Pharma companies pump billions of dollars into research of drugs that are designed to hit specific disease proteins, but the drugs often fail in clinical trials because they cause harm by hitting unintended targets. Swiss drug giant Novartis ($NVS) and researchers from the startup SeaChange Pharmaceuticals and the University of California, San Francisco, have shown that their computational strategy can predict such "off-target" problems before human testing.
In the study, published in Nature, the group forecasted the effects of 656 FDA-approved drugs on 73 targets associated with side effects. When the researchers attempted to confirm the off-target findings of the system with databases and lab tests, nearly half of 1,042 predictions were confirmed. For instance, the system showed that a synthetic estrogen drug for cancer inhibits COX-1, a target of anti-inflammatory drugs like aspirin, potentially explaining why the estrogen treatment causes abdominal pain.
Novartis has been using the computational approach for its own research for about a year, Laszlo Urban, global head of preclinical safety profiling at Novartis Institutes for Biomedical Research, told FierceBiotech IT.
Obviously, the researchers were unable to confirm many of the off-target effects that their system predicted. However, the computational strategy offers a systematic way for Novartis to analyze compounds for off-target behavior during the pre-clinical phase of development. This helps the company to tell quickly whether a compound is worthy of further investments in development, Urban says. Also, improvements can be made to the system over time as researchers build in additional knowledge about side-effect targets.
Regulators could use the technology to help spot side effects. As The Boston Globe reported, the computer system found 151 previously unknown interactions of the approved drugs. And such unknowns have previously led to toxicities that caused drugs such as the fen-phen treatment and Vioxx to get pulled from the market. Yet for Novartis the tech is most useful in preclinical research.
"I think this is an extremely powerful tool," Kenneth Kaitin, director of the Tufts Center for the Study of Drug Development, told the Globe. "One of the major challenges for the industry is that compounds enter clinical testing without a very good understanding of how the drugs work and what the side effects are going to be, so you're essentially entering blind."
- here's the Nature abstract
- check out the Globe's article