Issue Table of Contents
New Evidence for an AD Fingerprint in the Blood
Neurology Today
20 November 2007;
Volume 7(22);
p 31-32
Schuster, Larry
Outline
Links
ARTICLE IN BRIEF
✓ A panel of 18 proteins distinguished between AD and non-AD patients in 259 archived plasma samples from individuals with presymptomatic to late-stage AD and from controls.
Preliminary evidence of an Alzheimer disease (AD) fingerprint in the blood - a panel of 18 proteins that distinguished between AD and non-AD patients - suggests that a blood test for the diagnosis of AD might be possible, according to a new study.
The team, led by Tony Wyss-Coray, PhD, associate professor of neurology and neurological sciences at Stanford University, measured 120 known signaling proteins in 259 archived plasma samples from individuals with presymptomatic to late-stage AD and from controls.
Dr. Wyss-Coray explained that signaling proteins are soluable protein factors secreted by cells - that are carried in body fluids - and bind to specific receptors on cells where they trigger a cascade of reactions.
The investigators identified 18 with highly significant differences in expression in AD and non-demented control samples, the authors wrote in an Oct. 14 online edition of Nature Medicine.
Some of the 18 proteins are angiopoietin-2, granulocyte-colony stimulating factor, glial-derived neurotrophic factor, intercellular adhesion molecule-1, and interleukin.
PROTOCOLS AND RESULTS
Using those signaling proteins to classify AD or non-AD patients in a blinded set of 92 samples, investigators found the protein set agreed with clinical diagnosis in 89 percent of the cases (p< 0.001).
Figure. Dr. Tony Wyss-Coray: Some people have confused this as an assessment of risk for Alzheimer disease. That's not what it is. It detects the disease much earlier than it is usually diagnosed in the clinic.
For 47 patients with mild cognitive impairment (MCI), the biomarkers were 81 percent accurate in making an AD or non-AD diagnosis two to six years after initial MCI diagnosis (p< 0.001). The protein set also correctly classified as not Alzheimer disease eight of the 47 cases that progressed to other dementias such as frontotemporal dementia; three of the eight also had corticobasal degeneration.
Our data indicate that a highly specific plasma biomarker phenotype can characterize Alzheimer disease years before a clinical diagnosis can be made, the researchers wrote.
Dr. Wyss-Coray noted the current panel of signaling proteins is not a predictive tool. Some people have confused this as an assessment of risk for Alzheimer disease, he said. That's not what it is. It detects the disease much earlier than it is usually diagnosed in the clinic.
EXPERTS COMMENT
Although the numbers were small, the results have grabbed the attention of biomarker scientists and AD experts.
Figure. Dr. Steven T. DeKosky: This would put the diagnosis in the hands of the people who see most of the patients with dementia, the generalists.
I was surprised because, so far, plasma has been a dead end in the search for biomarkers, said Anne M. Fagan, PhD, research associate professor in the department of neurology at Washington University School of Medicine in Saint Louis. Dr. Fagan was first author on a 2006 paper suggesting that amyloid imaging and CSF amyloid-beta42 may potentially serve as biomarkers of presymptomatic AD (Ann Neurol 2006;59:512-519).
Dr. Fagan told Neurology Today: For me, the biggest impact (of this paper) is that there is any sort of signature that can be detected in blood and can discriminate clinical groups. I think that's amazing given that AD is a disease of the brain. That's wonderful.
Norman R. Relkin, MD, PhD, associate professor of clinical neurology and neuroscience at the Weill-Cornell Medical College and director of the Cornell Memory Disorders Program, said: The Stanford group has used an intelligent design and elegant techniques to identify their signaling protein panel.
In February, Dr. Relkin and colleagues reported promising results with a multiplexed panel of 23 CSF markers for AD (Ann Neurology 2007;61:120-129). [The study had appeared in an advance online publication in December 2006.]
Ideally, we need biomarkers that will predict disease development in any individual, regardless of whether they have MCI or are otherwise symptomatic at the time of testing, Dr. Relkin said. I can't say whether this test or any other will succeed in doing so.
The Stanford group so far has not attempted to study the biomarkers in presymptomatic, cognitively normal patients. That would be one of the many elements of future research in the development of a community-based test, Dr. Wyss-Coray said.
That would be a future goal …to make this a screening test, he added. But ideally that's what you will be able to do in a few years from now. Dr. Wyss-Coray is a co-founder of Satoris Inc., a Redwood City, CA, biotech company, which hopes to commercialize a blood test based on this research.
