The finding helps explain why Alzheimer's is a disease of aging. More importantly, it suggests a new weapon: drugs that give nature's cleanup crews a boost.

"It's a whole new way of thinking in the Alzheimer's field," said Dr. Andrew Dillin, a biologist at California's Salk Institute for Biological Studies who led the new research.

The discovery, published Thursday by the journal Science, was made in a tiny roundworm called C. elegans.

What do worms have to do with people? They're commonly used in age-related genetics research, and the new work involves a collection of genes that people harbor, too. Dillin's team from Salk and the neighboring Scripps Research Institute already is on the trail of potential drug candidates.

About 4.5 million Americans have Alzheimer's, a toll expected to more than triple by 2050 as the population grays. The creeping brain disease gradually robs sufferers of their memories and ability to care for themselves, eventually killing them. There is no known cure; today's drugs only temporarily alleviate symptoms.

Nor does anyone know what causes Alzheimer's. The lead suspect is a gooey protein called beta-amyloid. All brains contain it, although healthy cells somehow get rid of excess amounts. But beta-amyloid builds up in Alzheimer's patients, both inside their brain cells and forming clumps that coat the cells — plaque that is the disease's hallmark.

Thursday's study reveals one way that cells fend off amyloid buildup, and that natural aging gradually erodes that detoxification process.

"Every pathway we can discover that modifies amyloid provides us with new drug targets," said Dr. Sam Gandy, a neuroscientist at Philadelphia's Thomas Jefferson University and an Alzheimer's Association spokesman. "This now opens up a new pathway" for developing anti-Alzheimer's drugs.

Worms can't get Alzheimer's. So Dillin's team used roundworms that produce human beta-amyloid in the muscles of the body wall. As the worms age, amyloid builds up until it eventually paralyzes them; they can wiggle only their heads.

Then the researchers altered genes in a pathway called insulin/IGF-1, long known to be key in controlling lifespan. Making the worms live longer protected them from paralysis.

So in slowing down normal aging, something also slowed the buildup of toxic amyloid. But what?

Enter those cellular janitors, two proteins in that gene pathway.

One, named HSF-1, breaks apart amyloid and disposes of it, the researchers discovered. Natural aging slows HSF-1, so it can't keep up with the necessary detoxification.

Another protein called DAF-16 jumps in to help buy a little more time, by clumping extra amyloid together in a way that makes it less toxic.

That was a key finding, Dillin said: Until recently, scientists thought amyloid clumps, or plaques, were the bigger problem. His research supports more recent findings that smaller amyloid tendrils inside cells are the really poisonous form.

"We think probably the HSF-1 is the preferred way" to dispose of amyloid, Dillin said. "By the time you see the plaques, it's too late."

Mammals, including people, have these same proteins. Dillin now is repeating his experiment in mice to be sure they work the same way.

Scientists already are creating drugs to try to rid the brain of amyloid. These cleanup proteins point to a novel way to do that. The hope: Create drugs that boost their effects, and amyloid might not build up in the first place. Dillin said some initial drug attempts are showing promise in his worms.

The proteins won't be the brain's only natural amyloid scrubbers, noted Gandy, whose own research points toward involvement of another age-related gene.

The study is key for an additional reason, he added.

"We all knew that aging increases the risk for Alzheimer's," but not why, Gandy explained. "Now there's a direct link. ... It gives the molecular connection between aging and Alzheimer's."

And this process of "toxic aging" likely plays a role in still other neurodegenerative diseases, Dillin said, citing similar research with Huntington's disease.