An implantable stimulator from SetPoint Medical would deliver shocks to the vagus nerve. It has not yet been tested in humans. Credit: Patrick T. Fallon for Nature

Six times a day, Katrin pauses whatever she's doing, removes a small magnet from her pocket and touches it to a raised patch of skin just below her collar bone. For 60 seconds, she feels a soft vibration in her throat. Her voice quavers if she talks. Then, the sensation subsides.

The magnet switches on an implanted device that emits a series of electrical pulses — each about a milliamp, similar to the current drawn by a typical hearing aid. These pulses stimulate her vagus nerve, a tract of fibres that runs down the neck from the brainstem to several major organs, including the heart and gut.

The technique, called vagus-nerve stimulation, has been used since the 1990s to treat epilepsy, and since the early 2000s to treat depression. But Katrin, a 70-year-old fitness instructor in Amsterdam, who asked that her name be changed for this story, uses it to control rheumatoid arthritis, an autoimmune disorder that results in the destruction of cartilage around joints and other tissues. A clinical trial in which she enrolled five years ago is the first of its kind in humans, and it represents the culmination of two decades of research looking into the connection between the nervous and immune systems.

For Kevin Tracey, a neurosurgeon at the Feinstein Institute for Medical Research in Manhasset, New York, the vagus nerve is a major component of that connection, and he says that electrical stimulation could represent a better way to treat autoimmune diseases, such as lupus, Crohn's disease and more.

Several pharmaceutical companies are investing in 'electroceuticals' — devices that can modulate nerves — to treat cardiovascular and metabolic diseases. But Tracey's goal of controlling inflammation with such a device would represent a major leap forward, if it succeeds.

He is a pioneer who “got a lot of people onboard and doing research in this area”, says Dianne Lorton, a neuroscientist at Kent State University in Ohio, who has spent 30 years studying nerves that infiltrate immune organs such as the lymph nodes and spleen. But she and other observers caution that the neural circuits underlying anti-inflammatory effects are not yet well understood.

Tracey acknowledges this criticism, but still sees huge potential in electrical stimulation. “In our lifetime, we will see devices replacing some drugs,” he says. Delivering shocks to the vagus or other peripheral nerves could provide treatment for a host of diseases, he argues, from diabetes to high blood pressure and bleeding. “This is the beginning of a field.”

Shock value

It was only by accident that Tracey first wandered down the path of neuroimmunity. In 1998, he was studying an experimental drug designated CNI-1493, which curbed inflammation in animals by reducing levels of a potent immune protein called tumour-necrosis factor-α (TNF-α). CNI-1493 was usually administered through the bloodstream, but one day, Tracey decided to inject it into a rat's brain. He wanted to see whether it would lower TNF-α in the brain during a stroke. But what happened surprised him.

CNI-1493 in the brain reduced production of TNF-α throughout the animal's body. Other experiments showed that it did this about 100,000 times more potently than when injected straight into the bloodstream1. Tracey surmised that the drug was acting on neural signals.

His follow-up experiments supported this idea. Minutes after he injected CNI-1493 into the brain, Tracey saw a burst of activity rippling down the rat's vagus nerve2. This neural highway regulates a handful of involuntary functions, including heart rate, breathing and the muscle contractions that push food through the gut. Tracey reasoned that it might also control inflammation. When he severed the nerve and the drug's potent effect disappeared, he was convinced. “That was a game-changer,” says Tracey. The finding meant that if one could stimulate the vagus nerve, the drug wouldn't even be necessary.

And so he tried a pivotal experiment. He injected a rat with a fatal dose of endotoxin, a component of the bacterial cell wall that sends animals into a spiral of inflammation, organ failure and death. The drug's effects roughly mirror septic shock in humans. Then, Tracey stimulated the animal's vagus nerve using an electrode. The treated rats had only one-quarter as much TNF-α in the bloodstream as untreated animals, and they didn't go into shock3.

Tracey instantly saw medical potential for vagus-nerve stimulation as a way to block surges in TNF-α and other inflammatory molecules. Companies were already selling implantable stimulators to treat epilepsy. But to extend the technique to inflammatory conditions, Tracey would need to present a clearer picture of how it might work and what the side effects might be.

