Intragenic DNA methylation in transcriptional regulation, normal differentiation and cancer

Highlights

• The role of DNA methylation on gene regulation depends on the genomic context.

• Intragenic DNA methylation is involved in multiple gene regulation processes.

• Intragenic DNA methylation is modulated during cell differentiation and cancer.

Abstract

Ever since the discovery of DNA methylation at cytosine residues, the role of this so called fifth base has been extensively studied and debated. Until recently, the majority of DNA methylation studies focused on the analysis of CpG islands associated to promoter regions. However, with the upcoming possibilities to study DNA methylation in a genome-wide context, this epigenetic mark can now be studied in an unbiased manner. As a result, recent studies have shown that not only promoters but also intragenic and intergenic regions are widely modulated during physiological processes and disease. In particular, it is becoming increasingly clear that DNA methylation in the gene body is not just a passive witness of gene transcription but it seems to be actively involved in multiple gene regulation processes. In this review we discuss the potential role of intragenic DNA methylation in alternative promoter usage, regulation of short and long non-coding RNAs, alternative RNA processing, as well as enhancer activity. Furthermore, we summarize how the intragenic DNA methylome is modified both during normal cell differentiation and neoplastic transformation.

Introduction

Back in the 1970s, Riggs and Holliday independently suggested that methylation of cytosines in the context of CpG dinucleotides may represent an epigenetic mark associated with gene silencing [1], [2]. Years later, this hypothesis was experimentally demonstrated by several groups [3], [4], [5], [6], [7]. Interestingly, already in the 1980s it was recognized that CpG density is not evenly distributed within the genome, but rather shows a bimodal distribution. Regions with an elevated CpG content are called CpG islands (CGIs), and overlap with transcriptional start sites (TSSs) of approximately 60–70% of all human genes [8], [9]. In contrast, regions with low CpG density are frequently located outside TSS [10], [11]. Furthermore, it was noted in these studies that both DNA methylation and CpG density show a bimodal distribution. In general, CGIs are unmethylated whereas most regions with low CpG density are heavily methylated in normal tissues, something which is now considered general knowledge. Experimental evidence demonstrated that gene expression can be regulated by DNA methylation levels of CGIs in the proximity of TSSs [12], [13]. Based on these data, the majority of cancer-related DNA methylation studies focused on the role of CGI hypermethylation as a mechanism of tumor suppressor gene silencing [14], [15], [16], [17], [18]. Additionally, already in the 1980s, pioneer studies in cancer cells reported that neoplastic transformation was also associated with global and gene specific loss of DNA methylation [19], [20], [21], [22]. Collectively, the studies published so far in cancer epigenomics point to a massive disruption of the DNA methylome in tumor cells as compared to their normal cell counterparts.

DNA methylation can be analyzed by methylation sensitive restriction enzyme digestion, bisulfite conversion based techniques or affinity enrichment of methylated DNA [23], [24], [25]. A comprehensive overview of the advantages and disadvantages of these approaches and their applicability is given by Laird [26]. Until recently, these techniques were PCR or microarray-based to analyze single or multiple regions of interest, respectively [27], [28], [29], [30]. As a consequence, the great majority of studies designed to investigate DNA methylation dynamics were covering only a small fraction of the DNA methylome, namely CGIs and promoter regions. Nowadays, the use of high-density microarray analyses, e.g., comprehensive high-throughput arrays for relative methylation (CHARM) and Infinium [31], [32], and next-generation sequencing based DNA methylation analysis, e.g., methylated DNA immunoprecipitation followed by deep-sequencing (MeDIP-seq), CXXC affinity purification followed by deep-sequencing (CAP-seq), reduced representation bisulfite sequencing (RRBS) and whole genome bisulfite sequencing (WGBS) [33], [34], [35], [36], allows us to obtain an unbiased representation of DNA methylation maps throughout the genome. The use of these techniques has started to reveal that DNA methylation has an even broader function than previously anticipated and that its functions may vary in a context-dependent manner [37].

One of the major findings derived from whole-genome studies is that DNA methylation levels in the gene body, i.e. intragenic DNA methylation, widely change during cell differentiation and carcinogenesis. Although the precise role of DNA methylation within the gene body is still far from being well understood, recent publications indicate that it may be involved in the regulation of multiple processes, such as transcript elongation, expression of intragenic coding and non-coding transcripts, alternative splicing and enhancer activation. The goal of this review is to provide a comprehensive overview of the current knowledge on intragenic DNA methylation and its association with gene regulation in the context of cell differentiation and neoplastic transformation.