We often hear breathless news reports about breakthroughs in pharmaceutical research that could lead to powerful treatments for disease. But a closer look usually reveals that a wonder drug is years away from market — and often it gets scrapped or tripped up along the way as harmful side effects are discovered in clinical trials.
Now, however, computer models developed at the University of California, San Francisco could change the way new drugs are developed. As reported recently in the journal Nature, these programs can predict a wide range of potential side effects long before human testing — and they might even reveal possible benefits that a drug's developers hadn't anticipated.
"We tend to think of drugs as magic bullets," says Dr. Michael Kreiser, the study's co-author and a co-founder of SeaChange Pharmaceuticals in San Francisco. "But drugs are more like magic shotguns. More than one thing gets hit and that scattershot blast may hit something you never thought to look at before."
A drug's unintended interactions with proteins in the human body can cause serious side effects, but predicting such problems during development has long been a challenge for researchers. The new computer models change that by using a method called the "similarity ensemble approach." It identifies the molecules that a drug might bind to, based on a database of existing drugs and the proteins they are known to interact with.
The study applied this approach to 656 drugs already approved for human use and predicted their likelihood of binding to a specific set of proteins in the body. The model forecasted 1,600 potential interactions, only about half of which had been previously identified; 151 of these interactions were later validated by research at Novartis Institutes for Biomedical Research, a partner in the study.
"We were able to have more success than expected in predicting side effects ahead of time," Keiser says.
For example, the team's model predicted that a drug that was causing serious abdominal pain in some hormone-replacement therapy patients was binding to an enzyme to which aspirin also binds, causing similar pain. After verification by laboratory tests, researchers now know the cause of the side effect and are developing strategies to alleviate it.
The new computer models cannot replace laboratory or clinical testing. But they can give researchers a significant head start as they attempt to design drugs that are likely to be approved for human use. Pharmaceutical companies can invest hundreds of millions of dollars developing a new drug over many years of research, only to see the work derailed by the discovery of unanticipated side effects. "But if you can test a drug's safety earlier using these computer methods," Keiser says, "you can get it to people faster. I hope this will speed up the process with fewer false starts along the way."
(MORE: The U.S. vs. Alzheimer's: The Fight Heats Up)
A potentially enticing aspect of the computer modeling is the possibility of discovering unanticipated secondary benefits of drugs. "That's the flip side," Keiser says. "Sometimes drugs do useful things no one expected, and that could have real implications for a lot of patients." (Viagra, for example, was initially conceived as a treatment for hypertension.)
"We may realize that drugs can be used by a group of patients no one thought of before," he adds — and approval for those uses may come far more quickly because the drugs would have already undergone safety testing and clinical trials during their original development.
Next Avenue Editors Also Recommend:
- Cancer Treatment Progress Reported
- The U.S. vs. Alzheimer’s: The Fight Heats Up
- Is Your Sleeping Pill Shortening Your Life?
- Do You Really Need an Aspirin a Day?
Next Avenue brings you stories that are inspiring and change lives. We know that because we hear it from our readers every single day. One reader says,
"Every time I read a post, I feel like I'm able to take a single, clear lesson away from it, which is why I think it's so great."
Your generous donation will help us continue to bring you the information you care about. What story will you help make possible?