Racing Against Time
A trio of University of New Mexico scientists is using machine learning to cull through a “library” of thousands FDA-approved drugs to identify candidates that could be repurposed as treatments for the COVID-19 infection.
They’re in a race against time, as the novel coronavirus is mutating rapidly and spreading across the world, with devastating effect – and no clear-cut cure.
The effort was launched in late March, when the team started looking for – and testing – known drugs that might have antiviral properties, says Tudor Oprea, MD, PhD, professor and chief of the Translational Informatics Division in the UNM Department of Internal Medicine.
“We’re doing a multi-angle tackling of it,” Oprea says. The drugs, many of which were developed decades ago for very different health conditions, could be a game-changer at a time when there are no vaccines or broadly effective drug treatments for the disease.
The UNM team has been sharing its list of candidates with the National Center for Advancing Translational Sciences (NCATS) in the hope that NCATS can confirm their findings, Oprea says. “Every single chemical that we’re submitting is getting tested,” he says.
Oprea and his team are using computational methods to analyze the university’s DrugCentral database, which tracks drugs approved by the U.S. Food and Drug Administration, as well as those approved for use elsewhere in the world.
Oprea classifies each drug molecule according to which viral proteins it might target, enabling him to pinpoint the mechanism by which it might disable a virus – a critical first step, he says.
Oprea partnered with Larry Sklar, PhD, distinguished professor in the Department of Pathology and director UNM Center for Molecular Discovery, which maintains a collection of drug samples sponsored by the UNM Clinical & Translational Science Center.
“Tudor has been identifying molecules in the library that would be tested,” Sklar says. Samples are then passed on to Steven Bradfute, PhD, assistant professor in UNM’s Center for Global Health, who has been testing the drugs against the live virus in his Biosafety Level 3 laboratory.
Sklar, who has invented and patented unrelated technology that has transformed laboratory research, put his skills to work on Bradfute’s behalf.
“What we’re trying to do is help Steven and the people he’s working with make the testing more efficient,” Sklar says. “It’s a process called assay miniaturization that enables him to test a larger number of molecules at a time."
In keeping with Oprea’s strategy of leveling a “multiple-angle” attack on the virus, it will likely be necessary to administer two or more drugs simultaneously to achieve the desired effect, Sklar says.
“The other thing that we can do is to test drug combinations,” he says. “Because it’s likely that a single drug is not going to have the desired outcome on its own, we gave them the tools for making drug combinations that could also be tested in higher throughput.”
The team has already seen some preliminary success, Oprea says. Ciclesonide, a glucocorticoid drug that has been tried as an asthma treatment in Japan, appears to have a direct antiviral effect, he says.
The team has also studied moroxydine, an old antiviral drug that is chemically similar to metformin, commonly used to treat diabetes. “I think it has the potential to work as an antiviral, and potentially metformin has similar effects,” Oprea says.
And while there are feverish efforts around the world to develop a COVID-19 vaccine, RNA viruses – including the novel coronavirus – tend to mutate rapidly, meaning it may not be possible to develop a vaccine that prevents it from replicating, he says.
“This is why I think we need to find chemicals to hit it – and hit it hard,” Oprea says.