Stopped at a red light on his drive home from work, Karl Deisseroth contemplates one of his patients, a woman with depression so entrenched that she had been unresponsive to drugs and electroshock therapy for years. The red turns to green and Deisseroth accelerates, navigating roads and intersections with one part of his mind while another part considers a very different set of pathways that also can be regulated by a system of lights. In his lab at Stanford University’s Clark Center, Deisseroth is developing a remarkable way to switch brain cells off and on by exposing them to targeted green, yellow, or blue flashes. With that ability, he is learning how to regulate the flow of information in the brain.
Deisseroth’s technique, known broadly as optogenetics, could bring new hope to his most desperate patients. In a series of provocative experiments, he has already cured the symptoms of psychiatric disease in mice. Optogenetics also shows promise for defeating drug addiction. When Deisseroth exposed a set of test mice to cocaine and then flipped a switch, pulsing bright yellow light into their brains, the expected rush of euphoria—the prelude to addiction—was instantly blocked. Almost miraculously, they were immune to the cocaine high; the mice left the drug den as uninterested as if they had never been exposed.
Today, those breakthroughs have been demonstrated in only a small number of test animals. But as Deisseroth pulls into his driveway he is optimistic about what tomorrow’s work could bring: Human applications, and the relief they could deliver, may not be far off.
For all its complexity, the brain in some ways is a surprisingly simple device. Neurons switch off and on, causing signals to stop or go. Using optogenetics, Deisseroth can do that switching himself. He inserts light-sensitive proteins into brain cells. Those proteins let him turn a set of cells on or off just by shining the right kind of laser beam at the cells. That in turn makes it possible to highlight the exact neural pathways involved in the various forms of psychiatric disease. A disruption of one particular pathway, for instance, might cause anxiety. To test the possibility, Deisseroth engineers an animal with light-sensitive proteins in the brain cells lying along the suspected pathway. Then he illuminates those cells with a laser. If the animal begins cowering in a corner, he knows he is in the right place. And as Deisseroth and his colleagues illuminate more neural pathways, other researchers will be able to design increasingly targeted drugs and minimally invasive brain implants to treat psychiatric disease.
Optogenetics originally emerged from Bio-X, a multidisciplinary project spearheaded in part by Stephen Chu, then a Stanford physicist and now the U.S. Secretary of Energy. Bio-X takes some of Stanford’s best engineers, computer scientists, physicists, chemists, and clinicians and throws them together in the Clark Center, where an open, glass-clad structure makes communication unavoidable. Deisseroth, whose beat-up jeans and T-shirt practically define the universal academic wardrobe, proved a natural at working across disciplines. Over the past decade, his omnivorous quest has filtered far beyond Bio-X into a thousand institutions around the world.
Although his Bio-X work involves esoteric genetics and animal experiments, Deisseroth has never forgotten the human needs that motivated him in the first place. He still divides his time between his basement lab and the psychiatry patients who desperately need his research to pay off.
Psychiatry’s Core Dilemma
Karl Deisseroth was 27 when he first brushed past the curtains of the psychiatry ward at Palo Alto’s VA hospital in northern California. It was 1998 and he had just completed his first two years of Stanford Medical School, where he had earned a Ph.D. in brain cell physiology, exploring the electrical language of neuron communication. As part of his medical training, he was required to complete a rotation in psychiatry—a hazy specialty, he felt, much less compelling than the brain surgery that was his career goal.
Several patients in the ward lay in narrow beds lined up before him, awaiting a treatment called electroconvulsive therapy (ECT). After the anesthesiologist on duty put them under, the attending psychiatrist placed pads on the patients’ temples and walked from bed to bed, pressing a small button on each person’s control box, sending volts of electricity into their brains. Their bodies tensed and their brains rattled with seizures for a full minute. The recipients risked losing large swaths of memory, but if things went well, the current would reset their neurons, purging their depression and providing months of relief.
From that experience, Deisseroth determined that he would spend his life solving a core puzzle of psychiatric disease: A brain could appear undamaged, with no dead tissue or anatomical deformities, yet something could be so wrong it destroyed patients’ lives. Perhaps because the damage was invisible, the available therapies were shockingly crude. ect was lifesaving but usually temporary; although it was likely that just a small set of cells caused the patient’s troubles, the shock jolted neurons throughout the brain. Psychoactive drugs, targeting general brain regions and cell types, were too broad as well. And scientists were so uncertain about what chemical imbalances impacted which neural circuits that one-third of people with major depression did not respond to drugs at all.
