Researchers reported last month that an inexpensive, widely available pill substantially reduced hospitalizations and deaths in a large study of individuals with mild COVID symptoms who were at high risk for complications. It is the only existing oral medication with promising peer-reviewed data from multiple randomized COVID trials—and it is already used by millions of people worldwide. The drug is fluvoxamine, and it is approved in the U.S. for treating obsessive-compulsive disorder and depression. So how did this antidepressant end up in a trial for treating COVID?

“Drugs don’t know what their original indicated purpose was, they just do what they do, and they don’t usually do only one thing,” says Angela Reiersen, a child psychiatrist at Washington University in St. Louis (WashU). Along with her WashU psychiatry colleague Eric Lenze, Reiersen conducted a smaller randomized trial last year that suggested fluvoxamine could keep newly infected COVID patients from deteriorating.

Fluvoxamine is best understood for its impact on serotonin—a chemical messenger linked to mood and anxiety disorders. But the drug has other molecular targets. One is a protein called the sigma-1 receptor, which regulates the release of inflammatory molecules, including several that escalate in people with severe COVID. In a 2019 study, University of Virginia scientists chemically induced sepsis, a life-threatening infection complication, in mice. They observed that animals lacking the sigma-1 receptor developed severe inflammation, and many died. Yet in normal mice, a shot of fluvoxamine quieted the immunity overdrive and helped the animals survive.

That study caught Reiersen’s attention: the same overdrive immune response seen in those mice was a central feature observed in children who had come under her care because they had a rare genetic disorder—Wolfram syndrome—that led to psychiatric symptoms. When the pandemic hit and reports emerged that some people with COVID worsen because their inflammatory response goes haywire, Reiersen recalled the mouse study and wondered if fluvoxamine could also keep immune responses from spiraling out of control in these individuals.

She approached Lenze about testing fluvoxamine as an early COVID treatment, and he was game. Many modern mental health drugs in fact came from serendipitous observations of mood and behavior improvements with drugs developed for unrelated reasons—to fight nausea, for example, or to treat infectious diseases such as tuberculosis. So when Reiersen proposed repurposing fluvoxamine as a potential COVID treatment, “I remember saying, ‘Hey, you know, we owe psychopharmacology to the infectious disease field,’” Lenze says. “Wouldn’t it be good to pay them back on this one?”

Curiously enough, others who pivoted their research during the pandemic have also converged on the sigma-1 receptor. In a notable case, a team searched for drugs that block interactions between human proteins and proteins in SARS-CoV-2, the virus that causes COVID. One of the two sets of drugs that emerged from this systematic analysis, published last year in Nature, was compounds that bind the sigma-1 receptor.

Andrea Fekete, a nephrologist and CEO of SigmaDrugs, a biotech spin-off of Semmelweis University in Budapest, has used fluvoxamine and other sigma-1-receptor-targeting drugs to protect rats from fibrosis, or tissue scarring, in renal disease. She and her colleagues are now conducting a clinical trial to see if 200 milligrams of fluvoxamine daily can prevent lung fibrosis in hospitalized COVID patients with moderate disease. Recently, a small open-label study found that two weeks of fluvoxamine reduced mortality in COVID ICU patients who chose to receive the drug, compared with those who refused it, suggesting fluvoxamine could also help people with more advanced disease.

Other researchers have turned their attention to antidepressants more broadly. Early in the pandemic, Nicolas Hoertel, an associate professor of psychiatry at Paris University and a psychiatrist at Corentin Celton Hospital in France, noticed something intriguing: Mentally ill patients—even elderly ones—were not getting COVID nearly as seriously nor as often as hospital staff did. He called up his former colleagues in New York State, and they, too, noted that although emergency units were packed, there were surprisingly few psychiatric patients. In the spring of 2020, “I was thinking, what could explain this?” Hoertel says.

