Molecular phenotypes behind patient stratification based on TP53 and HPV status and potential for treatment personalization with first single-cell RNA sequencing atlas of penile cancer.

Abstract

8 Background: TP53 mutations and Human papillomavirus (HPV) have been suggested to impact penile squamous cell carcinoma (pSCC) prognosis and may predict treatment outcomes. Insights in differences in tumor micro-environment at single cell resolution may benefit personalized therapeutic development and are currently unavailable. Methods: We performed single cell RNA sequencing and targeted next generation sequencing (for TP53 loss-of-function mutations; p53LOF) on primary HPV-associated and HPV-independent (HPV+/-) pSCC lesions or healthy control samples of the coronal sulcus skin. Cell clusters were annotated based on known marker genes. Gene Set Enrichment Analysis (GSEA) and slingshot trajectory analysis was performed using R/Seurat package. Results: Twenty-two samples were analyzed: 4 tumors harbored p53LOF mutations (1 HPV+ p53LOF, 3 HPV- p53LOF), 6 HPV- p53WT, 6 HPV+ p53WT and 6 healthy controls. A total of 84.394 cells was annotated and we identified 68 distinct cell subtypes in pSCC. In epithelial cells, two distinct pseudotime trajectories stemming from the basal cells were identified characterized by either signals of partial epithelial-mesenchymal transition (pEMT; TGFB, ZEB2, SNAI2) or cornification ( ELF3, KRT19, PPL). Whilst the pEMT trajectory was enriched in p53LOF samples, cornification was enriched in HPV+ p53WT. HPV- p53WT samples displayed increase in metabolic changes and stemness hallmarks ( SOX2, HEY1, ZEB1). GSEA confirmed these findings and showed upregulation of pEMT gene sets in p53LOF; fatty acid metabolism and xenobiotic metabolism in HPV- p53WT, and cornification in HPV+ p53WT. We observed an immune suppressive environment in p53LOF patients with higher regulatory T-cell abundancy, and this correlated with abundancy in the epithelial pEMT clusters. Pseudo-bulk analysis on all T-cell subsets demonstrated a higher expression of inhibitory immune checkpoints ( PD-1, CTLA4, TIM3 and LAG3) in p53LOF and HPV- p53WT while HPV+ p53WT patients displayed higher abundancy of immunomodulatory and cytotoxic natural killer (NK) cells. Conclusions: We established the first comprehensive pSCC transcriptomic atlas at single cell resolution and illustrate differences in molecular phenotypes of pSCC according to TP53 mutation- and HPV status. p53LOF tumors are characterized by pEMT, which is a known driver of invasiveness, therapy resistance and metastasis. HPV- p53WT display hallmarks of stemness, which is related to and may explain chemotherapy-resistance often observed in these patients. p53LOF and HPV- p53WT display immune-evasion and exhaustion, while HPV+ p53WT are characterized by abundant cytotoxic innate-like lymphocytes. These observations provide deeper insights into cancer biology that will help advancing pSCC diagnosis and therapy.