A study suggests a way in which the dangerous side effects of drugs such as Avandia might be avoided. Credit: J. RAEDLE/GETTY

When the erstwhile blockbuster diabetes drug Avandia stood trial before the US Food and Drug Administration (FDA) last week, its sharpest critics pulled no punches. Among them was Steven Nissen, the cardiologist at Ohio's Cleveland Clinic who first exposed Avandia's heart-attack risk. Shortly before yielding the podium, he unleashed a parting salvo against not just Avandia, but also the drug's family.

Since Avandia (rosiglitazone) hit the market, he said, developers have wasted their time on at least 50 other drugs that act by a similar mechanism. All of them failed, Nissen said, some because they posed similar risks to the heart.

A paper in this week's issue of Nature (page 451) details a way in which drugs of this class might avoid such dismal fates. The results could help drug makers to tailor compounds that reduce blood sugar without harmful cardiovascular side effects. Some observers doubt that pharmaceutical companies will try, given the protracted battle over Avandia and the failure of its relatives. "At present, this class of drugs is tainted," says Steven Kliewer, a molecular biologist at the University of Texas Southwestern Medical Center at Dallas, who helped to develop rosiglitazone. "The big pharma companies have basically given up on them."

Nevertheless, there is plenty of incentive for industry to persist as the obesity epidemic spawns a growing pool of patients with type 2 diabetes. The market for diabetes treatments is expected to grow from US$20.2 billion in 2008 to $37.9 billion in 2018, according to the market-research firm Datamonitor.

Swelling demand such as this helped to make Avandia a blockbuster, achieving sales of £1.66 billion (some US$3 billion) for manufacturer GlaxoSmithKline in 2006. But after studies found that the drug boosted heart-attack risk, the FDA slapped it with a warning in 2007. Finally, at the FDA meeting on 13–14 July in Gaithersburg, Maryland, the majority of a 33-strong advisory panel recommended restricting the drug's distribution. Twelve wanted Avandia to be pulled from the market altogether. The FDA is not required to follow the advice of such panels, but often does.

Throughout the affair, the reasons for the drug family's unwanted side effects have been unclear. In 2000, a related drug, Rezulin (troglitazone), was yanked from the market for causing liver failure (see 'Meet the 'glitazones': Avandia's family history'). Although Actos (pioglitazone) remains on the market and does not seem to pose a heart-attack risk, both it and Avandia roughly double the risk of heart failure.

Each of these glitazones activates a receptor in cell nuclei called PPAR-γ, boosting a pathway that ups the body's sensitivity to insulin and lowers blood sugar levels. But new work by molecular biologist Bruce Spiegelman at Harvard Medical School in Boston, Massachusetts, and his colleagues suggests that Avandia's effects on blood sugar may not depend on PPAR-γ activation after all.

The team found that, in obese mice, a phosphate is attached to a specific site on PPAR-γ, altering the receptor's function and reducing insulin sensitivity. Furthermore, Avandia and a similar compound called MRL24 not only activated PPAR-γ, but also often blocked the addition of the phosphate. And MRL24, which only weakly activates the receptor, was just as effective at lowering blood sugar in mice as Avandia, a potent PPAR-γ activator.

The results, says Spiegelman, suggest that pharmaceutical companies focused on the wrong activity when screening for PPAR-γ-altering drugs. Drugs that are more specific and only block the addition of the phosphate to PPAR-γ may work just as well without the same risk of side effects, he says.

The door is open for companies to redo their screens. Robert Henry, an endocrinologist at the University of California, San Diego, believes that drug makers will be interested despite the glitazones' rocky history. "Obesity is such a horrific problem and its consequences are so widespread, this work is crucial."

But endocrinologist Clay Semenkovich of Washington University in St. Louis, Missouri, says that Spiegelman's work, although "interesting and provocative", is still preliminary. In particular, he notes, the human experiments didn't include a control group of people who didn't receive a drug. Instead, control samples were taken from each patient before they embarked on a six-month treatment schedule. "A lot can change about a person in six months," Semenkovich says.

In any case, drugs that work by different mechanisms are already on the market. One class, the gliptins, inhibits an enzyme that breaks down an insulin-releasing hormone — thereby increasing the insulin available to diabetics. And another, inhibitors of the protein SGLT2, currently in late-stage clinical trials, keep blood sugar levels down by blocking the uptake of glucose by the kidneys. As a result, excess sugar is excreted in the urine.

David Nathan, director of the Diabetes Center at Massachusetts General Hospital in Boston, says Avandia should never have been a blockbuster because even its predecessors — drugs such as insulin and metformin — control blood sugar better. "I was never convinced that this was a great class of drugs to start with," he says. "There was nothing magical about it. It was all a triumph of advertising."

figure a
figure d