A Peek Under the Hood of Embryonic Stem Cells

Just as teenagers struggle to define themselves, cells do too. For embryonic stem (ES) cells - the cells endowed with a seemingly magical ability to form almost any kind of tissue in the body - this process of self-discovery is particularly mysterious.

Findings reported in the April 21 issue of Cell clarify the molecular underpinnings of this unusual awakening and could help to unlock the basis of their regenerative powers.

A research team led by Broad Institute scientists has discovered unique molecular imprints coupled to DNA in mouse ES cells.

These imprints, or "signatures," appear near the master regulatory genes that control embryonic development and likely coordinate their activity during the initial phases of cell maturation.

Largely absent from other cell types, the signatures disappear from ES cells once they commit to a specific developmental path.

This exclusivity could be a crucial handle for understanding the rare and wide-ranging capacities of ES cells as well as the limited abilities that other cells have to repair or replace themselves.

"This is an entirely new and unexpected discovery," said Brad Bernstein, lead author of the study, assistant professor at Massachusetts General Hospital and Harvard Medical School, and a researcher in the Chemical Biology program at the Broad Institute.

"It has allowed us to glimpse the molecular strategies that cells use to maintain an almost infinite potential, which will have important applications to our understanding of normal biology and disease."

The DNA contained in cells is wrapped around a supportive scaffold called chromatin. But this framework, composed primarily of histone proteins, wields influence that extends beyond structural matters.

Chromatin selects the genes that can or cannot be active in a cell based on small chemical groups affixed to nearby histones.

For example, methyl groups joined to a particular type of histone exert opposing effects depending on their anchoring point within the protein: attachment to lysine 4 (abbreviated as "K4") stimulates gene activity, while attachment to lysine 27 (abbreviated as "K27") inhibits it.

Information gleaned from chromatin studies in ordinary cells indicates that these marks tend to be mutually exclusive, and rarely, if ever, act together. However, little is known of their disposition in ES cells.

Researchers examined the chromatin in ES cells and focused their efforts on the portions of the genome that have been highly conserved during evolution.

Here, they discovered unique overlapping patterns, which merge both K4 and K27 methylation. Such contradictory groupings - named "bivalent domains" for their bipartite structure - appear mainly near key developmental genes.

Because these nearby genes are silent, scientists reason that, when coupled together, the negative sway of K27 triumphs over the urgings of its companion.

However, the activating influence of K4 is not completely quashed and likely keeps genes poised for later activity.

"For genes, this is equivalent to resting one finger on the trigger," said Stuart Schreiber, an author of the Cell paper, the director of the Chemical Biology program at the Broad Institute, and professor at Harvard University.

"This approach could be a key strategy for keeping crucial genes quiet, but primed for when they will be most needed."

Though heredity is often viewed in terms of DNA and the genes encoded by it, chromatin also carries inherited instructions known as "epigenetic" information.

The chromatin scaffold, along with its bivalent domains, forms a sort of molecular memory that can be transferred, together with DNA, from a cell to its descendants.

Given the nature of ES cells and their status as the earliest cellular ancestors, the question remains unanswered as to how epigenetic history first begins.

"How the initial epigenetic state is established and then altered during development has implications not only for stem cell biology, but also for cancer and other diseases where epigenetic defects are implicated," Bernstein said.

The scientists found that the dual components of bivalent domains each correlate with distinct underlying DNA sequences.

K4 methylation coincides with regions enriched for "CG" nucleotide pairs, while K27 methylation is restricted to areas devoid of transposon sequences.

This striking link between the positioning of bivalent domains and DNA sequence suggests that epigenetic memory originates from the DNA itself.