PHILADELPHIA (GenomeWeb) – Several research groups using Epic Science's circulating tumor cell analysis platform presented data at the annual meeting of the American Association for Cancer Research here this week, ranging from new applications of the technology for sequencing individual CTCs to pharmaceutical research data hinting at future companion diagnostic applications.
Unlike many methods for isolating CTCs from plasma or serum, Epic's technology does not rely on antibody-based capture. Instead, the company lyses red blood cells from a sample and then spins the remaining nucleated cells into a pellet, which is then resuspended and spread out into a monolayer in which each cell is fixed in a particular location.
In this way, the technique creates a slide containing all of the normal circulating cells as well as any cells originating from a patient's tumor — a snapshot of the cellular content of the blood at the time of collection.
Epic President and CEO Murali Prahalad told GenomeWeb that a single blood tube generates about 12 replicate slides of this type. "We only use about two [slides] on average for assessment and then we can go back to the replicates for a single time point much later on if we need to. It's basically the FFPE equivalent for the blood," he said.
Once a slide is created, the company then uses a combination of immunofluorescent staining and a complex morphological analysis enabled by a high-speed scanner and computer software to pick out which cells in the grid are CTCs against the large background of normal cells.
At this point in the process, Epic can also apply a variety of other specific tests, not just to distinguish CTCs, but to identify particular biomarkers of interest. Individual CTCs can also be picked off the slides using automated micropipetting, and once the cells are isolated on their own, they are available for even more analyses, including genomic biomarker detection and even whole genome sequencing.
This downstream genomic analysis of isolated CTCs was the subject of one poster presentation at the meeting, and is the first public data describing this capability of Epic's platform.
In the study, company researchers led by Stephanie Green used the Epic platform and an Illumina HiSeq to perform next-generation sequencing to assess copy number variation in isolated CTCs.
The group took two healthy blood samples and spiked them with one of three different prostate cancer cell lines, creating six experimental samples in total. They then processed the samples using the Epic platform, and then isolated, lysed, whole genome amplified, and sequenced the identified tumor cells in replicate.
According to the authors, all the CTCs processed yielded successful sequencing libraries. Cells from duplicate samples and sequencing runs also demonstrated good correlation in their CNV profile, and they were positive for previously known CNVs among the three cell lines, including AR amplification, PTEN deletion, and Y chromosome null status.
Methods for comprehensive or targeted sequencing of tumor DNA in the blood have abounded in the last few years, but Epic believes its ability to perform genomic analysis at the single CTC level, rather than on an aggregate of circulating cell-free DNA or pooled CTCs, is a significant advantage.
"One of the biggest emerging issues in cancer is understanding temporal heterogeneity," Prahalad said. "Most existing methods look at cancer in terms of average measurements, for example in tissue biopsy you have a sample of cells and you do a genomic assessment averaging the different cells present. And in cell-free DNA you have a bulk measure of cellular turnover and [that is] an averaging in some sense of whatever may be there."
Prahalad added that Epic's is the only technology "that can unambiguously look at how morphology and other elements conform to the genomics of individual cells."
On the other hand, he admitted, sequencing a single cell also comes with limitations, namely the necessity of whole genome amplification, which can create biases that interfere with measurement of genome changes at a more detailed level.
However, according to Prahalad the CNV differences that the group showed it could assess accurately are by themselves an important initial measure of tumor heterogeneity.
"If you are trying to see if you have single clonal populations or multiple subclones you have to worry about, CNV is the important starting point," he said.
In the future, Epic is planning on to take advantage of the reproducibility of the biases that WGA introduces in order to develop focused mutational panels with robust coverage that can be performed on single CTCs isolated using its system.
Future CDx potential
In three other posters presented at the meeting, researchers also provided a glimpse of the fruits of some of Epic's behind-the-scenes work with pharmaceutical and academic partners in the context of specific biomarkers with clinical potential.
Prahalad told GenomeWeb that the company has worked with pharma partners mainly on protein biomarkers that can be analyzed over an entire slide, but also in some cases genomic analyses of isolated CTCs. With academic leaders it has also been looking at applications of the technology to both cancer genomics and proteomics.
In two related posters at the conference, researchers from Epic and Pfizer shared data from an ongoing collaboration looking at the expression of the fetal oncogene 5T4 in CTCs.
As part of a larger study in which Pfizer is investigating whether 5T4 expression can predict cancer patients' response to a 5T4-targeting antibody drug conjugate (ADC) currently in clinical development, it has partnered with Epic to develop a CTC-based 5T4 test.
Steven Pirie-Shepherd, a Pfizer researcher and the first author of one of the two posters, told GenomeWeb that as the company tests this drug further, it plans to collect matched tumor and CTC samples to evaluate if one or the other, or potentially both, may be a sensitive predictor of response to therapy using this ADC.
