NEW YORK – University of Michigan researchers hope that a new type of liquid biopsy test based on a graphene oxide (GO) microfluidic chip will provide a more accurate means of monitoring treatment response and predicting outcomes for patients with non-small cell lung cancer than currently available circulating tumor cell (CTC) tests.
CTC tests typically rely on the presence of epithelial cell adhesion molecule (EPCAM) on the cell surface to capture and detect tumor cells. However, as malignant cells undergo the epithelial-to-mesenchymal transition and become more mobile in the process, the population of cells entering blood circulation will include many that do not express EPCAM, according to Sunitha Nagrath, a professor of chemical and biomedical engineering at the University of Michigan.
"Using a marker to capture cells would not capture all types of cells that would reflect heterogeneity within the tumor," Nagrath said. Recognizing that, Nagrath's group sought to develop a device that would interrogate multiple markers.
They developed the GO chip more than a decade ago by anchoring graphene oxide nanosheets on top of a patterned gold surface. Linker molecules are immobilized on the graphene oxide sheets, to which capture antibodies can attach.
Nagrath's group chose graphene oxide as a material for the chip due to its favorable properties for biosensing, including ease of surface modification using PEG-based chemistry, the ability to manipulate the size of the sheets using common laboratory methods, and its unique optical properties. The resulting chip captured CTCs in blood samples drawn from breast, lung, and cancer patients with a sensitivity of 73 percent at a concentration of 3 to 5 cells per milliliter of blood.
In a subsequent study published in 2019, Nagrath and her team used the GO chip to measure PD-L1 expression in circulating tumor cells from patients with non-metastatic NSCLC during treatment. They found that the proportion of PD-L1-positive cells increased significantly after initiation of radiation therapy and that patients positive for PD-L1 at baseline had a shorter median progression-free survival. The results suggest that changes in CTCs monitored using the GO chip could predict immunotherapy response.
For the present study, Nagrath connected with University of Michigan radiation oncologist Shruti Jolly, who is conducting a clinical study to determine whether imaging biomarkers can detect early pathologic changes in heart and lung function from thoracic radiation in patients with stage III NSCLC.
Jolly and Nagrath used the GO chip to profile CTCs from patients in the study who are receiving six weeks of chemotherapy and radiation and then, in line with the current standard treatment in this setting, receiving AstraZeneca's PD-L1 inhibitor Imfinzi (durvalumab) as a maintenance therapy for one year. The investigators drew blood at six time points throughout the treatment period and profiled it for CTCs using a GO chip containing anti-EPCAM, anti-EGFR, and anti-CD133 capture antibodies. In addition to isolating the CTCs, they also molecularly characterized the captured cells and identified signatures that are predictive of progression-free survival.
In addition to EPCAM, Nagrath said her team chose EGFR because it is expressed by most lung cancer cells and CD133 because it is a commonly used marker for lung cancer stem cells. They felt that the combination of those three markers would "capture the entire gamut of [lung cancer] cells on the chip, which are not just epithelial tumor cells but also the cells that might be downregulating their epithelial phenotype," Nagrath said.
Jolly noted that patients in the clinical study were not screened for EGFR mutations, since that is not routinely done for stage III patients. "When they metastasize, it's at that point that we will get a lung cancer sample to see what different mutations they have such that we can guide therapy better," Jolly said.
Although patients with EGFR mutations may have different outcomes within the study, Jolly doesn't expect that their mutation status would impact their CTC results.
Nagrath and Jolly saw a statistically significant decrease in CTC counts among patients at week one, week 10, and week 30 during treatment compared to their pre-treatment baseline CTC count. Additionally, among patients who had samples taken at weeks 18 and 30 in conjunction with immunotherapy, 69 percent had a decrease in CTCs following the transition from chemotherapy to immunotherapy. They noted that "interestingly," there was no correlation between visible tumor volume and absolute CTC quantity at any time point.
The study authors found that changes in CTCs at week four correlated with progression-free survival. Specifically, patients whose CTCs increased at four weeks of treatment or decreased less than 75 percent were more likely to progress, while those whose CTCs decreased more than 75 percent did not progress at an average follow-up of about a year.
Upon extracting RNA from the captured CTCs and analyzing it via microarray, the researchers identified 97 differentially expressed genes between the pre-therapy and four-week samples. The CTCs from the four-week samples showed greater expression of genes associated with aggressive or proliferative phenotypes, including variations in RAB27A, TXN, ANXA3, and S100A12, suggesting that tumor cells surviving the first four weeks of chemoradiation therapy may be more aggressive and more metastatic.
They did not find differentially expressed genes when comparing pre-therapy samples and samples taken at 10 weeks of therapy. However, at the 10-week time point, expression of MTOR was predictive of shorter survival.
Nagrath and Jolly are now planning to validate these results in a larger patient group with the goal of implementing a GO chip-based liquid biopsy test in clinical studies and potentially using it as a tool to guide treatment for lung cancer patients. They also hope to separately validate the gene signatures identified in the study as predictive biomarkers in the same setting. Nagrath noted that the technology also has potential in early cancer detection.
Given the features of the GO chip, Nagrath believes it may offer a higher sensitivity than competing technologies, such as Menarini Silicon Biosystems' EPCAM capture antibody-based CellSearch system. "[CellSearch] has a low success [rate] in lung cancer because a lot of lung cancers downregulate these EPCAM molecules," Nagrath said, adding that the GO chip could also compare favorably to size-based microfluidic CTC technologies. The GO chip hasn't yet been tested head-to-head against these other technologies.
A Menarini spokesperson responded that the data reported in Nagrath and Jolly's publication are consistent with the prognostic significance of CTC enumeration across solid tumors using the CellSearch platform. "Numerous studies have demonstrated that when EPCAM is detected with the CellSearch system, this population has an inferior prognosis, poorer response to therapies, and shorter overall survival," the spokesperson said but also acknowledged that "EPCAM expression on CTCs is not ubiquitous in advanced NSCLC and is found in roughly 30 percent of patients."
The GO chip is currently offered as part of the liquid biopsy resources at the University of Michigan's Rogel Cancer Center and is being implemented in other clinical studies involving patients with solid tumors. "For commercialization, we are still looking out for good partners who would be interested in licensing this technology and developing it into a product," Nagrath said.