Over the next 15 years, Tracey's team performed a series of animal experiments to identify where and how vagus-nerve stimulation acted. They tried cutting the nerve in different places4 and using drugs that block specific neurotransmitters5. These experiments seemed to show that when the vagus is zapped with electricity, a signal pulses down it into the abdomen, and then through a second nerve into the spleen.

The spleen serves as an immunological truck stop of sorts, where circulating immune cells periodically park for a while before returning to the bloodstream. Tracey's team found that the nerve entering the spleen releases a neurotransmitter called noradrenaline6, which communicates directly with white blood cells in the spleen called T cells. The junctions between nerve and T cell actually resemble synapses between two nerve cells; the T cells are acting almost like neurons, Tracey says. When stimulated, the T cells release another neurotransmitter, called acetylcholine, which then binds to macrophages in the spleen. It is these immune cells that normally spew TNF-α into the bloodstream when an animal receives endotoxin. Exposure to acetylcholine, however, prevents macrophages from producing the inflammatory protein (see 'A shock to the immune system').

Tracey's findings lent new significance to research that had been going on for decades. In the 1980s and 1990s, David Felten, a neuroanatomist then at the University of Rochester in New York, captured microscopic images of hybrid neuron–T-cell synapses in various animals7 — not just in the spleen, where Tracey saw them, but also in the lymph nodes, thymus and gut. These neurons belong to what is called the sympathetic nervous system, which regulates body responses to certain stressors. Just as Tracey found in the spleen, Felten observed that these sympathetic neurons stimulate their T-cell partners by secreting noradrenaline — and often, this stimulation serves to blunt inflammation.

In 2014, neuroimmunologist Akiko Nakai of Osaka University in Japan reported evidence that sympathetic-nerve stimulation of T cells limits them from exiting the lymph nodes and entering the circulation, where they might stir up inflammation in other parts of the body8. But in many autoimmune diseases, this neural signalling is disrupted.

Lorton and her twin sister, neuroscientist Denise Bellinger of Loma Linda University in California, have found sympathetic-nerve pathways to be altered in rat models of autoimmune disorders9. The same is seen in humans. Sympathetic nerves are damaged by over-release of noradrenaline, which causes them to withdraw from the immune cells that they should be moderating. As the disease progresses, these nerves advance back into the tissues that they abandoned — but they do so in abnormal ways, making connections with different subsets of immune cells. These rearranged neural pathways actually maintain inflammation rather than dampen it9. It happens in places such as the spleen, lymph nodes and joints, and is causing a lot of pathology, says Bellinger.

But she, Lorton and others are sceptical of Tracey's account of the pathway by which vagus-nerve stimulation lowers inflammation. Robin McAllen, a neuroscientist at the University of Melbourne in Australia, has searched for connections between the vagus nerve and the nerve that stimulates T cells in the spleen — but so far, he has found none.

Vagal stimulation “is acting indirectly” through other nerves, says Bellinger. It's important that these neural circuits are properly mapped before moving onto treatment in people, she says. “The anatomy makes a big difference in what kind of side effects you might see.”

Yet, even these sceptics see potential in Tracey's methods. Bellinger points out that in many autoimmune diseases, not only do sympathetic nerves become overactive as they rearrange themselves into proinflammatory circuits, but also the vagus nerve, which opposes them, becomes underactive. Vagal stimulation might partially restore the balance between these two neural systems. “It's a first step,” she says. “I believe that they will introduce it to the clinic, and they will show remarkable effects.”

A patient approach

People given vagus-nerve stimulation for seizures or depression experience some side effects — pain and tightening in the larynx, or straining in their voice, for example; Katrin feels a minor version of this when she stimulates her vagus. Shocking this nerve can also lower the heart rate or increase stomach acid, among other effects.

In this respect, Tracey has cause for optimism. The human vagus nerve contains around 100,000 individual nerve fibres, which branch out to reach various organs. But the amount of electricity needed to trigger neural activity can vary from fibre to fibre by as much as 50-fold.

Yaakov Levine, a former graduate student of Tracey's, has worked out that the nerve fibres involved in reducing inflammation have a low activation threshold. They can be turned on with as little as 250-millionths of an amp — one-eighth the amount often used to suppress seizures. And although people treated for seizures require up to several hours of stimulation per day, animal experiments have suggested that a single, brief shock could control inflammation for a long time10. Macrophages hit by acetylcholine are unable to produce TNF-α for up to 24 hours, says Levine, who now works in Manhasset at SetPoint Medical, a company established to commercialize vagus-nerve stimulation as a medical treatment.

