Researchers have figured out how a promising cancer drug acts like a smart bomb, homing in on a very narrow range of potential targets in the cell. The compound, known as STI-571, has shown remarkable success in early clinical trials on patients with chronic myelogenous leukemia (CML). Knowing more details about its mode of action could aid in the design of similar weapons against cancer. "It's a very neat story," says cell biologist Tony Hunter of the Salk Institute for Biological Studies in La Jolla, California.

STI-571 was identified in the early 1990s by scientists at the pharmaceutical company Novartis, and subsequently shown to block the enzyme produced by abl, a so-called oncogene whose overactivity is thought to cause the massive proliferation of the leukemic cells in CML patients. Consistent with that, clinical trials conducted so far have shown that STI-571, in sufficient doses, can cause the apparent disappearance of the leukemic cells from the patients' blood, and the side effects of the treatment are relatively mild. But there's been a puzzle. Abl belongs to the protein kinases, a large family of enzymes that transfer a phosphate group from ATP to proteins. What was unknown is why STI-571 has such a limited range of action. In addition to Abl, it blocks only two others of the 50 or so protein kinases tested--which may be why it has so few side effects.

To explore the basis of STI-571's specificity, John Kuriyan and Thomas Schindler of the Rockefeller University in New York City and their colleagues crystallized the catalytic region of the human Abl protein together with a version of STI-571. They then used x-ray crystallography to determine the three-dimensional structure of the drug-protein complex. Abl, like many other kinases, won't kick into action until a phosphate is added to its "activation loop." This step alters the shape, or conformation, of the enzyme, opening it up so that the kinase can bind ATP and its target proteins. What the crystal structure reveals, the team reports in the 15 September issue of Science, is that STI-571 binds to the inactive conformation of Abl, effectively locking out the activating phosphate.

The finding also explains why so few other kinases are inhibited by the drug. "When the kinases are active, they look very similar," Kuriyan says. "But when they are turned off, they can look very different from one another." He notes that drug developers usually attempt to inhibit active enzymes. But for kinases, the inactive forms may make better targets, says Kuriyan. Meanwhile, clinical trials are continuing to see whether CML's early promise holds up.

--JEAN MARX

Related sites
The Kuriyan lab at Rockefeller
Information about leukemia from the National Cancer Institute