Scientists still aren't sure exactly what part of the complex immune response is necessary to successfully clear amyloid protein or what part triggers excessive inflammation, as seen in the first Elan trial. But second-generation vaccines that create more-targeted responses might soon answer that question.

William Bowers and Howard Federoff of the University of Rochester Medical Center, in New York, are working on a gene-therapy vaccine delivered via a stripped-down herpes-simplex virus. Their vaccine carries both the code for the toxic protein and the code for genes involved in different aspects of the immune response. "We can shape the response and evoke different kinds of antibodies and different immune responses," says Bowers, who presented his findings at the neuroscience meeting. In addition, the researchers can use different genetic "promoters"--genetic sequences that control where and when a gene is expressed--to target the vaccine to specific cell types. By testing different varieties of the vaccines, they hope to tease apart each component's effects.

"This is an example of how to deal with a promising pre-clinical therapy that didn't work in the clinic," says Marcelle Morrison-Bogorad, associate director of NIA's neuroscience and neuropsychology of aging program. "Don't give up--just go back to the bench."