In his magical-realist masterpiece One Hundred Years of Solitude, Colombian author Gabriel García Márquez takes the reader to the mythical jungle village of Macondo, where, in one oft-recounted scene, residents suffer from a disease that causes them to lose all memory. The malady erases “the name and notion of things and finally the identity of people.” The symptoms persist until a traveling gypsy turns up with a drink “of a gentle color” that returns them to health. In a 21st-century parallel to the townspeople of Macondo, a few hundred residents from Medellín, Colombia, and nearby coffee-growing areas have begun to assist in the search for something akin to a real-life version of the gypsy's concoction. Medellín and its environs are home to the world's largest contingent of individuals with a hereditary form of Alzheimer's disease. Members of 26 extended families, with more than 5,000 members, develop early-onset familial Alzheimer's, usually before the age of 50, if they harbor an aberrant version of a particular gene.

Familial Alzheimer's, passed down as a dominant genetic trait from only one parent, accounts for less than 1 percent of the more than 35 million cases of Alzheimer's and related dementias worldwide, but its hallmark brain lesions appear to be identical to those in the more common late-onset form of the disease, in which symptoms do not appear until after the age of 65.

The predictability of disease onset in the Medellín families has attracted the attention of a group of scientists and pharmaceutical companies who are considering a novel approach to research that will test drugs in patients before the first signs of dementia appear.

In recent years a number of drug candidates for treating mild or moderate Alzheimer's have failed, persuading researchers that much of the disease pathology—accretions of aberrant proteins and loss of brain cells or circuits—begins well before the memory loss becomes apparent. This growing realization, confirmed by new technologies that can track the disease years before the first symptoms, suggests that to be most successful, treatment must start during the many years when the insidious disease process is already under way, even though a patient's memory remains intact.

Consequently, a major thrust of much Alzheimer's research is shifting toward arresting the disease in advance of symptoms—not only with drugs but also with lifestyle measures that would be safer and less costly than filling a drug prescription for 10 or 20 years.

An Early Start
The Colombian Alzheimer's families stand in the vanguard of prevention research. Francisco Lopera, the neurologist who more than 30 years ago first came across the families who were later discovered to bear the paisa mutation (named after the moniker for the people of the region), has begun to contact hundreds of still healthy family members. He has recruited them to participate in a test of drugs that would remove or stop the buildup of toxic protein fragments, amyloid-beta peptides, that damage brain cells early in the disease process. “The contribution made by these families may shed a lot of light on the treatment and prevention of both early- and late-life Alzheimer's,” Lopera says.

In the $100-million clinical trial, which began treating patients in late 2013 and is part of a broader effort called the Alzheimer's Prevention Initiative (API), healthy, mutation-bearing family members around the age of 40 are receiving an antiamyloid drug called crenezumab, made by Genentech. Besides Genentech, the Banner Alzheimer's Institute in Phoenix is also playing a pivotal role—and the National Institutes of Health has made a major contribution to the funding. The trial in Colombia also complements a number of separate trials now starting up that are treating patients with drugs before symptoms begin.

The trial will evaluate whether a treatment can delay or stop the inexorable silent progression of the disease if administered seven years before the average age of diagnosis in family members who carry the gene. Beyond testing specific therapies, the designers of the Colombian trials also plan to see whether tracking of Alzheimer's-specific biomarkers can indicate whether an experimental treatment is working. (A biomarker is a measurable indicator—such as a concentration of a particular protein—that changes in concert with progression or regression of a disease.) A reliable set of biomarkers would allow drug researchers and clinicians caring for patients to evaluate the success of a therapy relatively quickly, by measuring changes in such silent benchmarks, instead of having to wait to assess overt symptoms. The API, under the aegis of Banner, is also undertaking a similar set of trials with a U.S.-based group made up of carriers of two copies of a gene variant, APOE e4, that increases susceptibility to Alzheimer's, although carriers are not guaranteed to get the disease.

The API would serve as a model for making biomarker-based Alzheimer's prevention trials commonplace. Proving that a drug prevents a disease takes much longer and costs much more than ascertaining whether it works in a patient who is already sick.

With a set of biomarkers in hand, a pharmaceutical company can test whether a drug changes levels of amyloid or another biomarker, in the same way that physicians test cholesterol levels as a gauge of whether a statin is helping to prevent heart disease. “We need to move presymptomatic treatments forward. Otherwise we could lose a whole generation,” says Eric M. Reiman, executive director of the Banner Alzheimer's Institute, who, along with colleague Pierre N. Tariot, launched the API.

The challenges of prevention trials still loom large: the drawbacks posed by inevitable drug side effects are more difficult to weigh against potential benefits in patients who do not yet have symptoms. Moreover, no one can predict whether a drug that proves helpful for familial Alzheimer's will also work in patients who lack the particular gene mutation that brings on the early-onset form of the disease. But the urgency of finding new treatments—and the lure of a multibillion-dollar drug—has given momentum to prevention strategies. Given the wave after wave of drug failures, pharmaceutical companies have started to realize that clinical trials before symptoms appear may be the only way to find effective therapies.

Some treatments for Alzheimer's do exist, but they do little to delay disease progression. A true disease-modifying therapy would meet with overwhelming patient demand. Statisticians predict that by the middle of the century, the global prevalence of Alzheimer's will nearly quadruple, reaching 107 million. A treatment that delays disease onset by even five years would halve the number of people who die from the disease.

