Woods’s and Waxman’s work suggested a potential target for a novel painkiller. Opioids target the parts of the brain that receive pain signals. A drug acting on sodium channels might mitigate the sending of pain signals.
“We know that, in radios and computers, electricity is carried by electrons through wires,” Chris Miller, a professor emeritus of biochemistry at Brandeis University, explained to me. “In biological systems, it’s carried by ions via ion channels.” Miller has spent decades studying how the channels work. “I don’t really care what these molecules do for human health—I just find them such fascinating entities. A nerve spike will zoom down an axon to the tune of one hundred metres per second.” He compared that with other bodily systems, like hormones, which effect changes over minutes to hours. It is only relatively recently that we began to understand in much detail how the channels in our nerves work. In August, 1939, the British physiologist Alan Hodgkin and his student Andrew Huxley (Aldous’s half brother) examined squid giant axons, which are up to a thousand times thicker than typical human nerve fibres and thus easier to study. Hodgkin and Huxley used fine electrodes to look for voltage differences across axons, and within a few weeks had exciting preliminary results—but then Hitler invaded Poland. Their work was put on hold for about seven years. (Hodgkin went into radar development.) In 1946, before modern computers or microelectrodes, Hodgkin and Huxley designed clever experiments from a few basic measurements that allowed them to conclude that the nerve cells must have ion channels embedded in them, regulating the flow of current. (We now know that there are channels specific to five kinds of ions—sodium, calcium, potassium, chloride, and hydrogen ions—that generate electrical signals in nerves and other cells.) “They couldn’t see the channels,” Waxman said, with admiration. “They had no idea of their structure. Yet they predicted their presence and their properties with great prescience.”
A decade earlier, an anesthetizing compound that acted on sodium channels had been found—though it wasn’t understood that it was sodium channels it was acting on. Researching a mutated strain of barley, scientists at Stockholm University tried synthesizing substances that lent the plant pest resistance. The testing method was of its time. “One of them tests a compound on his tongue, and his tongue goes numb,” John Wood, a professor of molecular neurobiology at University College London, whom Woods describes as “the doyen of sodium channels,” told me. During the war, the Swedish anesthesiologist Torsten Gordh ran a small trial using his medical students as subjects. As compensation, he offered them a choice of a copy of his Ph.D. dissertation or a pack of cigarettes. Half the students were given the compound, half were given the placebo, and most took the cigarettes. The results were conclusive: the substance killed pain. “That’s the origin of lidocaine,” Wood told me. “It’s a Swedish fairy tale.”
When applied locally, lidocaine was a marvellous anesthetic. It worked especially well for dental procedures. But, if you took enough of it to knock out pain in your whole body, it could kill you. Postwar anesthesiologists and dentists knew not to give the drug systemically, but they didn’t yet fully understand that it worked by acting on sodium channels, which are found in pain-sensing neurons, as well as in muscles in the heart, and in the brain. Lidocaine blocks all the sodium channels, everywhere in the body. Your heart muscles fail to contract, your brain goes quiet. Researchers realized that if you want to design a painkiller that you can administer systemically and safely, it needs to block only some kinds of channels, and only in specific locations.
The genetic mutations that the patients of both Waxman and Woods had affect a sodium channel called NaV1.7, which is predominantly found in peripheral pain-sensing neurons. A drug interrupting pain signalling before it ever reached the brain would likely lack the addictiveness of opioids. “We all went crazy, because people without NaV1.7 were pain-free but otherwise normal,” Wood, the doyen of sodium channels, told me. “It was unbelievably exciting.” All that researchers had to do was to make a compound that affected only that sodium channel. Well, actually, that would be very difficult, but still. “The genetic validation for NaV1.7 was knock-your-socks-off strong,” Waxman said. NaV1.7 was the perfect target. “But there’s a catch in the story,” Wood said.
