Abstract A1-50: Genomic and epigenomic alterations in human and high-risk HPV DNA can discriminate normal from cervical dysplasia patients in urine

Abstract
Human Papilloma Virus (HPV) testing is increasingly used for cervical cancer screening in conjunction with cervical cytology. Although privacy, cultural, and infrastructure issues challenge the effective implementation of HPV testing for cervical cancer screening worldwide, several countries have already implemented HPV testing in their screening protocols. There are currently no tests that can reliably identify the patients with abnormal cytology and positive oncogenic HPV results (HPV+) that need to be referred for colposcopy. A useful triage test from cytology to colposcopy must discriminate the patients with a Cervical Intraepithelial Neoplasia (CIN) lesion that are more likely to progress to cervical cancer (CIN2+), from those that are less likely to progress. We set out to identify a panel of methylated HPV and human genes that can discriminate between CIN2+ and normal/CIN1 patients, first in cytology samples and then in urine samples. The Reverse Line Blot Assay identified high-risk HPV types in study participants from a Discovery cohort (n=96): 31 women with normal cervical cytology and 65 women with dysplasia (LSIL, n=43) and (HSIL, n=22). We developed custom sequence capture pools of baits to pull down genomic (Roche SeqCap EZ) and bisulfite converted high-risk HPV DNA (Agilent SureSelect) from urine, before library prep for NGS in a Roche 454 and MiSeq instruments, respectively. We also optimzed a qPCR assay to identify high risk HPV types, and quantitative Methylation Specific PCR (qMSP) assays to examine methylated HPV and human genes, in Pap smear DNA and TrDNA samples from women with normal Pap and women with CIN1, CIN2 and CIN3 lesions (n=80). The human genes selected for the qMSP assays, ZNF516, FKBP6, and INTS1, were identified in a previously published genome-wide analysis of the cevical cancer methylome using Nimblegen 385K CpG plus Promoters oligonucleotide tiling arrays. Presence of high-risk HPV in cervical epithelium correctly classified 37% of HSIL when compared with LSIL patients, with 61% sensitivity and 22% specificity. Promoter methylation of INTS1 correctly classified 67% (31% sensitivity, 92% specificity), promoter methylation of ZNF516 correctly classified 54% (11% sensitivity, 84% specificity), promoter methylation of FKBP6 correctly classified 59% (19% sensitivity, 88% specificity), and methylation of the HPVL1 gene correctly classified 61% (22% sensitivity, 90% specificity) of HSIL patients when compared with LSIL patients. A molecular panel comprised of HPV16-L1 methylation and promoter methylation of INTS1, ZNF516 and FKBP6 correctly classified 70% of HSIL patients with an Area Under the Curve of 0.66, 44% sensitivity, 88% specificity and 72% Positive Predictive value. While HPV16-L1 methylation discriminated patients with normal cytology from women with cervical dysplasia with close to 100% sensitivity and specificity, the high-risk HPV TrDNA qPCR assay had a pre-capture and post-capture sensitivity of 79% with 50% specificity. The detection of circulating HPV DNA in urine is a cost-effective and non-invasive alternative for cervical cancer screening. We have shown that a panel of genomic and epigenomic alterations in human and high risk HPV DNA can discriminate normal from cervical dysplasia in cervical epithelium DNA and TrDNA isolated from patients seen in a high risk cervical cancer screening setting. Citation Format: Rafael Guerrero-Preston, Anne Jedlicka, Oluwasina Folawiyo, Francesca Pirini, Blanca L. Valle, Fahcina Lawson, Angelo Vergura, Gabriela Perez, Marisa Renehan, Carolina Guerrero-Díaz, Liliana Florea, Teresa Díaz-Montes, Josefina Romaguera, David Sidransky. Genomic and epigenomic alterations in human and high-risk HPV DNA can discriminate normal from cervical dysplasia patients in urine. [abstract]. In: Proceedings of the AACR Special Conference on Translation of the Cancer Genome; Feb 7-9, 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 1):Abstract nr A1-50.