Medicines that reshape or tamp down immune responses may be life-changing for patients with cancer and autoimmune disorders, but in some cases they can awaken a dormant virus and unleash a deadly brain disease. A new study suggests that the root of the problem is buried in our genetic code.
Researchers uncovered four gene variants connected with the immune system that significantly increase a person’s risk of developing the disease, known as progressive multifocal leukoencephalopathy or PML. These variants were far more common among those suffering from PML than in the general population. And they were conspicuously absent in multiple sclerosis patients who’d been receiving treatment for years without developing PML.
The study’s authors argue that these variants could be used to screen patients and ensure that those with a higher risk of PML aren’t given drugs linked to the disease — similar to how doctors already use genetic testing in some patients to determine the right dose and type of medication used to treat depression, high cholesterol, and other conditions. Scientists estimate that a four-variant test could cut PML cases by about 9.4%, and they’re now searching for additional variants that could extend the test’s protective power to additional patients.
The study, published on Wednesday in the journal Frontiers in Neurology, was funded by Population Bio and Emerald Lake Safety, companies focused on how DNA shapes our risk of developing certain side effects from medication. But outside experts found the findings compelling.
“This would be a major advance not only in preventing PML, but also in potentially reassuring patients that it is safe to take many of the drug therapies associated with PML risk,” said Revere Kinkel, a neurologist and director of the University of California, San Diego’s multiple sclerosis program.
PML is a rare disease triggered by a remarkably common virus: John Cunningham virus, or JC virus, named after the man in whom doctors first discovered it in 1971. Odds are that you, like most adults, carry the virus, which resides in the tonsils and kidneys, among other places.
These infections are almost always harmless. But immunosuppressive therapies or diseases sometimes rouse the virus from its slumber. In these cases, JC virus attacks the cells that produce myelin, the blubbery layer of fat and protein that swaddles the electrical signals that zip back and forth through neurons. The first symptoms are often mild — a bit of clumsiness, or issues speaking and seeing. But things get worse, and around 30% to 50% of people with PML die within the first few months of their diagnosis.
While PML affects only about 1 out of every 200,000 people overall, dozens of drugs include warnings about risk of the disease in their product labeling. In 2009, Genentech took a psoriasis drug off the market in the U.S. over concerns the therapy was triggering PML at unusually high levels. And in 2014, the Food and Drug Administration added a warning label to Tecfidera, a multiple sclerosis drug, after a confirmed case of a patient dying of the brain infection while undergoing treatment.
These regulatory precautions left a key question unanswered: Why did some patients develop PML while most didn’t?
Eli Hatchwell, chief scientific officer of Population Bio, suspected the answer could be found in our genetic code. So he and colleagues launched an initial study that compared the DNA sequences of 184 PML cases with those of the general population, focusing their search on genes linked to the immune system.
The researchers found 19 variants linked to PML. But the study had a key limitation, according to Hatchwell — researchers hadn’t used the right comparison group. That’s because people in the general population likely weren’t on medications that might induce PML. A better comparison would have been to sequence both PML cases and patients who hadn’t developed PML despite carrying the JC virus and taking drugs linked to the disease.
That’s what scientists did in this latest study, relying on patient samples from the University of California, San Francisco and the University of Toulouse in France. Researchers whittled down their original list of 19 variants to four, all found in genes associated with viral defense. Two of the gene variants are associated with X-linked lymphoproliferative disease, or XLP, a condition similar to PML in that it leaves some people vulnerable to a common and usually harmless herpes virus (in this case, Epstein-Barr virus).
Researchers then developed a simple genetic test using the four variants. A positive result gives a person a roughly 20% chance of developing PML while on treatment — far higher than the rate among the general population. And that’s information Hatchwell says could one day guide physicians to prescribe different drugs to such patients.
“This is an example of how precision medicine can give patients very specific advice in very specific situations,” he said. “We believe that tests such as ours will become more important when it comes to studying people on drugs.”