Protein 'fingerprint' in spinal fluid could spot Alzheimer's disease

RSS Feeds

Right now, physicians rely on their clinical judgment to decide whether a particular patient has Alzheimer's rather than some other form of dementia. In many cases, the diagnosis remains uncertain until brain tissue is examined at autopsy.

"Our study is the first to use sophisticated proteomic methods to hone in on a group of cerebrospinal fluid biomarkers that are specific to autopsy-proven Alzheimer's disease. Those postmortem tests confirmed that the panel is over 90 percent sensitive in identifying people with Alzheimer's disease," says Kelvin Lee, the Samuel C. and Nancy M. Fleming Professor of Molecular and Cell Biology and associate professor of chemical and biomolecular engineering at Cornell.

Researchers at a variety of centers have long sought biomarkers in blood or cerebrospinal fluid that identify the presence of Alzheimer's pathology and distinguish it from other conditions that cause dementia.

"Some of these studies have met with limited success, but most have correlated their findings with patient's clinical symptoms rather than working with the gold-standard of autopsy-proven Alzheimer's," notes Norman Relkin, M.D., associate professor of clinical neurology and neuroscience at Weill Cornell and director of the Memory Disorders Program at NewYork-Presbyterian Hospital/Weill Cornell Medical Center.

Erin Finehout, Ph.D. '05, the lead author on the research who had been a doctoral student in Lee's laboratory, said this has great potential to impact human health. "Typically, Alzheimer's disease is not diagnosed until the disease has already caused some amount of dementia," she said. "Having a chemical test available may allow patients to be diagnosed earlier in the course of the disease."

The Cornell study combined cutting edge "proteomics" technology, detailed image analysis, and complex computational and statistical analyses to simultaneously compare 2,000 cerebrospinal fluid proteins from 34 patients with autopsy-proven Alzheimer's disease to those of 34 age-matched controls without the disease.

"Just as the human genome reflects the array of genes a person possesses, the 'proteome' is the vast collection of proteins expressed by those genes," said Lee. "Essentially we used high-tech methods to contrast the proteomes of Alzheimer's patients against those of a control cohort that included people with other forms of dementia as well as healthy individuals, looking for key differences between the two groups."

This effort yielded intriguing results: 23 proteins that individually might not point to Alzheimer's but together formed an identifying pattern or "fingerprint" specific to the illness.

"Although it need not have turned out that way, several of the 23 markers that emerged from this analysis proved to be proteins with known links to the pathological mechanisms of Alzheimer's disease," said Relkin.

For example, some of the biomarkers are associated with proteins that clog the brains of Alzheimer's patients. Others molecules were linked to inflammation, also a part of Alzheimer's brain pathology. Still other proteins in the panel were linked to synaptic dysfunction - the breakdown of communication between brain cells that occurs as Alzheimer's disease progresses.

"A subsequent validation group of 10 patients with suspected Alzheimer's and 18 healthy and demented control subjects turned up similar results," said Relkin. "Based on their clinical symptoms, we found the new screen to have 93 percent sensitivity to probable cases of Alzheimer's and a 90 percent accuracy in avoiding false diagnoses."

Despite their excitement over the new findings, the researchers stress that the results still need to be replicated in much larger populations.

"There's a federally funded, multicenter trial going on right now that involves many more patients, in which many enrollees will undergo spinal taps and brain imaging," Relkin said. "We hope to link into that effort to expand on these findings."

The dream of a highly accurate Alzheimer's disease test that patients and doctors could use to diagnose illness and guide treatment is the ultimate goal. But the finding has important implications for current Alzheimer's research, as well.

"You might have a promising treatment for the disease, but how can you know for sure that it's impacting on the underlying pathology, rather than just easing outward symptoms as most of the drugs that we have now do"" Relkin said.

"We are hopeful that by monitoring changes in these cerebrospinal biomarkers, we can actually track the effectiveness - or lack thereof - of experimental drugs."

He continued: "In fact, we are now using this panel to study the effects of a promising new experimental treatment for Alzheimer's called IVIg (intravenous immunoglobulin). "Data gleaned from the use of this 23-protein screen suggests that IVIg may be having a positive impact on underlying disease processes."

And while a standard test for spotting early Alzheimer's disease is not in doctors' offices today, it could be in the not-too-distant future.

"These findings speak to the practicality of using biological markers for discerning whether symptoms are reflective of Alzheimer's, another dementia or normal aging," Relkin said. "And in this study, we're offering a much more sophisticated and better-validated approach than has ever been presented before."

Cornell University News Service