Cartoon by Natalie Horberg
Waxman’s lab started with a small trial of a drug that targeted the NaV1.7 sodium channels. Five people with inherited erythromelalgia participated. “We saw an encouraging response,” Waxman recalled. The drug advanced to a trial involving dozens of patients with other conditions at multiple sites. But, in the large trial, researchers “did not see a signal of efficacy,” Waxman said. It could be that the drug did block NaV1.7 channels, but that the dose was insufficient; or that the drug didn’t distribute to the right locations in the body; or that NaV1.7 blocking worked on some forms of pain but not others. And there was yet another possibility. “Pain is important for survival, so it makes sense that the mechanism of pain signalling has redundancy at the molecular level to make it robust,” Bruce Bean, a sodium-channel researcher at Harvard, told me. NaV1.7 was out of favor.
But it wasn’t the only promising sodium channel. A toxin found in the puffer fish, that marine creature that resembles a devilish massage-therapy ball, affects six of the nine known sodium channels. During their research into pain, Wood and his team discovered that mice in which they had disabled the gene for NaV1.8—a channel that the puffer-fish toxin does not block—felt much less pain. The researchers were thrilled. They formed a company and quickly raised eight million British pounds in support.
But they, too, encountered difficulties. Wood said, “We were all set to go into toxicity studies”—and then they ran out of money, then merged with another company, which also ran out of money. A further discouragement: by 2015, it became known that some people with Brugada syndrome, in which the heart may abruptly stop, had mutations in the gene that encodes NaV1.8. It wasn’t clear whether a substance that blocked NaV1.8 would precipitate such a problem, but it was a serious concern. “We thought, Oh, that’s no good,” Wood said. Many researchers put NaV1.8 behind them. But the cell biologist Paul Negulescu, who had started looking into it in 1998, continued working.
In college, at Berkeley, Negulescu had initially studied history. “Then, as a junior, I took a physiology class where a professor explained how the kidney worked,” he told me. “It was all about keeping sodium ions and chloride ions and potassium ions in balance.” The kidney, a tremendously under-celebrated organ, basically does four-dimensional sudoku with ions. “I was just in awe of the genius of nature. It just clicked in my head—this is amazing.” He volunteered in an ion-channel lab as an undergrad, and later, as a Ph.D. student in physiology, collaborated with the professor on research; when the professor started a company, Negulescu joined it, and in 2001 it was bought by Vertex Pharmaceuticals, where he is now a senior vice-president. In 2019, Negulescu’s team received F.D.A. approval for Trikafta, a drug for cystic fibrosis which works on the faulty chloride channels responsible for the disease. A patient who starts taking the drug as a teen-ager has a life expectancy of more than eighty years—nearly twice the span of someone whose disease is managed with supportive-care treatments only. “We like ion channels,” Negulescu said. “We think they’re really good drug targets. They just require a lot of care and attention to how you measure them.”
The papers that Wood’s team published on the role of the NaV1.8 channel in pain signalling were a major inspiration for Negulescu to turn his attention to sodium channels and pain. “Each sodium-channel type has its own personality,” he said. “They open at different voltages. They remain open for different lengths of time. They evolved to perform in certain ways in certain tissues.” NaV1.8 channels open and close up to twenty times a second. “So we had to catch them in the act,” he said.
In trying to find a molecule that would inhibit NaV1.8, one might surmise that likely compounds would have shapes similar to those of lidocaine or of other anesthetics. But, Negulescu said, “We didn’t want to rely only on our intuition about what chemical classes might work.” His team aimed to be “agnostic,” remaining open to unforeseen possibilities. This approach would not have been feasible even a few years earlier, because of limits on how many lab tests could be done in a reasonable window of time. But Negulescu’s team had developed a new technology that allowed them to screen compounds much more quickly; it was like buying tens of thousands of lottery tickets, instead of a few hundred. Eventually, they discovered a previously undescribed class of molecules that looked promising—a process that took about ten years.
