Israel and his colleagues are now working on ways of delivering gene therapies that last for years or even a lifetime, in the hope of developing long-lasting treatments for alcoholism. Most of the medications available now need to be taken at least once a day, and many alcoholics don't comply with the routine. A longer-lasting drug is likely to be more successful, Israel says.

Two of the three existing drug treatments approved for alcoholism by the Food and Drug Administration--naltrexone and acamprosate--limit the craving for alcohol. The other treatment--disulfiram--works in a similar way to Israel's gene therapy: by making patients sick if they drink. The trace of alcohol in mouthwash is enough to trigger a reaction, and most alcoholics "really dislike this medication," says Carolyn Drazinic, an assistant professor in the Department of Psychiatry and the Department of Genetics and Developmental Biology at the University of Connecticut. She was not involved in the gene-therapy research.

"All of these drugs," Israel says, "really require patients' compliance with their medication, which is rare."

Drazinic says, though, that a lifelong treatment that makes someone sick after a whisper of alcohol might not have too many takers. "There may be a lot of patients who would refuse something like this, if they've ever experienced a disulfiram reaction," she says. Drazinic believes that a more popular option might be a treatment that doesn't last a lifetime, but long enough not to be a daily hassle.

That, Koob says, would be better than "someone sitting there with a baseball bat telling you to take your Antabuse [the trade name for disulfiram] with your Wheaties."

Robert Swift, a professor of psychiatry and human behavior and the associate director of

Brown University's Center for Alcohol and Addiction Studies, says that the gene-therapy approach "is a very interesting technique, but it's not ready for prime time."

"There are a lot of medications that reduce drinking in animals but may not be as effective in humans," he says. "The question is, can you really make enough difference in the enzymes that humans will reduce their drinking?"

Gene therapy is risky, and if it's ever used to treat alcoholism in humans, it should be a last-ditch option for hardcore alcoholics, Swift says. However, those patients are often suffering from liver damage, and "if someone's got damaged liver cells, you've got a greater risk of complications from genetic treatment."

Israel's gene-therapy approach is "perfectly logical," says Raymond White, director of the University of California, San Francisco's Ernest Gallo Clinic and Research Center, where scientists study the biological basis of alcohol and substance abuse. But White adds that he'd be quite surprised "if this became a real therapy."