Drug Designer

In drug discovery, the recipe for success has long called for ingredients such as experience, chemistry and biology know-how, technology, and, of course, perseverance. Over the years, however, many companies have bulked up on technology while skimping on the chemistry know-how of their scientists.

This variation on the recipe has had its successes, but there's one quality that the reliance on technology may overlook: intuition. It's a word that quickly comes to mind when talking science with Julian Adams, an organic chemist who is president and chief scientific officer of Cambridge, Mass.-based Infinity Pharmaceuticals. Adams already has discovered two important drugs, Millennium Pharmaceuticals' cancer treatment Velcade and the Boehringer Ingelheim's HIV medication Viramune. In the next five years, he could add one or two more cancer drugs to his roster.

Adams' success seems to flow from a back-to-basics approach to drug discovery that combines time in the lab with an instinct for what makes a molecule a good drug. He has a fondness for complex molecules that can be derived from nature and for attacking targets that others deem intractable.

For example, drug companies had for years summarily rejected any drug that targeted the proteasome. The rationale was that blocking the activity of the proteasome, the cell's trash bin, would have toxic consequences for normal cells. But Adams tackled the proteasome with the small molecule Velcade, and the drug eventually sailed to market in record time.

Adams now is championing a new breed of cancer drugs that block heat shock protein 90 (Hsp90), a target that, like the proteasome, has had its detractors. Responsible for the care of more than 300 regulatory proteins, Hsp90 was for years believed to have too central of a cellular role to be promising in the long run. Furthermore, the first drug found to inhibit Hsp90, the natural product geldanamycin, posed serious structural challenges for synthetic chemists.

But such challenges are at the heart of drug discovery, Adams says. His idea is to turn a "tool" drug—a molecule that interacts with a target but has deficiencies—into a "real" drug. "The industry typically relies on a high-throughput screening approach and all kinds of numbers, which is like playing the lottery," he says. Carrying out that process and optimizing the hits that result from it requires an enormous amount of money and time.

Also, high-throughput screening can ignore many viable drug candidates. Roughly 30% of the approved drugs are natural products or derived from natural products, which tend to be highly three-dimensional in structure, stereochemistry, and selectivity. Such characteristics are hard to achieve, Adams says, with the flatter, heterocyclic molecules that generally inhabit the molecular collections used in high-throughput screening.

Infinity researchers scour the literature for intriguing natural products or tool drugs and then improve these molecules' potency, solubility, or pharmacokinetics to make a real drug. In the case of the company's Hsp90 program, the tool drug geldanamycin was too toxic and insoluble to be commercialized. Infinity poked and prodded the molecule until it became a real candidate, IPI-504, which has shown promise in Phase I trials.

This isn't to say Adams is against using new technology. Infinity used high-throughput screening to track down leads for its an oncology program based on the so-called hedgehog pathway. MedImmune is a partner in this program. "Where possible, we would like to circumvent high-throughput screening," Adams says. "We don't ignore it—one of our programs came out of it—but where possible, we like to take a directed chemistry approach."

[Top of Page]