By David N. Leff

More people die in bed than anywhere else. So the logical lesson is to stay away from beds.

Clinical investigators live in dread of falling into this statistical error of misplaced guilt by association. But the flip side of their coin is the fact that many biomedical correlations are so compelling — albeit, unproven — that they emerge as "risk factors."

Such a one is the connection between Alzheimer's disease (AD) and a huge protein in the blood that acts as circulation manager for cholesterol and other lipid-hustling apolipoproteins. Its name is apolipoprotein Ee4 (APOEe4) — its fame, that it pops up 50 percent more often in AD patients than in the population at large.

That discovery some five years ago by molecular neurologist Alan Roses, at Duke University, in Durham, N.C., put APOEe4 on the map as a major risk factor gene for susceptibility to AD. It led Athena Neurosciences Inc., of South San Francisco (since acquired by Elan Corp. plc, of Athlone, Ireland), to launch — under license from Duke — a diagnostic test for AD. (See BioWorld Today, March 27, 1996, p. 1.)

But as molecular neurogeneticist Corinne Lendon, at Washington University, in St. Louis, pointed out: "Having a predisposition gene doesn't mean a person is going to get the disease. That's the caution for any use of diagnostic testing for susceptibility of these risk factors."

Lendon is among the authors of an article in the January issue of Nature Genetics that announces an entirely new but related AD risk factor. It's based not on the APOEe4 gene or its protein product, but on the quantity of that protein expressed in the brain. The paper's title: "A polymorphism in the regulatory region of APOE associated with risk for Alzheimer's dementia."

That research reflects an unusual transatlantic collaboration between molecular neurologists in Madrid, Spain, and St. Louis. The Spanish group, supported by Boehringer Ingelheim España, was the first to identify key alleles (i.e., inherited variants) in the promoter region of the APOE gene, which ups the protein's expression in the brain.

The scientists then mapped these variants in a large cohort of AD patients — some diagnosed, others autopsied — compared with a closely matched population of controls.

"The APOE gene variant that we and the Spanish group have been looking at," Lendon said, "lies in the promoter region and appears to differentially bind nuclear proteins that alter transcription levels. In effect, it's controlling how much of the gene is being expressed."

Verification By Replication

"The group in Spain actually contacted us," Lendon told BioWorld Today, "after they had studied their sample families. One of the problems with case-controlled samples," she observed, "is it's very easy to generate Type I errors [i.e., perceiving unsupported differences] and spurious statistical associations.

"So when the Spanish investigators asked us to replicate their findings," Lendon continued, "we did. And because the data came out in exactly the same direction, with the same risk-factor effect, we became quite excited, and carried on to study the effect of the different allelic variants in an in vitro situation."

Into two cell lines, a hepatoma and an astrocytoma, transfected with APOE genes, they tested for one or another parental allele of the promotor region. "Astrocytes," Lendon explained, "are cells that synthesize the APOE protein in the brain. All neurons are supported on a bed of astrocytes, which made them the cells of choice for us. By adding two different allelic variants of the APOE gene, A and B, to this astrocytoma cell line, we got as close to physiological as one can get in the test tube."

She pointed out that senile amyloid plaques, one hallmark of the Alzheimer brain, form outside the neurons, while fibrillary tangles, the other AD signature, occur inside the lesioned neurons.

"What we found," she recounted, "was that a double chromosomal dose of the A allele, which we had seen to be associated with increased risk for AD in our two sample populations, increased the level of APOE expression. It was functional proof that the two variants produced a functional difference.

"And circumstantially also, it's supported by the reported occurrence of increased levels of messenger RNA for APOE in the brains of AD patients, which may be possibly causative." She added a caveat: "We're still at the stage of having to confirm this, but the good news is that we have some in vitro data to support it.

"Even though the Spanish and American ethnic origins are very different — although both are of Caucasian European descent — we got a second sample that produced the same results. And even as we speak," Lendon added, "we're trying to replicate it in yet two other case-controlled samples

Mice Suggest APOE-Amyloid Link In AD

An in vivo experiment reported in Nature Genetics for November 1997 added evidence of a functional link between APOE and the amyloid plaques of AD.

Its investigators crossed transgenic mice carrying a mutant human gene for amyloid precursor protein, which normally deposits amyloid-beta in the brain, with animals carrying one, two or zero copies of the APOE gene.

"What they found," neurogeneticist Alison Goate, at Washington University, told BioWorld Today, "was that the amount of amyloid deposition correlates with the number of copies of APOE, suggesting that the level of expression might be important in determining one of the factors that controls how much amyloid is deposited."

But Goate, who is the American senior author of the January paper in Nature Genetics, observed that the actual pathogenic function of APOE in AD "remains completely unknown."

"Now that we're looking at two different AD risk factors within the same gene," Goate went on, "it looks as if the APOE story is more complicated. Where I think it might be useful in the long-term future," she predicted, "is when we have a number of drugs available that could be used to treat AD. It might be that people with different genetic risk factors will respond differently to different drug treatments. Therefore, it will be very important, that in any clinical trials of such drugs, we collect the genotypic information on responders and nonresponders, so we can see whether there is some correlation between genetic risk factors and response to drugs." *