The Berkeley Science Review: Articles

Alzheimer's Unfolded
Taking a complex approach to diagnosing a complex disease (view PDF)
by James Walker

Every 72 seconds, a new case of Alzheimer's is diagnosed. There is still no cure for this progressive and fatal brain disease, which afflicts five million Americans and counting. The numbers are expected to rise as the baby boomer generation ages—it is predicted that almost eight million people will be diagnosed by 2030, and 16 million by 2050. Yet many of the fundamental questions about Alzheimer's remain unanswered. What is the exact cause? Who is predisposed to get it? How can it best be diagnosed? How can it be treated? Researchers in a variety of fields at UC Berkeley are applying their particular skills and tools to address these questions. Some are exploring the molecular basis of the disease, others are looking for ways to use brain-scanning tools to improve diagnosis, and still others are looking into the usefulness of genetic tests to identify those at increased risk. The approach is truly multidisciplinary.

A deadly plaque
The disease is named after Dr. Alois Alz-heimer, a physician who in 1906 presented the details of a disturbing case to a small medical meeting in Tubingen, Germany. The patient, Auguste D., had approached Dr. Alzheimer with the symptoms that today are closely associated with the disease: rapidly failing memory, confusion, disorientation, difficulty expressing thoughts, and paranoia about her family and hospital staff. By the end of her life four years later, she was bedridden and mute, her decline a haunting harbinger of the disease that today affects nearly half of Americans over the age of 85. An autopsy of her brain revealed a multitude of dead and dying nerve cells, along with microscopic deposits no one had ever seen before.

The root cause of Alzheimer's disease has not been determined, but researchers think it may involve a malfunction in the processing of a protein called amyloid precursor protein (APP). This protein is embedded in the membrane of many cells in the body, but is especially concentrated in neurons. While the function of APP is not fully understood, it is hypothesized to be involved in regulating the connections between neurons. For unknown reasons, APP is enzymatically cut into different pieces, called "peptides," by a complex of proteins known as gamma-secretase. Normally, this complex cuts at a specific place on the APP protein, producing a small peptide called amyloid beta 40 (Aβ40) as a byproduct. However, about 10 percent of the time in normal cells, it cuts at a different location and releases a different small peptide, amyloid beta 42 (Aβ42). Due to its unique structure, Aβ42 is much more likely to form fibrils than Aβ40. Thus, when particular genetic mutations—or other, as-yet-undiscovered factors that increase with age—cause increased production of Aβ42, it accumulates into large, dense plaques that are deposited on the outside of neurons. These were the deposits that Alzheimer identified over a century ago. The plaques disrupt communication between neurons at the synapses, and can eventually cause the neurons themselves to die. Though the brain is resilient, its gradual submission is manifest in the typical cognitive impairments that can be so terrifying for patients and their families.

On the molecular path
Two different groups at UC Berkeley are each looking into how and why the body produces Aβ42. Bing Jap, a senior staff scientist at Lawrence Berkeley National Laboratory, studies the structure and function of membrane proteins. He is intrigued by questions surrounding the gamma-secretase complex responsible for Aβ42 production. "We are interested in the regulation of gamma-secretase—what is its structure and how does it work?" says Jap. His group was the first to isolate the protein complex from human cells and separate it into its individual components. In the process, they identified a new regulatory protein called CD147, which had never before been implicated as part of the complex. CD147 is important in Alzheimer's because it naturally decreases the enzymatic activity of a protein called presenilin-1 (PS1), the part of the gamma-secretase complex actually responsible for cutting APP. Because mutations in the presenilin-1 gene (the primary cause of early-onset Alzheimer's) often result in an increase in the production of Aβ42, CD147's regulatory effect on PS1 may prove important for understanding and perhaps even treating the disease.

Jap is also interested in determining the three-dimensional structure of the protein complex, which he hopes will reveal the mechanism for cutting APP, and may eventually explain how Aβ42 and Aβ40 are produced. This basic biochemistry is important because "once you know how it works, you can try to block the activity with a peptide or other compound—some highly selective inhibitor," explains Jap. Though they are a long way from that point, Jap believes that the first steps toward understanding the basis of Alzheimer's disease will be taken at the protein level.

Professor Randy Schekman in the Department of Molecular and Cell Biology is also interested in understanding Alzheimer's at the protein level. His interest in the field began when he was invited to an Alzheimer's meeting at the Howard Hughes Medical Institute. There, he says, "it became clear to me that there were traffic issues no one was investigating." "Traffic" refers to the highly complex and thoroughly regulated movement of membrane proteins from where they are made, near the cell's nucleus, to their final destinations inside or outside the cell. Schekman's original hypothesis was that mutated APP and PS1 proteins do not make it to the cell membrane, and instead are stopped by a "quality control" process before reaching their destination. If PS1 were still able to act on APP as the proteins accumulated in the quality control area of the cell, an excess of the aggregating Aβ42 would be produced in this arrested state.

However, while many of the mutant PS1 proteins are indeed subjected to this quality control stop, Schekman's group found that PS1 is inactive in the arrested state, meaning no Aβ42 is produced. Furthermore, when they completely prevent the transport of PS1 to the outer membrane in neuronal cells, there is no increase in the amount of Aβ42 produced. This result hasn't discouraged Schekman, however, who believes there is much more to be learned, especially since "we don't know when or where in the cell processing of APP occurs." Current studies in his lab are designed to fill in these gaps, in the hopes that a full understanding of the normal situation will shed light on what goes wrong in Alzheimer's disease.

