In recent years, much of the research on the role of genes in Alzheimer's disease has focused on people with the rare and dramatic form of the illness that strikes before the age of 50. Now, however, biologists are getting a look at the genetic underpinnings of the far more common -- and more complicated -- form of Alzheimer's that appears in old age.

A study published today in the journal Nature Genetics suggests that a gene called APOE-e4, thought by some scientists to be central to Alzheimer's vulnerability, is instead important in determining when a person develops the disease, should they develop it at all. A second team of researchers, writing in the same journal, report that they have identified a new gene, called A2M-2, that seems to increase a person's risk of the dementing illness.

Neither of the two genes is essential for Alzheimer's, nor does either seem sufficient by itself to cause the disease. Nevertheless, the findings shed some light on an uncommonly murky corner of modern biology.

"This is just one more piece of the puzzle, and having this piece in hand will make it that much easier to find other pieces," said Deborah Blacker, a physician and biostatistician at Harvard Medical School, who helped direct one of the research teams.

Alzheimer's disease affects about 4 million Americans, and is the most common cause of dementia in old age. Beyond the age of 65, the prevalence of the disease doubles every five years. From 30 percent to 50 percent of people older than 85 are believed to have some degree of Alzheimer's dementia.

Over the last decade, biologists have identified three genes that appear to explain nearly all cases of Alzheimer's disease that occurs in people in their forties. That form of the illness is much more clearly inherited than the late-onset form. In the latter, both genes and environmental influences are at play, and probably not the same genes that are at work in younger patients.

In one of the new studies, researchers at Johns Hopkins University, the University of Utah and Duke University screened 4,932 elderly people for the APOE gene, using cells scraped from the inside of their cheeks. APOE comes in three slightly different versions -- e2, e3 and e4. The e4 version is the one that previous work had shown raises a person's risk of dementia.

The researchers identified 22 cases of Alzheimer's disease in 141 people who carried two copies of the e4 version (that is, one inherited from each parent). There were fewer cases in people carrying one copy -- 118 cases out of 1,452 people. The fewest cases of Alzheimer's occurred in people who carried the e2 or e3 versions of the gene, but not the e4 -- 80 out of 3,339.

Unexpectedly, when the population under study was followed over time, each group had a characteristic age at which the disease appeared. More precisely, each group had an age beyond which nobody got Alzheimer's. For the two-copy group, this was 84. For the one-copy group it was 94. For the group of people with no copies of e4, it was 95.

These findings suggest that a person's "dose" of APOE-e4 both contributes to the risk of developing Alzheimer's, and helps determine when the disease sets in. However, APOE-e4 clearly isn't the sole determinant of Alzheimer's. More significantly, it's obvious that many people are invulnerable to the gene's risky effects. What's responsible for that invulnerability is unknown.

In the second study, Blacker and colleagues at Harvard, Johns Hopkins University and the University of Alabama at Birmingham studied 286 families containing at least two people who developed Alzheimer's disease in old age. They looked for previously unrecognized genes associated with Alzheimer's. As it happens, they used knowledge gained by APOE researchers to get hints where to look.

The protein made under the direction of the APOE gene was known to attach to a cell-surface receptor named LRP. Blacker's team asked: Are there other proteins that bind with LRP that are either unusually common or unusually rare in Alzheimer's patients?

The answer was yes.

A protein made by a gene called A2M easily attached to the LRP receptor. As with APOE, the A2M gene came in more than one version. A version, denoted A2M-2, which was missing a fragment of the protein chain, was more common in people with Alzheimer's than in their family members without the disease. Having either one or two copies of the A2M-2 version of the gene, in fact, raised the risk of Alzheimer's dementia slightly more than threefold.

People with A2M-2 and APOE-e4 had the highest risk. But a person didn't have to have both to develop the disease -- or either.

"It is possible there are a lot of things [involved in disease susceptibility] that are additive," Blacker said. "There may be a large number of overlapping pathways that lead to the same end."