Deisseroth pondered the problem through a subsequent psychiatry residency, where he oversaw more than 200 ECT procedures over four years. Then, in 2004, he became a principal investigator at Stanford and was given his own lab. As a clinician treating patients, his arsenal was limited. But with his scientific imagination roaming free and a brand-new lab sparkling with empty chairs and beakers to fill, he began to envision elegant new strategies. One stood out: a concept first suggested by Francis Crick, the legendary genetics and consciousness researcher.
Crick’s idea was that light, with its unparalleled speed and precision, could be the ideal tool for controlling neurons and mapping the brain. “The idea of an energy interface instead of a physical interface to work with the brain was what was so exciting,” Deisseroth says. He thought creating a light-sensitive brain was probably impossible, but then an idea floated up: What about tapping the power of light-sensitive microbes, single-celled creatures that drift in water, turning toward or away from the sun to regulate energy intake? Such brainless creatures rely on signals from light-sensitive proteins called opsins. When sunlight hits the opsin, it instantly sends an electric signal through the microbe’s cell membrane, telling the tiny critter which way to turn in relation to the sun.
Deisseroth wondered if he could insert these opsins into targeted mammalian brain cells in order to make them light-sensitive too. If so, he could learn to control their behavior using light. Shining light into the brain could then become the tool Crick imagined, providing a way to control neurons without electric shocks or slow-acting, unfocused drugs.
Lighting the Brain
The necessary tools were already out there. The first opsin—the light-sensitive protein made by microbes—had been identified in 1971, the same year Deisseroth was born. Bacteriorhodopsin, as it was called, responded to green light, and scientists have since found it in microbes living in saltwater all over the world. The next opsin, halorhodopsin, which responds to yellow light, was discovered in 1977. Like bacteriorhodopsin, it was found in bacteria living in salty lakes and seas.
Deisseroth, who read everything he could about opsins, realized that light-sensitive microbes speak the same basic language as neurons: When light hits the opsin, gates in the cell membrane open, allowing charged particles called ions to flow in and out. In microbes, ion flow tells the organism which way to turn. In neurons, ions flowing through the cell wall initiate action, setting off a string of communications that tell organisms like us how to feel and behave. This similarity suggested to Deisseroth that opsins could be manipulated to switch brain cells on and off.
Deisseroth was still mulling this over in 2003, when German biologists Georg Nagel and Peter Hegemann announced a new light-sensitive microbe, a green alga called Chlamydomonas reinhardtii. The 10-micrometer-wide microbe has a small eyespot, which Deisseroth describes as “kind of cute,” that spins around to detect light. It makes a protein called channelrhodopsin-2 (ChR2) that acts as an antenna to receive blue light and convert it to an ion flow. When a light shines on ChR2, the cell becomes active and tells the microbe where to turn.
Deisseroth immediately wanted ChR2. The other opsins might do the trick, but because his goal—putting them into a brain and getting that brain to respond—was so tricky and success so improbable, he needed to try as many options as possible. He wrote to Nagel in the spring of 2004, requesting a copy of the gene and explaining he planned to try inserting it into neurons. “I was realistic enough to know it was worth testing but probably a long ways away from being useful,” he says. “If I’d told him I was going to cure depression with it, I’m sure he would have thought I was crazy.”
Deisseroth realized that even the first step of his plan—inserting the microbial opsin molecule into a mammalian neuron and getting the two to sync up—was a long shot. For one thing, there was a good chance the mammalian immune system would reject the foreign protein. Even if the opsin was tolerated, there was no way to know whether it could toggle mammalian cells in the same way it controlled algae. The opsin’s electric signals would need to fire and shut down within milliseconds of the stimulus to communicate as quickly as neurons; Deisseroth doubted that the simple biology of algae required such speed.
