NEW YORK (GenomeWeb) – Authors from Epic Sciences and Genentech published a study this week in PLoS One describing their development and validation of a methodology for single-cell sequencing to identify clonal heterogeneity in circulating tumor cells.
Speaking to the ultimate utility of this approach, the company also highlighted a poster this week that was presented this October at the annual congress of the European Society for Medical Oncology, which demonstrated that a phenotypic CTC-based biomarker (which Epic developed by matching CTC genomics to their physical features) can identify patients with metastatic castratation-resistant prostate cancer who are more likely to respond to treatment with a PARP inhibitor.
The company has been exploring single cell-CTC sequencing for several years, and first discussed data on its methods at the 2015 annual meeting of the American Association for Cancer Research.
In the new study in PLoS One, Epic demonstrated that in addition to profiling genome-wide copy number variation from single cells, it could also characterize the extent of genomic instability within individual CTCs.
Epic first reported on how this type of analysis could be translated to develop phenotypic biomarkers earlier this year at the annual meeting of the American Society of Clinical Oncology. In that presentation, researchers showed that they could train a classifier based on physical features of isolated CTCs compared to their genomics, which could pick out cells that had high genomic instability. They could then predict which metastatic prostate cancer patientswould have poor outcomes with anti-androgen treatment and taxane chemotherapy using the phenotypic features of cells alone.
In the newer poster at ESMO this fall, Epic, along with academic collaborators, studied samples from the NCI 9012 trial examining the efficacy of Abbvie's investigational PARP inhibitor veliparib in combination with the androgen receptor signaling inhibitor abiraterone. The researchers took their earlier phenotypic predictor linked to CTC genomic instability and then adapted it for prediction of PARP inhibitor response using response data from the trial.
The investigators showed that they could use the resulting phenotypic marker to predict patients' response to therapy. Those who had CTCs that fit the phenotypic profile had higher response to the combination of abiraterone and veliparib than to the hormonal therapy alone. There was a 93 percent overall response rate with the combination drug compared to 22 percent with the single androgen therapy in patients whose CTCs demonstrated the physical features in question.
Having a particular phenotype isn't necessarily the base mechanism whereby cancer cells are sensitive to PARP inhibition. Rather, this physical marker represents genomic features like homologous recombination DNA repair deficiency (HRD), which in turn confer PARP inhibitor response.
But based on the study results, according to Epic, it looks like phenotype alone can yield a sensitive prediction of response to these drugs without the need for genomic analysis of CTCs or tissue DNA.
PARP inhibitors, which belong to one of the largest classes of new cancer drugs in development, target genomic instability associated with HRD. In prostate cancer, and in other cancers, not all patients respond to PARP inhibition, and so strategies to identify which patients harbor HRD are gaining a lot of attention as potential companion or complementary diagnostics to guide use of these new drugs.
Tissue-based genomic methods are also being investigated to predict patient response to PARP inhibitors. Myriad Genetics, for example, recently filed the first module with the US Food and Drug Administration for a rolling premarket approval application for its myChoice HRD, which drugmaker Tesaro used in studies to identify best responders to its PARP inhibitor niraparib.
However, Epic says that it has collected data showing that its phenotypic CTC-based predictor appears to identify more patients likely to respond compared to what has been seen in experiments with tissue genomic approaches.
In the veliparib NCI 9012 trial discussed at ESMO, 29 percent of first-line mCRPC patients were positive for a new phenotypic CTC biomarker that the company defined specifically in regard to PARP inhibitor response. Epic argued that this is significantly higher that what has been seen with previous reports using sequencing of single-site, metastatic tissue biopsies to identify inactivating HRD mutations.
In addition to observing increased prevalence of biomarker-positive patients within the NCI 9012 cohort, the Epic researchers were also able to use their phenotypic marker to observe and track dynamic changes in HRD-positive CTCs throughout patients' treatment.
On average, patients receiving both abiraterone and veliparib saw a reduction of HRD+ CTCs during therapy. In contrast, patients given only abiraterone saw an increase in HRD+ CTCs, the investigators reported.
Ryan Dittamore, vice president of translational research & clinical affairs at Epic Sciences, and a co-author on the ESMO study as well as the PLoS One paper published this week, told GenomeWeb that Epic sees the ability to predict response to PARP inhibition without the need for relatively expensive and time-consuming genomic analyses as a potential "game changer."
"The feedback from both pharma and oncologists is that the time it takes to get the sequencing test back from tissue can be [too long]," he added. "When we think about the application, the ultimate goal of having a predictive test for PARP inhibitors, for us to be able to provide results back from a blood sample in a week or less, that seems incredibly enabling from a workflow perspective."
Although the ultimate product of its work in this area is a phenotypic predictor, Dittamore highlighted that it is the Epic platform's ability to interrogate the genomics of single cells, and then relate it to physical and chemical features, which has driven this work and will hopefully support other similar discoveries.
Coupling genomics in single CTCs with phenotypic characterization has also yielded other discoveries for Epic. For example, researchers from the company and from Memorial Sloan Kettering Cancer Center characterized phenotypes and copy number variation in CTCs from a set of patients with metastatic prostate cancer in a study presented at an ASCO conference in January.
The results of this paired analysis indicated that patients whose CTC population showed a high degree of heterogeneity — either in their phenotypes or their genetic clonality — did not respond as well to hormone therapy as those with less heterogeneity.
In the new PLoS One study, authors reported that their analyses of CTC copy number allowed them to detect a variety of genomic driver alterations and genomic instability events that may potentially affect patient selection and therapeutic efficacy.
"We observed both inter- and intra-patient heterogeneity across CTCs," the authors wrote, including a broad range of copy number changes in AR and PTEN in most patients analyzed.
"The increased resolution of single-cell CNV analysis allows for characterization of sub-clonal driver alterations, their clustering in sub-clonal populations … and the weighted average of targetable pathways within a single patient. These tools will greatly aid drug development and stratification of patients for therapeutic combination strategies and clinical trials," they added.
Dittamore could not discuss other similar areas where Epic may be able to harness its ability to assess both genomic heterogeneity of circulating tumor cells and their phenotypic features to generate promising biomarkers. However, the company believes the approach has broad potential, he said. To date, Epic has sequenced over 7,000 single CTC genomes from multiple cancer types.
"We are starting to partner up with many pharmaceutical companies developing PARP inhibitors to evaluate this signature in a variety of different studies and indications," Dittamore told GenomeWeb. Though he said he couldn't comment on companion diagnostics implications for this particular assay, he said that Epic is "focused on developing tools that can answer this question on a community level and access level that is unprecedented today."
"We think the data we have presented is supportive of going forward," he added.