HIV vaccine research is at a crossroads, as once-promising approaches have proved disappointing and uncertainty lingers as to where to go next. This was reiterated by discouraging results announced this month at the 2004 AIDS Vaccine Conference in Switzerland and in July at the 15th International AIDS Conference in Bangkok. "We're dealing with a very formidable scientific challenge," said Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases. "This is probably going to turn out to be the most difficult vaccine to develop in the history of vaccine research."

Nearly two-dozen prototype vaccines are in human trials, compared to just seven two years ago. Eleven of these are in the US government-sponsored HIV Vaccine Trials Network. Some now in trials include DNA vaccines, made up of bits of pure genetic HIV material stripped of the sections that would allow the virus to become active and reproduce itself. Others are "Trojan-horse" vaccines, created by packing other viruses with inactivated HIV proteins that researchers hope will kick the immune system into overdrive.

But many basic questions remain unanswered. For example, researchers do not know what immune responses best guard against HIV infection. And they cannot figure out how to develop a vaccine with neutralizing antibodies — critical to the effectiveness of every other vaccine — capable of killing HIV before it establishes itself in the body.

"Part of any investigation into the unknown is running into dead ends. That's how science is done," said Emilio Emini, director of vaccine research at the International AIDS Vaccine Initiative.

Missteps are inevitable because monkey studies of HIV vaccines are unable to predict how well the vaccines will work in humans. That has prompted a profound shift in research strategy: Researchers must now rely on costlier, small-scale human trials, said Lawrence Corey, director of HVTN.