To run the necessary tests, Deisseroth had to hire staff for his lab, and fast. Someone would have to provide expertise in handling viruses—specifically, a virus to serve as a vector, or Trojan horse, to cart algal genes into mammalian cells. The gene for the opsin would need to be inserted into the virus, which would infect the neurons, transferring the opsin gene to them. If all went as planned, mammalian neurons would then produce light-sensitive microbial opsins as if they were proteins of their own.
A Team Is Born
Luck was on Deisseroth’s side that summer of 2004. As a new Stanford faculty member, he had moved into an office that had been occupied by Steven Chu, who had recently left Stanford to become director of the Lawrence Berkeley National Laboratory. Deisseroth’s door still had Chu’s name on it. One afternoon, a disoriented young chemistry student named Feng Zhang wandered in, looking for Chu. “I can still remember looking at him—he was a little surprised to see me,” Deisseroth says. But the two started talking. Zhang wanted to understand the chemical imbalances underlying depression. He also had the skills to help Deisseroth with viruses: At age 15 he had started working with viral vectors, a project that won him the top prize at the Intel International Science and Engineering Fair. Now an aspiring Stanford Ph.D., he decided to join Deisseroth’s team.
Next, Deisseroth required someone skilled at patch clamping, a technique that uses an electrode to record ions passing through cells. This would allow him to record when neurons fired or shut down, indicating whether they were responding to light. For this he hired Ed Boyden, a newly minted neuroscience Ph.D. at Stanford. Boyden was brilliant and energetic, with an aggressiveness that was sometimes off-putting but was ideal for tackling nearly impossible experiments and getting them published. He also had expertise in electrophysiology, another skill required for Deisseroth’s nascent optogenetics project.
That fall Deisseroth set to work with his new team. First they inserted the ChR2 opsin into a harmless retrovirus that Zhang had harvested. Then they added the engineered virus to a culture of rat neurons in a petri dish. As hoped, Zhang’s virus penetrated the neurons and delivered the light-sensitive gene. The final step was observing whether the cell actually fired quickly in response to light. Boyden hooked up one neuron to a glass electrode that could also deliver light. The other end of the electrode was attached to a computer. When the cell was quiet, a steady line appeared on the computer screen; when it was active, the line jumped up in a spike.
To Deisseroth’s elation, the effort was a success: As Boyden poked the electrode into the cell, Deisseroth saw pulses of bright blue light in the culture dish and spikes precisely matching those pulses on the computer screen. “For the next nine months we worked frenetically to publish it. We wanted to move quickly,” Deisseroth says. The paper, published in Nature Neuroscience in August 2005, chronicled the first time anyone had managed to control brain cells with light.
The cell cultures still did not prove whether optogenetics would apply to brain cells inside living, freely moving mammals, however. The effort to find out required expanding the team. By 2006 Deisseroth had a tight-knit group of 15 who took frequent excursions to local Indian buffets and In-N-Out Burger when they were not working intensely side by side.
Cracking the Animal Code
In cell culture, only a small number of mild virus particles were needed to deliver the opsin gene to targeted neurons. But inserting genes into mammalian neurons inside an intact brain required a larger number of more virulent viruses. Zhang worked tirelessly on this challenge, developing a highly concentrated but still-safe retrovirus derived from HIV; in essence, he removed HIV's toxic genes and replaced them with a version of ChR2. He could brew the virus from scratch in just three days.
Deisseroth also needed a miniature flashlight that could be surgically inserted in the brain to turn cells on and off at close range. Mice weigh only about 20 grams, less than two tablespoons of sugar, so the device could not be big or heavy. And although the light needed to be 100 times as bright as room light, the system could not heat the brain as it delivered the beam. The team’s solution was to implant a fiber-optic cable in the brain and connect it to a miniature laser affixed to the animal’s head. The contraption was small and light enough to travel with the mouse wherever it went.
Finally, Deisseroth needed a way to tag the specific neurons he wanted to study so that only those cells would become activated in response to the light. Other brain researchers had identified certain cell types and areas of the brain associated with fear, reward, addiction, and depression. But they had no way of knowing exactly which neurons within these regions were driving a particular behavior. Deisseroth strove to find out. He used snippets of DNA called promoters to link ChR2 genes with DNA found only in the specific neurons he wanted to study. When he shined his light, it would not disturb the entire region but just the relevant cells.
