Search Continues for Biochemical Pathway That Leads to Onset of Alzheimer's Disease

Like explorers searching for the source of a long, expansive river, Alzheimer's disease (AD) researchers are seeking the origins of an exceedingly complex biochemical pathway, one that culminates in the onset of a debilitating condition that afflicts more than 4 million people in the United States alone. Along the way, investigators run into numerous tributaries, factors that, although proven to somehow contribute to disease pathogenesis, may or may not be connected or causally related to one another. Although recent findings in the field have been extremely encouraging, scientists are far from being able to precisely map and unify these factors into a synthesis that accurately describes exactly what's going on and why. In the early 1990s, researchers hotly contested the cellular and molecular mechanisms of the disease; they debated which of amyloid beta (Aß) deposits or neurofibrillary tangles composed of the protein tau, was the more likely culprit for the widespread neuron death characteristic of AD sufferers. Now, most agree that both are intricately involved. They also almost uniformly agree that the malfunctioning of amyloid beta-protein precursor (APP) on chromosome 21, has a crucial role. The parent protein of Aß, APP helps trigger deadly plaque formation. CRITICAL LINK: Studies have tied Par-4 to the chain of events that leads to neuron cell death in Alzheimer's patients, contends Mark Mattson. The AD field got a shot in the arm when, a few years ago, two genes, presenilin-1 on chromosome 14 (R. Sherrington et al., Nature, 375:754-60, 1995) and presenilin-2 on chromosome 1, (E. Levy-Lahad et al., Science, 269:970-73, 1995) were confirmed to have a role in the disease cascade even earlier than Aß. The former, mutated much more frequently, causes severe early onset AD; the latter triggers the disease at varying ages. Current findings do suggest one cascade of events that leads to AD onset: Genetic mutations in the presenilins and APP result in faulty APP processing and overproduction of Aß. But complications abound. These genes, while important, account for only about half of the occurrences of the early onset form of the disease, and for very little of the late-onset form. And these findings have brought with them an array of new targets of investigation. Numerous other cascade players, several in just the past few months, continue to crop up. In the August issue of Nature Medicine (Q. Guo et al., Nature Medicine, 4:957-62, Aug. 1998) a team from the University of Kentucky Chandler Medical Center reported that a protein called Par-4 (prostate apoptosis response 4) may contribute to cell death in persons suffering from AD and other neurodegenerative disorders; Par-4 was originally observed in prostate cancer cells undergoing apoptosis, or programmed cell death, following exposure to chemotherapeutic agents. When researchers exposed cultured hippocampal neurons to Aß peptides, levels of Par-4 proteins rapidly increased. And when they blocked the increase of Par-4 expression with antisense DNA, neurons staved off Aß-induced death. Researchers also found that cell cultures containing presenilin-1 mutations had higher Par-4 levels than controls had. TOO SOON: Scientists are still in an early stage of understanding risk factors, says Daniel Alkon. "These experimental studies suggest that Par-4 is a critical link in the chain of events that leads to neuron cell death in Alzheimer's patients," contends Mark Mattson, senior author and a professor of anatomy and neurobiology at Kentucky. However, exactly how and where Par-4 comes into the picture is still unclear. Mattson hopes to clarify things with a couple of mouse models currently in the works. Other potential neuron killers include a family of enzymes called caspases. Last month, a team composed of scientists from Yale University and Vertex Pharmaceuticals in Cambridge, Mass. reported the first in vivo evidence of Caspase-9 related apoptosis (K. Kuida, Cell, 94:325-37, Aug. 7, 1998). Transgenic mice missing the Caspase-9 enzyme avoided neuron death in the stages of early development. Both Par-4 and the caspases have been billed as therapeutic targets. However, understanding the mechanism of cell death in AD may require not only an examination of cell components in isolation, but, as in the case of cancer, a broader view of the variety of mechanisms that disrupt proper cell cycling as well. With cancer, too much uncontrolled cell cycling, too little cell death, spawns tumors. Recent studies indicate that with neurodegenerative disorders it's too much cell cycling that triggers neuron death. Hanging over Karl Herrup's desk is a large sign that says simply, "Why cycle? Why die?" For Herrup, a professor of neurology at Case Western Reserve University, the sign nicely sums up the questions that, as he puts it, "keep him up at night." In a study published this past April (J. Busser et al., Journal of Neuroscience, 18:2801-7, April 1998), Herrup and colleagues found that, before dying, the brain cells in AD patients unexpectedly enter into something that looks a lot like a cell cycle. According to Herrup, there's been growing recognition both in the Alzheimer's field in particular and in the cell biology field in general that the regulation of cell cycling and programmed cell death are very closely related. Rather than focusing on the role of specific cell cycle components, his team looked to the regions of the AD brain where neurons are at risk, and found within those neurons proteins called cyclins, proteins known to be unique to cell division. This, according to Herrup, is "pretty strong circumstantial evidence that something has reinitiated the urge of the cell to divide." He emphasizes that there's nothing wrong with the cells' signal transduction mechanisms. Rather, it's some external pressure