By 2011, SetPoint was ready to try the technique in humans, thanks to animal studies and Levine's optimization efforts. That first trial was overseen by Paul-Peter Tak, a rheumatologist at the University of Amsterdam and at the UK-based pharmaceutical company GlaxoSmithKline. Over the course of several years, 18 people with rheumatoid arthritis have been implanted with stimulators, including Katrin.

She and 11 other participants saw their symptoms improve over a period of 6 weeks. Lab tests showed that their blood levels of inflammatory molecules, such as TNF-α and interleukin-6, decreased. These improvements vanished when the devices were shut off for 14 days — and then returned when stimulation was resumed.

Katrin, who has continued to use the stimulator ever since, still takes weekly injections of the anti-rheumatic drug methotrexate, as well as a daily dose of an anti-inflammatory pill called diclofenac — but she was able to stop taking high-dose, immune-suppressive steroids, and her joints improved enough for her to return to work. The results of this trial were published last July in Proceedings of the National Academy of Sciences11.

In our lifetime, we will see devices replacing some drugs.

Findings from another vagal-stimulation trial were published around the same time12. Bruno Bonaz, a gastroenterologist at the University Hospital in Grenoble, France, implanted stimulators into seven people with Crohn's disease. Over a period of six months, five of them reported experiencing fewer symptoms, and endoscopies of their guts showed reduced tissue damage. SetPoint is also midway through a clinical trial of its own, using vagus-nerve stimulation to treat Crohn's disease.

Tracey and Bonaz aren't the only people looking to harness neural circuits to treat inflammation. Raul Coimbra, a trauma surgeon at the University of California, San Diego, is studying it as a way to treat septic shock, which affects hundreds of thousands of people each year. Many people who die from the condition are pushed past the point of no return by a singular event: the rapid deterioration of the gut lining, which releases bacteria into the body — triggering inflammation that damages organs, including the lungs and kidneys.

Like Tracey, Coimbra has successfully counteracted this fatal sequence in animals by stimulating the vagus nerve, either with electricity13 or by administering an experimental drug called CPSI-121 (ref. 14). Coimbra hopes to carry this work into a clinical trial. But his research has also unearthed another major challenge that vagus-nerve stimulation must overcome: unlike rats, some humans are probably resistant to the technique.

The human genome codes for an extra, non-functioning acetylcholine receptor protein not found in other animals. Todd Costantini, a collaborator of Coimbra's also at the University of California, San Diego, has discovered that if this abnormal receptor is produced in sufficient quantities, it can disrupt signalling and render macrophages unresponsive to acetylcholine. They may then continue releasing TNF-α despite vagal stimulation15. There's a 200-fold range in the amount of this protein that people produce, says Costantini. He plans to test people to determine whether high levels really block the anti-inflammatory effects of vagal stimulation. Anecdotal evidence suggests that this might be the case.

The small clinical trials run so far have revealed that some people don't respond to vagal stimulation. It may be that testing could determine who will benefit from the treatment before people receive implants.

Despite the uncertainties, however, the field of electroceuticals is starting to gain momentum. Last October, the US National Institutes of Health announced a programme called Stimulating Peripheral Activity to Relieve Conditions (SPARC), which will provide US$238 million in funding until 2021 to support research updating the maps of neural circuitry in the thoracic and abdominal cavities.

GlaxoSmithKline is also showing interest. It has invested in SetPoint, and it announced last year the formation of a joint venture with Google — called Galvani Bioelectronics — that will develop therapies for a range of conditions, including inflammatory diseases.

Whether vagus-nerve stimulation lives up to expectations remains to be seen. The number of people who have been treated so far is minuscule — just 25 individuals in 2 completed trials. And treatments often look promising in early trials such as these, but then flop in larger ones.

But people with autoimmune disorders are starting to take notice. Treatments for rheumatoid arthritis and Crohn's disease carry some risks, and they don't help everyone. Katrin was one of more than 1,000 people who inquired about the trial for vagal stimulation. “I had nothing else,” she says. “I wanted it.”