Anti-NMDA receptor encephalitis: new research reveals how immune system attacks brain

Some serious brain disorders begin with symptoms that are easy to misinterpret.

A person may become confused, forgetful or paranoid and be treated for mental illness before doctors realize the underlying problem isn’t psychological at all. Instead, the body’s immune system — meant to fight infection — is attacking the brain.

Researchers at Oregon Health & Science University say they have identified, in new detail, how that attack happens. Their findings, published last week in Science Advances, could lead to more precise treatments and earlier diagnosis of an autoimmune condition known as anti-NMDA receptor encephalitis.

The disease affects roughly one in a million people each year and often strikes young adults. It occurs when the immune system produces antibodies — proteins designed to fight disease — that mistakenly latch onto something called an NMDA receptor, which helps brain cells communicate and plays a key role in memory, learning and normal thinking.

Scientists have long known antibodies were involved but did not fully understand where they attached or why current treatments don’t always work.

“What we wanted to understand was where on the receptor the antibodies actually bind,” said co-author Dr. Gary Westbrook, a neurologist and senior scientist at OHSU’s Vollum institute. “That’s important because it provides a clue to how one could design more specific treatments to deal with the disorder.”

To find out, the OHSU team used a mouse model of the disease to examine the entire ensemble of antibodies involved, rather than studying one antibody one at a time.

The researchers also relied on an advanced imaging method called cryo-electron microscopy, which allows scientists to see biological structures at near-atomic detail. OHSU houses one of only three national cryo-electron microscopy centers at its South Waterfront campus.

What they found surprised them.

“We found that the antibodies don’t bind everywhere,” said Eric Gouaux, a senior scientist at OHSU and an investigator with the Howard Hughes Medical Institute. “They don’t coat the receptor like a layer of paint. Instead, they bind to just a few very specific areas.”

The antibodies don’t immediately shut down the receptor. Instead, they act more like glue, causing receptors to clump together and get pulled inside brain cells, where they can’t do their job. That’s when symptoms begin to escalate.

“It can begin in a rather occult way,” Westbrook said. “Someone may be behaving a little funny, and people don’t quite know what’s happening. It may be days or weeks before the full picture becomes clear.”

But Westbrook said diagnosing the disorder early is notoriously difficult because its symptoms can mimic psychiatric illness, viral infections or other neurological disorders. He said many patients are first referred to a psychiatrist before the immune cause is identified — a pattern described in the book and film Brain on Fire, which brought public attention to the disease.

“That delay can be dangerous,” Westbrook said. “By the time the disease is diagnosed, patients may already be very sick.”

If diagnosis is delayed, symptoms can worsen, leading to seizures or extreme states such as catatonia, in which patients may be unable to move or speak, according to researchers.

Westbrook said the mouse model used in the study helps overcome that limitation, by allowing

researchers to study the disease from its earliest moments — something that isn’t possible in people.

Gouaux said the findings are especially important for drug development. He said current treatments rely on broadly suppressing the immune system. While many patients improve, recovery can be slow and relapses are common, he said.

“Now that we know exactly how and where the antibodies bind, the most straightforward approach would be to simply block that interaction,” he said.

The research could also improve diagnosis by making antibody tests more specific. Gouaux said current blood tests can detect NMDA receptor antibodies but can’t show where they bind or reliably predict who will develop the disease — something the new findings could help address.

“This work simplifies strategies to test for the disease more precisely,” Gouaux said.

As testing improves and awareness grows, Westbrook said more cases could be identified.

“We were seeing this disorder long before we knew what it was,” he said. “As awareness improves, more cases are almost certainly going to be identified.”