Embryonic stem cells are generally thought of as uniform cells that can give rise to all cell types within the embryo proper yet can no longer produce the extra-embryonic tissues that form the placenta. But these cells might actually be more dynamic and plastic than previously thought. In a pair of recent papers, two Cambridge, UK, research groups identify single genes that modify genomes within embryonic stem cells, thus regulating how embryonic cells commit to lineages within and outside the embryo proper. The research establishes plasticity and diversity within pluripotent cells and hints at a potential way to control cell fate.

Azim Surani and his co-workers at the University of Cambridge's Gurdon Institute noticed that the germ cell lineage marker Stella is unevenly expressed in embryonic stem (ES) cells. Using transgenic mouse ES cell lines, the researchers showed that a stable equilibrium of 20–30% of ES cells express Stella, although individual cells can reversibly turn the gene on and off through histone modifications. What's more, Stella-positive ES cells were enriched in markers specific to the inner cell mass, whereas Stella-negative ES cells more closely resembled the epiblast cells that arise from the outer layer of the blastoderm1.

“ES cells generally display a spectrum of potentials, and although we rightly consider them to be pluripotent, they are by no means homogenous,” says Surani. Tilo Kunath, of the University of Edinburgh, UK, notes that other groups also recently reported ES cell populations with unequal expression of the genes Nanog2 and Rex1 (ref. 3). Collectively, this could help explain why ES cell lines vary in how well they become other cell types, he says. “The heterogeneity you see in differentiation may be a reflection of the heterogeneity in ES cells to begin with.” Conversely, the findings have implications for reprogramming adult cells to an ES-like state if ES cells themselves are naturally variable. “When we're trying to drive somatic cells backward toward pluripotency, it's important to appreciate what the endpoint is,” Surani says. “This tells us that there are several endpoints.”

Surani's team also found that, like epiblast cells, Stella-negative ES cells could differentiate into trophectoderm, and that the Stella locus is repressed in post-implantation epiblast stem cells by DNA methylation.

In the other paper, Myriam Hemberger of the Babraham Institute and her colleagues focused on the earliest crossroads in mammalian development. Cells within the blastocyst can yield ES cells, but only cells at the outer edge of the blastocyst form the trophectoderm, which produces trophoblast stem (TS) cells. Her team found that mouse ES cells deficient for DNA methyltransferase 1 could differentiate efficiently into TS cells, indicating that DNA methylation normally restricts these cells' potential.

After a genome-wide screen for promoters with higher methylation levels in ES cells than in TS cells, Hemberger identified the transcription factor E74-like factor 5 (Elf5), which acts as a gatekeeper — it reinforces TS cell development by activating trophoblast-specific transcription factors when hypomethylated, and it blocks the trophoblast pathway when hypermethylated4.

“This is really the first molecular example of an epigenetic restriction of lineage fate,” says Hemberger. “Elf5 really sits at the junction of lineage divergence between the embryonic and trophoblast lineages in early development.”

Thus, whereas Elf5 methylation demarcates ES and TS cell fates, Stella methylation may provide an epigenetic lineage boundary between ES cells and epiblast cells. Both papers show that DNA methylation prevents “swapping back and forth between lineages,” says Hemberger.

As both studies found that methylation affects the switch to trophectoderm, the big outstanding question, says Janet Rossant of the Hospital for Sick Children in Toronto, Canada, is whether Elf5 is also hypomethylated in the trophectoderm-forming Stella-negative cells. Hemberger thinks that this is a real possibility. Last year, Toru Nakano, of Osaka University in Japan, discovered that Stella prevents DNA demethylation5, Hemberger notes. Thus, it's reasonable to predict that Elf5 should be demethylated in Stella-deficient cells, although future experiments will be needed to confirm this link. “There are connections [between the two papers],” Rossant says, “but we don't have all the data to make those connections at this point.”