Scientists Discover 3 More Genes With Links to Alzheimer's Disease

In one of them, Williams and her collaborators scanned the genomes of about 4,000 people with Alzheimer's and 7,800 people without it, looking for patterns of variation in half a million locations on the DNA chain. They found three that were far more common in the people with dementia.

One was a variation in the APOE gene -- known since 1993 -- that leads to the overproduction of amyloid protein. The two other genes -- CLU and PICALM -- were new.

CLU appears to be involved in "chaperoning" newly formed amyloid molecules and helping suppress their deposition in the brain.

PICALM, however, seems to play a role in maintaining healthy synapses, the suction cup-like connections nerve cells make with each other to communicate. The loss of synapses is highly correlated with loss of mental function in Alzheimer patients.

A second research group, led by Philippe Amouyel, a neurologist and epidemiologist at the Institut Pasteur de Lille, in France, scanned the genomes of about 2,000 Alzheimer patients and 5,300 control cases.

In addition to APOE and CLU it found another gene, CR1, that is involved in the body's inflammatory response and may specifically play a role in capturing and clearing away amyloid molecules.

Some clinical studies of people with arthritis have found that those taking anti-inflammatory drugs for long periods may have lower rates of Alzheimer's disease.

"Fighting against inflammation may be a clue to fighting against Alzheimer's," Amouyel said, noting that at the moment that is just speculation.

Williams's group had also found CR1, and Amouyel's group had also found PICALM. In each case, however, the frequency of the variations were below the cut-off that defined a "hit." The complementary findings, however, suggest the two genes are truly contributors to a person's Alzheimer risk.

Overall, the effect of APOE is responsible for about 25 percent of Alzheimer cases, CLU 9 percent, PICALM 9 percent and CR1 4 percent. Three other genes are involved in early-onset Alzheimer's, which accounts for about 1 percent of cases.

The importance of various genes in individuals may be quite different from that of the population as a whole.

"The key question is what's in the hand of cards you're dealt, what is the combination of genes you have," said Michael J. Owen, also of Cardiff University, who led the study along with Williams and another scientist, Michael O'Donovan.

The team identified about a dozen other gene "signals" that weren't strong enough to be labeled hits, but that were nevertheless quite strong.

Williams said the team plans to scan 60,000 people next year. She hopes that will provide enough statistical power to identify a second tier of weaker, but still important, genetic variations contributing to a person's Alzheimer risk.