Given the functional and architectural complexity of the neurons that form the cerebral cortex, one might expect their development to require extensive external cues, such as growth signals and morphogens from surrounding cells. However, new research shows that pluripotent embryonic stem cells can differentiate into diverse subtypes of cortical neurons in culture through a developmental program that occurs spontaneously within embryonic stem cell–derived neural progenitors1. This points to the possibility of creating tailor-made cortical neurons to study or treat neural diseases.

Although researchers had previously shown that neural progenitor cells follow an internal program to develop into neurons, no one had, until now, found a way to make embryonic stem (ES) cells differentiate into the pyramidal neurons characteristic of the cerebral cortex. Pierre Vanderhaeghen of the Institute for Interdisciplinary Research at the University of Brussels, Belgium, and his colleagues first differentiated neural progenitor cells from mouse ES cells without added help from known cortical morphogens, and established that these progenitors were typical of the forebrain. To produce large numbers of neurons expressing markers exclusive to the dorsal forebrain, from which the cerebral cortex develops, the researchers then had to suppress the activity of endogenously produced Sonic hedgehog, a ubiquitous protein in early development that induces a ventral forebrain identity.

As these neural progenitor cells differentiated in culture, they developed the distinctive shape of pyramidal neurons. Over time they sequentially expressed molecular markers corresponding to the variety of pyramidal neurons found in the different layers of the cerebral cortex. “There is a sort of intrinsic program that drives the cells once they have become cortical progenitors,” Vanderhaeghen concludes. These ES cell–derived cells may not produce all the cell types of the cortex in a laboratory dish, he admits, but they nevertheless generate a large amount of the diversity.

As a final proof of identity, Vanderhaeghen's collaborator Afsaneh Gaillard of the University of Poitiers, France, grafted the cultured neurons into the frontal cerebral cortex of neonatal mice. After a month, axons from the cultured neurons had connected with a variety of targets, particularly projections from the visual cortex and the limbic system (thalamus), suggesting that these cells had taken on specific cortical identities without requiring extensive interaction from the rest of the brain. This could help scientists tease out the relative influences of intrinsic and extrinsic mechanisms that control how distinct cortical areas develop, a much-debated topic in neurobiology, says Vanderhaeghen.

If cell-replacement therapies for neural diseases are to work, “the ability to generate specific cell types from pluripotent stem cell sources will be very important,” says Arnold Kriegstein, who studies the embryonic development of the cerebral cortex at the Institute for Regeneration Medicine at the University of California, San Francisco. “This is the first step, or a beginning glimpse, into the possibility of doing that.”

Fundamental questions remain about the molecular basis of the differentiation sequence and whether similar results can be obtained with human cells. If so, these findings might lead to cellular model systems that could be used to tease out the mechanisms of neurodevelopmental diseases that are difficult to study in animals, says Vanderhaeghen.

There will, of course be signals to work out, predicts Kriegstein. The need to block Sonic hedgehog activity suggests that cells do rely on chemical signals to differentiate, perhaps coming from other cells within the culture. “My suspicion is that there are other factors that are being generated by cells that they [the authors] haven't come to grips with.” In other words, producing the diversity of neurons within the cortex is still a communal effort.