As Alzheimer's Studies Progress, Debate On Cause Persists

Sidebar: Genetics of Alzheimer's Disease -- Further Information At its annual meeting in November 1995, the Society for Neuroscience hosted a lively debate on whether ß-amyloid deposition is the cause of Alzheimer's disease. The debate was light in tone, but that masked real and often rancorous divisions in the field (R. Finn, The Scientist, Oct. 16, 1995, page 14.). Since then, significant advances in this rapidly developing discipline have generated considerable excitement among Alzheimer's researchers. These include the discovery that mutations in two "presenilin" genes directly cause the disease-apparently through a mechanism involving a certain form of ß-amyloid-and the development of no fewer than three realistic mouse models. Back in 1907, Alois Alzheimer first noticed the two main neuropathological hallmarks of the disease-amyloid plaques and neurofibrillary tangles-in the brain tissue of a 51-year-old demented woman. The idea that abnormal deposition of ß-amyloid (also called Ab) was the primary cause was given considerable weight by the 1991 discovery that a rare inherited form of early-onset Alzheimer's disease was caused by a mutant form of the gene encoding amyloid precursor protein (APP), which is found on chromosome 21. UNCONVINCED: Warren Strittmatter is skeptical that presenilin mutations work through an amyloid-related mechanism. But others focused on mechanisms involving neurofibrillary tangles, especially after the 1993 discovery of the role of apolipoprotein E (apoE) in Alzheimer's. There are three main alleles of apoE-E2, E3, and E4. Warren J. Strittmatter, Allen Roses, and their colleagues at Duke University Medical Center determined that individuals who inherit one or two copies of the E4 allele have a greater chance of developing the disease and tend to develop it earlier than those without the inherited allele (A.D. Roses, Journal of Neuropathology and Experimental Neurology, 53:429-37, 1994). According to Roses's hypothesis, apoE4 promotes the aggregation of microtubules into neurofibrillary tangles, leading to cell death. Roses regards amyloid plaques as secondary consequences of the disease process, not as a central cause of neuronal death and dementia. But recent results seem to be swinging the weight of the argument back to amyloid. "I think that debate from [1995] is already stale," comments Dennis J. Selkoe, director of the Center for Neurologic Diseases at Brigham and Women's Hospital in Boston. "We've moved beyond that. Many researchers in the field have come to the consensus that there is a cascade of events causing the disease. Probably amyloid deposition is early in that cascade, but that something comes later is not a pejorative term." ALZHEIMER’S AMYGDALA: High-power photomicrograph of a section of the amygdala from an Alzheimer’s patient, demonstrating two senile plaques consisting of compacted, spherical deposits of extracellular amyloid surrounded by a halo of silver-positive dystrophic neurites. Some of the pyramidal neurons in this field contain neurofibrillary tangles. Within the last few years, two genes have been discovered that seem to have a role even earlier in the cascade than amyloid. It's been known almost since their discovery that APP mutations account for only a tiny proportion of early-onset familial Alzheimer's disease (FAD). In 1992, Gerard D. Schellenberg and his colleagues at the University of Washington in Seattle identified a locus on chromosome 14 that seemed to account for more than 50 percent of early-onset FAD (G.D. Schellenberg et al., Science, 258:668-71, 1992). In July 1995, Peter H. St. George-Hyslop and his colleagues at the University of Toronto zeroed in on the specific gene, which codes for a protein now called presenilin-1, or PS-1 (R. Sherrington et al., Nature, 375:754-60, 1995). To date, more than 35 mutations on the PS-1 gene are known to cause FAD. But mutations to PS-1 did not account for one major family, referred to as the Volga-Germans, who suffered from FAD. Schellenberg and his colleagues tracked their mutation to a locus on chromosome 1. In 1995 a mutated gene, which codes for a protein now called presenilin-2, was discovered in that location. Although the normal functions of PS-1 and PS-2 are unknown, the two proteins are more than 50 percent homologous. PRESENILIN PROGRESS: Sangram Sisodia is studying how presenilin proteins act to produce the disease. It turns out that defects in PS-1 cause an aggressive onset of the disease, explains Sangram S. Sisodia, an associate professor of pathology and neuroscience at Johns Hopkins University School of Medicine. "In those pedigrees with presenilin-1 mutations, the onset is between 30 and 40 or 45. There are actually some patients who turned up clinically at 26 or 27 years of age." By the age of 50, all carriers of PS-1 mutations will develop Alzheimer's. "But the onset is quite variable with presenilin-2 mutations. It's between the age of 40 up to the age of 80." Sisodia and his Johns Hopkins colleague Donald L. Price are studying the structure and function of the presenilins and how they act to produce Alzheimer's disease. Structurally, the