Slipping by unnoticed is what circulating tumor cells (CTCs) do unless they are subjected to detection, isolation, and capture technology, which is currently wrangling with the task of enumeration, that is, the generation of CTC counts—not just grand totals (for CTCs of all kinds) but subtotals (for different kinds of CTCs).
As challenging as enumeration is proving to be, many scientists are already developing technology that does more with CTCs. Don’t just count CTCs, these scientists insist, preserve them for downstream applications. Culture them. Analyze their contents systematically. Use them to advance precision cancer medicine.
To accomplish these goals, scientists are trying to treat CTCs more gently. For example, CTCs that are separated from normal cells by microfluidic devices may be spared bruising experiences such as being targeted by immunoaffinity assays or marked by immunostaining procedures.
Several technologies capable of treating CTCs more tenderly were presented at a recent conference, “Circulating Tumor Cells: Understanding Their Biology and Clinical Significance.” This event, which was organized by the Cambridge Healthtech Institute and held in Washington, DC, emphasized that complexities such as CTC heterogeneity, with respect to both molecular markers and biophysical properties, need to be better understood if CTCs are to guide diagnoses and predict patient outcomes.
CTCs, or cells shed by a primary or metastatic tumor into the vasculature or lymphatics, are present at very low levels. Some types of CTCs are believed to be responsible for the spread of cancer, and they may become targets for individualized cancer treatments. However, little is known of the exact composition of CTCs, or the characteristics that distinguish CTCs of different types from each other. Detailed characterization of CTCs could lead to a better understanding of the drivers of metastasis.
Methods of confirming whether a cell is really a CTC or not were discussed by Siva A. Vanapalli, Ph.D., professor, chemical engineering, Texas Tech University. Existing methods, Dr. Vanapalli complained, are destructive to cells. These methods make use of immunostaining, which opens the cell membranes, killing them, ruling out their use in downstream applications.
Dr. Vanapalli suggested an alternative approach in his presentation, Microfluidics and CTCs: Detection, Metastatic Insights and Drug Testing. Microfluidics, he pointed out, may be used for label-free detection of tumor cells in blood. He described how his group combines inline digital holographic microscopy (DHM) with machine learning technology. The group uses inline DHM to obtain a fingerprint of every cell that flows through a microchannel, and it uses machine learning to distinguish tumor cells from background blood cells.
With inline DHM, a laser is directed onto the cell. The cell scatters the beam, and the interference of the incident laser beam and the scattered beam gives rise to diffraction patterns around the cell.
“We’re using light to look at the scattering pattern of the cell, and we use that to decode whether it’s a CTC or not,” explained Dr. Vanapalli. Early studies of cancer cells spiked into blood samples have shown that inline DHM can detect the cancer cells and differentiate them from other blood cells.
Another potential label-free identification approach is based on the deformability of the cell. When the cells are shed from the primary tumor and enter the circulatory system, they pass through narrow capillary vessels. The cells also deform when they pass through microfluidic devices, which may be used to obtain measures of deformability—measures that may serve as CTC markers.
According to Dr. Vanapalli, the key challenges of CTC analysis may be summarized as follows: First, figure out how to isolate CTCs in a label-free way. Second, figure out how to do drug assays on these cells. “That’s how we can really benefit the patient,” maintained Dr. Vanapalli. “I’m skeptical of using CTCs as a diagnostic tool, but they can have significant prognostic value.”
CTCs and background cells differ with respect to size, a quality that several laboratories are trying to exploit in novel detection, separation, and identification technologies. Size-based technologies could replace conventional alternatives such as immunoaffinity assays, which use antibodies to bind to antigens expressed on cell surfaces. The antibody-antigen interactions allow cells to be removed from the blood and concentrated. Size-based technologies might also be preferable to other biophysical methods, such as those which rely on density gradients or electrical gradients.
At the “Circulating Tumor Cells” conference, the advantages and disadvantages of size-based technologies were highlighted by Richard J. Cote, M.D., professor of pathology, biochemistry, and molecular biology, University of Miami. His presentation, Capture, Interrogation, Imaging, Automated Analysis and Culture of CTCs: Strategies for the Development of a Transformative Tool to Understand Cancer, described a microfilter approach for separating CTCs.
