Human pluripotent stem cells (hPSCs) are exciting for many reasons. Perhaps the most talked about is their potential for cell therapy. Equally compelling, however, is that they are a source of human cell types that are otherwise difficult to obtain and hence to study in the lab. The recent derivation of blood-brain barrier (BBB) endothelial cells from hPSCs by Eric Shusta and Sean Palecek at the University of Wisconsin–Madison is a case in point.

Shusta and colleagues have been developing animal and human BBB models in the culture dish for years. But animals, rodents in particular, do not yield substantial numbers of brain endothelial cells; furthermore, there are functional differences between animal and human BBB. Primary cells of the human BBB can be acquired from autopsied tissues or from resected brain samples of patients with epilepsy or tumors, but these specimens are not easily obtainable. Shusta hoped to solve all these problems by teaming up with Palecek and his group, who have expertise with human pluripotent stem cells.

The collaboration has proven fruitful. The researchers knew that signals from developing neural cells in the early embryonic brain are important for promoting the development of the specialized endothelial cells that form the BBB, the brain microvascular endothelial cells (BMECs). They reasoned that codifferentiation of hPSCs into neural and endothelial cells could mimic conditions in the embryonic brain and yield BMECs.

The researchers cultured hPSC lines under standard culture conditions, on Matrigel and in mTESR1 medium, and subsequently induced differentiation. Within a week or so, they observed a mixture of mainly neural progenitors, immature neurons and endothelial cells in the cultures. A subsequent switch to culture medium containing growth factors known to promote endothelial differentiation yielded up to 60% endothelial cells with a marker profile suggesting they were BMECs.

To test the function of these cells and establish their BMEC subtype beyond doubt, it was necessary to purify them. Shusta, Palecek and colleagues found that this was easily accomplished by passaging the cells onto a collagen-fibronectin extracellular matrix routinely used for culturing primary BMECs. The resulting pure population of cells formed tight endothelial monolayers with the expected marker profiles, high electrical resistance that increased upon exposure to astrocytes, and substrate transport properties typical of BMECs.

These combined differentiation and purification methods will yield sufficiently pure populations of functional endothelial cells for use in screens for drug permeability or modulators of BBB function. But that is not the only application of this system. “We hope to use this for blood-brain barrier science,” says Shusta. “One could model aspects of neurological disease or investigate blood-brain barrier development.” Indeed, the researchers used the coculture system to demonstrate that Wnts are a component of the cues with which neural cells promote differentiation of BMECs, in keeping with previous work.

Will this type of coculture model be useful for differentiation of hPSCs to other cell types? Sean Palecek thinks so. “This is a strategy that's fairly promising when you don't know the specific cues that are required to generate a certain cell type. You can create an environment in the culture dish similar to what they would see during development and systematically identify factors important for lineage specification,” he says. He cautions, however, that extending this strategy to cellular subtypes from organs that develop later than the brain may prove to be more of a challenge.