Only then was Deisseroth ready to test optogenetics in a living animal. He charged Zhang with conducting a study of hypocretin neurons, sleep-related cells located deep in the brain’s hypothalamus. The cells are crucial for arousal during sleep-wake cycles and are thought to play a key role in narcolepsy.
Zhang did the research at Stanford Sleep Center, where he could record brain waves of snoozing mice. He targeted ChR2 to the sleep cells and then, using optical fiber, delivered light directly to the mice’s brains. In early 2007 his team placed a ChR2-altered mouse in a sleeping chamber with two implants in its skull. One was the optical fiber; the other consisted of four wires that measured the animal’s brain waves.
Deisseroth vividly remembers the moment when an excited postdoc summoned him to the room. “I walked in and he whispered to me, ‘Be quiet.’ ” A mouse was peacefully dreaming in his chamber. But when the laser was turned on, they saw a slight change on the brain-wave monitor and the animal began to twitch. It was waking up in response to a light signal inside its brain. For the first time ever, Deisseroth’s team had used optogenetics to control behavior in a living animal.
Soon after, in March 2007, their results were more dramatic still. Deisseroth implanted an optical fiber in the cortex of a mouse with ChR2 in its motor neurons. When he flashed blue light through the cable, a meandering mouse began running to the left. When the laser was switched off, the mouse resumed wandering aimlessly. “You can turn it on and off and the animal isn’t distressed. It’s comfortable. You’re just reaching in there with the fiber-optic, controlling the cells, and you’re causing its behavior,” Deisseroth says. “That was the moment I knew this would be amazing.”
In the five years since, the Deisseroth lab, dubbed the D-lab, has expanded into an entire brain-control research center, with more than 40 scientists on the job. Molecular biologists, neuroscientists, engineers, and physicists from all over the world rush through his cavernous laboratories, tinkering with microscopes, lasers, viral soups, electrodes, and rodent brains. Located in the heart of Silicon Valley, the D-lab feels like an entrepreneurial start-up. Members enthusiastically talk among themselves, build and invent together—there is a palpable sense of enthusiasm and urgency.
One of the team’s greatest accomplishments was spearheaded by Kay Tye, a former postdoc who now works at MIT. In a lab near Deisseroth’s office, Tye inserted a fiber-optic cable into a mouse’s little brain at just the right spot, leaving enough slack for the animal to run around. Tye was studying anxiety circuits and needed to put the cable into a specific part of the amygdala. For decades, researchers have known that the amygdala is associated with fear and anxiety but did not know exactly which neurons in what part of the amygdala played a role. Tye used data from previous studies to home in on a likely circuit, then carefully positioned the cable to deliver light right there. As the targeted neurons were stimulated, she watched to see how the mouse’s behavior changed. If it suddenly became bolder, that would be a good sign that she had found a neuron set involved in anxiety.
Mice are naturally fearful of exploring open spaces, where they are vulnerable to predators. When placed in Tye’s four-armed maze, they would spend most of their time in the two arms protected by high walls, occasionally poking a nose out to explore. But when Tye switched on the light and activated the circuit in her subject’s brain, the mouse ventured out, exploring the open part of the maze with no visible anxiety. The results suggested that Tye had located an anxiety circuit in the brain that could someday be targeted by drugs.
Breaking the cycle of addictive behavior was another goal for the D-team. Again working with mice, they built a three-chamber cage in which one room became a designated drug den. Mice in that room received a shot of cocaine. Animals typically formed a positive association between the effects of the cocaine and the room, just as a person addicted to alcohol might form an association between feeling good and the pop of a cork. Left to their own devices, the mice hung around the room long after the cocaine wore off, even when they were free to wander elsewhere.
But when mice were injected with cocaine and also treated with halorhodopsins and light—in this case a yellow pulse sent directly to the brain’s reward center—the rush of euphoria was blocked. Those mice never formed a positive association between cocaine and the room and roamed freely around the cage.
