Development of a high-resolution mapping platform for HPV DNA integration in premalignant lesions

Abstract The vast majority of human cervical cancers are caused by human papillomaviruses (HPVs). These viruses are also implicated in a fraction of other types of cancer (head & neck, anus, vagina, vulva). Cervical cancer is the second leading cancer cause of death of women worldwide. Over 40 different types of HPV infect the genital tract, and nearly half of the human population is infected by an HPV at least once. However, the vast majority of HPV-infected people infected do not develop invasive tumors due to antiviral immune responses. Cervical carcinomas develop through a series of cervical intraepithelial neoplasia (CIN) steps, CIN1, CIN2, and CIN3, but only a minority of women even with CIN3 progress to invasive carcinomas. The HPV DNA genome replicates as a circular, extra-chromosomal episome with up to many thousands of copies per infected cell. However, in most invasive carcinomas, HPV DNA is integrated into human genomic DNA due to aberrant host cell DNA repair mechanisms. This results in the viral oncogenes (notably E6 and E7) becoming permanently associated with the host cell and its descendents. Usually, the viral DNA is integrated into a human oncogene, often as only a fraction of the viral genome. Integrated viral DNA alters oncogene expression resulting in clonal expansion of that cell. Cervical disease has traditionally been screened and monitored by Pap smears, but HPV testing is proving to have superior specificity and sensitivity, and is supplanting Pap smears as the primary tool. However, current HPV clinical testing generally detects only the most common HPV types, and often searches for only a subfraction of the viral genome. We propose here to develop a method for detection of a massive set of different HPV types that will succeed even when only a fraction of the viral genome is present. The assay will use hybridization capture by a DNA probe set for the entire a clade of HPVs (currently 143 types) to enrich for HPV DNA in tissue samples, followed by deep, next generation DNA sequencing. It will also use a unique biorepository of cervical CIN1-3, tumor and control samples that we established featuring the highly diverse Bronx patient population that we serve. In preliminary studies using a 15 HPV type probe set, our method detected 8 different HPV types in a set of 26 CIN1-3 lesions and tumors, and detected integrated HPV DNA in 22 of 24 CIN2/3?s and tumors. Our single assay will simultaneously 1) yield unambiguous HPV type specificity because of the extensive viral sequences obtained, 2) detect common and rare HPV types, 3) find HPV even when only part of the viral genome is present, 4) determine if integrated HPV DNA is present, and 5) discern if integrated HPV DNA is near a human oncogene. We further propose to develop a fluorescence microscopy approach (Junc-FISH) to detect patient-specific integrated HPV in clinical samples. Our proposed work should provide superior HPV detection methods with higher and much broader HPV type specificity and sensitivity that also yields disease-relevant insight about HPV DNA integration.