Figure. Dr. Anne M. Fagan: For me, the biggest impact (of this paper) is that there is any sort of signature that can be detected in blood and can discriminate clinical groups. I think that's amazing given that AD is a disease of the brain. That's wonderful.
AREAS FOR FURTHER RESEARCH
The possibility of a screening test, he said, depends on the ability to collect samples from a large population-based study, with perhaps 5,000 individuals followed over time.
Among the questions a study would have to answer is whether it could distinguish between AD and other dementias and whether any number of confounding factors would interfere with test results. And most important, researchers will want to know how well it would function in a community setting in identifying presymptomatic patients, where it would have the most value.
We don't know to what extent diet, medications, or comorbidity could affect the profile of biomarkers, study co-author Douglas R. Galasko, MD, professor in residence in the department of neurosciences at the University of California-San Diego, told Neurology Today.
The fact that samples came from a variety of medical research centers and were not drawn under fasting conditions may indicate that the plasma profile is likely to be robust even in the presence of potential complicating factors, Dr. Galasko said.
And researchers also want to know why agreement of the biomarker predictors with clinical diagnosis was not even higher. Some of the patients could have been erroneously diagnosed by the clinician, or they might have a subtype of AD that is not detected by the test, or the test may need further refinement, Dr. Wyss-Coray said.
If confirmed, though, the technology could be the basis of a cheap, efficient screening test to be used by non-specialists, researchers said.
This would put the diagnosis in the hands of the people who see most of the patients with dementia, the generalists, said Steven T. DeKosky, MD, director of the Alzheimer's Disease Research Center at the University of Pittsburgh.
It could be used as a public health measure for screening people, and would help doctors decide who to put on therapy earlier, while they are still cognitively normal, said Dr. DeKosky, who was not part of the study.
Will it hold up to be that accurate in the population? Dr. DeKosky asked. That's what you'd need to prove.
The most validated AD biomarker test to date involves CSF tau and amyloid-beta42. Dr. DeKosky said the tau and amyloid-beta42 tests are not experimental, and he offers them to some patients to provide biochemical evidence of AD in cases where there is some doubt.
If you have a low amyloid and high tau, there's a 90 percent chance the subject will develop AD within five years, he said.
Figure. Dr. Samuel Gandy: The new work is only surprising in that we now have the detection power to find signals in the plasma… This suggests that Alzheimer pathology may be more a 'firestorm' of signaling pathology than previously realized.
Even so, Dr. DeKosky said, You're not going to do lumbar punctures on everybody over 65. That is, unless a CSF biomarker test became the only accurate test to show which cognitively normal people might benefit from effective preventive medications.
And it is almost certain that amyloid imaging would be able to do this, although that is more expensive and less convenient, he added. There is no question that people at risk would have a scan, if there were a prevention medication to take!
As for the signaling proteins themselves and what they say about the disease, the authors wrote that their analysis pointed to an overall reduction in factors associated with hematopoiesis and inflammation during AD, as well as to deficits in neuroprotection, neurotrophic activity, phagocytosis, and energy homeostasis.
By identifying a panel of signaling proteins, this study suggests that interactions between neural and immune cells, some of which appear to travel from the bloodstream to the brain, may be important in mounting a response to AD pathology, Dr. Galasko told Neurology Today. The next question is whether elements of this response are beneficial, damaging, or neutral.
One curious lead in the panel of 18 signaling proteins, said Dr. Wyss-Coray, is granulocyte colony-stimulating factor, a hematopoietic growth factor, already used clinically to treat neutropenia. In a paper earlier this year (J Exp Med 2007;204:1273-1280) by a different group, researchers reported evidence the protein stimulated neurogenesis and improved memory in two mouse models of AD, Dr. Wyss-Coray noted.
Samuel Gandy, MD, PhD, professor of neurology and psychiatry at the Mount Sinai School of Medicine and the associate director of the Mount Sinai Alzheimer's Disease Research Center, told Neurology Today that the topic of signaling proteins goes back more than 20 years, when he was at Rockefeller University.
Dr. Gandy, who is also chair of the National Medical and Scientific Advisory Council of the Alzheimer's Association, said: Our group made the first discovery that signaling proteins controlled generation of amyloid. The new work is only surprising in that we now have the detection power to find signals in the plasma… This suggests that Alzheimer pathology may be more a 'firestorm' of signaling pathology than previously realized.
