Neurology: Alzheimer's Study Maps Alternate Route to Disease Cell Biology: Yeast Ramps Up Sugar Production to Survive Cold Clinical Research: Harvard Clinical Research Group Gains Reputation for Trial Design, Statistical Analysis Women's Health In Kass Lecture, Brundtland Points Way Toward Eliminating Global Health and Gender Disparities Regulator of Protein Degradation Emerges as Anticancer Target Flaws Revealed in Study Assailing Mammograms Inflammatory Pathway Uncovered with Tie to Early Atherosclerosis   New Full and Named Professors Thier Professorship Established for Work in Health Policy at BWH or MGH Hirsh Nominated for AAMC Humanism Award, Celebration Planned for Award Nominees Student Research Displayed at HST Forum HMS Again Takes Top Spot in U.S. News Rankings Honors and Advances News Briefs   Coalition Supports Haitian Immigrants   When It Comes to Drugs, Price Is Not the Real Problem   Front Page

NEUROLOGY Alzheimer's Study Maps Alternate Route to Disease Presenilins Linked to Loss-of-function Dementia In 1906, Alois Alzheimer was the first to correlate dementia with the presence of amyloid plaques, dense proteinaceous deposits that appear throughout specific regions of the brain, particularly the hypothalamus and cortex. Fast forward almost 100 years and scientists are still striving to piece together sometimes conflicting clues to the pathology of the disease that now bears his name. One theory suggests that the presenilin (PS) proteases that cleave beta-amyloid peptides from the amyloid precursor protein are overactive and release too many peptides for the brain to manage. The discovery that certain rare forms of the disease are caused by presenilin mutations supports this gain-of-function hypothesis, as does the finding that these mutants preferentially lop off the slightly longer and stickier of the two amyloid peptides, Abeta42. Conditionally knocking out presenilin genes in forebrain neurons leads to progressive neurodegeneration and to learning and memory deficits. Hippocampal dendrites in two-month-old double knockouts (cDKO) appear normal (left), but seven months later dendritic spines are significantly depleted compared to those from age-matched control animals (right). (Copyright 2004 Cell Press.) But do presenilin mutations really result in a gain of function? Maybe not, suggests Jie Shen, HMS assistant professor of neurology at Brigham and Women's Hospital. Reporting in the April 8 Neuron, Shen shows that presenilins are needed to prevent neurodegeneration and that they are essential for learning and memory, two of the brain functions hardest hit in Alzheimer's patients. "The role we've found for presenilins in the adult brain is almost exactly what one would expect of an Alzheimer gene," said Shen. Her data suggest that familial presenilin mutations that cause Alzheimer's may primarily result in a loss of function. Bart de Strooper, professor of molecular medicine at the University of Leuven, Belgium, agrees. De Strooper was one of the first to propose the gain-of-function hypothesis in the late '90s and has been commissioned by Neuron to write a review on Shen's work. "Shen's data add a new twist to our thinking about familial Alzheimer's disease," he said. "A year ago it was more or less accepted that familial Alzheimer's disease mutations were gain-of-'disease'-function; now it is apparent that they could also act via a loss of function." Precision Knockouts Shen spent eight painstaking years, with support from the Alzheimer's Association and the National Institute of Neurologic Disorders and Stroke, to arrive at this conclusion, developing mouse models in which either PS1 or both presenilin genes are inactive. Because presenilins are essential for development, double knockouts die as embryos. So to determine what role the presenilins have in adult mice, Shen and colleagues set about to make conditional knockouts, or cKOs. In these model animals, the gene of interest can be selectively ablated with spatial and temporal precision. Shen engineered double cKOs, or cDKO mice, so that both presenilins would be silenced only in forebrain neurons, and only postnatally, to avoid any impact on development. "A year ago it was more or less accepted that familial Alzheimer's disease mutations were gain-of-'disease'-function; now it is apparent that they could also act via a loss of function." --Bart de Strooper At two months old, the cDKOs appeared similar to control mice, and they had normal brain morphology. In tests of learning and memory, however, they were lacking. Their deficits were apparent when postdoctoral fellows Carlos Saura and Seema Malkani subjected the mice to the now famous Morris water maze test. This requires that mice, using visual clues, train themselves to find the safety of a hidden platform in a circular water tank. After five day's training, normal mice were taking about 17 seconds to find the platform. The PS conditional knockouts took considerably longer, almost 60 percent longer in fact, or around 27 seconds, indicating they had a spatial memory deficit. Further evidence of the memory impairment came from contextual fear conditioning, in which mice learn to associate an experimental chamber with a mild foot shock. Again, the cDKOs remembered the association less well than control mice. As in Alzheimer's disease, these memory impairments became much more severe as the animals aged. By six months of age, for example, the mutant mice took more than 50 seconds to find the platform in the water maze. Dementia Without Plaques So what was happening to these animals? To address this question, Shen enlisted the help of Alfredo Kirkwood, an electrophysiologist at Johns Hopkins University. Kirkwood, together with his postdoc Se-Young Choi and Dawei Zhang from Shen's lab, examined the mice for deficits in synaptic function that might alter learning and memory. Both these higher brain functions require synaptic plasticity, which is demonstrated by long-term potentiation, or the ability of a neuron to respond to repetitive stimulation by generating stronger and stronger electrical signals. Again, neurons from mutant mice were found wanting. Those from two-month-old mutant animals were about 20 percent weaker at long-term potentiation than normal neurons, while neurons from six-month-old animals were more than 35 percent weaker than those from age-matched controls. Presenilin mutations were thought to cause Alzheimer's disease through a gain-of-function mechanism. Now, Jie Shen and colleagues, including Dawei Zhang (left) and Vassilios Beglopoulos (right), show that learning and memory deficits in mice can actually be caused by the loss of these genes. (Photo by Leah Gourley) To understand what molecular changes could explain the learning and memory defects, Shen, together with Saura and postdoc Vassilios Beglopoulos, examined expression of two key players in long-term potentiation, the N-methyl-D-aspartate receptor and alpha calcium calmodulin-dependent kinase II. They found that both of these were significantly reduced in two-month-old mutant mice. But there were also other molecular ramifications of knocking out presenilins. CREB-binding protein, which together with CREB activates a whole host of genes, was also significantly repressed, and phosphorylated tau, the major constituent of the neurofibrillary tangles found in brain tissue from Alzheimer's patients, was elevated. These effects may underlie the substantial and progressive neurodegeneration that Shen observed in the cDKOs. In the neocortex, for example, neurons are decimated by nine months, while in the hippocampus, synapses and dendritic processes are extensively ablated (see figure). Taking the data as a whole, ablating presenilins seems to create many of the symptoms of Alzheimer's, but without the production of Abeta. How does this fit with current knowledge? "Gain of function is still important," said Shen, "because we know that PS mutants can preferentially produce Abeta42, and that has been shown to have effects on memory also. But we must also consider familial Alzheimer's disease in the context of loss of function." "This is the most interesting point," said de Strooper, "that loss of PS leads to synaptic problems and neurodegeneration, but with no Abeta in sight. Shen's work reemphasizes the fact that we have a major outstanding issue to be resolved, namely, there is more to Alzheimer's disease than Abeta." --Tom Fagan