Columbia University Medical Center

Going Retro: Obscure Protein Complex Furnishes Fresh Clues for Alzheimer’s Investigators

November is National Alzheimer’s Disease Awareness Month. This article was published originally in the 2014 P&S annual report.

By Andrea Crawford

Columbia neurologist Scott Small, MD, was hot on the trail of a relatively unknown cellular AD illocomponent known as retromer and sure he was on to something big—the molecular engine underlying the pathologies at the core of Alzheimer’s devastating symptoms—when a journalist investigating the root causes of dementia visited the Small lab in 2006. “I think that if you come back here in five years,” he told the reporter, “we’ll have drugs for retromer dysfunction.”

Though he did not quite make the five-year prediction, in an April 2014 paper in Nature Chemical Biology, Dr. Small revealed a “pharmaceutical chaperone” for retromer, an obscure complex of proteins that acts as a clearinghouse during intracellular transport. The finding, which resulted from a decade-long collaboration, offers the potential for targeted therapy for Alzheimer’s disease as well as Parkinson’s disease. In its report of the breakthrough, Science magazine noted that, while Dr. Small’s published work has thus far been carried out only in cells, “the new results have nonetheless impressed some veterans of the Alzheimer’s field.”

It’s an exciting lead for Columbia researchers—and one that has emerged from two directions. Richard Mayeux, MD, founding co-director of the Taub Institute for Research on Alzheimer’s Disease and the Aging Brain, has spent more than a decade compiling the genetic records of families with Alzheimer’s. Dr. Mayeux is part of an international team that has identified about 25 genes that play a role in late-onset Alzheimer’s; many are involved in retromer function as well. It is also an exciting lead for anyone touched by the intractable disease, which affects some 5 million Americans. That number is growing sharply as the population ages; last year the disease represented $210 billion in health care costs.

The retromer breakthrough comes at a time when the mood among researchers and the public alike has darkened, as a number of large-scale drug failures around the country have raised questions about the basic hypothesis of what causes Alzheimer’s disease. But that dark mood is not universal. Investigations by Taub scientists have uncovered where the disease begins, how it spreads, and what genes are associated with it. This year, the drug discovery program for retromer entered the preclinical stage by testing for efficacy and safety in animal models. In sum, P&S researchers are bringing an altogether new level of understanding to the disease and raising expectations that treatments may be within sight.

In 1906, in the autopsied brain of a woman who had suffered from severe dementia, the German physician Alois Alzheimer saw how amyloid-beta peptide—a-beta to the scientists investigating it—forms around neurons in dense clumps known as plaques and how neurofibrillary tangles, the pathological twisting of tau protein, builds up within nerve cells. In the 1980s, scientists discovered that a-beta is a byproduct of amyloid precursor protein (APP), which, like tau, is an important protein for tissue stability and cell biology and thus essential to human health. As Michael Shelanski, MD, PhD, the Delafield Professor of Pathology & Cell Biology, chair of pathology & cell biology, and founding co-director of the Taub Institute, puts it: “It turns out that in Alzheimer’s disease, as best we know, a number of proteins that are good friends of ours have taken on evil roles. The first question is ‘Why?’ The second is, ‘What can we do to return them to normal function?’”

The leading theory to address Dr. Shelanski’s first question emerged in the 1990s, when scientists identified a genetic mutation that triggers early-onset Alzheimer’s, a rare form of the disease. The mutation causes APP to be overproduced. Based on that finding, in 1992 scientists formulated the amyloid hypothesis: The accumulation of amyloid plaques lies at the root of the disorder.

“There are a lot of reasons why the amyloid hypothesis is taking body blows,” says Dr. Small, the Boris and Rose Katz Professor of Neurology and director of Columbia’s Alzheimer’s Disease Research Center. “The reasonable ones involve a couple of observations.” First is the failure of hundreds of drugs that were designed to clear the plaque; second is that the area of the brain with the highest density of plaques in Alzheimer’s is not the area with the highest dysfunction.

