By David N. Leff

MIAMI — "We spent a lot of money to demonstrate that mice are not human."

So spoke neuroscientist Alan Roses, addressing a session of the 30th Miami Winter Biotechnology Symposia on "Genetic Analysis of Complex Diseases." High on the roster of such ills is Alzheimer's disease (AD), to which Roses contributed the insight in 1992 that apolipoprotein E-4 (APOE-4) is a hallmark of AD lurking in the brain cells of its sufferers.

"APOE-4 is a 299 amino acid protein," Roses explained, "that trucks cholesterol around the bloodstream. But it also has functions of carrying lipid-like molecules, and it doesn't necessarily have to be in the bloodstream; it can do that in the brain. And what it does in the brain may be quite different than what it does in the bloodstream, and that's what we're studying."

After studying APOE-4 for years at Duke University, in Durham, N.C., Roses moved last May to nearby Glaxo Wellcome Research and Development, in Research Triangle Park, N. C., as vice president and worldwide director of genetics.

There he determined, at considerable expense, why workers had never been able to create a transgenic mouse expressing the human APOE-4 gene in its brain.

"The one thing that people don't understand about transgenic mice is that the mouse does not make APOE in its neurons, like humans do. It's not because of the gene itself; it's because of the downstream promoter that's present in human DNA. So taking a cDNA [a stripped-down coding region of the gene] and sticking it into a mouse and expecting to see human expression isn't going to work."

To circumvent this roadblock, he explained, "You have to take genomic DNA that has the downstream promoters. And there's a whole big argument in the literature as to whether what we found is true or not true, based on different transgenics.

"Whether it's human APOE or mouse APOE, if you knock in the gene alone it's not expressed in the neurons of the mice. But if you put in the rest of the human surrounding stuff, then you see it.

"Now, these are created mice," Roses told his chuckling audience: "All transgenic mice are not created equal."

As a prelude to trying to determine what APOE-4 is doing in AD brains, human or murine, Roses conducted a trial run of a new genome mapping technique based on single nucleotide polymorphisms (SNPs).

He explained: "SNPs are the difference between one base and another base that are common in people. Once in every 1,500 base pairs of DNA, there are differences that can be detected in populations. What we demonstrated was that that number is approximately true; that we were able to find them, to map them, and we are now able to do the genetic experiment, which we're in the process of doing.

"Here's how we do this proof of principle," he continued. "We take these ordered polymorphisms to look for, say, the APOE-4 single nucleotide polymorphism, which is one of these, to see whether we'll be able to detect it in AD, because we already know it's there. But how far on either side of that site, because of linkage disequilibrium, could we go, and still detect the positive peak that something was there?

"The idea was having an SNP map that goes over the human genome. What we did at the Glaxo Wellcome laboratories, between September 1997 and January 1998, was demonstrate that it is possible, and even practical, to locate the APOE-4 SNPs not on 100 kilobases on average, but 30 kilobases on average. We did in three months with SNPs what conventional linkage mapping would have taken three years.

"The continuation of our experiment now," he went on, "is to take these SNPs and test them on the populations of AD patients that are at Duke University. We have given all those SNPs we did in Glaxo Wellcome to our former colleagues, Margaret Percik-Vance and Jeffrey Vance, in Duke. They are doing the experiment and will report it out.

"And as soon as our paper gets published," Roses said, "we will put all the SNP data in the public domain, just like good boys and girls are supposed to do."

Alzheimer's Definition May Need Refining

In his presentation, Roses pointed out that there was an increased association in patients with AD having the APOE-4 allele. So that in a normal population, 15 percent of the chromosomes have the APOE-4 allele. In an AD population, it's 40 percent. But he emphasized that "just because one-third of the population carries an APOE-4 allele, they're not all going to get AD. It's an association, a susceptibility, not a cause."

To which he added: "If we all lived to the age of 140 or 150, we'd all get AD."

At some point in the future, Rose predicted, "it will be necessary to revise the definition of Alzheimer's disease."

"Now," he observed, "it's defined pathologically, which is as good as you can get today. But it's defined by having to have amyloid plaques and fibrillary tangles in the AD brain. This leaves out things without a lot of plaques. They can't be classified as AD.

"But if you look at the AD population," he continued, "it turns out it's enriched for APOE, too. So in fact what you're seeing is a later onset with less amyloid, but it very well may be the same process.

"What we would strive for," he concluded, "is what happened in other diseases, like asthma — defined as response to a medication. Well, if we get a medication that AD with APOE-4 responds to, but AD without APOE-4 also responds to, then maybe we'll redefine AD." *