NEW YORK (GenomeWeb) – A team led by researchers from King's College London and Proteome Sciences has identified a panel of 10 protein biomarkers that could help predict patients likely to progress from mild cognitive impairment to Alzheimer's disease.
The panel, which was described in a paper published this week in Alzheimer's & Dementia, could prove particularly useful for helping drug companies construct clinical trials for the disease, Ian Pike, Proteome Sciences' COO and an author on the study, told ProteoMonitor.
The researchers identified the panel by measuring levels of 26 candidate proteins in 1,148 subjects drawn from three multicenter cohorts composed of cognitively healthy elderly patients, individuals with mild cognitive impairment, and Alzheimer's patients.
These original 26 candidate proteins stemmed from various past discovery efforts undertaken by KCL and Proteome Sciences, Pike said. The researchers pared that group down to the final 10 via several analyses, including looking at which proteins appeared linked to hippocampal atrophy as measured by MRI and examining the relationship between protein levels and the rate of patient cognitive decline.
They used 75 percent of the 1,148 subjects to identify this subset of 10 that provided the best predictive power and then used the remaining 25 percent to test the performance of the panel.
Based on this analysis, they found that the 10 proteins plus APOE genotype, which is associated with dementia, predicted patients who would progress from MCI to Alzheimer's within 12 months with an accuracy of 87 percent, sensitivity of 85 percent, and specificity of 88 percent.
Proteome Sciences and KCL are now working to obtain additional patient samples to validate the test in larger cohorts, Pike said, noting that they have identified several thousand samples they hope to use. He added that he expects they will also be able to obtain samples through the Medical Research Council's UK Dementias Research Platform, a $27 million public-private partnership that aims to enable earlier detection and improved treatments for dementia.
In addition to testing the panel in larger cohorts, Proteome Sciences and KCL also plan to move the assays to a single diagnostics platform optimized for the test. In the A&D paper, the researchers used Luminex xMAP multiplex bead assays for all protein measurements.
Pike suggested that moving to an optimized platform could further improve the panel's accuracy, and said he hoped to draw interest from outside diagnostics firms in collaborating to turn the panel into commercial assays.
Proteome Sciences and KCL share intellectual property to the panel, with Proteome Sciences being in charge of its commercialization. Pike said that the panel's commercialization timeline would depend largely on how quickly the company could line up a partner to assist in platform development. He predicted it could take a little as two years or as many as four.
While Proteome Sciences specializes in mass spec-based proteomics, Pike said that the Alzheimer's panel would remain on an immunoassay platform, enabling it to be easily used in memory clinics and hospital psychiatry departments.
The primary use for the test, assuming its performance holds up to further validation, will be as a tool for selecting patients for Alzheimer's drug trials.
Clinical trial development is difficult in Alzheimer's because of the challenges of identifying and enrolling patients who have the potential to show clear benefit of treatment, Pike said.
"When we trial new drugs in Alzheimer's we usually do it in patients who have developed dementia, so even if [the drugs have an effect] their brain is already sufficiently damaged by the process that showing the benefit of treatment is very difficult," he said.
Likewise, patients in the placebo arm may not progress over the course of the study, which can also make it difficult to detect any benefits of treatment.
Proteome Sciences and its KCL collaborators hope that by allowing researchers to identify patients with MCI who are likely to progress to Alzheimer's within 12 months, their panel can give drug companies a better chance of detecting patient benefit from their therapies.
"We want to be able to enrich the population used for clinical trials for people who have mild memory deficit who are going to develop dementia over 12 to 18 months," Pike said. "And if we can do that for both the drug and placebo arm, then we know that the people in the placebo arm should progress over the same 12 to 18 month period. And so if we see a decrease in progression in the drug arm then we can assume that that is the therapeutic effect of the drug."
Much Alzheimer's protein biomarker research has revolved around cerebrospinal fluid levels of the proteins amyloid-β and tau, which, in fact, offer slightly better prediction of progression from MCI to Alzheimer's than the Proteome Sciences-KCL panel.
However, Pike said, the disadvantage of CSF-based measurements is that they require patients to undergo a lumbar puncture. Outside of the US and Scandinavia, patients tend to be reluctant to undergo this procedure, he said, noting that this has proven an obstacle to recruiting patients for clinical trials.
Additionally, Pike said, the CSF markers predict progression over a longer period of time – two to three years, as opposed to 12 months for the plasma panel. This, he said, means that trials enrolling patients based on these markers might have to run for a longer period of time, resulting in delays and higher expenses.
In an email to ProteoMonitor, Holly Soares, director of clinical neuroscience biomarkers at Bristol-Myers Squibb, called the findings "quite encouraging" noting that "there is some consistency with other studies that have identified [Alzheimer's] blood-based biomarkers," which, she said, supports "the utility of blood-based approaches in dementia diagnostics."
"The most exciting aspect about the study revolves around the ability to identify pre-dementia MCI patients who progress to dementia," added Soares, who was not involved in the study but who has participated in other efforts pursuing plasma markers for Alzheimer's.
She noted that the findings were "preliminary" and that "the next step will require prospective testing in a longitudinal cohort," but she called the study "an excellent good start."
In addition to Proteome Sciences and KCL, a number of other parties are investigating plasma markers for Alzheimer's. Last year, for instance, molecular neuropsychiatry firm Genomind signed an exclusive licensing agreement with Emory University to commercialize blood-based protein markers for the disease.
Those markers stemmed from a 2012 analysis of potential plasma protein in three separate patient cohorts done by a consortium of researchers under the Alzheimer's Disease Neuroimaging Initiative.
And while plasma markers are desirable given the inconvenience of collecting CSF, some previous efforts to identify such markers have shown poor reproducibility. Most notably, in a 2012 PLOS One study, Lund University researchers evaluated a panel of 18 plasma proteins previously identified as Alzheimer's biomarkers in a 2007 Nature Medicine study, but were unable to reproduce the original study's results. Additionally, they found that none of the 18 proteins could identify Alzheimer's patients with high diagnostic precision.