I also wanted to draw your attention to this post at RealClearScience by Nevan Krogan. When time is of the essence there’s probably no better way to identify a treatment for a newly emerging disease than by mining the existing data to find one and that’s what Dr. Krogan and his team are doing:
Facing this crisis, we assembled a team here at the Quantitative Biosciences Institute (QBI) at the University of California, San Francisco, to discover how the virus attacks cells. But instead of trying to create a new drug based on this information, we are first looking to see if there are any drugs available today that can disrupt these pathways and fight the coronavirus.
The team of 22 labs, that we named the QCRG, is working at breakneck speed – literally around the clock and in shifts – seven days a week. I imagine this is what it felt like to be in wartime efforts like the Enigma code-breaking group during World War II, and our team is similarly hoping to disarm our enemy by understanding its inner workings.
and
By March 2, we had a partial list of the human proteins that the coronavirus needs to thrive. These were the first clues we could use. A team member sent a message to our group, “First iteration, just 3 baits … next 5 baits coming.†The fight was on.
Once we had this list of molecular targets the virus needs to survive, members of the team raced to identify known compounds that might bind to these targets and prevent the virus from using them to replicate. If a compound can prevent the virus from copying itself in a person’s body, the infection stops. But you can’t simply interfere with cellular processes at will without potentially causing harm to the body. Our team needed to be sure the compounds we identified would be safe and nontoxic for people.
The traditional way to do this would involve years of pre-clinical studies and clinical trials costing millions of dollars. But there is a fast and basically free way around this: looking to the 20,000 FDA-approved drugs that have already been safety-tested. Maybe there is a drug in this large list that can fight the coronavirus.
Our chemists used a massive database to match the approved drugs and proteins they interact with to the proteins on our list. They found 10 candidate drugs last week. For example, one of the hits was a cancer drug called JQ1. While we cannot predict how this drug might affect the virus, it has a good chance of doing something. Through testing, we will know if that something helps patients.
Facing the threat of global border shutdowns, we immediately shipped boxes of these 10 drugs to three of the few labs in the world working with live coronavirus samples: two at the Pasteur Institute in Paris and Mount Sinai in New York. By March 13, the drugs were being tested in cells to see if they prevent the virus from reproducing.
Our team will soon learn from our collaborators at Mt. Sinai and the Pasteur Institute whether any of these first 10 drugs work against SARS-CoV-2 infections. Meanwhile, the team has continued fishing with viral baits, finding hundreds of additional human proteins that the coronavirus co-opts. We will be publishing the results in the online repository BioRxiv soon.
The good news is that so far, our team has found 50 existing drugs that bind the human proteins we’ve identified. This large number makes me hopeful that we’ll be able to find a drug to treat COVID-19. If we find an approved drug that even slows down the virus’s progression, doctors should be able to start getting it to patients quickly and save lives.
The great advantage to this approach is that all of their candidate drugs have already received FDA approval for other uses. The risks of using any of these drugs for off-label use are already known to some degree.
Epidemiology is, indeed, a science but it’s not the only science and, given COVID-19’s lengthy recovery time (as I’ve pointed out practically nobody who was sick with it three weeks ago has recovered) mining the existing data to identify effective treatments is a pretty darned good strategy.
“…mining the existing data to identify effective treatments is a pretty darned good strategy.”
Absolutely. Always is.
Reason. Thanks, Dave.
Big data has been, I think, a bit disappointing but this is one area where it makes sense. Hope they find something.
We are now at the point where we are making our own masks. I now get daily charts showing which material works best as a filter. Who would have known that cotton works better than silk? (Not me apparently.)
Steve