Nevertheless, the hypothesis is legitimate, if imperfect, he says. When APP is misprocessed, it is ultimately broken into multiple fragments, the final fragment of which is a-beta. It is becoming clear that keeping APP from breaking into these fragments is key. “There’s growing evidence that the intermediate fragments could be just as toxic,” Dr. Small says. In that case, if you clear only the amyloid plaques, “you’ve cleared the smoke, but the fire’s still burning.”

The burning fire appears to be in the endosome, an organelle within the cell that acts as the trafficking hub, directing proteins on their itineraries among cellular destinations. The two primary routes are toward the secretory pathway and the degradation pathway. Proteins can be recycled back to the cell surface, via the Golgi, where they are secreted outside the cell or sent to the lysosome—the garbage can of the cell—to be degraded.

Retromer is a key component of the endosome-to-Golgi pathway, packaging the cell’s cargo of proteins and lipids and thus playing a role in how proteins are sorted and transported to the cell surface. When Dr. Small first found retromer a decade ago, he says, “I knew absolutely nothing about it. But it was not that difficult to read everything because there wasn’t much to read.” Retromer had been discovered in yeast in 1998 by a Cambridge University scientist and confirmed in mammalian cells just a year before Dr. Small encountered it in 2004.

AD scientific illo

Alzheimer’s disease starts in the entorhinal cortex (shown in yellow in a mouse brain, left, and a human brain, right). Using fMRI, the researchers show how the disease spreads from the entorhinal cortex to other cortical regions: the perirhinal cortex (shown in red) and the posterior parietal cortex (shown in green). Image courtesy of Scott A. Small

At the time, Dr. Small’s research was focused on determining where Alzheimer’s begins. Using newly developed fMRI techniques to image the brains of living patients, he confirmed what was first suggested postmortem—that the entorhinal cortex is the site where the disease originates. “You probably could have intuited that from postmortem studies,” says Dr. Small, who in a December 2013 paper in Nature Neuroscience described more precisely where in the entorhinal cortex the damage begins, why it starts there, and how it spreads. “What turned out to be more important is identifying a neighboring area that was relatively unaffected by Alzheimer’s,” he says. What made that adjacent, unaffected area different were its higher levels of retromer.

Once Dr. Small had found those low levels of retromer in the area of the brain where Alzheimer’s originates, he hypothesized that when the protein complex malfunctions, it slows the movement of APP through the endosome—often enlarged in Alzheimer’s disease—where it is then broken down into the harmful amyloid-beta. “APP processing happens in the endosome, so the goal is to keep APP flowing quickly,” Dr. Small says. “When you have defects in the retromer-related proteins, APP remains longer than it should in the endosome—even milliseconds too long—and it’s there that it meets the enzyme that starts to cleave it.”

Dr. Small and a team of P&S collaborators took that insight to the lab, where they delved deeper. With Tae-Wan Kim, PhD, associate professor of pathology & cell biology (in the Taub Institute), he demonstrated retromer’s role in the APP-endosome process in cell culture. With Karen Duff, PhD, professor of pathology & cell biology (in psychiatry and in the Taub Institute), and Brian McCabe, PhD, assistant professor of pathology & cell biology and of neuroscience, he documented the pathway in animal models. Then, in what Dr. Small calls “the most powerful experiment, Mother Nature’s experiment,” Dr. Mayeux identified genetic variants in SORL1, which codes for the transport of APP through the cell by regulating retromer, in patients with Alzheimer’s disease. “I think the field now is comfortable in concluding that the retromer pathway plays a pathogenic role in late-onset Alzheimer’s,” says Dr. Small.