As described in one poster at the meeting, the researchers collected 45 blood samples from breast, ovarian, and non-small cell lung cancer patients who were on active therapy, processed them using the Epic platform, and analyzed both CTC number and 5T4 expression alongside pathology, staging, and treatment information.
Among non-apoptotic CTCs, increased 5T4 was detected in six of 10 breast cancer patient samples, with the highest percent seen in patients with corresponding ER-positive primary tumors.
CTCs expressing 5T4 were also similarly detected in a subset of the lung and ovarian cancer patients, the authors wrote. Overall, 5T4 expression was higher in NSCLC patients than in breast cancer or ovarian cancer.
In a second poster, Pirie-Shepherd and the Epic researchers reported on the results of work comparing 5T4 assays created for both CTCs and tumor tissue in a small cohort of patients with NSCLC.
Overall, the team found that the expression of 5T4 was robust and measurable in both the primary tumor and circulating tumor cell compartments. In addition, the researchers saw expression of 5T4 in both adenocarcinoma and squamous cell carcinoma, in all stages and grades of tumor, and with no specific correlation between expression and stage, grade, or pathology, the authors wrote.
Interestingly, they also found that there was no concordance between 5T4 expression in the primary tumor and the circulating tumor cell compartment, suggesting that one or the other, not both, is likely to be more closely linked to response to the investigational Pfizer ADC.
According to Pirie-Shepherd, its not clear yet whether measuring 5T4 in CTCs or in tumor tissue will be a better predictor of patient response, only that there appears to be discordance between measurement of marker in tissue biopsy and in circulation.
In upcoming clinical trials of its 5T4-targeted ADC, Pfizer is planning to measure 5T4 expression in CTCs using the Epic platform, and in tumor samples using an equivalent assay. As these clinical trials progress, the question of which sample type can best predict response will hopefully be answered.
Academic partnerships
Epic was involved in more than 35 clinical trials with 17 different pharmaceutical companies as of last year. But according to Prahalad, research collaborations with academic groups are also a main focus for the company.
In a separate poster at AACR, researchers from Epic and Memorial Sloan Kettering Cancer Center shared data on the development and early clinical validation of a method to measure AR N terminal, AR C terminal, and both AR-V7 protein and AR-V7 mRNA in CTCs in metastatic castration-resistant prostate cancer.
Resistance to androgen signaling directed therapies, including Abiraterone and Enzalutamide, is associated with the presence of the AR-V7 splice variant and, according to the authors, is also linked to other AR ligand binding domain alterations (LBDAs).
In light of evidence that AR-V7, as well as AR N and AR C, show marked heterogeneity in different cell types and differing protein expression in CTCs of prostate cancer patients, the group set out to use Epic's platform to understand the variability of expression of these markers in individual CTCs.
The team spiked cell lines expressing varying levels of these LBDAs into healthy blood samples and also collected 24 patient samples for CTC analysis using the Epic platform. They analyzed both sets for AR N, AR C, and AR-V7 immunofluorescence, and also looked at AR-V7 mRNA by FISH in a subset of samples.
According to the group, immunofluorescent expression of the three markers in the cell lines was consistent with published profiles. Meanwhile, patterns of the markers among CTCs in the patient samples led the team to conclude that measuring AR C may be more sensitive than AR-V7 alone.
Epic is not the only company interested in developing an AR-V7 test to potentially guide prostate cancer treatment. Researchers from Johns Hopkins are also currently working on a CTC AR-V7 test that they plan to offer as a laboratory-developed test.
The Hopkins group's method involves CTC isolation and enrichment using the AdnaTest Prostate Cancer Select kit from AdnaGen and custom designed multiplexed qRT-PCR primers to detect both full-length-AR and AR-V7.
In parallel, the Hopkins team is also involved in a collaboration in which they have licensed IP related to AR-V7 testing to Tokai Pharmaceuticals, which is working with Qiagen to develop a non-invasive AR-V7-based companion diagnostic for its investigational castration-resistant prostate cancer drug galeterone.
Prahalad did not provide any details about Epic's pipeline for clinical test commercialization. But the company recently received CLIA certification for its cancer diagnostics laboratory.
"It's a key milestone for us because it allows us to do a couple of things," Prahald said. "First, it allows us to run interventional clinical trials … to move from observational trials with our pharma partners in which we are looking at mechanisms of action and response in subpopulations based on CTCs, to being able to develop classifiers for responders versus non-responders and be the selection device for interventional trials going forward."
Secondly, he said, CLIA certification opens the door for Epic to take some of its promising academic partnerships and validate those projects toward developing its first clinical tests. "When the time is ready we now have the commercial outlet to offer those tests," he said.