Pan-cancer genomic characterization of human papillomavirus associated tumors

PROJECT SUMMARY/ABSTRACT Human papilloma virus (HPV) infection causes an estimated 610,000 human cancers globally every year. HPV accounts for nearly all cervical and anal squamous cell carcinomas, increasing rates of head and neck squamous cell carcinomas, and a subset of other anogenital squamous carcinomas – all with relatively heterogenous clinical

outcomes due to a lack of personalized care. My goal is to use a pan-cancer cohort of patients to define the molecular heterogeneity shared across HPV-associated anatomic sites. Analyses that integrate many genomic variables across patient cohorts, powered by next-generation sequencing, have successfully

ascertained genomic perturbations in individual tumor types. However, few studies have focused particularly on genomic changes across HPV(+) cancers of differing anatomical origin. Additionally, there is little genomic data available to analyze these differences in anogenital cancers. While large scale collaborative efforts

to investigate the molecular landscape of head and neck cancers have been successful, the specific genomic changes and clinical outcomes attributable to HPV infection lack clarity due to smaller HPV(+) cohorts and

confounders like smoking. The HPV lifecycle is tightly linked to differentiation of the epithelial cells it infects. The transformative potential of the oncogenic strains is classically attributed to expression of viral oncoproteins E6 and E7, which subsequently inactivate tumor suppressors TP53 and RB, respectively. However, while persistent

infection with high-risk HPV types and expression of E6/E7 proteins are capable of cellular immortalization in vitro and are necessary in the formation of tumors, these events alone are insufficient for cancer development in vivo. Thus, viral and host genomic changes must occur alongside HPV infection. It is my hypothesis that

the molecular heterogeneity of HPV-associated tumors is shared across anatomic sites and can be used to subclassify these cancers based on therapeutically meaningful pathways. To test this, integrative analyses of HPV(+) and HPV(-) tumors will be employed using hybrid-capture sequencing across extensive

cohorts of cervical, oral, and anogenital tumors. Specifically, paired-end DNA sequencing using probes specific to hundreds of genes commonly involved in cancer will be used to identify potential driver somatic mutations and copy number alterations in these tumors (Aim I). Novel genomic changes determined here will then be examined

in context with other known biologically pathways and molecular perturbations in HPV-associated cancers. Next, probes specific to the HPV genome will be used to assess viral properties such as strain, viral load, and integration events within and across tumor types (Aim II). Finally, the prognostic effect of oncogenic PIK3CA

alterations will be determined through integration of clinical variables and overall survival using multivariate modeling (Aim III). This unique training opportunity and exemplary mentoring environment will ensure my complete development as a physician-scientist in clinical genomics working to identify novel means of

subclassifying patients in order to tailor treatments and improve outcomes.