A group of Australian researchers has identified a set of plasma proteins that could help predict patient levels of the Alzheimer's disease-associated molecule beta-amyloid.
The protein panel, which was presented in a study published this week in Molecular Psychiatry, could prove useful in Alzheimer's research and drug development as a simple and inexpensive test for screening large numbers of individuals for high beta-amyloid levels, Samantha Burnham, a researcher at Australia's Commonwealth Scientific and Industrial Research Organisation and an author on the paper, told ProteoMonitor.
With roughly 5.2 million people in the US suffering from Alzheimer's currently and as many as 7.1 million predicted to have the disease in 2025, Alzheimer's has become a major area of focus for the pharmaceutical industry. It has also become a major area of frustration, with a number of large pharma firms like Eli Lilly, Pfizer, and Johnson & Johnson halting Alzheimer's drug development programs after disappointing clinical trial results.
A primary theory that has emerged regarding such failures is that the patient cohorts used in these trials were too far along in their disease progression for treatment to be effective, and, therefore, subjects must be identified earlier.
Alzheimer's patients "generally get picked up when they start having memory problems or other cognitive issues," Burnham said. "And by the time they have those issues they have already had some atrophy of the brain. As a field we don't know how we can actually replenish the brain cells, so it is going to be necessary to actually identify people before they actually have these changes to the brain."
One potential way of doing this is by measuring beta-amyloid levels, which can be elevated in Alzheimer's patients years in advance of the presentation of clinical symptoms.
Currently, researchers and clinicians measure these levels primarily via imaging techniques such as positron emission tomography or analysis of cerebrospinal fluid levels. These methods, however, have certain drawbacks: PET imaging, for instance, is relatively expensive and not widely available, while CSF analysis requires a lumbar puncture for sample collection.
This creates the need for a simpler, less expensive method – the plasma protein panel developed by Burnham and her colleagues, for instance – of measuring beta-amyloid levels. Such a test, Burnham noted, could be used as "a frontline population screening tool that could be followed up with confirmatory tests like brain scans or cognitive testing."
The researchers developed their protein panel using measurements of 176 plasma proteins and PET imaging data from 273 subjects from the Australian Imaging, Biomarkers and Lifestyle study. In cross-validation experiments in that cohort they found that a model using the six proteins — Ab1-42, chemokine ligand 13, IgM-1, IL-17, PPY, and VCAM-1 — along with patient age and APOE genotype predicted beta-amyloid status with sensitivity of 80 percent and specificity of 82 percent.
Using this test to analyze an independent cohort of 82 imaged patients from the Alzheimer's Disease Neuroimaging Initiative, they were able to predict beta-amyloid status with sensitivity of 79 percent and specificity of 76 percent.
The researchers also performed what Burnham called a "pseudo-validation" in 817 non-imaged subjects from the AIBL study, testing their model to see if it predicted the expected percentage of patients with high beta-amyloid burdens.
"In line with the literature, the model predicted all of [Alzheimer's patients], 87 percent of [mild cognitive impairment patients] and around 35 percent of [healthy controls]... to have high" beta-amyloid levels, the authors wrote. They are now going back and collecting imaging data on these patients, Burnham said, adding that they expect to have this data in hand within a year to 18 months.
The researchers have patented their panel and are currently in discussions with several pharmaceutical firms about using it for recruitment in clinical trials, she said. "It could be a frontline option [for enriching clinical trial cohorts] that could then be followed up with a brain scan or CSF test." She said, however, that the group had no concrete commercialization plans for the test at the moment.
In addition to convenience and accessibility, low cost would be a key advantage of such a test, Burnham noted. In Australia, for instance, PET imaging costs roughly A$1,500 (US$1,537) per sample compared to a likely cost in the range of "dollars per test" for the plasma protein assay, she said.
In an email to ProteoMonitor, Holly Soares, director of clinical neuroscience biomarkers at Bristol-Myers Squibb, said that if larger validation of the results of the Molecular Psychiatry study proved successful, "the findings would be highly significant, and there would be increased use of these types of blood tests to enrich for [Alzheimer's] and to serve as screening tools for more definitive testing."
However, cautioned Soares, who was not involved in the study but is involved in similar research into Alzheimer's plasma protein markers through her work with ADNI (PM 12/10/2010), the panel is still in the early stages, and the ADNI validation set used was "quite small." She added that while the addition of the blood markers appeared to improve "the ability to enrich for [beta-amyloid] in this study... caution should be applied as it is likely the models are over-fitting the data."
"It's a promising start," she said. But "the results will have to be reproduced more widely with validated blood assays."