He pointed out that cancer cells from solid tumors are larger than almost all normal cells found in the blood. According to Dr. Cote, normal red blood cells range from 5 to 7 microns, whereas tumor cells are 15 microns or more.
“What my lab developed,” said Dr. Cote, “was a way to produce microfilters that create a very standard and consistent pore size and distribution that allows for the capture of tumor cells, but almost all normal cells in the blood pass through.” Dr. Cote asserted that his method will capture 90 to 95% of tumor cells in a blood sample. The microfilter is made of a plastic material called parylene, which is optically transparent.
“Once we capture a CTC,” he explained, “we can analyze it on the filter, just as you analyze a microscopic slide.”
Unlike affinity-based technology, which is specific to a particular kind of cancer, Dr. Cote’s technology can be used for any type of solid tumor. His group has tested it in breast, colorectal cancer, kidney cancer, melanoma, prostate cancer, lung cancer, and other cancers.
In addition to capturing tumor cells, the filter can capture cancer-associated fibroblasts (CAFs). “Circulating CAF cells are associated with the stage of disease,” said Dr. Cote. “In addition to that, in a mouse model of breast cancer, we have shown that the presence of CTCs and circulating CAF cells is associated with progression of disease. The highest association of progression is when you find the co-clusters of CTCs and circulating CAF cells together in association with one another.”
Dr. Cote said that his group is investigating ways to automate the technology and developing novel methods to grow CTCs from patients’ blood.
Another approach to size-based separation of CTCs was described by Sunitha Nagrath Ph.D., associate professor of chemical engineering, University of Michigan. Dr. Nagrath said that when her group looked at size-based separations for CTCs, it determined that the methods that are being used, such as centrifugation, lack continuous throughput. Other size-based methods, such as filtration, have disadvantages such as clogged pores, pressure issues, and low throughput.
In her presentation (Microfluidic Labyrinth Chip for Monitoring Cancer Stem Cells), Dr. Nagrath described how she developed an alternative size-based method. She developed a microfluidic chip that was inspired by a legendary labyrinth, the one used to imprison the Minotaur. Dr. Nagrath’s labyrinth slows down the larger cells while letting the smaller ones move through more quickly.
The labyrinth is 500 microns wide and 100 microns high. It has 11 loops, 56 corners, and a total channel length of 637 nm. The channels, which have lengths ranging from 100 to 700 microns, allow blood to flow through at 2.5 mL/minute. “That’s a huge throughput for a microfluidic device,” Dr. Nagrath emphasized.
The labyrinth design includes curved channels and sharp corners to create a focusing effect. “Bigger cells experience different forces than smaller cells,” Dr. Nagrath explained. “Based on the forces they experience, they get different streamlines in the device, and we collect cells from the different streamlines.”
A CTC analysis on blood samples from 20 patients with pancreatic cancer yielded about 50 pancreatic CTCs/mL of blood, with less than 2 CTC-like cells/mL in healthy control samples. In another test in breast cancer patient samples, the average number of CTCs was 9.1/mL.
Dr. Nagrath said that an important advantage of the technology its ability to “focus” cells, achieving CTC isolates of higher purity by excluding contaminating cells. Applications for the isolated CTCs thus far have included looking for mutations in lung cancer for potential targeted therapies and expansion of the cells.
At present, Dr. Nagrath is working on incorporating the labyrinth device into some clinical studies. She said that the University of Michigan has two ongoing clinical trials that incorporate CTCs as a biomarker. Dr. Nagrath added that her group would also like to establish collaborations in breast and prostate cancer.
RNA-Based CTC Signatures
In assessments of cancer, a well-known kind of liquid biopsy focuses on circulating tumor DNA (ctDNA). This kind of liquid biopsy is attractive for several reasons, not the least of which is the ease with which ctDNA can be isolated. Although CTCs are harder to isolate, they can provide a basis for liquid biopsies, too. Despite this difficulty, CTC-based liquid biopsies may offer advantages over ctDNA-based liquid biopsies.
