Comprehensive characterization of human papillomavirus methylation and integration sites in cervical intraepithelial neoplasia and carcinoma by targeted, deep next‐generation sequencing. HPV methylation and integration

Invasive cervical cancer (ICC) and its premalignant phase (cervical intraepithelial neoplasia; CIN) can be distinguished by gynecological and pathological examination. However, precancerous lesions are destined to either persist, progress, or naturally regress, but there are no current methodologies to distinguish among these fates. Epigenetic aberrations in human papillomaviruses (HPV) and integration of viral DNA into the host genome are hallmarks of cervical carcinogenesis, yet they remain to be described in the full disease spectrum or less common HPV types. In this study, we investigated HPV methylation and integration in CIN and ICC samples harboring various HPV types. We used a novel HPV--]enrichment approach with next--]generation DNA sequencing of bisulfite- and non-bisulfite-treated samples. This approach allowed the screening of a large number of samples for complete methylation and integration profiles of up to 63 HPV types. Our data demonstrate that global viral methylation, across many HPV types, is significantly higher in ICC than CIN. Further, we show that methylation levels at individual CpG sites as a quantitative classifier achieves a sensitivity and specificity of >95% for detecting ICC in HPV 16-positive samples. Our HPV integration analysis revealed viral insertions favoring intronic regions of coding and non--]coding genes and highlighted 8q24.21, 3q28, and 17q23.1 as the most prevalent regions of integration in our samples. Interestingly, many integration sites occurred within or near noncoding RNA. Our results provide invaluable insights into methylation differences between CIN and ICC as well as into the sites where the virus integrates into the host genome.