Abstract 4022: HPV16 CpG methylation patterns determined by next-gen bisulfite sequencing and association with cervix precancer/cancer

Abstract

Abstract Epigenetic changes within host genes are associated with a variety of tumor types. Increased methylation of CpG sites within the HPV16 genome has been associated with precancer and cancer tissues compared to transient infections. The objective of the study was to evaluate the use of next generation (NG) sequencing of bisulfite treated DNA in comparison to pyrosequencing to quantitate HPV16 methylation of CpG sites. In addition, NG sequencing provides additional information about methylation at multiple sites in cis across a DNA fragment. This information might have additional predicative value for discriminating prevalent HPV16 infections associated with precancer and cancer. A case-control study of HPV16 infected women from a population-based HPV natural history study in Guanacaste, Costa Rica was performed. Controls were 30 women who had transient HPV16 infections; whereas, cases were 28 women with HPV16 associated CIN3, and 13 with HPV16 cervical cancer. Four regions within the HPV16 genome containing multiple CpG sites were studied. These fragments contained 3 CpG sites in the E6 region, 7 sites in the E2 region, 5 sites in the L2 region, and 4 sites in the L1 region. After treatment with bisulfite, samples were amplified with barcoded forward primers and sequenced on an Illumina HiSeq 2000. The percent methylation at CpG sites was determined by the ratio of C/C+T. Agreement with prior pyrosequencing data was evaluated using linear regression and a Pearson's product moment correlation coefficient (r). Receiver operating characteristic (ROC) analyses were used to evaluate how well methylation could separate women with CIN3+ from those with transient infection, and the performance of this test was assessed using the area under the ROC curve (AUC). Linear regression curves indicated good agreement between NG and pyrosequencing results for CpG methylation, with an overall r = 0.97 for the E6 region, r = 0.90 for the E2 region, r = 0.73 for the L2 region, and r = 0.90 for the L1 region. Given that prior studies showed that methylation of L2 was significantly associated with CIN3+, we focused on the L2 region for further analysis. Using the Chi-square Goodness of fit test, the observed distribution of methylation patterns across NG sequenced fragments was different than the expected distribution based on individual CpG site methylation levels and the assumption of independence. Therefore, an analysis using patterns of methylation within the L2 fragment was carried out by an ROC analysis. A fragment with 3 methylated CpG sites had an AUC of 86.4%, while the highest single site AUC was 77.2%. This study demonstrates that NG sequencing of bisulfite treated DNA is an accurate method to survey methylation patterns in HPV16. Moreover, the combination of methylation at multiple sites within single molecules provides additional information. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4022. doi:1538-7445.AM2012-4022