presenilins both seem to be integral membrane proteins, containing between seven and nine transmembrane regions. Functionally, they seem to be working by interfering with the metabolism of APP. In the course of its metabolism, APP is processed by a protease (or possibly more than one) called g-secretase. Sometimes g-secretase clips APP after its 40th amino acid, producing Ab40, and sometimes it clips APP after its 42nd amino acid, producing Ab42. "The 40-amino-acid peptide is very soluble," observes Sisodia. "It doesn't aggregate to anything else. But the 42-amino-acid peptide has a very high aggregation rate. It nucleates very rapidly, so one could imagine that it plays a role in initial plaque deposition. What appears to be going on is that the presenilins accelerate the production of this 42-amino-acid peptide, and so one can imagine then that it deposits earlier and hence disease onset is much earlier." A 'STALE' DEBATE? Dennis Selkoe contends "we’ve moved beyond" arguing whether amyloid plaques or neurofibrillary tangles cause Alzheimer's. "My model is that both wild-type and mutant presenilin molecules are expressed in certain vesicles in the cell, in endosomes, for example, that also contain APP," Selkoe conjectures. "The theory would be that APP is normally contacted by wild-type presenilin, or if not directly contacted, perhaps there are one or two protein intermediates that we don't know yet." When presenilin is mutant, Selkoe believes, direct contact is altered. "As a result, APP is broken down by a protease in a different way than it usually is." What is still unknown is whether the presenilins have a role not only in FAD but also in the far more common "sporadic" form of Alzheimer's disease, which does not run in families. "The way I like to think about it is that presenilin is expressed at one of the lowest levels of any protein that I know that's expressed in brain," notes Sisodia. "It's almost invisible. So the fact that you can express a molecule at such a low level and when mutated it gives you Alzheimer's disease tells you that it's got to be upstream of a lot of pathways that might lead to Alzheimer's disease. And if there is this convergence in terms of the pathology, it certainly indicates that identifying the intermediates in the pathway is going to be important in identifying the causes of later-onset familial disease and later-onset sporadic disease." STIRRING UP EXCITEMENT: Karen Hsiao’s mouse model demonstrated behavioral as well as neuropathological problems. The three newly developed mouse models of Alzheimer's disease constitute another milestone in identifying the causes of-and potential cures for-the disease. The first model was developed by a group now at Athena Neurosciences Inc. of South San Francisco, Calif.; the second was developed by a group at the University of Minnesota Medical Center led by Karen Hsiao; and the third was developed by a group at Sandoz Pharma AG of Basel, Switzerland. All three groups introduced mutant APP genes into the germlines of various strains of mouse. "I find [all three models] quite compelling," remarks Selkoe. "What I'm struck by is not so much the differences as the similarities. In three different models . . . when investigators express different APP transgenes with different promoters in somewhat different strains of mice, what they get is Ab deposition in the brain. They get Ab deposits, they get increasing numbers of Ab deposits over time; they're both the diffuse type and the fibrillary type as we see in humans." But Hsiao's model seems to be generating the most excitement because she has demonstrated that her mice have behavioral as well as neuropathological problems (K. Hsiao et al., Science, 274:99-103, 1996). Specifically, they have trouble remembering the location of a hidden platform in a water maze test. Dale B. Schenk, director of neurobiology at Athena, claims that the Athena mouse shows behavioral abnormalities as well, but he declines to discuss details in advance of the publication of a peer-reviewed paper. LOOKING AHEAD: Dale Schenk of Athena Neurosciences is "extremely optimistic on the future." According to Hsiao, who is a professor of neurology at Minnesota, the fact that she was able to produce a mouse model of Alzheimer's disease means "it's not magic, and it's not a mystery any more. When the Athena mouse became known, people thought that there was perhaps only one way and one way only in which you could get a mouse to produce amyloid plaques. . . . We have received over 60 requests from academicians for the mice, and we're in the process of making them available." For more information on the genetics of Alzheimer's disease, see G.D. Schellenberg, "Genetic dissection of Alzheimer's disease, a heterogeneous disorder," Proceedings of the National Academy of Sciences, 92:8552-9, 1995. The Alzheimer's Association (more formally known as the Alzheimer's Disease and Related Disorders Association) supports research on the disease and serves as a clearinghouse for Alzheimer's information. The association can be reached at P.O. Box 5675, Chicago, Ill. 60680; (800) 272-3900. Fax: (312) 335-1110. World Wide Web: http://www.alz.org/ In the