Integrative Analysis of Epigenome and Transcriptome Data Reveals Aberrantly Methylated Promoters and Enhancers in Hepatocellular Carcinoma

Abstract

DNA methylation is a key transcription regulator, whose aberration was ubiquitous and important in most cancers including hepatocellular carcinoma (HCC). Whole-genome bisulfite sequencing (WGBS) was conducted for comparison of DNA methylation in tumor and adjacent tissues from 33 HCC patients, accompanying RNA-seq to determine differentially methylated region-associated, differentially expressed genes (DMR-DEGs), which were independently replicated in the TCGA-LIHC cohort and experimentally validated via 5-aza-2-deoxycytidine (5-azadC) demethylation. A total of 9,867,700 CpG sites showed significantly differential methylation in HCC. Integrations of mRNA-seq, histone ChIP-seq, and WGBS data identified 611 high-confidence DMR-DEGs. Enrichment analysis demonstrated activation of multiple molecular pathways related to cell cycle and DNA repair, accompanying repression of several critical metabolism pathways such as tyrosine and monocarboxylic acid metabolism. In TCGA-LIHC, we replicated about 53% of identified DMR-DEGs and highlighted the prognostic significance of combinations of methylation and expression of nine DMR-DEGs, which were more efficient prognostic biomarkers than considering either type of data alone. Finally, we validated 22/23 (95.7%) DMR-DEGs in 5-azadC-treated LO2 and/or HepG2 cells. In conclusion, integration of epigenome and transcriptome data depicted activation of multiple pivotal cell cycle-related pathways and repression of several metabolic pathways triggered by aberrant DNA methylation of promoters and enhancers in HCC.

Predicting active enhancers with DNA methylation and histone modification

“…Moreover, the results obtained from cross-model predictions further support the contributions of DNA methylation and H3K27ac to eRNA transcription [ 29 – 31 ]. Our approach to predicting eRNA demonstrates good accuracy, utilizing only two omics datasets.…”

“…eRNA exhibits several distinct features that can aid in their prediction and identification. These features include: (1) low levels of DNA methylation [ 30 , 32 ]; (2) specific histone modifications at enhancer loci [ 31 , 33 ]; (3) accessible (open) chromatin [ 34 ]; (4) TF occupancy [ 35 – 37 ]; and (5) RNAP II occupancy [ 38 ]. To build a prediction method that is less dependent on omics data, DNA methylation, gene expression, H3K27ac, and H3K9ac were used as input features (as Fig.…”

Enhancer-associated regulatory network and gene signature based on transcriptome and methylation data to predict the survival of patients with lung adenocarcinoma

“…In addition to screening biomarkers, the combination of epigenetic and transcriptional data can also demonstrate the mechanism. For example, the aberrant methylation of promoters and enhancers could activate critical cell cycle-related pathways and inhibit several metabolic pathways, thus affecting the progression of HCC (Huang et al, 2021). In short, the integrative analysis of multi-omics data can help us find new and more effective function targets in various diseases.…”

Integrating the Epigenome and Transcriptome of Hepatocellular Carcinoma to Identify Systematic Enhancer Aberrations and Establish an Aberrant Enhancer-Related Prognostic Signature

“…The harvested count data for each strand were combined for methylation level estimation. Differentially methylated loci (DML) and differentially methylated regions (DMRs) were detected with customized R scripts like our previous WGBS study ( Huang et al, 2021 ). For RNAseq data, clean reads that passed quality control were aligned with the hg38 genome, and the reference transcriptome was downloaded from GENCODE (v. 29) ( Harrow et al, 2012 ) with STAR (v. 2.5.2a) ( Dobin et al, 2013 ).…”