Skip to main content
Premium Trial:

Request an Annual Quote

GMU-led Team Develops Proteomic Signatures for Classifying, Guiding Breast Cancer Treatment


NEW YORK – A team led by researchers at George Mason University has identified a series of proteomic and phosphoproteomics biomarkers that could inform selection of therapies for individuals with triple-negative breast cancer (TNBC).

In the study, detailed this month in Cell Reports Medicine, the scientists identified a phosphoproteomic signature that distinguished between TNBC patients likely to respond to anti-HER2 therapies and those likely to respond to immune checkpoint therapy.

The findings, which were generated as part of the I-SPY 2 TRIAL (Investigation of Serial Studies to Predict Your Therapeutic Response with Imaging and Molecular Analysis 2), could aid therapy selection in TNBC patients and specifically in a subset of TNBC patients without good options for molecularly guided treatments, said Emanuel Petricoin, codirector of the Center for Applied Proteomics and Molecular Medicine at George Mason University and senior author on the study.

Petricoin and his GMU colleagues have long focused on how phosphoproteomic measurements can inform cancer treatments, typically using laser capture microdissection (LCM) and reverse phase protein arrays (RPPA) to measure proteins and phosphoproteins in cancer cell lysates and assess the presence and activation levels of different cancer signaling pathways.

In previous work within I-SPY 2, Petricoin and his collaborators found that some TNBC patients with elevated phosphorylation levels of EGFR and HER2 responded to Puma Biotechnology's anti-HER2 tyrosine kinase inhibitor Nerlynx (neratinib) despite being HER2-negative based on standard pathology, suggesting that protein phosphorylation could be an important measure to consider along with protein expression levels when selecting treatment for such patients.

In the recent Cell Reports study, the researchers expanded their phosphoproteomic analysis of TNBC, using LCM and RPPA to measure 139 proteins and phosphoproteins pre-treatment in 736 patients enrolled in eight different treatment arms within I-SPY 2. These same patients had also undergone gene expression profiling, allowing the scientists to compare transcriptomic-based and phosphoproteomic-based patient stratification approaches.

A key finding of the work, Petricoin said, is that phosphoproteomic data appears to provide information that could aid in selecting treatment for a group of TNBC patients for whom gene-expression data offers little guidance.

This group, categorized by transcriptomic data as the HER2/Immune/DNA repair deficiency (DRD) subtype, "was kind of a question mark," in terms of molecularly guided treatment, Petricoin said. "It really had no therapeutic options that were rationally seen" based on its transcriptomic signature.

When the researchers applied the EGFR-HER2 phosphorylation signature they identified in their earlier I-SPY 2 work to this subset of patients, however, they found that roughly 45 percent exhibited this signature (named HARPS-positive), indicating that they could be good candidates for treatment with anti-HER2 TKIs like Nerlynx. Additionally, they found that HARPS-negative patients within this group responded more strongly to PD1 inhibitors than did HARPS-positive patients.

"HARPS-positive patients seem to be responsive to HER2 inhibitors and HARPS-negative patients seem to be responsive to [checkpoint] inhibitors," Petricoin said. "It's an interesting binary signature that can kind of bifurcate a [TNBC subgroup] that would previously be kind of a question mark in terms of [treatment]."

He suggested the signature might ultimately be useful as a sort of reflex test in cases where transcriptomic testing yielded little actionable information.

While Petricoin and his colleagues have established HER2 and EGFR phosphorylation as the key components of the HARPS-positive signature, they are still working to identify the proteins underlying the HARPS-negative signature and its responsiveness to checkpoint inhibitors.

Petricoin said the researchers have seen elevated activation of STAT1 and other immune components in these individuals but added that characterizing the HARPS-negative signature would require closer analysis of the stromal tissue around the tumor as opposed to the tumor tissue itself, which was the focus of the work described in the Cell Reports paper.

He said that the researchers are currently analyzing this stromal tissue and plan to publish on it separately.

"I think we'll have some interesting markers that will help provide a rationale for why the HARPS-negative population seems to be so responsive to [checkpoint inhibitors]," he said.

In addition to exploring the utility of the HARPS signature for guiding treatment in the HER2/Immune/DRD transcriptomic subgroup, the researchers generated 11 phosphoproteomic subtypes for the 736 patients analyzed.

Based on the activation of protein signaling pathways, these subtypes showed little overlap with established TNBC subtypes, Petricoin said. "It's providing a different view of tumor biology, and because of that I think it could be very helpful in identifying subsets of patients that may be most responsive to other inhibitors."

"We do find [patient] clusters that are driven by mTOR activity. We do find clusters that are driven by Aurora kinase activity," he said. "And we would suggest that maybe those patients would be more sensitive to those therapies."

The researchers also identified a signature consisting of elevated levels of the proteins cyclin D1, ERα, and AR S650 that Petricoin said appears linked to treatment resistance across all the TNBC subtypes included among the 736 patients in the study. He suggested that this points toward the possibility of using therapies targeting these proteins with the aim of sensitizing them to other treatments.

Moving forward, the researchers hope to have their findings more fully incorporated into future work within the I-SPY 2 trial, which would give them the opportunity to further validate them.

"We don't have enough clinical utility data yet," Petricoin said. "HARPS is not being used to stratify patients in I-SPY 2 right now. We are trying to work that in prospectively to get HARPS into [the trial's] block design where we can test this."

Petricoin and his GMU colleagues are also working with Rutgers Cancer Institute of New Jersey under a $1.3 million grant from the US Department of Defense to explore whether the HARP signature is able to predict patient outcomes in a retrospective study set.

He added that Golden, Colorado-based proteomics firm Theralink, which uses the LCM and RPPA technologies to aid cancer treatment decisions, is also generating data on the HARP signature and other activation signatures through its commercial work. Petricoin is a consultant for Theralink and chair of its science advisory board. The company's lead product is its Test for Advanced Breast Cancer, which measures the abundance and activation of 32 protein markers.