O02 Neoantigens from actinic keratosis are predicted to be more immunogenic than those from cutaneous squamous cell carcinoma – a strategy for immune escape?

Abstract

Abstract Actinic keratosis (AK) is a precursor of cutaneous squamous cell carcinoma (cSCC). More than 70% of cSCC arise from AKs, but fewer than 0.6% of AKs progress to cSCC. We have shown that the molecular landscapes of AK and cSCC are similar in terms of tumour mutational burden (TMB), copy number burden, mutational signatures and driver gene mutations but differ in levels of intrasample heterogeneity. We have also demonstrated that alterations in signalling pathways, including transforming growth factor-β, may be important for AK-to-SCC progression. However, the pathomechanisms underlying progression are not well characterized and it remains unclear why the majority of AKs do not progress, and most will regress. There is some evidence that differences in the neoantigen landscapes of AK and cSCC may play a role, and a previous study examining neoantigens in AK and cSCC from nine patients was the first to suggest that there is downregulation of the expression of highly immunogenic neoantigens in the immune escape of cSCC. In this study we further explore, in a larger data set, the hypothesis that the neoantigen recognition potential of AK vs. cSCC is a key determinant of AK-to-cSCC progression. We generated exome and transcriptome data from AK (n = 5) and cSCC (n = 15) with matched normal exome data (from blood). Human leucocyte antigen typing was performed using the OptiType algorithm and neoantigens restricted to major histocompatibility complex (MHC) class I were predicted using the NeoPredPipe workflow. Fitness of neoantigens, defined as the overall effectiveness of a neoantigen at stimulating an immune response, was determined by combining measures of MHC binding affinity and T-cell receptor recognition potential for the predicted neoantigen, and an overall neoantigen recognition potential score was derived for each neoantigen. The difference in median total neoantigen burden for AK and cSCC was not statistically significant [11 224 (range 30–17 330) and 1680 (range 3159–67 357), respectively (P = 0.07)]. Neoantigen burden was strongly correlated with TMB in both AK and cSCC. We found that, compared with cSCC, AK had predicted neoantigens with significantly higher mean MHC-binding affinity (P = 0.049), mean T-cell receptor recognition potential (P = 0.0093) and overall neoantigen recognition potential (P = 0.02). Taken together, our results confirm that neoantigens in AKs are predicted to be more effective at eliciting an immune response than those in cSCC. This suggests there may be a selection pressure in AKs for mutations that result in fewer immunogenic neoantigens resulting in immune escape and progression to cSCC. Finally, our results underscore the critical requirement for an intact immune system in prevention of AK-to-cSCC progression.