Clinical Laboratory Strategies: November 8, 2007

Finding a Biomarker for Alzheimer’s Disease
Research Points to Differences in Concentrations of Signaling Proteins
By Phil Kibak


According to the Alzheimer’s Association, someone develops Alzheimer’s disease every 72 seconds. Currently, a definitive diagnosis is only possible at autopsy; therefore, new diagnostic tools to accurately detect Alzheimer’s disease as early as possible are vitally important. Reliable diagnostic tools would not only enable those affected by the disease to receive the most appropriate therapies, it also would permit physicians investigating new treatments to differentiate between people expected to develop the disease and those who will not. This issue of Strategies examines the results of a recent study that identifies changes in the concentrations of a group of signaling proteins as a potential blood test for Alzheimer’s disease.

Diagnosing Alzheimer’s disease, an irreversible and incurable form of dementia that affects an estimated 5 million older Americans, relies on the results of neuropsychological tests that measure memory, language skills, and other brain functions, and by excluding other causes of dementia. Diagnosing the disease in its initial stages would provide patients early access to existing treatments that may improve symptoms, but finding a reliable disease marker has proven very difficult. Now, an international team of scientists led by Tony Wyss-Coray, PhD, associate professor of neurology and neurological sciences at Stanford University School of Medicine (Palo Alto, Calif.), has discovered that plasma concentrations of a number of signaling proteins differ considerably between people with Alzheimer’s disease and healthy people who do not have dementia. Their study, which made national headlines, was published October 14 in an advance online issue of Nature Medicine (Nature advance online publication 14 October 2007 (DOI 10.1038/nm1653)).

Finding an Alzheimer’s-specific Signature

The brain controls many body functions via the release of signaling proteins. Because central and peripheral immune and inflammatory mechanisms are increasingly implicated in Alzheimer’s and related diseases, Wyss-Coray and his colleagues hypothesized that the pathological processes leading to Alzheimer’s would cause changes in the concentrations of signaling proteins in the blood, generating a detectable disease-specific molecular phenotype.

“By ‘listening’ to the chatter of these different proteins, we’re measuring whether something is going wrong in the cells,” said Wyss-Coray. “It’s not that the cells are using new ‘words’ when something goes wrong. It’s just that the some ‘words’ are much stronger and some are much weaker. The chatter has a different tone.”

The scientists collected 259 archived blood samples from U.S., Swedish, Polish, and Italian individuals who had symptoms ranging from nothing abnormal to mild cognitive impairment to advanced dementia. Using an algorithm called predictive analysis of microarrays (PAM), they identified 18 of 120 proteins that correlated with the presence of Alzheimer’s disease. In a subset of 92 people with symptoms ranging from no impairment to full dementia, the protein analysis matched the clinical diagnosis for 90% of the individuals. The researchers also used PAM to test blood samples collected between 2 and 6 years earlier from 47 people with mild cognitive impairment to see if alterations in these proteins could “predict” the appearance of the disease in individuals later diagnosed with it. On clinical follow-up, 22 of these people were found to have Alzheimer’s disease, eight had some other form of dementia, and 17 remained mildly impaired. The scientists observed that the PAM diagnosis of Alzheimer’s disease agreed with the clinical diagnosis in 91% of cases.

“Our analysis identified two independent regulatory networks connecting the 18 signaling proteins,” Wyss-Coray and his colleagues reported. “One network centered on tumor necrosis factor-α and monocyte-colony stimulating factor, whereas the other centered on epidermal growth factor. Consistent with these findings, gene ontology (Kyoto Encyclopedia of Genes and Genomes, www.genome.jp/kegg) and BioCarta pathway analysis (www.biocarta.com) indicated involvement of the 18 markers in immune response, hematopoiesis, and apoptosis.” Previous studies have suggested that hematopoietic cells may enter the brain and help modulate the disease. Dysfunction of pathways involved in programmed cell death (apoptosis) also has been linked with the disease.

Will It Be Applicable?

“The results of the study are very interesting,” said William Thies, PhD, Vice President for Medical and Scientific Relations with the Alzheimer’s Association. “However, it’s pretty early on in the development of this assay, and like any preliminary study, it was done on a limited population. The real test of this or any other potential biomarker is how well it performs when it’s tried on a broader population.”

Many efforts are being made to find a chemical test to diagnose Alzheimer’s disease. But, added Thies, even more important is the development of a test to allow clinicians and researchers to track the impact of disease-modifying medications. “This test is in that category,” he says. “It’s different from other tests because it’s not specifically related to what we know about the pathology of Alzheimer’s disease – it’s an observation about a series of proteins and there’s incomplete knowledge about what these proteins do in Alzheimer’s or how they’re involved in the disease process. But they clearly change, so the question is: Is this marker going to be useful for tracking the progression of the disease and tracking the impact of therapies?”

A problem in testing disease-modifying therapies for Alzheimer’s is that the illness and its symptoms progress more rapidly in some people than in others. From a clinical perspective, a test that could identify such individuals would be useful because researchers could then selectively place these rapid-progressers in clinical trials to better see the effects of a target medication.

“There’s no way to tell when a commercially available assay will be available,” noted Thies. “It would be much easier to show the value of a biomarker in the presence of an effective therapy. Conversely, it’s also easier to show the value of a therapy in the presence of an effective biomarker, so how quickly one emerges is partially dependent on the other.”

Wyss-Coray and his colleagues focused on signaling proteins rather than the entire plasma proteome. And this concept of proteomic analysis will be instrumental in the development of biomarkers. “We’ve gone through a long period where we’ve unraveled the human genome and that’s useful, but it’s like getting a phone book that has all the numbers but not the names or addresses,” noted Thies. “We not only have to get the genome, we have to learn what proteins the genes make and we have to learn what those proteins do. These proteomic analyses are the next step and the scientists at Stanford deserve credit for developing a functional test out of this theory.”

This research was supported in part by Satoris, Inc (San Francisco, Calif.). Wyss-Coray is a paid consultant for Satoris and a founder of the company, which hopes to commercialize the assay.

Was this issue of Strategies of interest to you? Please help us improve by rating this article.

Page Access: