Biotech Takes on Toxic Drugs
Nothing robs pharma officials of more sleep than unexpected toxicity linked to one of their companies’ medicines. Recently they’ve lost more shuteye than ever as reports on cardiac risks of widely-used diabetes drugs have fanned public anxiety about drug hazards. Now Congress is moving to intensify FDA surveillance of ill effects possibly caused by medicines. “We’re going to find adverse events associated with every drug,” says Keith Elliston, CEO of Genstruct, a Cambridge-based pharmaceutical consultant. That means the industry will urgently need better ways to tell whether such events really are caused by drugs, and, if so, to identify at-risk patients early. “To do that,” he adds, “we have to understand more about drugs’ mechanisms of action.”
As you’ve probably guessed, Genstruct’s specialty is figuring out how medicines do their things. The closely held company uses systems biology, an emerging discipline in which scientists try to capture enough of the complexity of living systems in computer models to predict what they will do when perturbed by drugs or illness. If Elliston is right, systems biology can prevent toxicity issues from erasing billions of dollars of pharmaceutical sales and turning patients away from needed therapies.
A Genstruct collaboration with Pfizer supports his thesis. Like many drug companies, Pfizer has been developing PDE4 inhibitors, anti-inflammatory compounds that may alleviate everything from asthma to rheumatoid arthritis. Unfortunately, one of its promising PDE4 inhibitors was found to cause severe blood-vessel damage in rats. That ordinarily would have killed the drug. But when Pfizer enlisted Genstruct to analyze the toxicity, says Elliston, systems biology revived the project.
Genstruct assembled all the available data on PDE4 inhibition in a computerized model comprising a network of interacting genes, proteins, and other molecules. The model highlighted key cause-and-effect relationships, revealing that the vascular injury was probably a rodent-specific effect—among other things, the model riveted attention on a central mediator of the damage that turned out to be an enzyme found at much higher levels in rats than in humans. The analysis also identified early-warning indicators of the toxic effect that might be used in human trials of PDE4 inhibitors to prevent drug-induced harm.
Genestruct is not the only biotech company betting on the market created by vanishing public tolerance for drug risks. In Mountain View, CA, Perlegen Sciences is focused on finding gene variants that make certain people react badly to specific drugs. One of its flagship projects seems prescient in the wake of reports that diabetes drugs made by GlaxoSmithKline (Avandia), and by Takeda Pharmaceuticals/Eli Lilly (Actos), elevate risk of heart attacks. Hoping to develop a competing drug in the same family as Avandia and Actos, Perlegen earlier collected 3,000 DNA samples from patients treated with the drugs in an effort to identify gene variants associated with the medicines’ tendency to cause fluid retention—a side effect that may aggravate heart failure. Perlegen hoped its competing drug could be offered with a proprietary genetic test that would enable the medicine to be targeted at patients unlikely to suffer harm from the side effect.
The privately held company’s first pass at the problem proved disappointing. “We’d hoped to find [gene variants] accounting for 50 percent” of fluid-retention cases on the drugs, says David R. Cox, Perlegen’s chief scientific officer. “But what we found accounted for only 16 percent.” Perlegen hasn’t given up—it’s now re-analyzing the patients’ DNA with “high-throughput” gene sequencers from 454 Life Sciences, recently acquired by Roche Group. The new technology “enables you to cost-effectively explore for rare gene variations,” says Cox, which often underlie the most troublesome adverse events—those afflicting a small fraction of patients, making them extremely difficult to detect before drugs reach the market and are prescribed to many thousands of people.
Cox says Perlegen doesn’t plan to try resurrecting canceled drugs. But it does hope to develop genetic tests for targeted prescribing of existing drugs that many people already depend on but that have come under a cloud due to unexpected adverse reactions. “In another setting,” he says, such reactions “are called icebergs. We now have the ability [to address their risks] in an expeditious way.”
Pharmaceutical executives no doubt hope he’s right—they can’t afford very many more Titanics.