Integrated analyses of multi-omics reveal global patterns of methylation and hydroxymethylation and screen the tumor suppressive roles of HADHB in colorectal cancer

Abstract

Background: DNA methylation is an important epigenetic modification, associated with gene expression. 5-Methylcytosine and 5-hydroxymethylcytosine are two epigenetic hallmarks that maintain the equilibrium of epigenetic reprogramming. Disequilibrium in genomic methylation leads to carcinogenesis. The purpose of this study was to elucidate the epigenetic mechanisms of DNA methylation and hydroxymethylation in the carcinogenesis of colorectal cancer.

Methods: Genome-wide patterns of DNA methylation and hydroxymethylation in six paired colorectal tumor tissues and corresponding normal tissues were determined using immunoprecipitation and sequencing. Transcriptional expression was determined by RNA sequencing (RNA-Seq). Groupwise differential methylation regions (DMR), differential hydroxymethylation regions (DhMR), and differentially expressed gene (DEG) regions were identified. Epigenetic biomarkers were screened by integrating DMR, DhMR, and DEGs and confirmed using functional analysis.

Results: We identified a genome-wide distinct hydroxymethylation pattern that could be used as an epigenetic biomarker for clearly differentiating colorectal tumor tissues from normal tissues. We identified 59,249 DMRs, 187,172 DhMRs, and 948 DEGs by comparing between tumors and normal tissues. After cross-matching genes containing DMRs or DhMRs with DEGs, we screened seven genes that were aberrantly regulated by DNA methylation in tumors. Furthermore, hypermethylation of the HADHB gene was persistently found to be correlated with downregulation of its transcription in colorectal cancer (CRC). These findings were confirmed in other patients of colorectal cancer. Tumor functional analysis indicated that HADHB reduced cancer cell migration and invasiveness. These findings suggested its possible role as a tumor suppressor gene (TSG).

Conclusion: This study reveals the global patterns of methylation and hydroxymethylation in CRC. Several CRC-associated genes were screened with multi-omic analysis. Aberrant methylation and hydroxymethylation were found to be in the carcinogenesis of CRC.

Keywords: Colorectal cancer; DNA hydroxymethylation; DNA methylation; Epigenetic; Sequencing.

Fig. 1

Global patterns of methylome and hydroxymethylome. The Pearson’s correlation between DNA methylation and hydroxymethlation levels and features of autosomes (chromosomes 1–12) and sex chromosomes X and Y. The read depth was plotted against the length, repeat density, gene density, SNP density, GC content, and CpGo/e ratio of the individual chromosome. The line represents linear regression. Different colors represent the different modifications in tumor and normal tissues (red is Medip in normal tissue, blue is Medip in tumor tissue, gold is hMedip in normal tissue, and green is hMedip in tumor tissue)

Fig. 2

Different distributions of methylation and hydroxymethylation. a, b Principal component (PC) analysis based on methylation and hydroxymethylation levels for 0.5-kb tiles across 6 normal and tumor tissue samples. Coloring indicates classification of samples into subgroups. The red area denotes tumor tissues and the green area means normal tissue. c A large number of regions of differential hydroxymethylation (DhMRs) and methylation (DMRs) occur in colorectal cancer. Regions that gain a mark (“hyper-”) are represented by blue bars, whereas losses (“hypo-”) are red. d Visualization of DMRs (left) and DhMRs (right) patterns across all normal and tumor tissues (blue and red bars in the box). Boxed regions with green are candidate DMRs or DhMRs

Fig. 3

Association between epigenomic modification and gene expression. a, b Correlations between epigenomic modification (a is 5-mC, b is 5-hmC) and gene expression in TSS (defined as − 500 bp to + 500 bp across TSS) and genebody respectively. For Medip, there is a clear anti-correlation of gene expression in TSS and a positive correlation in the genebody. For hMedip, only positive correlations in the genebody can be observed. c 5-mC is enriched in the TSS flanking region (defined as − 500 bp to + 500 bp across TSS) of lowly expressed genes in both despite normal and tumor tissues. d 5-hmC is enriched in the genebody region of highly expressed genes in both normal and tumor tissues

Fig. 4

Expression and epigenetic modification validation of HADHB in additional samples and public functional genomics data. a Expression level of HADHB determined by real-time PCR in 15 additional pairs of colorectal tumors and their adjacent normal tissues. b Expression data of HADHB from the GEO database. c, d Expression level of HADHB from the TCGA database. The results derived from different platform (c is Hiseq200 and d is G450 array) are shown respectively. e, f, Methylation and hydroxymethylation levels were determined by real-time PCR in 15 additional pairs of colorectal tumors and their adjacent normal tissues

Fig. 5

Functional experiments of HADHB on CRC cell line. a Transcriptional and translational levels of HADHB in seven cell lines detected by RT-PCR and western blot. b HADHB expression in knockdown and overexpression cell lines of HT29 and HCT8. c Cell growth curves in the knockdown and overexpression cell lines of HT29 and HCT8. d Cells migration and invasion in the knockdown and overexpression cell lines of HT29 and HCT8