that's forcing otherwise genetically normal cells to try to divide--a decision that's lethal. Right now, however, the most viable targets for AD treatment may actually involve the original players, the ones that started it all: Aß and tau. "The priority has to be determining which of these many proteins implicated in the disease by either by genetics or by molecular biology ... are going to be meaningful targets," says Rudy Tanzi, an associate professor of neurology at Harvard Medical School and the director of the Genetics and Aging Unit at Massachusetts General Hospital. He adds that because cell death is a healthy process in many parts of the body, investigators should tread very carefully when developing strategies aimed at halting apoptosis. His group and others have done preliminary work on understanding how the presenilins, caspases, and Aß interact to lead to brain cell suicide. According to Tanzi, most AD researchers are in general agreement that, based on current knowledge, therapeutic targets are still best aimed at curbing Aß accumulation and preventing tangle formation. PERILOUS WATERS: Investigators should tread very carefully when developing strategies aimed at halting apoptosis, says Rudy Tanzi. Daniel Alkon, chief of the Laboratory of Adaptive Systems at the Neurologic Disease and Stroke Institute (NDSI), agrees. "I don't think we're near a synthesis," Alkon says of the prospect of placing a variety of targets of AD research into one cascade. "But I do think that no one would doubt that beta-amyloid, amyloid precursor protein, and tau protein are important in Alzheimer's disease." Researchers have known for some time that AD brains were polluted with damaging neurofibrillary tangles composed of tau, and that understanding tangle formation, as well as Aß accumulation, was key to understanding how AD came about. Until very recently, however, the tau gene had been pushed to the wayside by many who considered it a secondary player, a nonviable target for treatment in the AD disease cascade; they instead preferred to look for causative agents further "upstream." Tau protein normally stabilizes the internal structure of neurons and makes up the channels that carry nutrients and other molecules through cells. In AD sufferers, disrupted tau causes a communication breakdown between neurons. Several teams' findings published in the last few months suggest that mutations in tau cause hereditary forms of fronto-temporal dementia (FTD), dementias related to AD. The findings affect AD research because they're the first to show that tau complications lead to neuron death. Says Michael Hutton , lead author of one study (M. Hutton et al., Nature, 393:702-5, June 18,1998) and assistant professor of neurobiology at the Mayo Clinic in Jacksonville, Fla., "It suggests that the tau pathology we see in Alzheimer's disease is not just a secondary effect of the disease process, but it's actually part of the pathogenic cascade." Hutton and colleague John Hardy, a professor of pharmacology at the Mayo Clinic in Jacksonville, have presented a controversial hypothesis suggesting that tau and Aß actually interact (J. Hardy et al., Nature Neuroscience, 1:355-8, Sept. 1998). They contend that APP and presenilin mutations cause increased amounts of A ß , which leads to tau dysfunction and the formation of tangles. "When you have deposits of amyloid, you do get tangles. But when you get tau mutations, you don't get amyloid," explains Hardy. Says Tanzi, "It may be that in the end of all these pathways, the convergence is just abnormal tau and tangles, and prevention of the disease can be accomplished by preventing the aggregation of tau and the formation of the tangles." The Tale of A2M: This past July, a team from Massachusetts General Hospital (MGH) in Boston found a new genetic risk factor for late-onset Alzheimer's disease (AD). Or did they? Announced at the Sixth Annual International Conference on Alzheimer's Disease in Amsterdam, Netherlands, the findings include statistical genetic evidence that the harboring of a mutation for the gene that encodes the protein alpha2macroglobulin (A2M) indicates an increased risk for AD. The study, which has since been published (D. Blacker, Nature Genetics, 19:357-60, August 1998), reports that A2M confers as much of a risk (about threefold) of developing late-onset AD as does the confirmed risk factor apolipoproteinE4 (ApoE4), an allele of the lipid-ferrying protein ApoE. But others aren't so sure. Groups from Duke University, the University of Toronto, and Washington University in St. Louis are among the teams that have been searching for a late-onset familial AD risk factor on chromosome 12 since the Duke group first reported evidence for such a gene a year ago (M.D. Pericak-Vance et al., JAMA-Journal of the American Medical Association, 278:1237-41, 1997). Peter St. George-Hyslop, the leader of the Toronto group and the director of the Centre for Research in Neurodegenerative Diseases at the University of Toronto disputes the MGH team's conclusions saying that other A2M tests using other data sets have yielded different results. Hyslop also argues that their methods may have been inadequate; they based their results on a sib-pair family based association method, but got negative results using the lod (log of odds) score method, which estimates genetic linkage between family members. But MGH team leader and senior author Rudy Tanzi, an associate professor of neurology at Harvard Medical School and the director of the genetics and aging unit at MGH, is sticking to his guns. He emphasizes that it's not necessary for every late-onset AD gene to yield similar results in all genetic tests, and that the lod scores for ApoE4 are "not particularly impressive." He's also quick to note that data presented by the Washington University, Duke, and Toronto groups indicated positive lod scores for genetic markers next to