By David N. Leff

At least four million Americans suffer the dementia and disorientation of Alzheimer's disease (AD), but the number of mice with the same affliction amounts to a scant 100.

These newly fledged, high-value rodents came by their AD not via familial inheritance or the brain-cell degeneration of old age, but by genetic engineering.

They are the latest in a succession of transgenic mice designed to mimic the behavioral and neuronal ravages of authentic Alzheimer's disease. Authentic, because a 100 percent unequivocal diagnosis of AD requires a postmortem of the brain, rather than clinical symptoms in the living patient.

Only cerebral autopsy can reveal the amyloid neuritic plaques and neurofibrillary tangles that are the neuronal hallmarks setting AD apart from other--treatable--senile dementias.

That's why an animal model that reproduces both the amyloid deposits around AD brain neurons, and the loss of memory and cognition that marks the clinical disease, is so much to be desired by researchers seeking psychoactive drugs to treat the untreatable AD.

In the last two years, BioWorld Today has reported the creation of at least three transgenic mice described as mimics of AD symptoms and neurology. (See BioWorld Today, June 6, 1995, Oct. 8, 1996, and Feb. 19, 1997.)

The latest of these AD look-alike and act-alike animals makes its bow in today's Nature, dated May 29, 1997, under the title: "Impaired learning and LTP [long-term potentiation] in mice expressing the carboxy terminus of the Alzheimer amyloid precursor protein."

Its lead author is molecular biologist Josephine Nalbantoglu, on the faculty of McGill University's department of neurology and neurosurgery, in Montreal.

"In Alzheimer's disease," Nalbantoglu told BioWorld Today, "aggregated ß-amyloid is deposited in extra-cellular neuritic plaques of the brain. Unlike human AD patients," she pointed out, "mice and rats don't get plaques and tangles, but the gene for them is almost identical in rodents and humans."

To construct her transgenic AD mice, Nalbantoglu recounted, "We put together a vector containing cDNA for only the last 104 amino acids of the human precursor gene for ß-amyloid, beginning at its C-terminus. This we coupled to the gene for neurofilament-like protein, as a promoter highly expressed in neurons. Then we microinjected this package into fertilized mouse eggs."

She and her co-authors jumped the resulting AD transgenic progeny through two hoops, behavioral and neuronal, to test their fidelity to the human disease.

To gauge the acquisition and loss of learning and memory, mouse runners rely largely on the Morris water maze. This is a spacious circular pool filled with opaque water, in which they place a submerged mouse-friendly platform. Helpful visual cues around the pool's perimeter point to the hidden location of the platform.

The elapsed time it takes an animal to swim to this goal post is an estimate of its spatial learning and memory.

The McGill group trained 10 of its transgenics and 11 control animals to find the submarine platform, then compared their performance.

After five days of training, control mice learned to escape to the hidden platform in about 20 seconds. Transgenics took much longer * 36 to 45 seconds. But when the platform was placed above water, the transgenics swam to it just as quickly as controls; their vision, unlike their brains, was not impaired.

Paradoxically, when the platform's underwater location was moved, the transgenics found it more quickly than control, because they swam aimlessly back and forth, intersecting the new site by chance, not memory.

"We gave the mice their first behavioral tests at eight months of age," Nalbantoglu said. "Their results worsened in their second or third year, which is very old for a mouse. It compares to AD's late onset in humans, in their 70s and 80s.

"The other mouse models," Nalbantoglu observed, "have not really done behavior testing, except for the one by Karen Hsiao at the University of Minnesota. [See BioWorld Today, Oct. 8, 1996, p. 5.] They use similar water-maze tests, but their mice were also impaired when the platform was visible. But if they don't see as well, or swim as well, they're going to have differences compared to controls that have nothing to do with memory."

Another difference between previous animal models and McGill's, she mentioned, was that "most people have used the ß-amyloid precursor protein in mutated form. We used the intact sequence, but not its full length--just the 104 amino acid-expressing fragment."

At the level of the brain itself, rather than its behavioral manifestation, the McGill group compared the neuronal mechanism of memory between its transgenic mice and controls.

"We measured LTP * the long-term potentiation activity of nerve cells in vitro," Nalbantoglu said. "One thought is that it might be a model for how we acquire memory and how we store memory."

She explained: "When we're asked to access an acquired memory, we already have synapses in neuronal pathways that have been activated, because we went through the process once. In the lab, we used pulses of electricity to activate a single cell, then looked at the response of another cell that's downstream from this previous one. What we see is that this second cell becomes activated too, because there is communication between them across synapses."

Nalbantoglu continued: "In our mouse-brain experiments, when we activated one cell, its downstream counterpart cell was also activated. Control animals maintained that activation for about 65 to 90 minutes. This strengthening of the synapses meant that communication between these cells is occurring faster than usual. In the transgenics, we got the same activation, but it decayed right away; the strengthening wasn't maintained."

About a year and a half ago, the president and CEO of Montreal-based Advanced Bioconcept Ltd. (ABL), Lloyd Segal, heard Nalbantoglu deliver a seminar introducing her transgenic AD mice. The company began funding her project and filed for a patent on the animals.

"ABL," Segal said, "is the assignee and owner of the patent, holding all commercial rights to the animals. It's a standard animal-model patent," he added, "claiming the various elements of its genetic engineering, including the promoter, the animals' pathological and behavioral characteristics, and the invention's application for the kinds of uses we foresee, which includes screening for potential AD therapies."

The patent's co-inventors are Nalbantoglu and psychologist Matthew Shapiro, senior author of the Nature paper.

McGill's animal research facility maintains a colony of about 100 of their AD transgenic mice.

"We expect that biopharmaceutical companies licensing the animals will be taking breeding pairs," Segal said, "breeding their own colonies, then subjecting these to relative treatments, comparing transgenics and controls to test efficacy."

He concluded: "Our animal is going to be an absolutely breakthrough tool. We're happy to make it available as widely as we can." *