Variation in proportion of methylated CpGs between human cell types

Question

[Rubal]

Hello Everyone,

I am trying to ascertain the extent to which the proportion of methylated CpGs varies between cell types in the human body. I have looked through some of the methylation literature, particularly looking at whole-genome bisulphite sequencing data but have found it challenging to get a sense for the scale of variation between tissue types. Could anyone with experience in this field provide a sense of the scale of the variation in the proportion of methylated CpGs between human cell types, preferably with some suggested references? As far as I understand it approximately 70% of CpGs are methylated in the human genome and that it varies across cell types, but on what order of magnitude is this variation? My current understanding is that it only varies a few percent between most tissue types but I would greatly appreciate any input form anyone with experience in the field or a good grasp of the literature. Especially if there is a good review paper addressing this question.

Many thanks in advance for your time and input.

Answer 1

[Ventrilocus]

Hi Rubal,

the quick answer is that it depends on the tissue. If you compare differentiated blood cells with differentiated cells of the liver, you will probably see similar global methylation state. If you want to see higher differences in global methylation state, try employing WGBS data from embryonic stem cells, sperm or hematopoietic stem cells. Plenty of data available!

Best,

Ventri-locus.

[Rubal]

thanks Ventri, that makes good sense. I come across claims that some stem cells are hypo or hypermethylated but rarely find a statistic for the global methylation of CpGs compared to the approx 70% avg, I'm curious if some human cell types are lower than say 30% methylated CpGs or if hypomethlation could be anything lower than 60%. For instance ive seen claims that spermatogonia are hypo and hypermethlated (depending on the differentiation stage) but not found a metric of what this means in terms of % of methylated CpGs.

[Ventrilocus]

Dear Rubal,

indeed, there seems to be a lower bound for global methylation. This is probably due to the involvement of DNA methylation in the repression of selfish elements in the genome (e.g. retrovirus, retrotransposons, etc) that conforms a large percentage of the human genome and hence a big component in the computation of the average methylation. If the methylation gets low enough, such deleterious elements may cause cellular mayhem followed by cell death.

When you artificially induce hypomethylation in the lab via treatment with azacitidine or decitabine (so called hypomethylating agents), you do observe an increase in expression of dormant selfish elements [1]. The lowest global methylation percentage you may find may be from samples obtained upon treatment with hypomethylating agents (though not "naturally" occuring).

Also, I think you may find this other paper interesting [2]. Although it is microarray-based, you can observe how the (highly biased) estimate of average methylation varies between individuals, twins and tissues (whole blood, cord blood, adipose tissue and buccal epithelium).

[1] The double-edged sword of (re)expression of genes by hypomethylating agents: from viral mimicry to exploitation as priming agents for targeted immune checkpoint modulation. Cell Commun Signal. 2017; 15: 13. [2] Genome-wide average DNA methylation is determined in utero. Int J Epidemiol 2018 Jun 1;47(3):908-916.

[Rubal]

Thanks for this informative reply and also for the references, I will look into these. Much appreciated!