NEW YORK (GenomeWeb News) – The process of reprogramming somatic cells to make induced pluripotent stem cells leads to some distinct methylation patterns that persist even after these iPSCs are differentiated once again, according to a study appearing online today in Nature.
By characterizing CG- and non-CG methylation patterns to single base resolution in iPSCs, embryonic stem cells, and differentiated cells, a California and Wisconsin-based research team found evidence that methylation is not completely reset to a stem-cell like state in iPSCs and cells generated from them. Instead, these cells contain differentially methylated regions distinct from those found in ES cells and, in some cases, from the parental adult cells as well.
"From these first comprehensive whole-genome, base-resolution methylome maps it seems clear that iPSCs are fundamentally distinct from ES cells, insofar as they manifest common, quantifiable epigenomic differences," senior author Joseph Ecker, a researcher with the Salk Institute for Biological Studies' Genomic Analysis Laboratory, and co-authors wrote.
Some of the regions appear to represent a cellular "memory" of methylation patterns found in the parental cells from which the iPSCs are derived, Ecker told GenomeWeb Daily News, while others — dubbed iPSC-specific differentially methylated regions, or iDMRs — differ from both parental and ES cells.
Induced pluripotent stem cells have garnered considerable attention because they possess many of the same characteristics of ES cells but can be obtained without destroying human embryos, skirting ethical concerns related to this process.
For the current study, researchers began by studying methylation patterns in iPSC generated from fat tissue, which also contains adult stem cells, Ecker explained.
The researchers used the Illumina GAIIx and HiSeq 2000 to do MethylC-Seq bisulfite sequencing of iPSC lines derived from fat tissue, as well as adipose-derived stem cells and differentiated adipose cells.
They subsequently compared the methylation patterns in these cells with those found in other iPSC types, ES cells, and cells differentiated from either ES cells or iPSCs.
While they found that the methylation patterns in iPSCs are quite similar to those seen in ES cells overall, there were still distinctive differences. For instance, the researchers found 1,175 CG-methylation sites that were differentially methylated in at least one of the iPSC lines tested.
And while most CG islands near gene promoters do seem to get reprogrammed in the iPSCs, a large percentage of all differentially methylated CG sites, dubbed "CG-DMRs," did fall in these CG islands. That, in turn, suggests such methylation differences could affect the expression of some genes, the team noted.
Moreover, the researchers found that non-CG methylation — a type of methylation that seems to largely absent in differentiated cells but relatively common in stem cells — was not completely restored in iPSCs. Instead, the team found so called "non-CG mega-DMRs" spanning millions of bases that do not show the sorts of non-CG methylation found in stem cells.
"If you take differentiated cells and reprogram them, you see very nicely that the non-CG methylation comes back," Ecker said. "But it doesn't come back in these non-CG mega-DMR regions."
And, he added, the non-CG mega-DMRs tend to cluster around telomeres and centromeres — parts of the genome containing poorly expressed genes with hyper-CG-methylated promoters.
"Here you have absence of reprogramming of the non-CG methylation, hyper-methylation of the CG islands of the promoters of genes that contain CG islands, and reduced expression of those genes," Ecker said, adding that at least one chromatin mark typically found in differentiated cells also tends to stick around in these regions.
Moreover, when the researchers focused in on the iPSC-specific DMRs as a group, they found that some of these DMRs are shared across iPSCs tested, while others vary depending on the type of cells from which iPSCs are derived.
"iDMRs, as a group, have [differentially methylated regions] that are unique to the iPSC line, so they're not like the parent, they're not like the ES line, and they're not like the other iPS lines," Ecker explained. "And that's interesting. But maybe more interesting are ones that you actually see in most, if not all, of the iPSC lines but not in the parent and not in ES cells."
Together, the findings suggest it will be necessary to consider the functional effects of such DMRs, if any, before using iPSCs therapeutically or to create models of specific diseases.
"We don't know the consequences of these differences," Ecker said, explaining that while it's known that iPSCs can be coaxed into becoming many different cell types, it's still unclear whether these differentiated cells behave the same as typical cells or not.
But now, he says, "We have an assay to begin to look at the consequences of these reprogramming events.
"Whether they will have phenotypes, we're not sure," he added, "but it certainly could affect the long-term potential of some of the differentiation processes and protocols that you might use."