‘Brain Balls’ Grown From Skin Cells Spark With Electricity

Globe-like agglomerations of human neural cells look and act strikingly like real, living grey matter.
Crosssection of a human cortical spheroid showing the specific arrangement of radial glial processes orthogonal to the...
Cross-section of a human cortical spheroid showing the specific arrangement of radial glial processes, orthogonal to the lumen of a ventricular-like zone. Radial glia are labeled with an anti-GFAP antibody (red); nuclei labeled with DAPI (blue).Stanford University

Scientists shouldn’t be allowed to name their own creations. Today, researchers at Stanford announced a new way of creating gobbets of human brain cells that look and act like real, living grey matter. The researchers took this striking result and named their product "human cortical spheroids," or hCSs. Which is terrible. C’mon guys, tell it like it is: You’re making brain balls.

In recent years, physiologists have learned to make and grow neural cells that look more and more like the real thing---most recently, by moving cell cultures beyond flat layers on the bottom of a Petri dish and into the third dimension. (Is this sounding like an ad for a 3-D movie?) A group out of Japan’s RIKEN Institute, led by the late Yoshiki Sasai, recently developed a cerebellum-like 3D culture. Jorgen Knoblich’s group at the Austrian Academy of Sciences created what they’re calling "cerebral organoids." (Again, really: brain balls.)

The spheroids made by Sergiu Paşca’s group at Stanford aren’t the first 3-D neural cultures, then. But they are the first that neuroscientists have been able to study functionally, looking at the electrical workings of their structure as a whole. Nobody understands the workings of the entire brain as it fires, but at least they can begin to figure out how these simplified 5-millimeter globes of cells work.

To grow their spheroids, the group started with stem cells, cells (derived in this case from skin) that---with a little tweaking--grow into any kind of cell a researcher wants. It's a property called "pluripotency." But those cells won't grow on their own; the team used a mixture of neuron-fertilizing molecules in a fluid bath.

It worked. And the neurons didn't just divide and grow: They actually echoed some of what would happen to cortical neurons in a real, live brain. The tiny, growing brain balls curled inward, developing multiple layers of neurons both deep and superficial, just like the human cortex does.

Critically, after a certain amount of time the brain balls also started growing a type of cell called an astrocyte. These star-shaped cells provide physical and perhaps chemical support to neighboring neurons, so some of these brain balls stayed alive (and kept growing) much longer than they typically would---as old as 300 days. The cells also are critical for the formation of synapses, the junctions where neurons trade electrical messages. Because Paşca’s team was able to grow astrocytes alongside these cortical neurons, almost 90 percent of all the neurons inside the spheroids had active synapses, spontaneously sending electric missives to the network around them. They weren't "thinking," but they were doing something.

Cross-section of a human cortical spheroid showing a ventricular zone (high nuclear density) and surrounding neurons as shown by the expression of the neuronal marker MAP2 (red). ). Nuclei labeled with HOECHST (cyan).

Stanford University

That meant the team was able to do something other culturing methods don't allow: slicing and studying brain balls like actual brains. When neuroscientists study neural networks in mice, “you take the mouse brain and slice it into thin slices,” says Paşca. “What we have done is take these spheroids and slice them like you would with a rodent brain and do slice recordings.”

Electrophysiological recording is a big deal. “The studies presented here, particularly the electrophysiology, raise the hope that organoid systems can be used for modeling neuronal network activity," Knoblich, creator of those other organoids, wrote in an email. "They demonstrate a response of the neurons to external stimuli---an observation that has not been described before."

Paşca talks about these brain balls, now growing by the thousands in his lab, like a mother hen protecting its eggs. Each dish, filled with its nourishing growth factors, supports 50 to 100 spheroids. And each is a different age, cultured from a different person’s induced stem cells. So each dish has its own identity, in a way---one that needs to be protected by changing that nutrient bath every four days and applying antibiotics to resist infection.

Maybe Paşca knows how this sounds, because he finds it necessary to interject, without my asking: “This is not a small human brain in a dish,” he says. “I don’t have any interest in building that.”

That would be creepy. A lot of people do, though. The ultimate goal for these 3-D cultures is to mimic the brain’s actual cytoarchitecture as closely as possible. “The thing that everybody has been waiting for is ‘can we build a circuit in a dish?'” says Paşca. “We’re not there yet, but for now we have built a very complex neural network.”

To make that a reality, a lot of work still needs to happen. “The method suffers from the same weaknesses that we are all trying hard to overcome,” writes Knoblich. It’s tough for a neural network to grow very large or sustain itself without a blood supply, which is why the spheroids are unlikely to grow much larger than 5 millimeters across (and why the researchers have to constantly switch out the nutrient broth, like cleaning a fish tank). More importantly, perhaps, while a certain amount of structure developed spontaneously in the capsules (as other culturing techniques have achieved as well), the brain balls still don't develop the complex convolutions on the surface of the brain called gyri.

Another difference between brains and brain balls: For now, the method only produces excitatory neurons, the kind that tell other neurons to do more, more often. But there are plenty of other kinds of neurons: oligodendrocytes and microglia and inhibitory neurons and interneurons. That's how a brain programs itself. In a way, that limitation on brain balls is good---biologists know little enough about how human neurons interact with each other that it makes sense to study one class on its own before expanding into others. But if you really wanted to recapitulate neural development in a dish, you’d need the other types. “It will be very complicated,” says Paşca. “There are so many developmental cues that we don’t even know of yet that lead to all of those different cell types.”

For now, Paşca will keeping growing his spheroids---and use them to study excitatory neurons cultured from patients with different disorders of the brain. “We already have a very broad collection of induced pluripotent stem cells from patients with autism or schizophrenia,” says Paşca, “and we’re developing spheroids from those cells.” Maybe they'll come up with a better name next time.