ANN ARBOR, Mich. – There are numerous clinical trials underway across the country and around the world in hopes of finding an effective treatment for the coronavirus (COVID-19). Researchers at the University of Michigan are trying to speed up the process of identifying drugs that might be helpful to the cause.
Developing a brand-new drug typically takes a decade. That’s why researchers at the University of Michigan are focusing their attention on the drugs we already have.
“We’re screening all FDA-approved drugs, which is about 1,200 in our collection, and we also have compounds that are experimental that have been in clinical trials,” said Dr. Jonny Sexton, who leads the U of M Center for Drug Repurposing.
Altogether, that’s nearly 8,000 potential treatment options, Sexton said. Luckily, they have technology to help with the process.
“We use our artificial intelligence methods to help us sift through all of those to find that needle in a haystack,” Sexton said.
In the lab, researchers infect human lung cells with the virus that causes COVID-19. Then, they expose those cells to the various drugs and compounds.
“During the course of this experiment, we collect millions of images -- just too many images that we can inspect visually, and so what we do is we train,” Sexton said. “We train computers to look at the images for us and tell us and indicate to where those efficacious drugs are working.”
They’re looking for pictures in which red areas of infection have been eliminated. There are benefits to using drugs that are already approved.
“It can have an immediate impact,” Sexton said. “It can actually be used very shortly, so we’re just in this experimental process right now. We’ll have results actually today and all through the rest of this week, so we’re very excited that we’re going to have potential candidates that actually can be tested in the clinic very rapidly.”
Their goal is to reduce the severity of symptoms and save lives.
“We’re hoping to find compounds with enough efficacy that we can keep patients from dying,” Sexton said.
He said the most effective treatment might actually be a cocktail of existing drugs. The artificial intelligence can help narrow down the best combinations, too.