The matter of the mind
The plaques formed by Aβ42, the hallmark of Alzheimer's disease, have to date only been identifiable by a post-mortem autopsy, which, as Andrew Szeri of the Department of Mechanical Engineering notes, "is a bit drastic as a diagnostic tool." Mental status tests—patients are often asked to draw a clock showing particular times—are good indicators, but singling out Alzheimer's as the culprit is often more difficult. Blood tests might reveal an underactive thyroid, liver failure, or a variety of other conditions that can cause impaired cognition but are unrelated to the disease. Similarly, brain scans are used to test for tumors and strokes that can occasionally result in Alzheimer's-like symptoms.

To address these problems, Szeri worked with graduate student (now PhD) Mark Kramer to find a non-invasive, quantitative method for analyzing the brains of patients that, when combined with other diagnostic tools, would give a more confident diagnosis of Alzheimer's disease before death. Kramer began a collaboration with physicians at Indiana University Hospital who were doing electroencephalogram (EEG) measurements on patients. An EEG is performed by placing a net of electrodes on the head and measuring the voltage at different points to provide a readout of brain activity. To detect evidence of Alzheimer's, Kramer and Szeri developed an EEG algorithm that looks for synchronicity between two different points in the brain. The algorithm searches for situations in which two "events," or spikes of activity, always follow one other with exact regularity. While they found many such synchronized events in the EEGs of normal patients, those of Alzheimer's patients had significantly fewer. The uncoupled spikes suggest decreased neural communication, especially between the left and right halves of the brain.

Both researchers stressed that their sample size was very small and more research is needed, but Kramer says this method "is consistent with what's known, the idea that Alzheimer's disconnects areas of the brain." However, he continues, "our measure is like a thermometer—we know that something is wrong, but not why." Next, Szeri would like to follow individuals through the progression of the disease to see if their EEG measure correlates with the decline in cognitive abilities. Eventually, Kramer and Szeri hope their work will lead to a less invasive, more conclusive diagnosis. Ideally, says Szeri, "you could just slap an EEG net on your head and 20 minutes later you have your data set. But we're a long way from that."

Professor Bill Jagust, of the Helen Wills Neuroscience Institute and the Alzheimer's Disease Neuroimaging Initiative, is taking a more direct approach to diagnosis. Because Aβ42 plaques build up before any signs of neurological impairment, a method that would specifically identify these plaques would be extremely useful to researchers and clinicians. Researchers at the University of Pittsburgh developed a compound that, when injected into the body, binds specifically to Aβ42 plaques and can be seen using positron emission tomography (PET) scans. Jagust's research focuses on whether this imaging technique can be used to differentiate between different causes of cognitive impairment, of which Alzheimer's is only one, because accurate diagnosis is a crucial first step in the process of analyzing a disease. He would also like to know if and how Aβ42 plaques appear in normal people as they age. Eventually, he hopes that, like Szeri and Kramer, his method will both offer early detection and help measure the efficacy of preventative efforts or treatments, should any become available.

To test or not to test?
Because current diagnoses of Alzheimer's aren't 100 percent accurate and rely on symptoms that appear at a late stage of the disease's progression, many investigators have been searching for genetic factors that would help predict its onset. Though there are many mutations that have been linked with early-onset Alzheimer's, such as those in PS1 and APP, only one gene has so far been linked to an increase in the chances of developing Alzheimer's late in life. This gene is known as APOE, and one form of it, called the e4 allele, increases the likelihood of developing Alzheimer's (all other things being equal). Individuals with one copy of the e4 allele are approximately three times more likely to get the disease, and those with two copies of the allele are 15 times more likely. However, since the presence of the ApoE e4 allele only increases the odds, testing for it is usually not recommended for people who have no other symptoms.

Given that there is no treatment for Alzheimer's, and that the ApoE e4 genetic test has very limited predictive value, why would anyone be interested in the results? This is exactly what Berkeley researcher Holly Gooding wanted to know. Gooding is interested in genetic counseling as part of the medical process. As part of an Alzheimer's study that incorporated a genetic test for the ApoE e4 allele, she conducted interviews with the volunteers, all of whom had at least one parent with the disease. She found that, of those people who wanted to know the results of their test, the idea of control was very important. "Alzheimer's disease itself can seem like the ultimate lack of control. Many of the participants were looking for a way to exert control over the disease, and the test was one way for them to do so. The act of knowing was empowering." Most of the people in this category also tended to seek out information whenever they were stressed, rather than hide from it. Some of the subjects said that a positive test would motivate them to "do more, such as get their finances together, write their will, or get long-term care insurance," says Gooding, although the study revealed that those without the bad version of the gene weren't any less likely to do these things, as one might expect. Overall, she found that even without prevention or treatment options, getting a genetic test might be a useful coping strategy for some individuals at risk for the disease, improving their emotional outlook and helping them deal with the issue of their health. As Gooding puts it, "Some people were so grateful that they just had something to do."

Attacking from all sides
The ultimate goal of all Alzheimer's research is to decrease the cost of this terrible disease, both in terms of human lives and the economic burden it places on the loved ones of those afflicted. Perhaps, as Jap learns more about the structure of the gamma-secretase complex, this knowledge may help Schekman design more specific experiments to determine how and when this structure is finalized. And if either one discovers a way of preventing or reversing Ab42 buildup, it may well be Szeri's EEG method that enables doctors to identify those requiring treatment, and Jagust's PET scans that verify its efficacy in live patients.

James Walker is a graduate student in molecular and cell biology.

Comments on this article? Drop us a line at with 'letter to the editor' in the subject!