From mice to men: Broad researchers develop a human model for studying DNA methylation

It’s not every day that scientists get to offer their colleagues a model system that will enable a wave of future research. “Often you make a discovery, you describe it, and that’s the end of the story,” said Alexander Meissner, a Senior Associate Member at the Broad Institute of MIT and Harvard....

It’s not every day that scientists get to offer their colleagues a model system that will enable a wave of future research. “Often you make a discovery, you describe it, and that’s the end of the story,” said Alexander Meissner, a Senior Associate Member at the Broad Institute of MIT and Harvard. “Here it’s not the end of something, it’s just the beginning.”

Two postdoctoral fellows in Meissner’s lab, Jing Liao and Rahul Karnik, are first authors on a new paper in the journal Nature Genetics in which they report the generation of human embryonic stem cells (ESCs) specifically designed to interrogate the biological implications of a process called DNA methylation. As a layer of the epigenome, or the chemical instructions that determine which genes are expressed in each unique cell type in the body, methylation is absolutely essential for normal development. The process involves turning genes on or off through the addition of carbon atoms to specific sites along the genome.

“The DNA sequence doesn’t change from cell type to cell type,” said Meissner. “So something else has to make sure that your skin cells never turn into neurons.” DNA methylation is believed to play a central role in this regulatory process as well as in a host of associated processes. But until now, there was no straightforward method for examining the biological implications of methylation and its dynamic regulation in human cells.

Three enzymes are responsible for methylating DNA and they’re fittingly called DNA methyltransferases, or DNMTs. In previous work, other scientists had eliminated each of the enzymes (DNMT1, DNMT3A, and DNMT3B) from mouse ESCs to build a powerful tool for studying their roles and functions: By knocking out one of the enzymes and observing the resulting cellular behavior, researchers could begin to understand what that enzyme does when it’s intact. Using this approach, researchers have determined that in mouse cells DNMT3A and DNMT3B are responsible for creating new methylations along the genome where none existed before. DNMT1’s purpose is to maintain the existing status quo. While these general insights are conserved in mouse and human cells, more detailed examinations of the targets and specific regulation require a human model.

“For some things, it’s good enough to have the mouse model,” said Meissner. But for many studies — like those examining how DNA is deregulated in diseases — human cells become necessary. While the mouse models have been around for more than a decade, there was no equivalent system for human ESCs.

Meissner believes the reason for this gap in the field was due not to a lack of interest, but rather a lack of tools. A recently developed method called the CRISPR/Cas9 system allows researchers to edit the genome with unmatched precision. This, and another method called TALENS, were the missing tools that allowed Meissner’s team to effectively knock out the DNMT enzymes in human ESCs.

First the team deleted the genes that code for DNMT3A and DNMT3B. They also created a double knockout by taking some DNMT3A-deficient cells and further editing their genomes to also eliminate DNMT3B.

Interestingly, when the team tried to knock out DNMT1, the resulting cells did not survive. This was surprising because the same thing doesn’t happen in mouse ESCs. About a decade ago, researchers managed to construct human colon cancer cells devoid of DNMT1, but those also died, suggesting early on that this particular enzyme is essential to human ESC survival. “DNMT1 is a maintenance enzyme so its knockout has a much more dramatic effect,” said Meissner. “If you block maintenance you block all methylation throughout the genome, and that is incompatible with any somatic cell type.”

In order to get around this, Meissner and his colleagues created a system that rescues cells lacking DNMT1. To do so, they introduced an inducible bit of DNA into the cells’ genome that expresses DNMT1 on command at levels just sufficient for survival. As soon as that bit of DNA is turned off and DNMT1 stops being expressed, the cells begin to die within a few days. During this window, they confirmed that DNMT1 is critical for normal cellular differentiation and that loss of DNMT1 causes damage to the DNA.

The fact that DNMT1 isn’t critical for mouse ESC survival but is so for human ESCs highlights the significance of the new system. It is speculated that mouse ESCs exist in a different developmental phase than human ESCs, which is not yet dependent on methylation. This fact, which was borne out in the present study, makes mouse ESCs a more accessible experimental system, but one that doesn’t perfectly parallel the human system in which researchers are most interested.

That’s why the new human ESC model system is so valuable. It consists of three new cell types: one each that is missing DNMT3A or DNMT3B, and one that is missing both enzymes. By offering the community cells that have the experimental advantages of mouse ESCs but retain the general characteristics of human ESCs, Meissner and his team have opened the doors to a host of new research opportunities.

“The work has impacts on multiple different fields,” Meissner said. “Each one will get different things out of this.” For instance, the stem cell community can use the system to gain a better understanding of how cells differentiate and all the associated processes that methylation is presumed to be involved in.

The work has already begun to enhance the epigenomics community’s understanding of the basic principles and mechanisms of DNA methylation: Meissner’s team used the cells created in the study to map the entire landscape of methylation along the genome, providing an unprecedented view of the shared and unique genomic targets of the two DNMTs responsible for depositing new methylation patterns along the genome. This was the first time such an effort had been reported—even for mouse ESCs. Now the community has a reference of all the regions in the genome that are targeted by DNMT3A and DNMT3B.

Finally, various disease research communities, such as those that use blood progenitor cells in leukemia studies or those that use neural progenitors in studies of amyotrophic lateral sclerosis (ALS), will also benefit. Researchers will now be able to study the effects of methylation on, for example, blood development and mutation in the context of the human cell.

“Especially in recent years, one paper after another comes out that finds mutations in DNMT3A in particular in cancers and many other diseases,” Meissner said. “So having a human model to be able to create these disease relevant cell types and study the implications of the condition is important.” Thanks to his efforts and those of his team, the community now has that model.

Other Broad researchers involved in the work include: Hongcang Gu, Michael Ziller, Kendell Clement, Alexander Tsankov, Veronika Akopian, Casey Gifford, Julie Donaghey, Christina Galonska, Ramona Pop, William Mallard, John Rinn, and Andreas Gnirke.

Paper cited: Liao, J., Karnik, R., et al. Targeted disruption of DNMT1, DNMT3A and DNMT3B in human embryonic stem cells, Nature Genetics (2015). Doi:10.1038/ng.3258