Ideally, one wants a drug that is highly selective—like Cinderella’s glass slipper, it fits the intended target and not a whole range of feet—and potent. An early version of an NaV1.8 blocker developed by Negulescu’s team was selective and fairly potent. But, in drug development, adverbs like “fairly” won’t do. Years of “optimization” followed. When I asked Negulescu to explain what optimization was like, day by day, he said, “Painful. It’s iterative learning. There’s the hypothesis: this is what we think would improve the potency of the molecule, or the selectivity of it.” Synthetic chemists then make the compounds they think might improve efficacy, and the lab team tests them quickly—“within hours”—then sends the data back to the synthetic chemists. I asked Negulescu how many compounds his team screened. “Hundreds of thousands,” he said. Then he said it again. “Hundreds of thousands.” Millions were screened to find the class of molecules, and then there were another ten thousand or so screenings done in the optimization process. Negulescu recalled encountering one of the chemists holding a tray in the hallway outside a lab: “I asked him, ‘Are there some important compounds in there?’ He looked at me and said, ‘Paul, they’re all important.’ ” After more than twenty years, they had a potent and extremely selective compound, called suzetrigine. And it wasn’t making people sick. The time had come to bring it to a large-scale clinical trial.
Establishing a painkiller’s efficacy is trickier than, for example, seeing whether a blood-pressure drug is effective. There’s a reason that the McGill Pain Questionnaire had seventy-eight words. Todd Bertoch ran the Phase III clinical trials for suzetrigine. “It’s a very high bar in pain research, to show effectiveness,” Bertoch said. “Some of the drugs don’t reach that bar, not because they’re not great drugs but because the models are imperfect and our statistical approaches are imperfect.” Terms like “moderate” and “uncomfortable” don’t offer the precision of, say, 135 and 150. As Negulescu put it, “There’s no pain-o-meter.”
Two large-scale Phase III clinical trials on suzetrigine have been completed so far. One looked at 1,118 patients following an abdominoplasty, and another at 1,073 patients following a bunionectomy; both are procedures after which people experience acute pain. Participants were given either suzetrigine, Vicodin, or a placebo, and were monitored for forty-eight hours. A smaller trial looked at suzetrigine versus a placebo in two hundred and two patients with sciatica, a nerve pain. In the sciatica study, suzetrigine worked about the same as the placebo. However, for the abdominoplasty and bunionectomy patients, suzetrigine worked as well as Vicodin and better than a placebo. And more patients reported side effects on the placebo than on suzetrigine. In January, suzetrigine, under the name Journavx, became the first new non-opioid painkiller in more than twenty years to receive F.D.A. approval for acute-pain treatment.
This has occasioned enormous celebration, which can at first glance be difficult to understand, since the results seem modest: the comparison is to a relatively weak opioid, and it remains unclear if Journavx will be helpful with chronic pain, cancer pain, or neuropathic pain. Additionally, the drug costs fifteen dollars a pill. Insurance plans and assistance programs can lower the price, but it is still much more expensive than the pennies-per-pill option of a generic opioid.
Yet scientists working in pain research described the underlying scientific achievement as “a magisterial first step,” “just marvellous,” and “the holy grail.” “This proves the concept,” Waxman told me. “My expectation is that there may be next-generation medications that work even better.” Painkillers that alleviate chronic and neuropathic pain are especially needed. A Phase III clinical trial of suzetrigine for diabetic peripheral neuropathy is under way, and the F.D.A. granted the drug a Breakthrough Therapy designation for the treatment of such pain, which should speed the drug’s potential approval.
“I don’t think there’s a miracle drug that’s going to replace opioids—and suzetrigine isn’t that drug—but what we’re doing is chipping away,” Bertoch said. “Before suzetrigine, if acetaminophen and an NSAID were insufficient, my next step was a mild to moderate-strength opioid. Now I can kick the opioid can down the road.” Bertoch said that early in his career a mentor told him, about opioids, that, “as long as someone had real pain, they can’t become addicted. Obviously, that’s been proven completely wrong.” And the correction on opioid prescribing has precipitated a new problem—pain going undertreated or untreated. “We need something else to fill that gap,” Bertoch said. “We’re not just talking about addiction—we’re talking about people who are suffering and can’t get the pain medicine they need.” He went on, “Ultimately, I think we are going to be able to find a place where, if opioids are needed, it’s going to be rare.”
