Read our COVID-19 research and news.

Opposite approaches to altering the activity of vasopressin in the brain improved some social deficits in people with autism.

Drotyk Roman/

Can manipulating a ‘social’ hormone’s activity treat autism?

Many people with autism have trouble making eye contact, reading the emotions in other faces, and sharing affection. And no drugs are approved to treat such social impairments. Now, results from a small academic clinical trial suggest boosting levels of vasopressin—a hormone active in the brain that’s known to promote bonding in many animals—can improve social deficits in children with autism. But in a confusing twist, a larger, company-sponsored trial that took the reverse approach, tamping down vasopressin’s effects, also found some improvements in adults with autism.

“I’ve never seen this before,” Kevin Pelphrey, a neuroscientist who studies autism at the University of Virginia in Charlottesville, says of the conflicting results. He and others say the vasopressin-blocking approach doesn’t have much support from previous animal research. The new study showed some benefits but failed to meet the main endpoint set out by investigators. Still, he says, both studies suggest vasopressin’s signaling in the brain plays a key role in autism and “give me a lot of renewed excitement” for treating the condition.

Though vasopressin seems to stimulate social bonding in animals, the hormone’s activity in the brain isn’t fully understood, and its effects vary by species and context. Blocking its activity in the brains of some rodents prevents them from forming an attraction to a mate. But in a species of asocial hamster, injecting it into a male’s brain seems to stimulate aggression.

Another brain signaling molecule with a very similar structure, oxytocin, is already under investigation as an autism treatment. But vasopressin has gotten less attention. Recently, Karen Parker, a neuroscientist at Stanford University in Palo Alto, California, and her colleagues turned up evidence that monkeys that were less social—that tended to keep their distance from their peers, for example—also had lower levels of vasopressin in their cerebrospinal fluid. The group also found that children with autism who had the most social impairment also had the lowest vasopressin levels.

So the Stanford team gave a nasal spray containing vasopressin to 17 children with autism, aged 6 to 12 years. Another 13 children with autism served as a control group and got a placebo spray. Before and after the 4-week treatment, the research team asked parents to rate the children on a questionnaire called the Social Responsiveness Scale (SRS-2), which asks how often the children “would rather be alone than with others,” for example, and how often they “avoid eye contact.” This 65-question scale results in gender-adjusted scores of 37 to 90 in girls and 38 to 90 for boys.

Vasopressin-treated children showed significantly more improvement—a seven-point-greater reduction—on the SRS-2 than those in the placebo group, the team reported today in Science Translational Medicine (STM).

Those results “are very exciting,” especially because the team didn’t see major side effects, says Angela Sirigu, a neuroscientist at CNRS, the French national research agency, in Bron who is also investigating neurohormones for autism treatment.

Children with higher levels of vasopressin in their blood at the start of the Stanford study saw greater improvements. That’s counterintuitive, Sirigu says—you’d expect children who were most deficient in the hormone to benefit most from the boost. These children may have needed a higher dose or longer treatment course to see maximum benefits, Parker speculates. Or maybe blood vasopressin is a marker for some other, yet-unknown feature of the children that predicts how much they’ll improve with the treatment.

The only weakness Pelphrey found with the Stanford study was its sample size: “If it were double, I would be jumping up and down.” Still, he says, “I’m kind of rocking back and forth in my chair with excitement.”

The second trial, meanwhile, tested a very different hypothesis about vasopressin. Although many people with autism have trouble interpreting and reciprocating emotions, they also sometimes have overactive emotional responses, says neurologist Paulo Fontoura of Roche Pharmaceuticals in Basel, Switzerland. And vasopressin could drive those responses, he suggests. In a rat model of autism, blocking a vasopressin receptor appeared to reduce the animal’s abnormally heightened brain response to the presence of a juvenile rat. In the new trial, the Roche team recruited 223 men with autism to test a compound called Balovaptan that blocks a vasopressin receptor in the brain. The 148 men who took the drug improved according to an SRS-2 assessment, but those who got a placebo improved just as much, the team reports today in STM. Fontoura and his colleagues say the strong placebo effect in the study suggests the SRS-2 isn’t the best way to measure social improvements. Men receiving Balovaptan did show significant improvements over a placebo group on a different test of autism symptoms, called the Vineland-II.

The authors of both studies propose that there could be important differences in their participants. Maybe some people with autism could benefit from boosting vasopressin and others from blocking it.

The Stanford group is running another trial of vasopressin that aims to enroll 100 children. Roche has two trials of Balovaptan underway—an initial test in children, and a larger study in adults meant to demonstrate the drug’s effectiveness to regulators. “It’s early to make a head-to-head comparison,” Fontoura says. “It’s only when we do see more data that we’ll be able to draw conclusions about which way is the right way.”