A2M. Hyslop's group has just recently reported that, according to data from a recent study (E. Rogaeva et al., JAMA, 280:614-18, Aug. 19, 1998), there is a risk factor on chromosome 12 somewhere within a specified area of about 60 centimorgans, an area that actually includes A2M--but he doesn't think that A2M's the culprit. Interestingly, a paper by a group from Washington University, which appeared simultaneously in JAMA with the Toronto group's (W.S. Wu et al., JAMA, 280:619-22, Aug. 19, 1998), uses nearly identical data to reach a much different conclusion. According to their interpretation, the data does not support the presence of any risk factor on chromosome 12. Contributing author John Hardy, a professor of pharmacology at the Mayo Clinic in Jacksonville, Fla., attributes the contradictory conclusions to the particularly poor resolution methods used by all groups--the goings-on of the genes on chromosome 12 are simply still too hard to make out regardless of the method employed. With regard to the MGH team's findings on A2M in particular, Hardy contends that, although he has confirmed Tanzi's linkage data with the same set of families that Tanzi studied, other data coming from several sources suggests that variants of A2M are not more common in AD cases than they are in controls. "We have these two discordant pieces of data ... neither piece of data is wrong," explains Hardy. "It's matter of interpretation at the moment." Hardy's best guess is that Tanzi's results are flawed due to type one error. "In other words, it's just a statistical freak," he says. But, while Tanzi concedes that type one error is possible, he asserts that it's very unlikely. Using the sib-pair family based association method in his study rather than a case-control approach circumvents the most common cause of type one error: detecting a difference in allele frequency due to the distributions of ethnic groups in a given data set rather than due to disease status. Tanzi suggests that if his results are proven wrong, the most likely source of error will have been his team's detection of an association with a gene right next to A2M. However, only a sib-pair analysis done with a new set of pedigrees will convince Tanzi that he's erred. He notes that the methods used in the August JAMA studies were not sensitive enough to pick up definitive proof of risk factors, and points out that certain genetic analyses are not appropriate for certain tasks. "They're welcome to keep looking on chromosome 12 for what they think is the real gene," maintains Tanzi. "We've already begun our biological experiments on A2M." In a study due out this November in Human Molecular Genetics, Tanzi and colleague Brad Hyman, an associate professor of neurology at Harvard, report a new A2M polymorphism associated with AD. And Tanzi's group has now moved on to functional studies to discover how an A2M deletion operates in the AD cascade. "I think everyone is aware that you can make a good biological story for A2M," Hyslop counters. "But there are obviously flies in the ointment." Both Hardy and Hyslop expect that, in all liklihood, studies done in the next few months will provide sufficient data to overturn Tanzi's findings. Whether A2M is a risk factor or not, perhaps the more pertinent question is, why does it matter? Risk factors may, in the future, prove useful in deciphering the cascade of events that leads to AD, but right now they only provide an estimate for the chances of developing the disease. "Fundamentally what we're waiting for ... is a successful treatment," says Michael Hutton, contributing author of the recent Washington University JAMA paper and an assistant professor of neurobiology at the Mayo Clinic. "The issue then of assessing one's risk factors for developing the disease will become increasingly useful." "Risk factors, to my mind, ... are still at an early stage of understanding," says Daniel Alkon, chief of the Laboratory of Adaptive Systems at the Neurologic Disease and Stroke Institute (NDSI). "They're just telling us that there's an association. Those kinds of things are interesting, but [they] are far from revealing the nitty-gritty of what's going on." Alkon's team hopes to directly address the most universal symptom of AD: loss of memory. He and others have chosen to focus their efforts on the maleffects of Aß on potassium channels, channels in neurons believed to be critical for learning and memory. Alkon's lab first identified defects of potassium channels and calcium homeostasis in the cells of patients with AD several years ago. They then showed that in vitro cells could develop the same defects after exposure to low levels of Aß His group has mapped out an elaborate molecular cascade (due to be published in a review article in the October issue of Trends in Neuroscience), which they believe governs the regulation of potassium channels. Alkon hopes to use the model as a diagnostic tool for developing drugs to curtail Aß 's damaging effects on memory. Risk factors, Alkon emphasizes, represent much less direct links to disease onset. Rather, they may set up conditions that predispose an individual to AD. For example, problems with ApoE, which removes cholesterol from the blood, may lead to increased incidence of artherosclerosis. And artherosclerosis may lead to reduced levels of red blood cells and oxygen deprivation in the brain. Says Hyslop a bit more optimistically, "It probably in the long run is going to be helpful to have multiple genetic entry points in the path that leads to Alzheimer's Disease." He goes on to caution, however, that, although genetic risk factors are the hot topic of late because of the available technologies, potential environmental risk factors like toxins have not yet been adequately investigated.

Search Continues for Biochemical Pathway That Leads to Onset of Alzheimer's Disease

Interested in reading more?

Become a Member of