Cancer-associated fibroblasts (CAFs) are a heterogeneous cell type found in the tumour microenvironment. CAFs can support tumour growth and metastasis and contribute to therapeutic resistance, making them a potential therapeutic target. Here, we aim to identify neoantigens resulting from somatic mutations in CAFs. HLA genotyping, a critical step for neoantigen prediction, can be performed using DNA sequencing data with various tools available. Claeys et al. (PMID:37161318) found that a majority voting approach improved HLA typing performance. No end-to-end pipeline exists to apply this approach, making it difficult for non-informaticians to implement. The objectives of this study were to: 1) develop a Nextflow pipeline implementing majority voting for MHC class I typing from DNA sequencing, and 2) use HLA calls from this pipeline to identify neoantigens in CAFs.

CAFs and corresponding tumour-associated normal fibroblasts (TANs) were cultured from the tissue of 11 patients with breast cancer (10 Luminal A, one triple-negative) and 10 patients with lung cancer (six adenocarcinoma, two squamous cell carcinoma, two of unknown subtype). Whole-exome sequencing (WES) and bulk RNA sequencing were carried out on all samples. Using our pipeline, nf-hlamajority, we carried out HLA typing on our patient WES data and WES data from 12 NCI-60 Human Tumor Cell lines using Optitype, Polysolver, HLA-LA, and Kourami. For each HLA gene, the pipeline assigned the HLA genotype called by the highest number of tools. These HLA genotypes were used as input to Landscape of Effective Neoantigens Software (LENS), along with the WES and RNA-sequencing data, to identify CAF-specific neoantigens (PMID:37184881).

Results from the NCI-60 dataset showed 97% accuracy, with 68 out of 70 HLA calls matching PCR-based genotyping. LENS identified potential neoantigens resulting from missense mutations, with more high-confidence expressed mutations observed in lung cancer CAFs compared to breast cancer CAFs (Welch’s Two Sample t-test, p = 0.017). All missense mutations were private, although two lung cancer CAF samples had a mutation in the COASY gene. Interestingly, this gene and other genes harbouring mutations are implicated in lipid metabolic pathways. CAFs contribute to lipid metabolism within the TME, thus impacting cancer progression and tumour immunogenicity.

In this study, we have developed an automated pipeline for consensus HLA genotyping which we envisage will be useful to the research community. nf-hlamajority has helped us identify candidate neoantigens in breast and lung cancer CAFs. Future work will focus on validation using T-cell immunogenicity assays and investigating the CAF subpopulation distribution of our candidate neoantigens using single-cell RNA sequencing. This will improve our understanding of the potential of targeting CAF neoantigens to enhance the efficacy of anti-cancer therapies.