United Kingdom, the equivalent organization is the Alzheimer's Disease Society, 10 Greencoat Place, London SW1P 1PH England; 44-01-71-306-0606. World Wide Web: http://www.vois.org.uk/alzheimers/ The Alliance for Aging Research is a nonprofit advocacy organization that works to increase the priority of scientific and medical research on the human aging process and associated diseases. The alliance can be reached at 2021 K St., N.W., Suite 305, Washington, D.C. 20006; (202) 293-2856. Three other sites on the World Wide Web may prove especially valuable to researchers interested in Alzheimer's disease: The Alzheimer Research Forum (http://www.alzforum.org/) is a Web site "established for the purpose of supporting the information needs of researchers and to promote openness and collaboration with colleagues worldwide to accelerate their common search for effective treatments for Alzheimer's disease." The Alzheimer Web Home Page: http://werple.mira.net.au/~dhs/ad.html The Massachusetts General Hospital Neurology Web-Forum: http://dem0nmac.mgh.harvard.edu/neurowebforum/neurowebforum.html Despite this interest, Hsiao thinks the Alzheimer's disease community may be focusing on the wrong mouse. A year earlier she announced a different Alzheimer's model she produced by introducing mutant APP into another mouse strain. These mice show behavioral impairment but have no amyloid plaques. "I found I learned much more from the first model than from the second," she says. "They taught me that maybe I should open my mind a bit more. The absence of plaques might mean there's something more subtle going on." Selkoe expresses skepticism about the value of Hsiao's earlier mouse model: "My view is that in all cases of Alzheimer's disease we do see abundant amyloid deposits in the brain. I would view [that mouse] as less interesting because it lacks a very major feature of the Alzheimer phenotype. . . . There are many ways to perturb neurons and get a behavioral phenotype." But Duke's Strittmatter, who is closely identified with apoE-related hypotheses, has a different view. Strittmatter, a professor of medicine and neurobiology, says that Hsiao's earlier mouse model "may dissociate the production of a microscopic lesion from the primary clinical phenotype of dementia. We need a drug that will help the dementia, which may or may not alter the pathology that we see under the microscope. The fact that alteration in APP expression can cause behavior without causing the formation of the senile plaque argues that APP is important in this process, and it also argues that the microscopic plaque may not be important in the process of dementia." While stating, "I think it's wrong to polarize the field," Strittmatter remains unconvinced that presenilin mutations are producing Alzheimer's through an amyloid-related mechanism. "The presenilins could be causing Alzheimer's disease by a mechanism independent of changes in APP processing, while still causing changes in APP processing," he notes. "The more parsimonious hypothesis is that it's working through changing APP processing, but it doesn't have to be that way." And he's also skeptical about the value of the three main mouse models: "I think the mouse models are reproducing some of the pathological changes seen, specifically the senile plaques. They are not reproducing the neurofibrillary tangle." Despite the current enthusiasm and debate over presenilin and the mouse models, Selkoe expects that there are exciting discoveries yet to be made. "My sense is that in science, as in all fields of human endeavor, there is a bit of fashion," he notes wryly. "My prediction is that when the next gene is identified and cloned that is a risk factor for Alzheimer's or is an actual causative gene like presenilin, then the boat, so to speak, will tilt to the other side. The whale will be on the other side of the boat, and people will run over to see that whale." Selkoe's overall assessment of the progress in Alzheimer's research over the last year or two is "very good. The only way to increase my enthusiasm from 'very good' to 'excellent' would be if someone announced a drug that blocked a key step in the pathogenesis. He notes the need for a synthesis of scientific thought: "We should stop saying that 'my finding in my lab is definitely the most important finding, and the other guy's finding is secondary or tertiary.' I don't think that's a good way to think of it. There are a number of steps that are necessary, but each by itself is not sufficient to produce the disease. I would put amyloid squarely in that camp." "I'm extremely optimistic on the future, and I can't wait to test out the amyloid hypothesis in the clinic," declares Athena's Schenk, referring to the prospect of clinical trials of drugs that somehow interfere with Ab42 production or deposition. And Sisodia concludes, "One could say that this is the heyday of Alzheimer's disease research in terms of identifying the cellular and molecular mechanisms of disease." Robert Finn, a freelance science writer based in Long Beach, Calif., is online at finn@nasw.org.

As Alzheimer's Studies Progress, Debate On Cause Persists

Interested in reading more?

Become a Member of