NEW YORK (GenomeWeb News) – Genome-wide demethylation occurs not only in primordial germ cells and pre-implantation embryos, but also during red blood cell differentiation, according to a mouse study appearing online today in Science.
A Massachusetts research team used several methods — including reduced representation bisulfite sequencing — to track DNA methylation patterns in differentiating red blood cells from a mouse fetal liver tissue. The results of their experiments, first looking at specific sites in the genome and then genome-wide, suggest that the process of red blood cell differentiation, or erythropoiesis, is accompanied by a progressive loss of methylation across the genome.
"People thought a global loss of methylation only happened in early development and didn't happen in somatic cells," senior author Merav Socolovsky, a pediatrics and cancer biology researcher at the University of Massachusetts Medical School, told GenomeWeb Daily News, noting that global demethylation in somatic cells had only been detected in certain cancers.
But the presence of demethylation across the genome during erythropoiesis suggests it can occur in normal somatic cell processes as well, she explained, since red blood cell differentiation occurs about once a month in adult mice and every 120 days or so in adult humans.
"This is not just something that's happening in the embryo," Socolovsky said. "It's probably happening all the time."
The study grew out of the group's previous work on erythropoiesis, including research showing that this red blood cell differentiation process relies on a switch involving precursor cells that are synchronized in an S-phase, a cell cycle stage marked by DNA synthesis and replication.
Based on those findings, the team decided to investigate whether the S-phase dependent transition to mature red blood cells corresponded to specific chromatin changes, Socolovsky explained.
After nabbing cells at different stages of red blood cell differentiation from mouse fetal liver tissue samples using flow cytometry, the team did a series of experiments to check out chromatin patterns in the differentiating cells, including methylation analyses.
Instead of seeing the local changes to methylation that they expected around red blood cell-related genes, though, the researchers' enzyme-linked immunosorbent assays, luminometric methylation assays, and other experiments pointed to more extensive demethylation: the majority of loci tested showed methylation losses, including imprinted genes and retrotransposons that normally remain methylated.
"Almost any gene we looked at, whether it was up-regulated or down-regulated was losing methylation," Socolovsky said.
Next, the team sequenced mouse cells at different stages of erythropoeisis with representation bisulfite sequencing, a strategy that enriches for sites in the genome where cytosine and guanine neighbor one another. Using that approach, the team generated sequence covering about 10 percent of the genome overall, but which represented a broad range of genomic elements that are normally methylated.
From those experiments, researchers saw that demethylation was happening across the genome as erythropoiesis progressed, with more differentiated cells showing less and less methylation in their genomes.
"What we found is really a global loss of methylation with differentiation," Socolovsky said. "As we proceed from the least differentiated cell types that we isolate, which we call S0, all the way to the most differentiated cell type, which we call S5, we see a gradual loss of methylation at almost every locus that we're looking at."
Despite the genome-wide decreases in methylation, though, the team did not find evidence of rampant gene expression at sites where methylation had been lost. Instead, demethylation was found both at genes showing decreased expression and at genes with elevated expression.
On the other hand, when they prevented this demethylation from occurring, researchers found that the expression of red blood cell-related genes was impaired, leading them to hypothesize that the epigenetic shift may help red blood cells produce the massive amount of hemoglobin protein needed for them to function properly.
"It may be that you need this global loss of DNA methylation in order to accelerate hemoglobin production," Socolovsky noted. "Certainly when we prevent the demethylation, we slow down the hemoglobin production."
The demethylation detected also seems to rely on the flurry of DNA synthesis that characterizes the switch from precursor cells to red blood cells, she explained. By artificially slowing down DNA replication with a DNA polymerase enzyme inhibitor, the researchers prevented widespread demethylation in the differentiating cells.
"It seems that DNA synthesis in that S-phase happens a lot faster than in previous [cell] cycles — about 50 percent faster," Socolovsky said. "If we slow it down, back to the rate it used to be in previous cycles, we no longer get the global demethylation."
So far such erythropoiesis-related demethylation has only been detected in mouse cells, she said, but researchers are gearing up to see if similar methylation muting occurs during red blood cell differentiation in humans as well. In addition, they plan to look at whether such shifts are red blood cell specific or whether they occur in other blood lineages or cell types as well.
Because global demethylation in somatic cells has been linked to some cancers, the researchers also hope to learn more about how red blood cell demethylation mechanisms compare to those found in cancer.