Scientists have genetically engineered a new strain of mice that may be a model of the early stages of Alzheimer's disease. Even before the brains of human patients develop a hallmark of the disease--the protein clumps called plaques--their memories deteriorate and they lose neurons. The new mice, described in tomorrow's issue of Nature, also show neuron and memory loss without plaques, suggesting that they might provide a testing ground for drugs designed to slow the progression of the disease in humans. But not all experts are convinced that the team has created an accurate model for early Alzheimer's.
In recent years, plaques, which build up over time from sticky, insoluble proteins called amyloids, have become the prime suspect for causing the debilitating symptoms of Alzheimer's. But some researchers have come to suspect that another protein, produced as an intermediate step when cells chop up a long molecule called amyloid precursor protein (APP) to form amyloid, could harm neurons even before plaques scar the brain. Test tube studies, for example, suggested that the APP fragment is toxic to neurons.
Now Jo Nalbantoglu, a molecular biologist at McGill University in Montreal, and her colleagues have created a line of mice which have brains that are riddled with these precursor fragments. To do so, they added just the piece of the APP gene for the -amyloid precursor to mouse embryos. When they grew into adults, the experimental group had much more trouble remembering how to solve a maze than did mice in a control group.
The researchers then dissected the brain to see why. When they electrically stimulated live slices of the hippocampus, a center for memory, the signal persisted in known neural pathways for 90 minutes in the control mice, but died out after about 15 minutes in the mice with copious amounts of the -amyloid precursor fragment. What's more, 18-month-old experimental mice had a 20% lower density of neurons in the CA1 region of the hippocampus--an effect seen in human patients with Alzheimer's. Previous mouse models, by contrast, have failed to exhibit this characteristic nerve cell death.
Neuroscientist Mark Mattson of the University of Kentucky, Lexington, calls the work "an important advance." But Sangram Sisodia of Johns Hopkins University School of Medicine, who works on other Alzheimer's mouse models, thinks the neuronal loss could be an artifact of the way the Canadian team counted neurons. To be a true Alzheimer's model, the mice should go on to produce amyloid and develop plaques--and they show no biochemical evidence of either. Next up, Nalbantoglu plans to look for those plaques in older mice while investigating how the precursor proteins might be causing memory loss.