Read our COVID-19 research and news.

Methods to influence rodent neurons using light have not translated smoothly into primate brains.


Controlling monkey brains with light could get easier thanks to open data project

When neuroscientist Sébastien Tremblay set out to manipulate monkeys’ brains with light, colleagues had sobering advice: “It’s more difficult than it sounds.” Tremblay, who works in neuroscientist Michael Platt’s lab at the University of Pennsylvania, uses light to activate or silence precise groups of neurons and probe their role in brain function. The method, called optogenetics, works well in rodents, but studies in nonhuman primates are critical if it’s ever going to become a therapy for humans—to suppress seizures, for example, disrupt tremors in Parkinson’s disease, or even project images into the brain of a blind person.

But in spite of more than 10 years of work, progress has been slow. The tools for rendering cells light sensitive were largely refined in rodents and behave unpredictably in monkeys. It’s hard to illuminate enough tissue in large primate brains to reliably change animals’ behavior. Researchers have devised their approaches by trial and error, often without knowing what had or hadn’t worked for others.

Tremblay, Platt, and colleagues from 45 primate optogenetics labs in nine countries hope to change that with the Nonhuman Primate Optogenetics Open Database, which published its first results last week. The database contains minute details of successes and failures, many of which have gone unpublished. And if it can be sustained, it may soon include tests in monkeys of promising new optogenetic tools. The open-data approach “is tremendously powerful, tremendously useful to the community,” says Hongkui Zeng, a neuroscientist who develops optogenetic tools for mice at the Allen Institute for Brain Science and was not involved in the project.

In optogenetics, researchers endow brain cells with a gene for one of several opsins, light-sensitive proteins from microbes. These proteins can influence the flow of ions in and out of a neuron to control whether it fires an electrical signal. Depending on the opsin, researchers can excite or inhibit neurons by shining light on them, usually via an implanted optical fiber.

Strains of mice have been genetically engineered to express opsins in their brains from birth. But for now, getting an opsin into monkey neurons means infecting the cells with a virus injected through a hole in the skull. Along with opsin DNA, the virus typically carries a sequence called a promoter, which restricts the opsin’s expression to certain cell types.

There’s no proven formula for getting monkey brain cells to make opsins. In the hunt for the right combination of viral strains and promoters, “we kind of entered this voodoo land,” says Arash Afraz, a neuroscientist at the U.S. National Institute of Mental Health. Scientists relied on rumors of other labs’ successes and failures, he says, and were afraid to vary a recipe once they got it working. Unlike with plentiful mice, researchers couldn’t afford to use lots of monkeys to hone their technique, he adds. “We value them more. They have names. We view them as our colleagues, in a sense.”

Afraz hopes the database, which he contributed to, will minimize wasted effort by pooling the field’s failures. It catalogs 1042 viral injections performed in nonhuman primates, 552 of them previously unpublished. Seven-tenths of the experiments were in rhesus macaque monkeys. Tremblay can’t be sure the database is exhaustive, but the 66 groups he invited to contribute—identified through publications and referrals from colleagues—represent the majority of labs active in the field, he says.

In a 19 October paper in Neuron introducing the database, the team estimates the success rate of the most commonly used vectors, promoters, and opsins in the data set. About half of the experiments in monkey brains looked for changes in neural activity after cells were hit with light; 69% found a strong effect. Of the 20% of experiments that aimed to influence an animal’s behavior—to prompt an eye or hand movement, for example—nearly half saw a weak effect or none at all.

Scaling up

Monkey brains have been harder to manipulate with light than mouse brains, in part because they’re much larger.

Macaque Mouse Optical fiber

Failure likely discouraged some researchers from publishing studies, says Julio Martinez-Trujillo, a neurophysiologist at Western University and contributor to the project. His group has tried, without success, to evoke eye movements in one macaque and to impair working memory function in another. “This is the first paper that shows our experience,” he says.

Such attempts probably fail in part because the virus doesn’t reach enough of the brain, Tremblay says. A single injection can infect about 1 cubic millimeter of tissue—a broad swath of a mouse’s brain, but a puny fraction of a monkey’s. And scientists want to avoid multiple injections that could cause excessive tissue damage. Instead, some labs are trying to send the virus farther by injecting it at high volumes and pressures, a technique called convection-enhanced delivery.

Others hope to eliminate the need for brain injections by designing viruses that are small enough to cross into the brain via its tiny capillaries after being infused into a vein. In a June preprint on bioRxiv, neuroscientist and bioengineer Viviana Gradinaru and her team at the California Institute of Technology describe such an engineered virus that selectively infects the neurons of a marmoset.

Delivering light to large brains is a hurdle as well. “Say I am using a 200-micron-diameter fiber optic for stimulating my mouse brain,” Afraz explains. “To scale that up, I’d have to stick a flashlight in the monkey’s head.” In a bioRxiv preprint last month, Afraz and colleagues describe a possible alternative: a 5-square-millimeter array of 24 light-emitting diodes (LEDs), each of which can produce as much light as a typical optical fiber. By laying this array over a monkey’s cortex, researchers might illuminate a relatively broad brain area without multiple implanted fibers, Afraz says. They can also use individual LEDs to excite separate parts of the cortex in precise patterns.

Other groups are developing more sensitive opsins so that weaker light can affect more distant tissue. In a study in mice reported on 5 October in Nature Biotechnology, a group led by Stanford University neuroscientist Karl Deisseroth—one of the original developers of optogenetics—used a highly sensitive opsin called ChRmine to activate neurons several millimeters below the brain’s surface with light from outside the rodent’s skull.

“I can’t wait to test them,” Laval University molecular biologist Marie-Ève Paquet says of these ultrasensitive opsins. She’s part of a Canadian collaboration that tests and disseminates emerging optogenetics tools. As opsins, promoters, and viruses make their way to participating research groups, Paquet’s team plans to upload its results to the new database.

To keep the database up to date, she says, “the community really has to be motivated,” especially because she expects the next few years to bring a boom in studies to influence and understand the brain circuits of some of our closest animal relatives.