A Booby Trap for HIV

Although so-called triple drug therapy can keep a patient's HIV count under control for years, just one or two mutations in the virus's genome can cause it to develop resistance and begin multiplying again. A paper in today's Nature Medicine describes a potential drug that may be much harder to outwit: a killer protein outfitted with a molecular booby trap to destroy infected cells.

The most effective HIV drugs block an enzyme called protease that helps the virus build new daughter particles. A group led by Steven Dowdy, a molecular oncologist at Washington University in St. Louis, decided to try an alternative approach: Instead of blocking protease, they hope to turn it against HIV. Their strategy is to trick the virus into using its protease to open a deadly package in cells it infects. "We designed a Trojan Horse molecule," Dowdy says.

The deadly package contains a protein called caspase 3, an enzyme that, when cleaved apart, can switch on a cell's suicide machinery. To limit this destruction to infected cells, the researchers changed the cleavage sites to those recognized by the HIV protease. They also included a molecular skeleton key, a protein called TAT that allows the package to enter the cells. In cultures of human T cells, the molecule killed 75% of HIV-infected cells within 16 hours. The team has already licensed development to IDUN Pharmaceuticals in La Jolla, California.

Some researchers believe, however, that the molecular Trojan Horse may have a fatal weakness. The protein would have to be injected into the bloodstream, where it "is likely to induce an immune response and be eliminated before it gets to the infected cells," says virologist Nava Sarver of the National Institute of Allergy and Infectious Diseases. Sarver is more optimistic about Dowdy's idea to adapt the strategy for localized cancer treatment, by using a TAT/caspase that can only be activated by tumor cells.