Skip to main content
Premium Trial:

Request an Annual Quote

Scientists ID Histone Modification Patterns That Predict Embryonic Cell Differentiation


By Adam Bonislawski

A research group led by scientists at the University of Pennsylvania has identified patterns of histone modification in multipotent embryonic cells that enable them to distinguish between those poised to differentiate into liver elements and those poised to become pancreatic cells.

The findings, which were detailed in a paper published this week in Science, represent the first demonstration that embryonic endoderm cells have distinguishing patterns of histone modification, and provide insights that could in the future help scientists control stem cell differentiation for purposes like tissue regeneration, said Kenneth Zaret, a professor of cell and microbiology at Penn and leader of the study.

While considerable work has been done on histone modification patterns, much of it has focused on highly differentiated cells and cultured cell lines, Zaret told ProteoMonitor. The Science study, he said, emerged out of preliminary research by his team indicating differences between the chromatin – of which histones are the primary protein component – of embryonic stem cells differentiated to endoderm compared to that of native embryonic endoderm.

Given those findings, Zaret said, his team "thought it was worth the effort of determining how [histone modification] works in the mammalian early embryo and using that as a benchmark for analysis of embryonic stem cell-derived cultures."

Using fluorescence-activated cell sorting with the ENDM1 antibody, the researchers isolated ventral foregut endoderm cells from mouse embryos just before the induction of hepatic and pancreatic fates. They then examined the histone patterns of these cells via chromatin immunoprecipitation assays, using 15 chromatin-specific antibodies to screen for a variety of known histone modifications. These assays revealed distinct patterns of modification for liver and pancreatic-specific regulatory elements, a result Zaret said was surprising.

"I had assumed that genes which have the potential to be activated in the endoderm lineage would look more or less the same at the level of chromatin features, and the surprise is that they didn't," he said. "Regulatory features for liver-specific genes looked different in terms of chromatin features than the regulatory elements for the pancreatic gene. That [indicated] that even in the undifferentiated progenitor [cells] there was already a prepattern set up which we presume causes the genes to respond differently to inductive signals in the embryo."

Previously, Zaret said, most scientists had assumed such differentiation came later in the developmental process, but "we find that already there is some difference in how the genes are set up before the inductive signal appears."

Additionally, the researchers identified roles in cell differentiation played by the enzymes histone acetyltransferase P300 and histone methyltransferase Ezh2, pinpointing "a three- to five-hour window in development where these enzymes play a crucial role in controlling the cell fate decisions of these endoderm cells," Zaret said, adding that this raises the future possibility of using or targeting these enzymes pharmacologically to control cell differentiation.

The study "is a very nice example of how cell states can be modulated by chromatin modifiers," said Joanna Wysocka, an epigenetics researcher at Stanford University who was not involved in the work. "It's yet another example showing that perhaps looking at epigenetic patterns can give us really more information than looking at gene expression because looking at epigenetic patterns will give us information as to the developmental potential of the cell, not only as to its current state."

Among the most interesting aspects of the study "is that they've done it on progenitor cells isolated directly from embryos," Wysocka told ProteoMonitor. "This is a true in vivo approach."

Among the main challenges to study histone patterns in such cells, Zaret said, was developing a low-cell number ChIP assay capable of working with the small samples available.

"The entire embryo is one millimeter at the stage we're looking at, and we're getting about 400 cells of the relevant type per embryo," he said, noting that conventional ChIP assays use on the order of a million cells. "We worked out conditions to get [ChIP] to work on several thousand cells, so that was a lot of work to get that working reproducibly."

This process consisted primarily of "scaling down volumes to keep concentrations of reagents the same as for a large-scale procedure and minimizing the vessel surface areas" to minimize loss of samples, Zaret said, adding that now his lab hoped to adapt the procedure for ChIP-Seq assays, which would allow them to sequence the DNA linked to the histones pulled down via immunoprecipitation.

He said he also hopes to develop mass spec assays that would allow for a broader, unbiased investigation into the chromatin complexes of these embryonic cells. Histones are a growing area of interest for proteomics researchers, particularly those using top-down mass spec techniques, but, Zaret said, obtaining enough sample for such work would be a challenge.

"You're starting with such small amounts [of sample] to begin with," he said, which makes it difficult to get "the material as pure as you would want without contaminating complexes."

It "could be very exciting," however, "to do this at the discovery level" using mass spec instead of antibodies, Zaret said. "Right now, we're limited to the chromatin modifications for which we have good antibodies. There are a lot of other protein domains where there could be very interesting changes and modifications going on."

Discovery mass spec work on such small sample sizes would be difficult, agreed Northwestern researcher Neil Kelleher, whose lab has done considerable work using mass spec to study histones. But, he said, it's not impossible.

"Certainly we're doing it on 10,000 cells, and it's feasible to do it on 1,000 [cells]," he told ProteoMonitor. "It's an expert post-doc for a year or two really pounding on it, but it's not impossible."

There would be limitations to this approach, though, Keller said, noting that, for instance, "you can't get all the [modification] combinations in full discovery mode with five orders of magnitude dynamic range for histone forms present in cells."

An easier method, Kelleher suggested, would be a "targeted proteomics approach using triple-quad mass spectrometers" or a "higher-end [Thermo Fisher Scientific] Orbitrap instrument," which would enable researchers to detect "all the major epigenetic marks" though not all the combinations of histone modifications.

"If you want all the combinations of all the histone [modifications], it's not impossible to do that with a top-down, full-discovery approach," he said. "It's just even one click more challenging."

Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.