Pacing the Heart with Light
Later in 2010, Deisseroth teamed with neuroscientist Anatol Kreitzer at the University of California, San Francisco, to investigate Parkinson’s disease—an important step toward using optogenetics to target a neurodegenerative disease. The ultimate cause of Parkinson’s is unknown but clearly involves the loss of a set of neurons that control voluntary movement. The basal ganglia are the brain’s action control center. One pathway there sends signals to “go,” as in go ahead and perform this action, and one sends “stop” signals. In Parkinson’s the pathways are thought to be out of balance, with interrupted motor cells causing the debilitating tremors and loss of movement control symptomatic of the disease.
Although this theory of Parkinson’s had been widely considered since the 1980s, there was no way to probe the circuit directly until optogenetics came along. Working with mice, Deisseroth and Kreitzer activated the “go” and “stop” circuits with light, confirming that one in fact facilitates movement while the other inhibits it. Next they tested a more nuanced hypothesis: Might Parkinson’s result from an overactive stop circuit? Deisseroth and Kreitzer tagged that circuit with ChR2 and delivered blue light directly into the brains of mice. When the light turned on, movement slowed and the mice had trouble walking, both symptoms of Parkinson’s.
What the researchers really wanted, though, was insight into how to treat the disease. They thought activating the go pathway could rebalance the overactive stop network. When they targeted the go circuit, that approach worked even better than expected. The mice began walking normally again, their movement indistinguishable from the way they had moved in their healthy state. Today’s leading treatment for Parkinson’s—deep brain stimulation—involves inserting a large electrode deep within the patient’s brain and zapping all surrounding tissue. Deisseroth hopes that his findings will bring a more targeted treatment soon.
Indeed, by combining opsins, including ChR2, which turns cells on, and halorhodopsin and bacteriorhodopsin, which turn cells off, Deisseroth can ask ever more nuanced questions about complex diseases: Epilepsy, autism, sleep disorders, and schizophrenia may all require this combination approach.
Turning cells on and off efficiently allows a whole range of new, more detailed experiments: Now Deisseroth can tell neurons to fire and shut down quickly, so they can be ready to receive the next signal telling them what to do. Using multiple opsins as well as blue, yellow, and green light, he can experiment with various combinations of activation in hopes of eliminating symptoms of disease.
Pacing the Heart with Light
Despite the fact that Deisseroth has focused on animal brains, the first optogenetic implants—which could be ready for human trials in as little as a decade—will almost surely focus on other organs, where applications are less risky. Early therapies could take the form of a heart pacemaker that uses light to activate heart cells and keep them firing on time. There has been talk of optogenetics for the blind, implanting opsins in vision cells and developing special glasses that shine light into them.
In the fall of 2011, Deisseroth cofounded a company in Menlo Park, around the corner from Stanford, dedicated to translating optogenetics research into therapies. One focus is peripheral nerve disorder, in which messages between the brain and the rest of the body are interrupted. It is often caused by spinal cord injuries, multiple sclerosis, and other nervous system disorders.
“It’s not very glamorous, but there’s a very large population of people who have peripheral nerve defects that keep them from having good bowel and bladder control,” Deisseroth says. “And what’s interesting is, if you ask the people who have paralysis if they could choose one thing, to be able to walk or to have bowel and bladder control, they essentially all pick bowel and bladder control, because it’s the most limiting for them. It is a problem well suited to optogenetics.” Bladder control requires both a contraction of the bladder and a relaxation of the sphincter, and optogenetics can both stimulate and inhibit those different neurons at the same time. Deisseroth hopes to introduce opsins to the crucial peripheral nerves outside the brain and then use simple LED implants to switch function back on.
Once someone has figured out how to get opsins inside the brains of primates and humans—Zhang at MIT is working on the problem now—optogenetic therapies targeting the brain can begin. The possibility also opens the door to Orwellian fears. If Deisseroth can control the brains of mice with light, what is to stop human mind control? The most cogent answer is this: Creating transgenic people by sending a retrovirus into healthy brains will never be allowed. Besides, the potential for healing is too great to ignore—starting with a better implant for those who suffer from Parkinson’s, a neurodegenerative disease already treated with electrodes in the brain.
