The skin is exposed to viral pathogens, but whether they contribute to the oncogenesis of skin cancers has not been systematically explored. Here we investigated 19 skin tumor types by analyzing off-target reads from commonly available next-generation sequencing data for viral pathogens. We identified human papillomavirus 42 (HPV42) in 96% (n = 45/47) of digital papillary adenocarcinoma (DPA), an aggressive cancer occurring on the fingers and toes. We show that HPV42, so far considered a nononcogenic, "low-risk" HPV, recapitulates the molecular hallmarks of oncogenic, "high-risk" HPVs. Using machine learning, we find that HPV-driven transformation elicits a germ cell-like transcriptional program conserved throughout all HPV-driven cancers (DPA, cervical carcinoma, and head and neck cancer). We further show that this germ cell-like transcriptional program, even when reduced to the top two genes (CDKN2A and SYCP2), serves as a fingerprint of oncogenic HPVs with implications for early detection, diagnosis, and therapy of all HPV-driven cancers.
Significance: We identify HPV42 as a uniform driver of DPA and add a new member to the short list of tumorigenic viruses in humans. We discover that all oncogenic HPVs evoke a germ cell-like transcriptional program with important implications for detecting, diagnosing, and treating all HPV-driven cancers. See related commentary by Starrett et al., p. 17. This article is highlighted in the In This Issue feature, p. 1.
©2022 The Authors; Published by the American Association for Cancer Research.