An "Insulin Pill"?

A lowly fungus that grows deep in the African forests near Kinshasa could soon be a pharmacological celebrity. In this week's issue of Science, scientists at Merck Research Laboratories in Rahway, New Jersey, report that the fungus, Pseudomassaria, produces a small, five-ringed molecule that can mimic insulin. Unlike insulin, the molecule is not a protein--so it could likely withstand the body's potent digestive juices, which suggests it could lead to a new type of antidiabetes pill. That would be good news for the millions of diabetics who now inject themselves with insulin or resort to a few orally taken drugs with serious side effects.

To look for an insulin mimic, Merck scientists Bei Zhang and David Moller used hamster ovary cells engineered to produce the human insulin receptor. Then, they divided the cells among thousands of miniature petri dishes. After trying some 50,000 mixes of synthetic chemicals and natural extracts on the hamster cells, the investigators scored a major hit with an extract from Pseudomassaria, a fungus found years ago in the rainforest. Merck chemist Gino Salituro isolated the active agent, a quinone. In tests on cultured cells, the Pseudomassaria product, known as L-783,281, stimulated insulin receptors as much as 100 times more than other natural compounds tested.

Preliminary animal tests look promising. The Merck team tried the compound in two mutant mouse strains that have classic diabetes symptoms. In both strains it suppressed the skyrocketing blood sugar levels by up to 50%--comparable to the reduction seen with current oral antidiabetic therapies, Moller says. And its effects appear to be specific: L-783,281 didn't stimulate a range of similar receptors. Achieving such specificity has always been "an elusive goal," says Zhang. Other antidiabetes drugs work in various ways, such as increasing insulin production by the pancreas or binding to the outer portion of the insulin receptor, but they may have serious side effects, such as excessively low blood sugar or blood pH, gastrointestinal problems, or liver failure.

If further animal trials confirm that L-783,281 or chemical variants resembling it are both effective in lowering blood sugar concentrations and safe, Merck says clinical trials might be feasible. It "could become a drug that may be able to be given by mouth," says endocrinologist Arthur Rubenstein, a diabetes expert at Mount Sinai Hospital in New York. "The potential is enormous."