About a decade ago, Dr. Mayeux started collecting data from families throughout the United States who had multiple members with late-onset Alzheimer’s. Scientists had already identified the risk-factor genes for early onset forms of the disease, which account for just 1 percent of patients. “With a common disease, there could be more than one gene involved, so you have to have a large enough sample to detect all of the genes, even though they might have small effects,” says Dr. Mayeux, the Gertrude H. Sergievsky Professor of Neurology, Psychiatry, and Epidemiology (in the Sergievsky Center and the Taub Institute), chair of neurology, director of the Sergievsky Center, and co-director of the Taub Institute. Partnering with other institutions, he collected DNA data on 1,600 families and from the beginning gave any qualified scientist access to the data. Subsequently, the federal government asked Dr. Mayeux and others with data on families to devise a project, now known as the Alzheimer’s Disease Sequencing Project, to sequence a large number of individuals, with the hope that by identifying the genetic substructure of the disease, the scientists could start to see how DNA fits into the picture.

“The limitation of that approach is that you basically get a signal near the gene, but it may not tell you what’s wrong with the gene,” Dr. Mayeux says. A group of Taub scientists is now doing targeted sequencing of the 25 known genes in an attempt to identify mutations. “I can’t imagine that each of them is causal,” he says. “My guess is that they work in collaborative pathways, and so far it looks like there are probably three major pathways.” In addition to the genes associated with retromer function, another likely culprit is the pathway involved in a localized inflammatory process in the brain. A third has to do with lipid transport.

“It is thought that some of these mutations just don’t give you enough SORL protein for retromer to actually perform its task, but there must be a half dozen genes that are all part of that family, and they all act about the same way,” Dr. Mayeux says. “Why do you need a half dozen? No one knows yet, but it’s a very interesting story and once we start defining these things, we’ll have a much better idea of how to treat the condition.”

The work of the Taub researchers “seems to come together like a jigsaw,” Dr. Small says. “Basically, what it said to us was that if the endosome is the problem, let’s start thinking about developing drugs to fix it.” In the Nature Chemical Biology paper, Dr. Small and his co-authors showed that the pharmacological “chaperone” they had developed increases levels of retromer proteins, shifts APP away from endosomes, and decreases the transformation of APP into amyloid-beta.

The elucidation of protein-processing functions in Alzheimer’s disease has been important in refining the long-held assumption that amyloid-beta is a key driver of the disease. “It’s a known fact that people with early-onset mutations produce way too much amyloid,” says Dr. Mayeux. Stopping overproduction “is important in early-onset disease. It may also be important in late-onset disease, although in late-onset disease none of these genes produces too much amyloid. But they don’t clear it, and over time that can be just as bad.”

A new study on gene profiling at P&S will investigate the differences between early- and late-onset forms of the disease. Columbia is one of 18 centers around the world taking part in the Dominantly Inherited Alzheimer Network—DIAN—study to identify people with early-onset mutations who are unaffected. The team has just finished a study on biomarkers and has begun a clinical trial to vaccinate these patients or give them an infusion that prevents amyloid from depositing in their brain. Another study involves people who have evidence of disease based on PET scans that reveal amyloid deposition but who are asymptomatic; they will be randomized for the vaccine.

“We know, based on studies done here and elsewhere, that people have pathological changes in their brain starting as much as 10 to 15 years before they show any symptoms,” says Dr. Mayeux. In this way, Alzheimer’s disease is not unlike other chronic diseases, such as atherosclerosis, in which the process that leads to the stroke or cardiac event develops over an extended period of time. Late-onset Alzheimer’s doesn’t typically occur until a person reaches his or her 80s, so to catch the disease early and slow its progress by even a few years would be transformational, he says. “If you could give people another five to 10 years of good quality life, that would be a good outcome.”

How close are experts to being able to offer something like that to patients? Dr. Small and his colleagues believe they are quite close: “We are where we want to be now,” he says. Over the past 15 years, “the goal was to understand the disorder to the point where we could develop novel therapies. We have our novel therapies, and now we need to see if they work.”


Read about a study led by Dr. Small showing that dietary cocoa flavanols—naturally occurring bioactives found in cocoa—reversed age-related memory decline in healthy older adults.

Read the full P&S annual report, an edition of Columbia Medicine magazine, here.