Moving Away From Target-Based Discovery

"I worry that most of the easy drug targets are gone," says Derek Lowe, a Ph.D. drug discovery chemist with a major pharma company in the U.S. and a blogger. Lowe's blog, "In the Pipeline," has between 3,000 and 4,000 visitors per day each weekday. In this blog, he talks about news and trends in the pharma industry, chemistry, lab work, and legal issues, among other things.

Lowe posts to his blog at night after his kids are asleep. For his day job, Lowe works "at the very beginning of drug discovery, just after the random screening," he says. His lab takes validated lead compounds and makes as many variations on the molecules as it can while getting continuous feedback from biology colleagues. (He would prefer that C&EN not identify his employer, because, he stresses, he is speaking only for himself, and his company has graciously permitted him to post his blog.)

"I'm fond of pointing out that there aren't many 'home runs' like ACE [angiotensin-converting enzyme involved in blood pressure regulation] out there these days," Lowe says. He reflects on the past: "Ah, ACE—an enzyme of a class that we know we can inhibit with small molecules, that's right in the middle of a key pathway for a condition affecting millions and millions of people." Lowe continues, "You could say similar glowing things about HMG [3-hydroxy-3-methylglutaryl] CoA reductase, the target of the statins."

Lowe says that one of the hopes of genome sequencing and mining was that there would be some more of these great targets out there that no one had yet discovered. "This expectation has not aged well," he tells C&EN.

What his company is looking for is drug activity for a specific biological target. "These days, almost every program in the industry is geared at some specific target, an enzyme or receptor that we have reason to think may be disease-relevant," he says.

But this drug discovery system may be breaking down. "Cancer may have 10,000 targets; it may turn out to be several thousand orphan diseases, each of which has a relatively small patient population," Lowe says. The cost and complexity of screening thousands of targets against higher systems like cell cultures or an animal model that mimics a real organism would be overwhelming.

He believes that pharma companies may start putting more of their resources back into what would be considered "old-fashioned" programs to screen huge numbers of compounds, hoping for a hit. One way that might work, he says, is to find "an animal or near-animal model for diseases and run as many compounds as you can through it." The point of this screening would be for researchers to find some molecule that actually works against a disease, regardless of whether they know what its target or targets are.

"This is how drug discovery was done in the 1940s and 1950s. We would need to do this in a high-throughput fashion because we have so many more compounds than those guys did then, and we'd like to test them all," he says.

Lowe is also hoping that the future holds something, anything, that would increase the number of compounds that make it out of clinical trials. Many fail because of toxicology problems. "If we were to improve just a few percent here and a few percent there, we'd be in much better shape. We are in such bad shape in a lot of these areas that all we have to do is go from a 90% failure rate to an 80% failure rate, and the sun would be shining."