By Ben Butkus
In an effort to create a decentralized genetic test for breast cancer intrinsic subtyping, virtual company Bioclassifier has exclusively licensed a 50-gene signature to NanoString, the companies said this week.
Under the agreement, Seattle-based NanoString will attempt to commercialize a test based on the so-called PAM50 gene signature for its nCounter multiplex gene expression platform, with the hope of eventually achieving regulatory approval in the US and other regions of the world, NanoString said.
In the meantime, ARUP Laboratories will likely make available a laboratory-developed qPCR test based on the gene signature in about a month, Bioclassifier's co-founders said.
Bioclassifier is a partnership of four breast cancer experts from leading research institutions, and was founded to develop a breast cancer intrinsic subtyping test based on 50 genes, the varying expression levels of which can help assess prognosis and treatment options for breast cancer patients. The gene set derives its name from Prediction Analysis of Microarray, or PAM, a prediction algorithm used in the development of the signature.
The company's founders include Charles Perou of the University of North Carolina at Chapel Hill; Matthew Ellis of the Washington University School of Medicine; Philip Bernard of the University of Utah and Huntsman Cancer Institute; and Torsten Nielsen of the BC Cancer Agency.
Bioclassifier owns the commercial rights to the PAM50 signature, but each of the co-founders' academic employers retains the right to use the signature internally for research or other purposes, Bernard told PCR Insider this week.
Because ARUP Laboratories is a wholly owned subsidiary of the University of Utah and one of the largest reference labs in the US, it was in a position to develop and offer a test based on the signature, which it will do next month, Bernard said.
"We're just putting in all of the informatics and stuff in place, but the technical validation is completed," Bernard said. "We do want the test to be out there and available to patients as quickly as possible, and that's really where the CLIA-type, laboratory-developed test labs sit: first to market with a high-quality test, but it doesn’t necessarily have to be FDA approved."
Bernard said that ARUP's assay will use Beckman Coulter automation and sample prep modules to extract RNA from formalin-fixed, paraffin-embedded samples, cDNA synthesis, and sample distribution into 384-well plates. The assay will run on the Roche LightCycler 480 with SYBR Green as a fluorescence detection dye. "Specificity is done by melting curve analysis and custom software has been developed for quality control," Bernard added.
However, Bioclassifier's founders retained a vision of making a PAM50-based test available to hospitals and pathology laboratories worldwide, and found that qPCR was not really the appropriate platform for such decentralization.
"For any breast cancer molecular test, to be able to have a distributed test, a kit, that everybody can use, as opposed to a single-site cleared assay, would be a huge advantage," Ellis told PCR Insider.
"If a single lab has got the energy and the will to do a homebrew qPCR test for 50 genes; and they can very carefully control every aspect of the test, then I think qPCR can be used," Ellis added. "But the problem we faced was the need to make a distributed test, where many hundreds of labs are running a gene expression profile. And the experience of trying to get the same qPCR result in just two labs suggested that would never be feasible using the sort of liquid-handling robot, 384-well plate, Roche LightCycler-type approach."
Enter NanoString, whose nCounter gene expression platform and associated CodeSets nucleic acid probes bridge the gap between qPCR and microarrays.
"With fewer genes, I think that qPCR is a suitable platform," Ellis said. "For larger numbers of genes, I don't think that the NanoString platform is suitable, either. But I think signatures that are in the range of a few dozen to the low hundreds of genes — that's a sweet spot for the NanoString approach, and that happens to be the number of genes in many of these multi-gene signatures."
In addition, NanoString scores points for ease of use, Ellis said, adding that "all the probes go into one tube. They're all hybridized, at the same time, in the same tube. That just creates a much more robust scenario. In addition, NanoString does not require an enzyme, so there's no variation. You end up with an enzyme-free, one-tube assay. For distributed testing, that's more likely to be successful."
Bernard added that the PAM50 signature was derived from PCR to begin with, and that Bioclassifier "planned all along" to commercialize it using qPCR. However, "it wasn't the right platform for decentralization. And we've shown, although we haven't published yet, that the concordance in the subtype calls between the PCR and NanoString are very close."
Under its licensing agreement with Bioclassifier, the financial terms of which were not disclosed, NanoString will develop a breast cancer intrinsic subtyping assay based on PAM50 to measure the expression levels of the genes in FFPE breast tumor tissue samples.
The company said that the test may provide clinically useful information for a broader range of breast cancer subtypes, including classification of tumors from patients with estrogen receptor negative, or node positive forms.
NanoString also said that it intends to collaborate with leading breast cancer researchers to clinically validate the gene signature on the nCounter system, and plans to seek regulatory approvals from the FDA "and other relevant agencies outside the US." According to a report this week by PCR Insider sister newsletter BioArray News, the company hopes to gain FDA approval by 2012.
"We think these tests can coexist, but I'll probably start phasing over toward the NanoString test after FDA approval, and that's in our contracts," Bernard said. "There may be situations on very severely compromised samples that we can't get a good reading by NanoString, at which point I would default to qPCR."
Have topics you'd like to see covered in PCR Insider? Contact the editor at bbutkus [at] genomeweb [.] com.