Investigating Molecular Drivers of Human Papillomavirus (HPV)-Related Oropharyngeal Cancer

Abstract

There has been a growing epidemic of human papillomavirus (HPV)-induced oropharyngeal squamous cell carcinoma (OPSCC) for the last few decades. This is an underappreciated disease that can have a devastating effect on the lives of otherwise healthy young patients. Patients with HPV-positive OPSCC generally have a good prognosis, but still 20-30% fail to respond to therapy or later recur for unknown reasons. Our hypothesis is that the type of interactions between viral DNA and the human host DNA may determine a patient’s disease progression. HPV can remain its circular episomal shape, but often linearizes and integrates into the human genome, either into intergenic loci or into genes. The process of HPV integration is of particular interest as a potential driver of HPV-positive OPSCC because it is thought to be a marker of disease progression in cervical cancer and is reported in a large proportion of head and neck tumors with estimates ranging from 50-70%.

Integration can lead to large structural variations, disrupt cellular genes and alter gene expression both locally and genome wide, but the exact effects of HPV integration on OPSCC progression are unclear. Previous studies that assessed survival differences between HPV integration positive versus negative patients demonstrated mixed results, so we aimed to clarify whether this process impacts patient outcomes. Using a polymerase chain reaction (PCR)-based approach, we found that HPV-positive, integration-positive patients had higher levels of the HPV oncogenes E6/E7 and a survival advantage over HPV-positive, integration-negative patients. The underlying mechanism for this improved outcome is unclear, but this work provides evidence that HPV integration could serve as a prognostic biomarker. We also utilized this methodology to investigate the clonal nature of HPV integration events, as it is clear that this process affects cell biology given the survival differences we discovered, but how cells containing these viral-human fusions may be selected for during tumor evolution is unclear. We explored the clonality of integration events in bilateral HPV+ tonsillar tumors and found evidence that these tumors often form as a result of clonal expansion from one tonsil to another given that we found shared integration sites across samples. These results indicate that integration events provide a survival advantage to tumor cells which are then selected for and expanded such that they are able to metastasize elsewhere.

Finally, to overcome limitations of previous integration calling methodologies, we optimized a new targeted capture sequencing and analysis pipeline called SearcHPV. Through integrated analysis of HPV+ models by SearcHPV and genome-wide linked read sequencing, we demonstrated that HPV integration sites were found not only adjacent to known cancer-related genes such as TP63 and MYC, but also near regions of large structural variation in the tumor genome. Further, analysis of SearcHPV-assembled junction contigs demonstrated that the tool can be used to accurately identify viral-host junction sequences and showed that viral integration occurs through a variety of DNA repair mechanisms including non-homologous end joining, alternative end joining and microhomology mediated repair. Together, these studies highlight HPV genomic integration as an important contributor to cancer progression, and with new tools available, we believe the field is now primed to make major advances in the understanding of HPV-driven pathogenesis, some of which may lead to the development of novel biomarkers and/or treatment paradigms.