ATLANTA Genentech and Roche are developing a drug-diagnostic companion product for advanced ovarian cancer and are relying on the first of a new class of Her2 inhibitors together with an unorthodox approach to identifying response biomarkers.
The nascent companion product includes a Her2-dimerization inhibitor, or HDI, and a quantitative real-time PCR gene-expression test whose final version is meant to identify cancer tissue samples that will likely respond to the treatment by inferring Her2's activation status. If the companies are able to successfully develop the theranostic, the diagnostic component may eventually play a role in helping predict the response of other kinds of HDIs.
"The idea would be that [the qRT-PCR test] would be submitted to the FDA at the same time as the drug," Lukas Amler, director of molecular diagnostics at Genentech, told Pharmacogenomics Reporter in an interview last week at the American Society of Clinical Oncology, held here. He declined to say when this might happen.
Genentech's drug, called Omnitarg (pertuzumab), is an antibody that inhibits cancer growth by blocking Her2 proteins from coupling with other members of the ErbB family. The drug is currently being tested as an ovarian cancer treatment alone and in combination with traditional chemotherapy.
"We're basically going into a setting where the target is not validated because it has a low expression level, and it really gets activation with this interaction with the co-receptors."
Roche's companion diagnostic, meantime, represents an unorthodox approach to discovering whether the therapeutic target is activated: Instead of testing tissue with antibodies directed at phosphorylated versions of Her2, the assay uses relative levels of gene expression to infer Her2 phosphorylation.
According to results of a phase II Omnitarg study by Amler and colleagues presented at ASCO this week, these genes will include Her2, EGFR, Her3 and unnamed Her ligands. In the study, Amler and colleagues used ELISA to analyze 28 samples and found eight that tested positive for Her2 phosphorylation. They also found that these eight reflected a time to progression of 20.9 weeks versus 5.8 weeks for those testing negative.
The companies plan to develop an assay that can work in the routine clinical setting, Amler said during his presentation. They have already developed the assay as a microarray expression profile for fresh tissue samples and are now working on a qRT-PCR test that will work with formalin-fixed paraffin embedded samples. However, this particular qRT-PCR assay, whether used in fresh or formalin-fixed tissue samples, can accurately predict Her2 phosphorylation in only about 75 percent of samples, he said.
Asked whether Genentech expects to stick with its gene-expression qRT-PCR diagnostic as a companion to Omnitarg, Amler said, "That's our main focus in the current randomized trial that's an endpoint. We're assessing other things, but that's the main goal."
Other methods, including ELISA and gene-expression microarrays, are not as simple as PCR for routine clinical use. And techniques such as immunohistochemistry, which are often employed to establish the phosphorylation level of proteins present in tissue samples, do not reliably survive formalin fixing, Amler said. Also, levels of protein phosphorylation are often too low for IHC detection, he added.
So Roche and Genentech will stick with qRT-PCR for the moment, and are currently conducting a randomized phase II trial of archived ovarian cancer tissue to confirm that their gene-expression biomarkers predict patient benefit, said Amler. The next step will be a phase III trial to validate the biomarkers' predictive power, he said.
Amler said that the Omnitarg theranostic program is one of several that Genentech is engaged in. "We have a program now [in which] basically all the drugs [in phase II] or earlier have companion diagnostic programs," said Amler.
In an interview, a Genentech spokesperson declined to specify whether any other programs at the company involved similar diagnostics or other dimerization inhibitors.
Roche, which owns a majority share in Genentech, was not able to reply to questions before deadline.
Other Diagnostic Approaches
QRT-PCR isn't Roche and Genentech's only option. There are companies that are developing technologies that aim to detect dimerization directly rather than using qRT-PCR to infer it. Monogram Biosciences is one, and its CFO Alf Merriweather claims the company has no direct competitors in the area. "We have an IP position around that," he said this week in an interview.
Monogram claims its eTag assays, which are not yet on the market, are able to quantitatively measure specific dimers and levels of protein expression in FFPE tissue. Merriweather said the assay is intended to measure the targets of drugs attacking the full ErbB family, including Herceptin, Tykerb, Omnitarg, and other compounds in development.
"We have a range of assays that we're working on for all the dimers of Her1, Her2, and Her3," and Monogram is involved in clinical studies to identify and validate correlations between those dimers and clinical response, he said.
Asked whether Monogram has worked with Genentech or Roche on Her2 dimerization inhibitors, Merriweather said he "can't comment on whether Omnitarg was looked at, but we had a collaboration early on with Genentech."
Chris Womack ([email protected])