The human DNA methylation atlas: a new tool to study our tissues
Figure 1 – Development of the Methylation Atlas
Abstract
DNA methylation is a key epigenetic modification that plays crucial role in gene regulation. Currently available DNA methylome datasets are often limited to a fraction of methylation sites. Moreover, many of them have been generated from cell lines, which may contain significant changes in methylation patterns occurring in vitro (Ziller et al., 2013). In a recent study (Loyfer et al., 2023), Loyfer and his colleagues have developed a new human DNA methylome atlas, whose data were obtained by the analysis – using Whole Genome Bisulfite Sequencing (WGBS) – of 39 cell types isolated from 205 normal tissue. They found cell type-specific hypomethylated and hypermethylated regions. Hypomethylated regions are often found in enhancers and contain binding sites for tissue-specific transcription factors, while hypermethylated regions are rich in CpG islands, Polycomb targets and CTCF binding sites, assuming that they play a crucial role in chromatin remodeling and organization. This atlas is valuable both for medicine and research, providing biomarkers for liquid biopsies and tools to study gene regulation and disease.
Review
Introduction
The methylation of the DNA consists in the addition of a methyl group to cytosine bases, forming 5-methylcytosine (5mC). This modification is mediated by DNA methyltransferases (DNMTs), and mostly occurs in a CpG dinucleotide context. Regions rich in CpG sites are called “CpG islands”, and they can be typically found near gene promoters. DNA methylation is responsible for chromatin organization and gene expression and is important in determining the cell identity during the differentiation process. Hypermethylation of promoters is associated with gene silencing, while hypomethylation is associated with gene activation. The totality of all the DNA methylation in our genome is called the “methylome”. Many of the currently available DNA methylome datasets are limited only to some methylation sites, and have been generated either from cultured cell lines, which might be very different from the physiological condition, or from tissues, which contain unknown mixtures of cell types.
To overcome these limitations, Loyfer et al. generated the most comprehensive atlas of human DNA methylation to date. The atlas was obtained by deep whole-genome bisulfite sequencing (WGBS) of 39 “pure” cell types from 205 tissue. The analysed cells encompass the major human cell types, providing a view of the DNA methylome at the organ level.
Figure 2. Cell types hierarchy and characteristics of hypo and hypermethylated regions
Generation of the human DNA methylome atlas
After collecting the tissue samples (during routine surgical procedures), the foreign tissue was removed and the cells from the tissue of interest were dispersed by applying enzymatic cell type-specific protocols. The resulting suspensions were incubated with tissue-specific antibodies and classified using Fluorescence Activated Cell Sorting (FACS). The purity of the sorted cells was determined by qRT-PCR analysis of the major cell type-specific genes. After this first phase, the DNA methylation profiles of the isolated cells was obtained by WGBS. Libraries were sequenced using the Illumina NovaSeq machine, which performed a paired-end sequencing at 150 bp-long reads.
Using a multichannel segmentation algorithm, the authors identified the cell type-specific differentially methylated regions. Specifically, the algorithm takes in input simultaneously all the different cell types that have to be analysed, and aims to group genomic regions with similar methylation patterns across cells. Moreover, by unsupervised hierarchical clustering of the generated DNA methylome profiles, the authors showed how DNA methylation patterns were closely tracking the developmental history of the analysed cell types (Figure 1A).
Analysis of cell type-specific hyper and hypo-methylated regions
The analysis of cell type-specific hypermethylated regions showed that they are enriched for CpG islands, and are characterized by the presence of histone modifications such as H3K27me3 and occupancy of the Polycomb protein complex in the other cell types. (Straussman et al.,2009) (Figure 1C).
Moreover, in agreement with DNA methylation preventing CTCF binding, the analysis of in vivo CTCF occupancy obtained from public ChIP-seq data showed selective absence of CTCF binding at cell-type-specific hypermethylated loci, as the authors demonstrated for colon-specific regions.
Next, by analysing the hypomethylated regions specific to the various cell types, they identified characteristic features (Figure 1B). In particular, they showed that these regions are enriched within enhancers, and present consensus sequences for tissue-specific transcription factors. The association between cell-type-specific unmethylated regions and transcription factors binding sequences can help in identifying new gene regulatory circuits influencing the expression of cell-type specific genes.
Interindividual DNA methylation variation
After the identification of hypermethylated and hypomethylated regions, researchers wanted to investigate how and how much DNA methylation was variable in different individuals. To do so, they measured the average percentage of methylation clusters with more than 3 CpG sites (of each replicate) for 37 cell types, and they considered only those methylation clusters with a 50% methylation difference in the replicates. They observed that the variability of the methylation patterns of the same cell types from different samples is really low: this means that the same cell type has the same methylation pattern, even in different organisms. This high similarity in DNA methylation between donors also reflects the estimated inter-individual variability of the genome sequence. (The 1000 Genomes Project Consortium, 2015)
cfDNA methylation analysis in COVID-19
Circulating free DNA (cfDNA) is an important biomarker of disease conditions. The newly generated DNA methylation atlas was used by the authors to improve the determination of the tissue of origin from which cfDNA is derived. An important example of this application was the use of the atlas to analyse in greater depth the shallow WGBS data of 52 patients hospitalized due to COVID-19 (Cheng et al.,2021). In particular, Loyfer and his colleagues found the presence of a large amount of cfDNA derived from vascular endothelial cells, which had not been identified before because a reference methylome for these cells was missing. They also determined that the concentration of endothelial cell-derived cfDNA was higher in patients with severe disease than in those with milder disease. These results confirm that vascular endothelial cell death plays an important role in the pathogenesis of COVID-19, and highlight the advantage of using a complete cell type-specific methylation atlas for cfDNA methylation analysis.
Discussion and future perspectives
This study has important applications in both basic research and the medical field. Despite missing some cell types due to limited material, the atlas is a living database intended for future updates. The authors anticipate that the missing information of the remaining cell types can be supplemented by repeating the same procedures on samples of other cell types. The atlas enhances the understanding of composite tissues and aids in identifying additional cell types’ methylomes. The authors also provided tools for fragment-level analysis of mixed cell samples, beneficial for detecting the tissue of origin in cfDNA, which has relevant applications for the improved diagnosis of major human diseases, including cancer. By analyzing the cell-type-specific differentially methylated regions, the authors showed how DNA methylation acts as a record of progenitor cell methylomes, and is preserved through developmental transitions. Finally, the atlas can be applied to the comparative analysis of the methylome of pathological cells in solid tumors, to identify new targets for therapies.
References
- Loyfer, N., Magenheim, J., Peretz, A. et al.A DNA methylation atlas of normal human cell types. Nature 613, 355–364 (2023).
- Ziller, M. J. et al. Charting a dynamic DNA methylation landscape of the human genome. Nature500, 477–481 (2013).
- Kundaje, A. et al. Integrative analysis of 111 reference human epigenomes. Nature518, 317–330 (2015).
- The 1000 Genomes Project Consortium. A global reference for human genetic variation. Nature 526, 68–74 (2015)
- Straussman, R. et al. Developmental programming of CpG island methylation profiles in the human genome. Nat. Struct. Mol. Biol. 16, 564–571 (2009)
- Cheng, A. P. et al. Cell-free DNA tissues of origin by methylation profiling reveals significant cell, tissue, and organ-specific injury related to COVID-19 severity. Med (N Y) 2, 411–422 (2021)