Getting into the Human Brain
Deisseroth’s great insight has spawned research around the world. Every two weeks, scientists come from universities in the United States and abroad to spend a week at the D-lab learning the secrets of optogenetics, mastering everything from mouse surgery to cooking up viruses. At the end of the week they present their plans for research of their own. Deisseroth slouches in his seat, wearing coffee-stained jeans, clogs, and a short-sleeve button-down shirt that he has not tucked in. The laissez-faire demeanor is deceptive: Deisseroth is fully engaged and always on, often jumping in during a presentation to ask questions or offer suggestions. The waiting list to attend his workshop is more than a dozen labs long.
One notable alumna is Ana Domingos, who flew in from New York’s Rockefeller University a few years back. She was investigating weight loss and wanted to use optogenetics to trigger dopamine, a mood-enhancing neurotransmitter, whenever mice drank water laced with an artificial sweetener, causing them to ignore their usual preference, a sugar-spiked drink. Domingos hopes to use her findings to develop weight loss therapies. “The first time I saw the mouse bingeing on water with sweetener, I got goosebumps,” Domingos says. “I couldn’t sleep. Karl gave me the tools to play god.”
Following these presentations, Deisseroth grabs lunch before attending his weekly patient psychiatry sessions. He picks an outdoor seat at a nearby café swarming with people on a sunny, 75-degree day in mid-January. It’s a rare moment of downtime for Deisseroth, who readily admits he needs to relax more.
Even with his lab in high gear, Deisseroth is constantly busy trying to help his psychiatry patients. One of them, Alicia A., has tried nearly every medication, ECT, and various electrical implants to keep her depression under control. She drives seven hours once a month to visit Deisseroth, and together they have found a successful combination of electrical nerve stimulation and antidepressant drugs that has allowed her to return to work and enjoy life. Yet she intently follows Deisseroth’s optogenetics work and is adamant that if he ever starts human trials, she will be the first in line.
As much as Alicia A.’s life has improved from sessions with Deisseroth, the electrical stimulation is often uncomfortable, and her treatment requires constant monitoring. Deisseroth has an entirely different therapy possibility in mind for her. From his experience with ECT, he knows inducing a seizure with electricity resets individual neurons in the brain just like rebooting a computer, so those neurons fire all at once in a different order than before. But something peculiar and fascinating happens to the patient: When the therapy is over everything about the person—memories, priorities, the sense of self—comes back. Apparently these things are not generated by neurons but arise from the brain’s physical structure and wiring. The wires are like superhighways, roads of activity where circuits of neurons constantly communicate, but sometimes the road might be gridlocked or icy, and the messenger can’t get through.
At one level, optogenetics is nothing more than using light to control a targeted population of cells. But how these cells are wired up is a huge puzzle in itself and, to Deisseroth, one that lies at the true root of future psychiatric cures. To turn his wiring insights into therapy, he wants to use optogenetics to narrow down which circuits dictate which specific behaviors. Then, if he can determine whether the circuits are somehow impeded or blocked, he can try to physically shift them and normalize activity flow.
The Magnetic Cure
Deisseroth isn’t certain which tools will allow him to study these connections—it’s a capability beyond the reach of optogenetics—so he is once again on the edge of something big and unprecedented. A type of brain imaging, called diffusion tensor imaging, allows doctors to scan patients and produce vibrantly colored images of the brain’s wiring. These connections vary from individual to individual. When abnormalities are detected, a machine therapy called transcranial magnetic stimulation (tms) can send into the brain magnetic pulses powerful enough to shift and rewire those connections so their function is improved. tms is already used to treat ailments like Parkinson’s disease, migraines, and depression.
Years ago as a psychiatry resident, Deisseroth assisted with the clinical trial that got the therapy FDA-approved. He plans to continue using optogenetics to pin down circuits of brain cells responsible for disease and to combine that knowledge with the colorful circuit images to home in on which wires need to go where to establish normal communication. Then tms can move the wires precisely where they need to go to cure any particular illness. If it works, scientists would have a complete understanding of an individual patient’s brain.
The concept may sound extraordinary, too grand to work, but this is the type of challenge Deisseroth loves most. “I want to come up with totally new things, so I don’t want to be affected by too many preconceptions,” he says. Conveniently, Deisseroth’s own brain is wired to generate its best ideas in moments of isolation. “I can remember a couple key insights just driving in my car. For me, that’s meditative. I rarely solve a problem by thinking about it. The insights usually come from out of the blue, like a bolt.”
