With DNA-based affinity agents gaining prominence as tools for protein capture, scientists and biotech firms are exploring biomarker assays that simultaneously report both genetic and proteomic data.
Such assays, researchers told ProteoMonitor, could enable better investigation of mechanisms of protein dysregulation in disease states and offer the potential for improved diagnostic sensitivity and specificity.
"When you multiplex [similar] markers, you always get the problem that increasing sensitivity decreases your specificity, and vice-versa," said Tony Shuber, co-founder and chief technology officer of diagnostics firm Predictive Biosciences. Combining data from two classes of markers, he said, is one potential way around that problem.
The company is developing tests intended to "increase specificity and sensitivity simultaneously," Shuber said. "And with one marker – DNA – being binary and the second marker – which is protein – being quantitative, you can achieve that goal."
"If you choose the right DNA marker, it can have extremely high specificity [for a disease]," Shuber said, but the sensitivity might only be in the neighborhood of 40 percent to 50 percent — an issue that can be addressed with protein markers.
"We hit the population with the DNA [markers] to pluck out as many people with cancer as possible with the very high positive predictive value associated with high specificity," he said. They then combine this information with data obtained using protein markers optimized for sensitivity, which allows for identification of patients highly unlikely to have the cancer.
Predictive's CertNDx bladder cancer diagnostic and grading assays use a combination of ELISAs and quantitative PCR to test for the protein Ki-67 and variants of the gene FGFR3, but, Shuber said, by shifting to aptamers for protein detection, the company aims to move the assays to a single quantitative PCR-based platform.
In June, the company received a US patent for its Multi-Analyte Diagnostic Readout, or MADR, technique, which uses aptamers combined with qPCR for simultaneous measurements of gene and protein biomarkers. According to Shuber, it aims to have tests available on the platform by the middle of 2012.
The notion of moving the two measurements to one platform emerged out of a desire to streamline the assay for potential users. The CertNDx assays are currently performed out of Predictive's CLIA laboratories, but, Shuber said, the company hopes eventually to package its tests as US Food & Drug Administration-approved IVD kits.
"We'd hear this from a number of people," he said. "The protein person would say, 'Well, I'm used to doing ELISAs,' or the nucleic acid person would say, 'I'm used to doing qPCR. So we asked, how could we convert ELISAs to a DNA platform?"
The company settled on aptamers primarily due to the low cost and ease of producing them, Shuber said. Techniques like DNA-antibody chimeras that use DNA linked to monoclonal antibodies could also achieve the goal of combining gene and protein detection in a single platform, but, he said, "we chose aptamers because ultimately we think that will be the cheapest way to go."
"Once you've gone through the process of selecting which aptamer sequence is best, it's a very large-scale synthesis, so it's really cheap," Shuber said. Another consideration, he said, is the industry's level of comfort with the synthesis of such reagents.
"The Roches of the world have gotten to the point commercially where they've figured out how to mass-produce oligos that are really cheap and have high purity," he said. "If we walk into Roche, I don't have to say, 'By the way, your product development team will have to figure out how to scale this up and get the cost down.' They can just pick up the phone to New Jersey and order whatever oligos they need."
In addition to CertNDx, Predictive is looking to apply the MADR platform to its work on prostate cancer diagnostics. There, Shuber said, the addition of protein markers could add significant utility to "some DNA markers that have been reported to have extremely high specificity but don't have tremendous specificity."
The company is also working to move the tests from a qPCR platform to next-generation sequencing, using for this research a Roche 454 GS Junior instrument.
Predictive isn't alone in its interest in using NGS for protein quantitation, as firms like Pronota and Olink Biosciences are also pursuing this goal (PM 5/6/2011). Unlike Predictive, these companies are both focused solely on protein biomarkers, but in an interview with ProteoMonitor in May, Pronota's then-chief scientific officer Koen Kas raised the potential for assays capable of simultaneously measuring genetic and proteomic data, citing prenatal testing as an example.
"What are the other things that you want to screen for in pregnant women?" he said. "Not everything you want to measure in a pregnant woman can be measured at the DNA level. It might be handy to measure something at the protein level. So I see [protein biomarker measurements] in the clinic combined with DNA and RNA-based measurements to give a comprehensive analysis in one single device."
Like Predictive, Pronota is pursuing an aptamer-based strategy. Olink and the lab of its founder and second-largest shareholder, Uppsala University researcher Ulf Landegren, are taking a different approach, however, exploring the use of proximity ligation assays, which use pairs of antibodies attached to unique DNA sequences to detect proteins of interest. When the antibodies bind their target, the attached DNA strands are brought into proximity and ligate, forming a new DNA amplicon that can then be quantified.
Landegren's research has focused on using NGS for PLA read-outs, which today are typically done via qPCR. In August, though, a team of scientists published a paper in the journal Neuro-Oncology, detailing a PLA-based technique enabling simultaneous measurement of gene and protein levels in single cells using fluorescent in situ hybridization.
Using Olink's PLA reagents, the scientists used FISH to simultaneously measure both EGFR gene status and EGFR protein expression in glioblastoma cells. According to Jaclyn Renfrow, a researcher at Northwestern University and author on the paper, the assay, called fluorescent in situ gene protein assay, or FIGPA, could prove useful in research into cancer diagnostics and therapeutic design.
"Cancer drugs are targeted therapies," Renfrow told ProteoMonitor. "Almost every cancer drug that is out there is based on inhibiting a protein in one way or another. So, for clinical translation, the correlation of the genomic side and the proteomic side of cancer need to feed into each other. It's important to understand how genomic changes end up causing protein dysregulation."
The technique could be prognostically useful, as well, she said, citing HER-2 testing in breast cancer patients where "there is debate currently whether the best testing is a genomic-based test or a protein-based assay."
The researchers chose PLA as opposed to other DNA-based protein detection techniques for its high sensitivity, Renfrow said. Also key, she added, is that PLA offers quantitative output, "in that each signal represents a single molecule, and when we quantify the number of [signals] it makes clinical translation potentially easier."
The grading systems currently used in traditional IHC leave room for "subjectivity between different evaluators," she said. "But if you have a computer program that can simply count the number of dots that are there, that may eliminate that bias and make it easier to establish what is biologically relevant where protein levels are concerned."
The researchers have thus far applied the technique to wildtype EGFR and a mutant form of EGFR and are currently working to extend it to other proteins. It's also potentially multiplexible, to allow for read-out of multiple markers, but, Renfrow said, that's not currently a focus.
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.