NEW YORK (GenomeWeb) – Researchers associated with the I-SPY 2 trial have identified a number of proteomic and phosphoproteomic markers that can predict the response of breast cancer patients to Puma Biotechnology's tyrosine kinase inhibitor neratinib.
In particular, they identified one marker, phosphorylated EGFR 1173, that has the potential to improve upon existing markers like HER2 and HR status while also expanding the population of patients considered likely candidates for neratinib treatment, Emanuel Petricoin, co-director of George Mason University's Center for Applied Proteomics and Molecular Medicine and an I-SPY 2 researcher, told GenomeWeb.
The I-SPY 2 trial aims to combine molecular measurements with screening of multiple agents from multiple companies in an effort to improve the efficiency of the drug development process. Patients are given treatments most likely to benefit them based on their molecular profile and, when a drug under investigation demonstrates superiority in one of the 10 molecular signatures being investigated, that drug is graduated in that patient population from the Phase II I-SPY 2 to the Phase III I-SPY 3 trial.
The trial in April graduated neratinib in the hormone receptor-negative, HER2-positive population.
The I-SPY signatures consist primarily of established markers like HER2 and HR status as measured by US Food and Drug Administration-approved assays like FISH and IHC. Yet, while biomarker discovery is not the stated purpose of the trial, it has offered opportunities for researchers to investigate whether other markers could add power in predicting patient response to the drugs being tested.
In the case of neratinib, Petricoin and his colleagues looked at 18 total protein and protein phosphorylation markers in the HER signaling family using reverse phase protein arrays. In an analysis of 168 I-SPY 2 participants, they identified seven markers that predicted response to neratinib.
Most of these markers did not provide additional information beyond the HR-negative, HER2-positive signature. However, one marker, phosphorylated EGFR 1173, both added predictive information beyond that signature and on its own provided better prediction of patient response than the HR-negative, HER2-positive signature.
In addition, the patient population exhibiting phosphorylated EGFR 1173 was significantly larger than the HR-negative, HER2-positive population, suggesting that the marker could enable wider use of the neratinib.
"HR-negative, HER2-positive is actually [one of the] least prevalent breast cancer populations, only about 6 percent of patients," Petricoin said. "So neratinib graduated [to I-SPY 3] in what is a very small population."
Phosphorylated EGFR 1173, on the other hand, shows up in around 25 to 30 percent of breast cancer patients, he noted. In their analysis, which the researchers detailed in a poster presented last week at the San Antonio Breast Cancer Symposium, 85 of the 168 patients they studied had high phosphorylated EGFR 1173, compared to 28 with HR-negative, HER2-positive status.
"So now if you're Puma [Biotechnology] it becomes very attractive." Petricoin said.
The researchers' analysis determined that phosphorylated EGFR 1173 alone was more predictive of phase III trial success than the HR-negative, HER2-positive signature. And when the phosphorylated EGFR 1173 population is added to the HR-negative, HER2-positive population, likelihood of phase III trial success rises to 90 percent from 87 percent for the HR-negative, HER2-positive signature alone while increasing the eligible patient population by 50 percent.
In addition to expanding the potential patient population, adding phosphorylated EGFR 1173 to the HR-negative, HER2-positive signature reduced false positives by around 50 percent, Petricoin said. "So you still find all your responders, but you drop your non-responding population by half, which means you [can] enroll [fewer patients] in your Phase III trial because many more of the patients you are bringing in you expect to respond."
As the researchers noted, these findings need to be replicated in larger trial cohorts, but, Petricoin said, they provide evidence of the potential usefulness of protein phosphorylation as a biomarker.
"It shows good evidence that measuring phosphorylation is really predicting clinical response," he said. "Phosphorylation of a single marker like EGFR is outperforming FDA-approved ways of measuring the graduating success, so this primes the pump to get these phosphoproteins measured [using] robust clinical methods."
Phosphorylated EGFR 1173, Petricoin noted, is part of a CLIA breast cancer biomarker panel offered by Theranostics Health, a phosphoproteomics-based cancer profiling firm he co-founded.
"So it's not like this is some abstraction," he said. "We could go to Puma and say that we have a CLIA-based test right now that measures this marker."
Indeed, Petricoin said he is currently talking with the drugmaker about additional trials in the expanded high-phosphorylated EGFR 1173 populations identified in the I-SPY 2 work, such as HER2-negative and triple-negative patients. He added that they are preparing research on use of phosphorylated EGFR 1173 for identifying neratinib responders in the triple-negative population for presentation at the American Society of Clinical Oncology annual meeting in May.
The researchers will also be looking at phosphorylated EGFR 1173 prospectively in I-SPY 3, which, Petricoin said, will provide them with additional data on the marker and validation in a larger population. They also plan independently to look at whether different combinations of the proteomic and phosphoproteomic markers they identified as predictive of neratinib response could provide additional information.
In 2012, Petricoin and his GMU colleagues published a study in Clinical Cancer Research that had similar implications for expanding the patient population of targeted breast cancer drugs. In that study they identified a subset of roughly 40 patients who, while negative for HER2 according to conventional IHC and FISH tests, showed phosphorylated HER2 and activation of HER2 signaling pathways.
Another interesting result to emerge from the I-SPY 2 biomarker research, Petricoin said, was the finding that total HER2 also provided predictive information in terms of neratinib effectiveness beyond that provided by the HR-negative, HER2-positive signature.
On the face of it, he noted, this was a strange finding in that the HR-negative, HER2-positive patients had already been identified as having high HER2 expression, so total HER2 levels wouldn't be expected to add additional predictive value.
However, he said, the HER2-positive population as determined by conventional FISH or IHC assays contains a "huge dynamic range" of HER2 levels. And the more precise quantitation provided by the RPPA data suggests that these HER2-positive populations can be further stratified by expression levels.