Opko Health last week presented data at the Alzheimer's Association International Conference in Vancouver demonstrating that its antibody-based Alzheimer's diagnostic identified cases of the disease with 100 percent specificity in a set of 36 autopsy-confirmed plasma samples.
According to Thomas Kodadek, the diagnostic's developer as well as a professor at the Scripps Research Institute and director of chemistry and molecular biology for Opko, the blood-based test is able to diagnose established Alzheimer's with performance approaching that of methods that combine MRIs with cerebrospinal fluid measurements of the Alzheimer's protein biomarkers Aβ42 and tau.
The test demonstrated only 67 percent sensitivity in the data presented at AAIC, but, Kodadek told ProteoMonitor, this score included results from a number of early-stage patients. While the 36 patients examined were all autopsy confirmed, their blood was collected at different points in the disease process.
"The 67 percent [sensitivity] is an aggregate number from samples that ranged all the way from pre-symptomatic to significantly demented," Kodadek said, adding that the test performed significantly better when the numbers for samples collected from later-stage patients were broken out, although he didn't provide exact figures, noting the small sample size of the study. "So the implication of [these results] is that our test is really quite good at picking up established Alzheimer's, but it's probably not that sensitive for very early-stage disease."
Laboratory Corporation of America signed a license for rights to the company's Alzheimer's diagnostic technology in April and is working with Opko to commercialize the test for diagnosis of later-stage Alzheimer's patients (PM 4/6/2012). Additionally, Opko is collaborating with LabCorp and Bristol-Myers Squibb on a second-generation version of the test intended as an early detection test for diagnosing the disease and facilitating drug development.
Opko and Kodadek are also at work revamping the test's format, attempting to make it better suited to high-throughput clinical applications without sacrificing performance.
The diagnostic uses synthetic molecules called peptoids to screen for autoantibodies in blood associated with Alzheimer's. In a study on the technique published in Cell in 2010, Kodadek and his colleagues screened patient blood samples against libraries of roughly 40,000 peptides arrayed on glass slides. In an interview with ProteoMonitor following release of the paper, Kodadek said his team had synthesized a library approaching roughly 1 million peptides (PM 1/7/2011).
Since then, however, the researchers have moved away from the glass slide array format to an ELISA plate format due to the difficulty of reproducibly synthesizing the glass slide arrays. This has allowed them to increase their throughput by roughly 100-fold, lifting it into the realm needed for clinical applications.
The change in format, however, has come with its own problem, Kodadek said: The lower density of the peptoids in the ELISA well format leads to weaker binding of the Alzheimer's-specific autoantibodies and a subsequent loss in signal.
"The performance [of the test] on the ELISA plates is not nearly as good as on the glass slides," he said, noting that the 100 percent specificity and 67 percent sensitivity figures presented at AAIC were generated using the ELISA plates.
"When we analyze those samples on the glass slide platform, our results are much better," Kodadek said. "More like 90 percent sensitivity and specificity, and at least in preliminary experiments they suggest we're perfectly capable of picking up Alzheimer's at the [mild cognitive impairment] stage."
Despite its significantly better performance, though, the glass slide array format is "a non-starter" for clinical work, Kodadek said, noting that it took his team roughly 18 months to run a 105-sample validation set on the platform. Instead, they are now attempting to improve the performance of the ELISA platform by developing higher-affinity peptoids that can bind the target antibodies effectively despite the format's lower density.
Even with the drawbacks of the existing ELISA format, Opko and LabCorp will "likely put out a test based on it," Kodadek said. "It's a good enough [test] to put out on the market. The frustration is that I know it's not as good as we can do."
"Basically, if someone is showing clear signs of dementia and you want to categorize it – that's what this [initial] test will be for," he said. "But we all know that that's really limited, and what we all want is a very early-stage diagnostic. I doubt if our first-generation ELISA test will be capable of doing that."
With that in mind, he said, the company is "trying to pivot as quickly as possible and put resources into getting these better [peptoid] compounds and getting the second-generation [test] in hand as soon as possible"
The company is also moving ahead in the meantime with research performed on both the glass slide array platform and the first-generation ELISA platform, currently using both for a 130-patient study looking at controls; advanced Alzheimer's patients; MCI patients who progressed to Alzheimer's; and MCI patients who did not progress to Alzheimer's.
Kodadek described the efforts as "kind of a two-tiered process" wherein the researchers were carrying on validation work using the old platforms while at the same time developing the new one.
In addition to the Alzheimer's data, Kodadek also presented data at AAIC on a peptoid-based diagnostic for neuromyelitis optica, or NMO, showing that the test could detect the disease with 90 percent sensitivity and 100 percent specificity.
The company plans to commercialize the test as a diagnostic for distinguishing NMO from multiple sclerosis, which it can resemble. Kodadek noted as well, that because the target of the autoantibodies in NMO is known to be the protein aquaporin 4, the test provided an opportunity to validate the peptoid-based approach.
"In theory, when we screen [NMO] patients we ought to find peptoids that bind the anti-aquaporin 4 antibodies, and, in fact, we did," he said. "So that was nice because it validated the technology in a well-defined human system."
The researchers also found that they were able to obtain good results for the NMO test using the first-generation ELISA platform, likely, Kodadek said, because of the high abundance of the NMO-specific antibodies.
Opko is also using the peptoid technology to develop tests for lung and colon cancer. Both of those programs are currently in the discovery phase.