In a position paper published this week in the online edition of The Lancet Neurology, a group of Alzheimer's disease experts proposed new definitions for the disease that would in part rely on the detection of various protein biomarkers.
The paper, submitted by the International Working Group for New Research Criteria for the Diagnosis of Alzheimer's Disease, marks an effort to update the definition of Alzheimer's for the first time since 1984 and incorporates findings from recent work on the use of protein biomarkers in cerebrospinal fluid to develop diagnostic and prognostic assays for the disease.
Traditionally, the diagnosis of Alzheimer's has required the presence of a clinical phenotype involving the presence of progressive dementia and specific neuropathological change usually accompanied by synaptic loss and vascular amyloid deposits.
Because, however, neuropathology cannot be examined until after death, Alzheimers' diagnoses in practice are typically probabilistic and based primarily on clinical symptoms. Research into biomarkers tied to the disease now suggests a more precise definition may be possible.
"Because the current diagnostic criteria were proposed in 1984, before we knew about the existence of biomarkers, at that time the diagnosis was based on very vague criteria," Bruno Dubois, professor of neurology at the Neurological Institute of the University Salpetriere Hospital and an author on the paper, told ProteoMonitor. "What [we] propose to now do is to clean up this very old concept of Alzheimer's disease."
The proposal expands upon a 2007 paper by the International Working Group suggesting the adoption of a new definition of Alzheimer's incorporating in vivo biological evidence of the disease. Whether all or any of its guidelines will be adopted by the research and clinical community remains to be seen — and, in fact, a number of the biomarkers included in the definition have yet to be clinically validated — but it highlights the growing importance of protein biomarkers in Alzheimer's research and therapies.
The new definition would allow for the diagnosis of Alzheimer's based on the presence of clinical symptoms, including hippocampal memory loss, as well as the presence of biomarkers, including the CSF protein biomarkers amyloid-beta, total tau protein, and phospho-tau.
This definition, Dubois noted, would allow physicians to diagnosis Alzheimer's before patients begin showing signs of dementia.
"We should no longer wait until the patient is demented," he said. "Because the biomarker can identify the specific lesion and signature of the disease, we can now make the diagnosis with the first symptoms because there is a specific pattern of memory dysfunction plus a specific [biomarker] signature."
Such an expanded definition could have important implications for the development of therapeutics for the disease, Dubois suggested.
"[M]ost of the [drug] studies that are negative today may be negative because the patients that are included in the trials are too late in the course of the disease," he said. "So this opens the window to include patients at the prodromal stage — they are clinically positive but not demented."
Indeed, several large pharmaceutical firms have recently reported disappointing results for Alzhiemer's therapies. In August, Eli Lilly halted late-stage development of a compound called semagacestat after studies showed it actually made patients worse and was associated with an increased risk of skin cancer. Prior to that, in March, Pfizer announced that its Alzheimer's drug Dimebon appeared to be no better than a placebo at treating the disease.
However, characterizing patients with biomarkers in order to detect the disease earlier "may help in determining drug response," Lon Schneider, director of the University of Southern California's State of California Alzheimer's Disease Research and Clinical Center, told ProteoMonitor. "The drugs may work earlier rather than later on. If you use them too late when there may be a substantial amyloid burden, there may be too much substrate for the drugs to work."
Such a redefinition could also be significant from a regulatory perspective, Schneider noted. "[Drugmakers] need to be able to label the conditions of use that drug will be used in. So it's possible that the definition of Alzheimer's for some future drug could be cognitive impairment with episodic memory impairment progressive over six months and a low CSF amyloid-beta," he said. "That would mean that you have to do a [lumbar puncture] to prescribe the drug, and if the [CSF amyloid-beta] level isn't low then you can't prescribe it."
In a comment included with the Dubois paper, Schneider raised questions, however, about whether enough research has been done on Alzheimer's biomarkers to make them part of a formal definition of the disease. He noted that, while refining and expanding the diagnosis of Alzheimer's through the use of biomarkers is an appealing idea, biomarkers like CSF amyloid-beta, total tau protein, and phospho-tau have yet to be clinically validated.
"Basically what I'm saying is, 'We're not there yet,'" he said. "We don't have longitudinal work over some 20 years in order to be able to truly map out the course of these biomarkers."
[ pagebreak ]
"Everyone agrees now that biomarkers are very important in making diagnoses for Alzheimer's disease," Neil Buckholtz, chief of dementias of aging at the National Institute on Aging, told ProteoMonitor. However, like Schneider, he said more work was needed to incorporate them into a clinical definition — particularly with regard to earlier stages of the disease.
"We don't think [the markers cited in the paper] are ready to go into the clinic yet. We feel they're more for research purposes at the current time," he said. "As we get more information on the biomarkers and the relationships of the biomarker change to clinical change, that may change."
The inclusion of the CSF protein biomarkers amyloid-beta, total tau protein, and phospho-tau in the proposed definition is based on a number of biomarker studies performed over the last decade, including work led by Les Shaw and John Trojanowski, researchers at the University of Pennsylvania School of Medicine and co-directors of the National Institutes of Health's Alzheimer's Disease Neuroimaging Initiative Biomarker Core.
Initial studies by these researchers determined that a pathological CSF biomarker signature of Alzheimer's could be defined by the combination of amyloid-beta1-42 and total tau protein levels. Using the signature, they were able to predict roughly 90 percent of the time which patients would progress from mild cognitive impairment to Alzheimer's (06/11/2010).
There has been difficulty in reproducing CSF protein biomarker measurements across laboratories, however, and work to clinically validate these biomarkers is still ongoing.
"None of these are [US Food & Drug Administration] approved as surrogate markers for clinical trials nor for routine clinical diagnoses," Trojanowski told ProteoMonitor this week.
Alzheimer's biomarker work, though, "is all converging on the kind of definitions and nomenclature they're proposing here," he said. "This is an exciting position paper presenting an exciting lexicon that could indeed be the lexicon we use two, three, five years from now."