Inside Your Head
An Alzheimer's prevention trial based on biomarkers, dismissed as a fantasy a decade ago, may come to fruition because imaging and other technologies, now flourishing worldwide, can track biomarkers to reveal the nature of the underlying disease process. In the U.S. since 2004, the Alzheimer's Disease Neuroimaging Initiative (ADNI), a collaboration among pharmaceutical companies, academics and the National Institutes of Health, has been developing methods to better assess the effectiveness of drugs tested in individuals suffering from the disease, a goal that quickly expanded to look at what is happening during the time before an actual diagnosis is made.

One intriguing report of progress in the field came in 2010, when Clifford R. Jack, head of the group within ADNI studying biomarkers that can be detected with magnetic resonance imaging, described a model of how the disease likely progresses and paired it with biomarkers that seem able to track this pathology. Jack presented his work, which also appeared in a technical paper, to an online audience of more than 100 people during a Web seminar on Alzforum, a gathering that included many leading researchers in the field. Co-founded by June Kinoshita, a former Scientific American editor, Alzforum is a meeting place for the exchange of ideas, a repository of research information and a source of perhaps the most in-depth journalism anywhere on Alzheimer's research.

At the Web seminar, Jack noted that the biomarker measurements have demonstrated that the disease process begins years before the defining symptoms that allow a diagnosis to be made. During this time (estimated to range from five to 20 years), a particular type of amyloid peptide begins to aggregate outside of brain cells and damage synapses, the contact points between neurons. A radioactive tracer molecule, such as Pittsburgh imaging compound B (PIB), can bind to amyloid in a patient's brain and then be imaged using positron-emission tomography (PET). The imaging technique, abbreviated PIB-PET, has shown that this aggregation process starts to level off before definitive symptoms arise.

Later on, but also before a diagnosis, a class of proteins called tau, which normally assist in providing structural support to neurons, become detached from the cells' scaffolding and clump into tangles, wreaking havoc inside the cells. The tau buildup can be detected by examining a sample of cerebrospinal fluid. This test can also look for decreasing levels of amyloid-beta, which occur as the peptides get removed from the fluid to form deposits in the brain. Together, decreasing levels of amyloid-beta and an increase in tau in the cerebrospinal fluid give a strong signal that the disease process is advancing.

Anywhere from one to four years before a person is diagnosed with Alzheimer's, a phase called mild cognitive impairment sets in. It is characterized by symptoms that range from memory lapses to poor decision making. Mild cognitive impairment can arise from causes other than Alzheimer's, but in those who are on the road to Alzheimer's dementia, mild cognitive impairment occurs because neurons in certain brain areas are damaged or dying—a loss that accelerates over time. (If memory problems are the primary symptom, the patient often progresses to Alzheimer's.) This stage can be tracked with a form of imaging called volumetric MRI, which measures shrinkage of the brain as neurons expire. The cascade of events, including the early accretion of amyloid, disrupts cell metabolism and can be monitored with a form of PET, fluorodeoxyglucose-PET (FDG-PET), that gauges the metabolic status of neurons.

But Does the Patient Get Better?
Using biomarkers as the basis of clinical trials for prevention poses a set of challenges to both pharmaceutical companies and regulators—and constitutes a barrier to moving ahead with the API and other prevention efforts. To be approved, an Alzheimer's drug needs to show that it provides cognitive benefits for the patient (in memory, language or a related measure) better than a placebo does.

If a biomarker is tracked instead of symptoms in a prevention study, researchers need to be sure that the measurements truly presage whether a subject is likely to develop dementia. For instance, investigators do not yet know whether changing amyloid-beta levels will ultimately prevent dementia, despite the large body of evidence suggesting that amyloid-beta contributes to disease development.

In fact, in one early trial of an amyloid therapy, the levels of the peptide decreased in some patients, but there was almost no evidence that cognition improved. “We're concerned that we might have a drug that affects a marker in the way that we predict but that it doesn't affect patients' clinical picture,” says Russell Katz, former director of the division of neurology products for the Food and Drug Administration. “In other words, their disease continues to progress, and they don't get any better.” Katz says a better approach to incorporating biomarkers in clinical trials would be to first show that reducing levels of amyloid or another biomarker benefits patients who have mild cognitive impairment or who are newly diagnosed with Alzheimer's and to attempt to use biomarkers in people without symptoms only afterward. “The best way to get there in my opinion is to start with patients who have symptoms, maybe very, very early on, and then work backward,” Katz says.

But the researchers in the Colombian prevention trials assert that they may already be capable of using biomarkers to detect subtle changes in memory, thereby allaying Katz's concern. And Reiman cites work from his group that might offer another way to help ease regulator concerns. In that study, carriers of the APOE e4 gene variant showed a small decline in scores for memory on psychological tests many years before any cognitive deficit became noticeable. This level of sensitivity, Reiman says, means that applying a cognitive test along with a biomarker measure in a prevention trial might suffice to indicate, say, whether prospects for avoiding dementia really do improve as amyloid levels drop.

Some companies are already trying to gain a better understanding of how to use biomarkers. The ability to target only the patients who are more likely to get a diagnosis of Alzheimer's could facilitate assessment of whether the drug really works, results that would be less clear if the trial had included participants who have little chance of getting the disease. As prevention research grows in sophistication, interventions could begin even earlier, at the first appearance of the pathological proteins that signal the start of the relentless progression toward an eventual diagnosis of dementia.