HPV-Associated transcriptional landscape in cervical and head and neck cancers

Human papillomavirus (HPV) is the principal cause of cervical cancer (CC) and an increasingly significant contributor to head and neck cancer (HNC). Although these malignancies arise at different anatomical sites, they share a common viral aetiology, and the molecular alterations associated with HPV across both cancers remain incompletely understood. The present study aimed to identify transcriptional alterations common to HPV-positive CC and HNC using integrated bulk and single-cell transcriptomic analyses. Differential expression analysis and weighted gene co-expression network analysis were performed on HPV-positive tumours and normal tissues from both cancers to identify shared tumour-associated genes. To distinguish HPV-associated alterations from general tumour-related changes, these genes were further refined by comparison with HPV-positive and HPV-negative tumours in both cancers. The resulting candidate genes were prioritised using a machine learning-based feature selection strategy and subsequently validated in independent transcriptomic cohorts and single-cell RNA-sequencing datasets. A total of 52 shared HPV-associated genes were identified by integrating differential expression and co-expression network analyses, of which six genes (CDKN2A, CDKN2B-AS1, MCM2, KCNS1, EFNB3, and COL28A1) were prioritised by machine learning. The six-gene panel consistently distinguished HPV-positive from HPV-negative tumours in both CC and HNC and showed similar expression patterns in independent validation cohorts as well. Single-cell RNA-sequencing analysis further demonstrated that these genes were predominantly expressed in epithelial cells, with CDKN2A, CDKN2B-AS1, KCNS1, and MCM2 showing increased expression in HPV-positive samples, whereas EFNB3 and COL28A1 were reduced. These genes are associated with biological processes relevant to tumour progression, including cell-cycle regulation, DNA replication, extracellular matrix organisation and cell–cell communication. Correlation with HPV viral transcript levels revealed modest associations for the identified genes, particularly CDKN2A and MCM2. Shared transcriptional alterations were identified across HPV-associated CC and HNC, providing further insight into the molecular landscape of HPV-driven carcinogenesis. These findings offer a foundation for future studies aimed at understanding HPV-associated tumour biology and identifying potential biomarkers.