Abstract 4998: Genome-wide methylation profiling in hepatocellular carcinoma

Abstract

Abstract Hepatocellular carcinoma (HCC) is one of the most common cancers and frequently presents with an advanced disease. Although it has been shown that a variety of epigenetic alterations can cause molecular heterogeneity of liver tumors, these studies have focused on specific genes, especially oncogenes and tumor suppressor genes. There is limited knowledge of the genome-wide methylation status in liver tumors and how this may potentially contribute to carcinogenesis and prognosis. We conducted genome-wide methylation profiling in a set of liver tumors. A pilot study was first performed on DNA from paired fresh frozen tumor and adjacent normal tissues from 5 patients. Genome-wide methylation profiles were created using Infinium Human Methylation 450K BeadChips. GenomeStudio was used to normalize data and to identify differentially methylated (DM) genes in tumor vs. normal tissue with False Discovery Rate (FDR) adjustment. Biological functions of the DM genes were assigned using the Ingenuity Pathway Analysis (IPA). Of the 485,577 CpG loci studied, significant DM loci were observed between normal and tumor tissues at 4% (17,612 CpGs) with an adjusted p-value < 0.001. Interestingly, 92% (15,712 CpGs) of the observed DNA methylation changes corresponded to general hypomethylation in tumor tissues, which was focused on promoter regions (741 out of 1,450 CpGs, 51%) and especially in promoter CpG islands (n=553, 74%). Conversely, hypomethylated CpGs were mostly observed in gene body and intergenic regions (12,343 out of 15,712 CpGs, 78%). Since the CpG islands in promoter regions can directly regulate gene expression levels, 428 genes including 12 miRNAs co-located with the significant DM CpGs (553 hypermethylated and 102 hypomethylated CpGs) in promoter region CpG islands were characterized for their biological function and potential role in liver cancer. IPA analysis showed Tissue and Cellular Development (p=1.0e-41), Cellular Growth and Proliferation (p=1.0e-14), and Gene Expression and Cell cycle (p=1.0e-12) as top networks. In particular, a top canonical pathway indicated 24 genes associated with the G-protein Coupled Receptor Signaling (p=3.12e-4) which is a key factor in transactivation of the commonly altered EGFR pathway in HCC development. Moreover, Nutritional Disease, with genes related to obesity (12 genes), weight loss (6 genes) and weight gain (4 genes) was observed as a top biological functional category (p<0.05). In addition, HCC related genes including TP73, HOXA9, TRIL, RASL10A and IGF2BP1 were differentially methylated between tumors and normals. In summary, we are performing genome-wide profiling of differentially methylated genes in HCC using DNA from tumor and normal tissue samples contributed to our biorepository. We expect to confirm the pathways preliminarily identified as being involved in liver carcinogenesis by obtaining methylation profiles from a larger set of patients with liver tumors. 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 4998. doi:1538-7445.AM2012-4998