Long-read sequencing reveals oncogenic mechanism of HPV-human fusion transcripts in cervical cancer

Integration of high-risk human papillomavirus (HPV) into the host genome is a crucial event for the development of cervical cancer, however, the underlying mechanism of HPV integration-driven carcinogenesis remains unknown. Here, we performed long-read RNA sequencing on 12 high-grade squamous intraepithelial lesions (HSIL) and cervical cancer patients, including 3 pairs of cervical cancer and corresponding para-cancerous tissue samples to investigate the full-length landscape of cross-species genome integrations. In addition to massive unannotated isoforms, transcriptional regulatory events, and gene chimerism, more importantly, we found that HPV-human fusion events were prevalent in HPV-associated cervical cancers. Combined with the genome data, we revealed the existence of a universal transcription pattern in these fusion events, whereby structurally similar fusion transcripts were generated by specific splicing in E6 and a canonical splicing donor site in E1 linking to various human splicing acceptors. Highly expressed HPV-human fusion transcripts, eg, HPV16 E6*I-E7-E1_SD880-human gene, were the key driver of cervical carcinogenesis, which could trigger overexpression of E6*I and E7, and destroy the transcription of tumor suppressor genes CMAHP, TP63 and P3H2. Finally, evidence from in vitro and in vivo experiments demonstrates that the novel read-through fusion gene mRNA, E1-CMAHP (E1C, formed by the integration of HPV58 E1 with CMAHP), existed in the fusion transcript can promote malignant transformation of cervical epithelial cells via regulating downstream oncogenes to participate in various biological processes. Taken together, we reveal a previously unknown mechanism of HPV integration-driven carcinogenesis and provide a novel target for the diagnosis and treatment of cervical cancer.