ABOVE: Histological stain of mouse motor cortex Corinne Lee-Kubli, Chalasani Lab

More than a decade ago, scientists developed optogenetics, a method to turn cells on and off with light. The technique allows scientists to spur or suppress cells' electrical activity with just the flip of a switch to tease apart the roles of specific cell types. But because light doesn’t penetrate deep into tissues, scientists need to surgically implant light sources to illuminate cells below the surface of the skin or skull.

In a new study published today (February 9) in Nature Communicationsresearchers report they’ve found a way to use ultrasound to noninvasively activate mouse neurons, both in culture and in the brains of living animals. The technique, which the authors call sonogenetics, elicits electrical activity in a subset of brain cells that have been genetically engineered to respond to sound waves. 

“We know that ultrasound is safe,” study coauthor Sreekanth Chalasani, a neuroscientist in Salk’s Molecular Neurobiology Laboratory, tells The Scientist. “The potential for neuronal control is huge. It has applications for pacemakers, insulin pumps, and other therapies that we’re not even thinking about. 

Jamie Tyler, a biomedical engineer at the University of Alabama at Birmingham who was not involved in the study but has previously collaborated with some of its authors, tells The Scientist that the work represents “more than just a step forward” in being able to use ultrasound to control neural activity: “It shows that sonogenetics is a viable technique in mammalian cells.”

Chalasani and his colleagues pioneered a similar technique seven years ago in nematode worms, Caenorhabditis elegans. The researchers took advantage of a protein called TRP-4, a mechanosensitive ion channel that helps the worms sense when their bodies are stretching—and one that is apparently sensitive to ultrasound. In C. elegans, adding TRP-4 to neurons that didn’t typically produce the channels made the neurons ultrasound sensitive. 

See “Stimulating Neurons with Sound

The next step was to try adding TRP-4 to mammalian cells, but when the research team tried that, “nothing happened,” Chalasani says. 

For the next six years, Chalasani’s lab continued to strive toward getting sonogenetics to work in mammals. The team wanted to get cells to respond to 7-MHz waves, which they consider a safe frequency for living tissue. 

“Sound is mechanical energy, so we looked at mechanically sensitive proteins,” Chalasani says. “We spent 18 months testing each one of them.”

One by one, the researchers engineered 300 proteins into human epithelial kidney (HEK) cells, a commonly used cell line that is normally not sensitive to ultrasound. Their aim was to engineer HEK cells to produce mechanosensitive receptors on their membranes that, when activated with ultrasound, would allow positively charged ions to pass through, resulting in an accumulation of positive charge inside the cell. In neurons, this rush of ions would produce an action potential.

Eventually, the researchers found one mechanosensitive ion channel that was responsive to ultrasound stimulation. The TRPA1 receptor, also known as the wasabi receptor, is a nonselective ion channel found naturally on many mammalian cells. Present in the gut, colon, stomach, esophagus, brain, and heart of many mammals, the receptor is thought to help sense pain, cold, and touch.  

The researchers genetically engineered mouse neurons in a dish to produce hsTRPA1 and found that doing so could induce the cells to respond to ultrasound. The next step was to use a combination of transgenics and virally delivered genes to deliver hsTRPA1 to neurons deep in the motor cortex in live mice. Using histology, the researchers showed that these mice only expressed hsTRPA1 in cortical motor neurons, showing that, like with optogenetics, the researchers could modify only a subset of cells. Ultrasonic stimulation at 7 MHz produced movement in the mice’s front and back legs, indicating that ultrasound stimulation was likely activating the altered cells in the motor cortex.  

“I found it an exciting and brilliant study,” David Maresca, a biophysicist at Delft University of Technology in the Netherlands who was not involved in the work, tells The Scientist. He adds that one of its strengths is that the researchers found proteins “that worked at high ultrasound frequencies. And by working at high frequencies, they don’t disrupt healthy neurons in any way.” According to Maresca, low-frequency ultrasound has “a lot of strange effects” on brain function.

The researchers don’t yet fully understand how hsTRPA1 senses ultrasound. Unlike other channels in the TRP family, the hsTRPA1 receptor is not “traditionally mechanosensitive,” says Chalasani. “This was a big surprise. This meant that ultrasound wasn’t actually just a mechanical stimuli. It was doing something else to the cell.”

The researchers did identify a part of the protein that seems to be important for ultrasound sensitivity. And they found that the structural protein actin is likely involved, as actin-degrading compounds decrease cells’ sensitivity to ultrasound. 

The cell’s inner membrane is attached to actin, Chalasani explains, so “our prediction is that . . . ultrasound is shifting the outer membrane [of cells] without affecting the inner membrane.” This, he notes, adds space between the membranes, possibly making the cells more electrically active. “What we don’t know is whether sticking TRPA1 in these membranes would allow the membrane to move even more or less.” 

Maresca says that finding the mechanism behind hsTRPA1 ultrasound sensitivity is “the million-dollar question for the field.” The authors “try to give a few hints about this mechanism, but I think the community at large still does not really understand how an ultrasound wave activates a neuron,” he says.

The potential for neuronal control is huge.

—Sreekanth Chalasani, Salk Institute

The researchers say they hope that sonogenetics will someday be used therapeutically in humans.

“The big prize will be to replace deep brain stimulation,” Chalasani says. Deep brain stimulation, a treatment for major depression and Parkinson’s disease, is currently performed by implanting electrodes deep into the brain that stimulate cells with electricity. Hypothetically, sonogenetics could allow clinicians to stimulate the deep centers of the brain noninvasively. A similar principle might work with the vagus nerve, Chalasani adds, a cluster of neurons in the neck that carries information to and from the brain and is targeted in treatments for seizures, PTSD, and depression.

Tyler says that he’s excited to see future developments in this project, given how much optogenetics has moved forward in the past decade. “There’s a whole new realm of possibilities,” he says. “It’s a new toolbox we can start to use.”

Maresca says he’s also excited about the potential future applications of this technology, saying that, with their approach, the researchers could “create a library or a toolset of similar genetic proteins that could be used at different frequencies, for different applications, and in different tissues, from the skin to the brain.”