Opko Health said this week that it has expanded its collaboration with drugmaker Bristol-Myers Squibb to use its protein biomarker discovery platform to identify markers predictive of patient response to several BMS therapeutics.
The expanded collaboration will initially focus on identifying biomarkers predicting response to the rheumatoid arthritis drug abatacept, marketed by BMS as Orencia, Thomas Kodadek, professor of chemistry and cancer biology at the Scripps Research Institute and a scientific director at Opko, told ProteoMonitor this week.
The agreement extends the companies' original 2010 collaboration, which has focused on using Opko's technology to discover biomarkers for diagnosing Alzheimer's disease and identifying individuals with mild cognitive impairment likely to progress to Alzheimer's. That work will continue under the expanded deal.
Both projects will use a modified version of the discovery platform, which uses arrays of synthetic molecules to screen for autoantibodies in blood associated with various disease states. This is the second time in the last year that Opko and Kodadek have reworked the platform in hopes of achieving higher throughput and sensitivity.
The original version of the platform, which Kodadek and colleagues published on in Cell in 2011, used molecules called peptoids – synthetic N-substituted oligoglycines – arrayed on glass slides to screen patient blood samples for antibodies generated in response to disease-related antigens (PM 1/7/2011).
Such an approach had several potential advantages, Kodadek said in an interview following publication of the Cell paper. Because "antibodies are a lot more stable than most serum proteins, how you collect the sample, how you freeze it isn't terribly important," he noted. "Whereas for the guys who are doing standard proteomics, that sort of stuff just plagues them."
More significantly, he added, "the adaptive immune response against an initiating antigen results in millions-fold amplification of the antibodies that recognize it." That amplification makes the antibodies much easier to detect and measure than typical protein biomarkers, many of which are of low abundance.
Despite these theoretical advantages, however, Kodadek and Opko have struggled to put the platform into practice. The initial version used dense assemblages of peptoids arrayed on glass slides. Making these slides, however, proved too laborious for the system to be practical as a validation or diagnostic platform, and so the researchers turned to an ELISA well format that provided the necessary higher throughput (PM 6/27/2012).
Kodadek told ProteoMonitor this week, however, that the ELISA well version of the platform had not proved sensitive enough to be effective. This, he said, was due to the fact that the peptoids could not be arrayed as densely in the ELISA well format as they could on the original glass plates and that, at this reduced density, they couldn't bind the target antibodies effectively enough to make for high-sensitivity tests.
As a result, Kodadek said, the researchers have moved away from peptoids to a new type of ligand that, he said, has shown much stronger binding.
"To be frank, we've largely given up on peptoids. They just weren't good enough," he said. The new compounds, he said, are similar to peptoids, but are more rigid, which results in a higher binding affinity to the target antibodies, meaning they can be used in the low-density, but high-throughput, ELISA well format.
"We're finding that [the new compounds] are maybe 100-fold better ligands than the equivalent peptoids for a given protein," Kodadek said. "Just recently we've completed a few initial screens with these things, and they are really working much better."
Kodadek said Opko was using this new version of the platform to go back and repeat much of the Alzheimer's work it had previously done on the peptoid-based system. In data presented last year at the Alzheimer's Association International Conference, an Alzheimer's test Opko developed using the old platform identified cases of the disease with 100 percent specificity in a set of 36 autopsy-confirmed plasma samples, but showed sensitivity of only 67 percent.
A subset of that data suggested, however, that the peptoid-based test could prove effective as a diagnostic for late-stage Alzheimer's, Kodadek said. He added that Opko was still trying to develop this test on the peptoid platform as a short-term solution because its partner on the test – Laboratory Corporation of America, which signed a license for rights to the company's Alzheimer's diagnostic technology in April of last year (PM 4/6/2012) – wants "to get something as rapidly as possible."
Longer term, though, Opko's efforts will revolve around the new modified version of the platform. In addition to the late-stage Alzheimer's diagnostic, Opko is working with LabCorp and BMS on an early detection test to diagnose the disease and facilitate drug development.
Kodadek said he hoped the company could "[regain] our lost ground on Alzheimer's" and have a test for the disease ready by the end of the year.
With the expansion of the BMS deal, Opko will also be turning its efforts to identifying biomarkers of patient response to the rheumatoid arthritis drug Orencia.
TNF-alpha inhibitors "work very well for some fraction of patients and not so well for others, so we're going to look to see if there are antibody markers that might predict the efficacy of response to Orencia," Kodadek said.
"There will be more [drugs] to come after that," he added. "We signed an agreement to tackle a certain number of projects, but they haven't all been decided upon yet."
A BMS official contacted by ProteoMonitor declined to comment on the deal.
Opko and BMS are not alone in looking for protein biomarkers that could predict rheumatoid arthritis patient response to TNF-alpha inhibitors like Abbott's Humira (adalimumab), Amgen/Pfizer's Enbrel (etanercept), and Janssen Biotech's Remicade (infliximab). In particular, Crescendo Bioscience's protein biomarker test Vectra DA has shown potential utility for identifying responders to such drugs.
A study published last year in the journal Arthritis Care & Research suggested that Vectra DA might prove useful in identifying responders to a given drug early in treatment, finding that the test identified significant changes in clinical responders to methotrexate and anti-TNF therapies as early as two weeks after the start of treatment (PM 7/6/2012).
Also last year Crescendo presented posters at the European League Against Rheumatism's Annual European Congress of Rheumatology indicating that Vectra DA tracks RA patient response to anti-TNF therapy and patient response to treatment with the Jak inhibitor tofacitinib.
The company presented data as well at the 2011 American College of Rheumatology and Association of Rheumatology Health Professionals’ annual meeting demonstrating that the test could identify patients responding to anti-TNF drugs.
Vectra DA is an immunoassay-based laboratory-developed test that measures levels of 12 proteins in blood to assess RA activity. Crescendo released the test in November 2010, and it is currently available through the company's CLIA laboratory. In 2011, Myriad Genetics made a $25 million debt investment in Crescendo, securing a three-year option to purchase the firm. This month the company raised $28 million in a Series D financing round.
Opko is also not alone in its antibody profiling approach. Indeed, Stephen Johnston, a researcher at Arizona State University's Biodesign Center and a former collaborator of Kodadek's when both were at University of Texas Southwestern Medical Center, has pursued a similar approach, using synthetic peptide arrays to detect antibodies associated with various diseases (PM 2/4/2011).
Last year, Johnston founded a company, Healthtell, to commercialize the technology. His ASU lab also last year won a four-year, $30 million grant from the US Department of Defense to develop the technology for in-field monitoring of soldiers' health and early detection of pathogen infection (PM 4/27/2012).
Like Kodadek and Opko, Johnston and his team have found producing the arrays in a high-throughput manner to be a challenge, with Johnston telling ProteoMonitor in a 2012 interview that producing high-content, high-quality chips has been among the largest hurdles they have faced in developing the method.
"There's a lot of chemistry we had to deal with," he said. "If you're making something on these little features, how do you look at that chemically to see whether it's right or not? Those technologies really weren't there, so we have to develop them ourselves in order to make progress on the [array] synthesis."
Beyond its Alzheimer's and TNF-alpha work, Opko is also using its discovery platform to identify biomarkers for detecting pancreatic, lung, and colon cancer, Kodadek said. The pancreatic cancer work, in particular, looks promising, he said, adding that his team is preparing for large blinded studies this spring or summer investigating potential markers for the disease.
He noted, though, that given the troubles of the past versions of the platform, he is "trying to learn my lesson about not being overly optimistic."