MIT-led study maps how a mouse embryo builds itself, cell by cell

The study, conducted with Zachary Smith's group at Yale University and the Max Planck Institute for Molecular Genetics, set out to record not just individual moments in embryonic development but the entire process as it unfolded.

"This is really a view of how the information in the genome plays out over space and time to go from linear DNA to a three-dimensional animal," Weissman said.

The team used a technology developed in the Weissman Lab called PEtracer, which the researchers describe as acting somewhat like a flight recorder inside a cell. PEtracer makes small, heritable changes to DNA that are passed down as cells divide; by reading those changes later, scientists can work backward to reconstruct which cells came from which ancestors, building a family tree for the cells in a developing embryo.

The researchers combined these family trees with information about which genes were turned on in each cell and where the cells were located in the embryo. They analyzed more than a million cells from mouse embryos collected at multiple points during development, allowing them to follow not only where cells ended up but when they began heading toward particular fates.

A separate study from Jay Shendure's group at the University of Washington, published concurrently in the journal Science, uses a related approach called DNA Typewriter to track cell divisions in developing mouse embryos and build a record of their cellular history. Both studies provide a more comprehensive look at mammalian development than existed previously, according to the researchers.

The Weissman team studied embryos at closely spaced stages of development, looking at multiple embryos at each stage. That approach, the researchers say, gave them a more detailed picture of how development unfolds and allowed them to ask how consistently cells follow the same paths from one embryo to another. With the detailed timeline, they could also determine when cells commit to particular fates, how many early cells go on to form a tissue, and how those cells ultimately contribute to a developing organ.

The map could have implications for understanding birth defects and cancer. Many birth defects arise when cells change course as tissues and organs are being assembled, and a detailed map of typical development could help researchers pinpoint where and when those processes go awry. In cancer, the same approach could help scientists trace when cells acquire traits such as the ability to spread or resist treatment.

"Development and cancer both depend on decisions cells make about what they will become," said Luke Koblan, a postdoc in the Weissman Lab and co-first author of the paper. "By tracing how cells develop in mouse embryos, we hope to better understand how cancer cells change in ways that help them spread or resist treatment."

The work could also serve as a guide for stem cell research. Scientists have spent decades trying to coax stem cells into becoming mature, functional versions of particular cell types, but reproducing the precise conditions in an embryo is difficult.

"If we could have a step-by-step recipe for how to create every cell type in the body," said graduate student and co-first author William Colgan, "that would be incredibly powerful for stem cell biology."

Weissman compared the work to the Human Genome Project, saying that knowing the sequence of the human genome gave scientists the ingredients but not the instructions for how those ingredients are used.

"The genome is like the ingredient list for a four-course meal," he said. "It tells you what the ingredients are, but it doesn't tell you which ones you use when, how much you use, or where you put them. Without the recipe, it's impossible to know how those ingredients can become the meal."

The researchers describe the new map as a resource that can be used to explore mammalian development in a new way, allowing scientists to see how separate developmental processes fit together into one unfolding story rather than studying them as a collection of separate processes.