By that time scientists were zeroing in on overexuberant inflammation as a key feature of severe COVID. As it turns out, people with depression have high levels of some of the inflammatory proteins that are elevated in people with COVID, and many antidepressants have anti-inflammatory properties. Hoertel wondered if COVID patients who happened to be taking antidepressants had better outcomes.

Following that hunch, his team combed through databases at dozens of Paris-area hospitals and showed that, indeed, COVID patients admitted to the hospital who were already taking an antidepressant had a 44 percent reduced risk of intubation or death. In a separate analysis, researchers tracked comorbidities and outcomes in people hospitalized for COVID in Germany and found that depression was the only condition associated with lower odds of death.

It was hard to explain these observations, however. In the Paris study, several—but not all—of the drugs associated with better COVID outcomes were selective serotonin reuptake inhibitors (SSRIs). Several—but not all—activated the sigma-1 receptor. Hoertel and his co-workers were puzzled as to why certain antidepressants appeared to help COVID patients while others seemed to have little effect. Last November he received an e-mail with a possible answer.

The message came from Erich Gulbins, a molecular biologist at the University of Duisburg-Essen in Germany. In it, Gulbins described a paper his team was about to publish showing that the five antidepressants associated with a COVID benefit in Paris hospitals also block SARS-CoV-2’s entry into human nasal epithelial cells grown in a lab dish. And those drugs have something in common: they inhibit acid sphingomyelinase, the key enzyme in a biochemical pathway that processes fats into forms that can help viruses enter cells. Some antidepressants are among the dozens of so-called FIASMA (functional inhibitor of acid sphingomyelinase) drugs known to block this pathway. When Hoertel and his colleagues scoured hospital records again, they found that patients who were taking FIASMA medications (a broader drug class than antidepressants) had lower odds of intubation or death despite being generally older and having more advanced disease.

It is not yet clear if fluvoxamine’s COVID benefit derives from its binding to the sigma-1 receptor, its effect on the acid sphingomyelinase or a combination of these and other potential mechanisms. Some wonder if fluvoxamine might also help treat COVID by directly suppressing SARS-CoV-2—similar to oral drugs recently developed by Merck and Pfizer that have not yet been authorized for use in the U.S.

Matt O’Meara, a computational pharmacologist at the University of Michigan Medical School, who was one of the authors of the human-viral protein analysis published in Nature, thinks the mouse sepsis study points to the sigma-1 receptor as a primary way fluvoxamine helps against COVID. By showing that clinically relevant concentrations of fluvoxamine could block a strong inflammatory immune response in living animals, O’Meara says, “that’s very strong evidence that this is a reasonable mechanism.”

In addition, he and his colleagues have conducted lab-dish experiments in which they have assessed the growth of SARS-CoV-2-infected cells mixed with increasing concentrations of fluvoxamine. Their data, still unpublished, shows that fluvoxamine had little effect on viral infection even at concentrations high enough to be toxic to the cells, O’Meara says.

Hoertel and his colleagues’ studies on antidepressants and FIASMA drugs raise another question: Could the antidepressant fluoxetine (Prozac)—a drug that, like fluvoxamine, binds the sigma-1 receptor—have a similar benefit in treating COVID? That would be an attractive possibility because fluoxetine is better tolerated and more widely prescribed than fluvoxamine. If targeting the sigma-1 receptor is the prime mechanism, some think it is unlikely that fluoxetine would help against COVID because it binds to that receptor about 10 times more weakly than fluvoxamine.

Fluoxetine has not yet been investigated in a randomized trial for treating COVID. But researchers at Stanford University and the University of California, San Francisco, have a forthcoming paper that will report their analysis of electronic health records to look for a potential link between preexisting fluoxetine use and COVID outcomes among more than 80,000 patients nationwide.

While these questions about how antidepressants help against COVID are being investigated, “what I think, in fact, is the most likely [answer] is that we have a contribution of different mechanisms,” Hoertel says. “I mean, we know that, in biology, nothing is simple.”