Researchers have long used still images of proteins known to be related to recurring cancers in an attempt to understand exactly why these proteins make some chemotherapies fail.
Now, biochemists at Southern Methodist University are using a 3D computer model of the human protein P-glycoprotein — believed to play a pivotal role in the failure of chemotherapy in many recurring cancers — to screen more than 8 million potential drug compounds in the hunt for one that will help stop this failure.
“This has been a good proof-of-principle,” biochemist John G. Wise said in a school news release. “We’ve seen that running the compounds through the computational model is an effective way to rapidly and economically screen massive numbers of compounds to find a small number that can then be tested in the wet lab.”
Wise’s supercomputer search, which he describes in the journal Biochemistry, has already yielded a few hundred “interesting” compounds, 30 of which he and colleague Pia Vogel have begun testing in their lab. Of those 30, they have found a handful that inhibit the protein, so they’re now trying to test those further.
While this digital screening has allowed Wise to run more than 10 million computation hours in the past three years (Wise says they currently screen about 40,000 compounds per day on SMU’s High Performance Computer), the odds are stacked against researchers trying to find a compound that will work.
“Out of a hundred good inhibitors that we might find, 99 of them might be extremely toxic and can’t be used,” Wise admits. “They metabolize too quickly, or they’re too toxic, or they’re not soluble enough in the acceptable solvents for humans. There are many different reasons why a drug can fail. Finding a handful has been a great confirmation that we’re on the right track, but I would be totally amazed if one of the first we’ve tested was the one we’re looking for.”
Still, at the rate of screening 40,000 compounds a day, they just might stumble upon that needle in the haystack some day.