This possibility was discussed by David Miyamoto, M.D., Ph.D., assistant professor of radiation oncology at Harvard Medical School and Massachusetts General Hospital. In his presentation (RNA-Based Circulating Tumor Cell Signatures for Precision Cancer Medicine), Dr. Miyamoto emphasized that ctDNA- and CTC-based liquid biopsies could complement each other.
According to Dr. Miyamoto, a limitation of using ctDNA is that it offers a limited view. It looks only at DNA. Using CTCs, however, offers access not only to DNA, but also to RNA and protein. “You get a better understanding of the biology of the tumor as it enters circulation,” Dr. Miyamoto insisted.
In collaboration with a multidisciplinary team of bioengineers, molecular biologists, and clinician scientists at Massachusetts General hospital, Dr. Miyamoto’s group developed a microfluidic technology to isolate CTCs from blood. “Not all CTC isolation technologies are equal,” Dr. Miyamoto noted. “One of the main limitations of the technologies out there is that RNA is often not very well preserved in CTCs.” RNA is a labile molecule that degrades quickly.
The Massachusetts General Hospital scientists developed the CTC iChip, a microfluidic device that separates CTCs while leaving the cells intact and unlabeled. By sparing the CTCs rough treatment, the CTC iChip preserves CTC RNA.
In a recent study in prostate cancer, the scientists developed an assay using digital PCR to study specific genes expressed in prostate cancer. In patients with metastatic castration-resistant prostate-cancer, they were able to use CTCs to predict which patients would do well when treated with abiraterone. “We found that this had very good predictive ability in determining up front in patients whether or not abiraterone will work,” asserted Dr. Miyamoto.
That was proof of principle, Dr. Miyamoto said. The next step is to try it with other drugs and replicate the results in larger cohorts. Dr. Miyamoto reported that in men with metastatic prostate cancer that had already spread mostly to the bone, the number of patients with a detectable CTC signal was low.
“In those few patients who had a high CTC score, there was actually a higher incidence of pathologic spread of cancer beyond the prostate found at the time of surgery,” Dr. Miyamoto said. That suggests that CTC scoring could be used to predict ahead of time whether patients have early dissemination of disease and need more aggressive local therapies, or additional systemic therapies up front.
Nuclear AR-V7 Protein Expression in CTCs Validates as a Predictive Biomarker
In the last eight years, five new FDA-approved drugs have demonstrated an improvement in overall survival (OS) for patients with metastatic castration resistant prostate cancer (mCRPC), yet most patients only receive two or three of these therapies.
The question of how to sequence the therapies to achieve the best possible outcome in an individual patient is one of the most important clinical decisions faced by physicians treating mCRPC. In particular, doctors need to decide which patients should receive a less toxic, oral androgen receptor (AR)-directed therapy and which ones should receive a more toxic, cheaper, IV chemotherapy.
To address this question, Epic Sciences has developed a nuclear AR-V7 blood-based test that has the ability to predict which patients will not respond to AR-directed therapy and have better OS with chemotherapy. The test examines the expression of the AR-V7 protein in the nucleus of circulating tumor cells (CTCs) in a peripheral blood draw from patients at the time of therapy change. The test is powered by the Epic Sciences No Cell Left BehindTM technology that is precise enough to identify and characterize all the CTCs in a sample, and link their response to a range of drug classes.
A study to clinically validate AR-V7 was performed by Howard Scher, M.D., and his team at Memorial Sloan Kettering, along with other collaborators. Results showed that nuclear localized AR-V7+ patients had longer OS when the patient switched to chemotherapy. Patients who were negative had equal or better survival when treatment with an AR-directed therapy was continued. Importantly, the test has been validated as a predictive biomarker for therapy selection in patients with mCRPC.
The test, available through Genomic Health and Epic Sciences as Oncotype Dx AR-V7 Nucleus Detect, has received draft coverage from Centers for Medicare & Medicaid Services (CMS) and is expected to be the first broadly utilized and reimbursed CTC test for clinical decision making in oncology.
Ryan Dittamore is chief of medical innovation at Epic Sciences.