The protein amyloid beta (A-beta) forms insoluble, sticky deposits in the brains of people with Alzheimer’s disease and was widely thought to be involved in triggering the disease process. But recent clinical trials have suggested that in Alzheimer’s patients, clearing amyloid deposits from the brain comes too late to do any good. By the time people show symptoms, severe brain damage has already occurred and multiple processes of neurodegeneration appear to be underway. Many researchers now suspect that any treatment at this stage is unlikely to stop the disease.
“The clinical trial folks are just now starting to understand that we need to treat people earlier than what we’ve been doing,” says Anne Fagan, an expert on Alzheimer’s diagnostic technology at Washington University in St. Louis, Mo.
Developing methods to predict Alzheimer’s well ahead of symptoms has become a major research priority. It will take years for this shift in scientific emphasis to bear fruit, but recent progress suggests that the pre-symptomatic prediction of Alzheimer’s could soon be routine.
Memory and cognition tests such as the Mini-Mental State Exam are already used to help diagnose Alzheimer’s. But some researchers think it may be possible to develop a memory test so sensitive that it predicts a disease-related mental decline years before it otherwise became evident.
At the Mayo Clinic in Rochester, Minn., neurologists are using a test known as the Free and Cued Selective Reminding Test (FCSRT) to gauge memory impairment in people suspected of having either Alzheimer’s or an earlier, more moderate condition known as mild cognitive ompairment (MCI). In the FCSRT, subjects view a set of images, and then after a brief interval must recall as many as possible. In a second pass, they view verbal cues to the images they haven’t yet recalled, and try to recall them, and so on for six passes. “The number of passes that it takes to get to complete recall is a very sensitive indicator of memory impairment,” says Michael Wenger, a memory researcher at Penn State.
With colleagues at the Mayo Clinic, Wenger recently reported a modification to the FCSRT that appears to make it even more sensitive. After recording subjects’ reaction times to different kinds of cues, Wenger fed these and other FCSRT variables into a sophisticated statistical program, which yielded a measure relating to the subjects’ memory capacity.
“What jumps out at you is the very large effect of adding the cues,” says Wenger. “So with the addition of one kind of cue you slow them down, and with the addition of another you speed them up.” The increased volatility of reaction times in response to this change suggests a reduced memory capacity and “distinguishes MCI individuals very cleanly,” says Wenger. Patients’ scores with this method also match more strongly with the amount of shrinkage of a disease-affected brain region known as the hippocampus, measured by high-resolution magnetic resonance imaging.
Wenger and his colleagues now hope to prove the technique’s usefulness in a large-scale prospective study of elderly patients.
Fluid and imaging biomarkers
Researchers hope that the earliest warning signs of the Alzheimer’s process will be detectable even before they can measure any impact on memory and cognition. Two candidate signs, or “biomarkers,” are the focus of much new research, says John Trojanowski at the University of Pennsylvania, a senior investigator on a recently completed, $60 million study known as the Alzheimer’s Disease Neuroimaging Initiative (ADNI).
As the Alzheimer’s process starts in a middle-aged or elderly person, says Trojanowski, “the first thing that seems to [appear] is A-beta amyloid in the brain as visualized by positron emission tomography (PET) imaging; and at the same time, and perhaps even earlier, A-beta levels measured in the cerebrospinal fluid (CSF) change. And this is while a person is still cognitively normal.”
A group at the University of Washington, including Fagan, found in a recent close study of one patient who eventually developed Alzheimer’s that the sharp change in his CSF A-beta levels came before the appearance of significant A-beta brain deposits on PET scans. CSF A-beta is “number one on our hit list right now,” says Fagan, who was lead author of a recent review article on biomarkers.
Fagan and Trojanowski say that it may take another year or two for ongoing and planned studies, including a follow-on to ADNI known as ADNI-2, to resolve the question of which of these biomarkers serves as the earliest warning.
Remarkably, the level of A-beta in CSF does not increase as A-beta is deposited in the brain; studies by Fagan’s group among others indicate that it sharply decreases. The effect might be due to the blocking of the normal diffusion of A-beta to the CSF, or to the impairment of mechanisms that normally keep A-beta in a soluble state, or both. “We suspect that somehow you reach this tipping poing where you overload the system,” says Fagan.
One of the other pathological hallmarks of Alzheimer’s is the presence within dying neurons of amyloid deposits known as neurofibrillary tangles, made up primarily of the protein tau. The CSF level of this form of tau turns out to be a relatively poor early warning signal for Alzheimer’s, because its rise tends to coincide with, rather than predict, the onset of dementia. But Fagan thinks that the combination of rising tau levels and falling A-beta levels in CSF could be used to predict dementia with some accuracy a year or two in advance of symptoms. “That really seems to mark the tipping point,” she says.
Untangling A-beta species
Intense interest also has developed recently in what are known as “oligomeric” forms of A-beta, a form in which a small number of individual A-beta proteins link up and are still soluble. There is some evidence that these forms, rather than the insoluble aggregates familiar to pathologists, are the true “toxic species” that impair synapses and kill neurons in Alzheimer’s.
If so, then CSF levels of these toxic, oligomeric species of A-beta could be the most directly relevant biomarkers of all. So far, no one has developed a test that reliably distinguishes oligomeric A-beta in CSF from single A-beta proteins, says Fagan. “But I happen to know several groups that are working very hard on this.”