Abstract 1562: Clusters of circulating tumor cells were selectively isolated in the blood of 12/12 epithelial ovarian cancer patients using facile gravity-flow-based filtration method adapted to clinical use

Abstract

Abstract Background The presence of circulating tumor cells (CTCs) in blood is correlated with disease progression in many cancers. Their prognosis value in ovarian cancer is still under debate.1 CTCs are heterogeneous in size and marker expression, and sub-populations with various metastatic potentials have been identified. CTC clusters, although rare and difficult to isolate, have emerged as a possible driver of metastasis owing to ~50-time higher metastatic potential than single CTCs.2 Only few methods have emerged to capture clusters, and are often complex and cumbersome, limiting our understanding of the role of clusters in metastasis. Here, we present a new filtration method for the selective capture of CTC clusters from blood and found clusters in 12/12 epithelial ovarian cancer (EOC) patients. Method Cluster capture was performed by filtration using a 3D printed cartridge3 and filters4 with pore diameters of 8, 10, 12, 15, 20 or 28 μm. We developed a gravity-driven process, generating reduced shear stress, and optimized capture using blood (1:6, v/v, in PBS) spiked with OV-90 and OVCAR-3 ovarian cancer single cells and clusters. Blood samples from 12 EOC patients were filtered. Clusters can be stained and imaged on the filter, or released for downstream analysis. Results Using the gravity-setup, we were able to selectively capture clusters with good integrity and with a rate that outperforms other technologies to the best of our knowledge. Viable CTC clusters, with 2 to >100 cells, were captured from 12/12 EOC patients. Their size distribution was surprisingly similar between patients. Small clusters (2-3 cells) were the most frequent, and this frequency decreased as their size increased. The molecular characterization of the captured clusters revealed a low and localized, heterogeneous expression of EpCAM (epithelial cell adhesion molecule), in combination with a widespread expression of c-MET (hepatocyte growth factor receptor) in all patients, suggesting a mesenchymal-like profile. Conclusion Using the gravity-filtration setup, CTC clusters were captured from the blood of all patients tested, suggesting that clusters are much more widespread than anticipated, and are in fact the norm rather than the exception. The cluster size distribution was conserved between patients with small clusters dominating, and some rare, very large clusters. Cluster staining revealed a mesenchymal profile, in agreement with a higher metastatic potential. Together, these results suggest that clusters should significantly contribute to disease progression, a hypothesis, which may be explored using our facile and selective method.