizing the tumour mutation burden, identifying neoantigens for the first time, comparing neoantigens between disease stages and determining neoantigen clonality. Our hypothesis is that a high tumour mutation burden in MF will result in a large number of neoantigens. However, these neoantigens are likely to be mostly subclonal, rendering them less responsive to immunotherapies. Methods We tested our hypothesis through a translational study utilizing bioinformatics. We obtained whole exome and whole transcriptome sequences from 24 MF samples (16 plaque, 8 tumour) from 13 patients. We used bioinformatics software (Mutect2, OptiType, MuPeXi) for mutation calling, HLA typing, and neoantigen prediction respectively. Using PhyloWGS, we subdivided malignant cells into stem and clades, to which neoantigens were matched to determine their clonality. Results We demonstrated that MF has a much higher tumour mutation burden than previously described (median 3217 nonsynonymous mutations). We found that MF expressed a significant number of total neoantigens (median 1309 per sample) including a significant number of neoantigens with high binding strength to MHC (median 328). In early disease most neoantigens were clonal but with stage progression, 75% of lesions had >50% subclonal neoantigens. There was very little overlap in neoantigens across patients or between different lesions on the same patient, indicating a high degree of heterogeneity. Conclusions The neoantigen landscape of MF is characterized by a high neoantigen load. Disease progression is associated with an increase in the number of neoantigens and an increase in neoantigen subclonality. Neoantigen subclonality may be the critical factor limiting the efficacy of immunotherapy in patients with advanced disease. Stratification of patients by neoantigen load and clonality may be useful to select suitable candidates for immune checkpoint inhibitor trials.