There are currently no drugs to halt or reverse the spread of Alzheimer's disease, and researchers have tread carefully in the wake of a failed vaccine trial six years ago that was stopped after 18 patients developed potentially fatal brain inflammation and two of them suffered strokes.

One experimental vaccine, however, shows promise. It not only prevented the buildup of amyloid beta (Aß), a sticky protein linked to Alzheimer's, but it also does not appear to produce the dangerous side effects of earlier versions tested in humans.

Scientists say the key may be in the way the vaccine is designed. This vaccine uses an inactive herpes virus (stripped of its viral genes) to transport a small amount of Aß as well as another protein called interleukin-4 that may help prevent brain inflammation. As hoped, this vaccine triggered the immune system to produce antibodies to clear out Aß accumulations, but thanks to interleukin-4 did not cause swelling.

This "is believed to be the safest way to derive an Aß-specific immune response in someone with Alzheimer's," says William Bowers, a University of Rochester neurologist who helped develop the vaccine.

Before inoculating 24 mice with the new compound, researchers trained them to run through a maze. The mice were then split into four groups of six: one received shots of saltwater, another was injected with the inactive herpes virus, a third received shots of Aß only, and the fourth set got the vaccine containing Aß and interleukin-4.

The mice given the combination vaccine showed the most improvement in learning and memory, navigating the maze in less time and making fewer errors than their counterparts. During autopsies, researchers found that their brains had the lowest amount of amyloid deposits and contained the most antibodies.

"From these vaccinations, we saw mice brains were devoid of amyloid and they showed cognitive improvement. There were no signs of inflammation," Bowers says.

Scientists say that the new design may eliminate the problems with the failed 2002 vaccine (manufactured by Elan Corporation, an Ireland-based drug company). They caution, however, that Elan's experiments in mice also showed promise initially: The vaccine triggered antibodies to fight Aß and removed amyloid plaques from the animal's brains. And in humans, about 20 percent of patients, ages 50 to 85, produced antibodies against Aß. Those who did also showed improved memory.

The 2002 trials were arrested, however, after  6 percent of patients who received either one, two or three injections over 12 months developed serious brain inflammation.

"The problem with the previous trial of Aß vaccine was it led to a toxic response caused by activation of an inappropriate immune response," says Michael Wolfe, a neurologist at Harvard University who researches Alzheimer's disease. "But this new vaccine sounds promising. I am most optimistic about this approach among all the things currently in trials right now." Wolfe acknowledges, however, that we won't know what the long-term safety issues are until a large-scale human trial is conducted, which researchers hope to begin in the next three years.

Bowers recognizes potential problems with the vaccine, as well. People naturally produce a form of Aß that does not cause Alzheimer's, and researchers aren't sure what it actually does. It may turn out that the natural form of Aß has a positive function that could be hampered by the treatment, causing a different set of problems.

"Targeting Aß may slow the progression of Alzheimer's," Bowers says. But if the natural form of Aß does something good, clearing it from the body may have harmful effects, he says.

Cynthia Lemere, a neurologist at Brigham and Women's Hospital and Harvard Medical School who researches Alzheimer's, points out another potential issue with the vaccine: Even with the novel design, it may not be enough to avoid the more aggressive immune response.

Lemere suggests that the solution may be a vaccine that uses a shorter segment of Aß. One end of the protein is important for producing antibodies against it, whereas another part of Aß triggers the more aggressive immune response. A vaccine that only uses the fragment of Aß that produces antibodies may eliminate the chance of the harsher immune reaction.

The ultimate test will be whether the vaccine works well in humans. Until then, Bowers plans to refine the vaccine in more mouse trials to make it as safe as possible. The study, funded by the National Institutes of Health, were published in the journal Molecular Therapy.