Cancer cells with hybrid epithelial and mesenchymal phenotypes for deformability-based classification

Abstract public domain (NIH RePORTER)

Pancreatic ductal adenocarcinoma (PDAC) is one of the deadliest tumors, highlighting the need for tools and prognostic markers capable of detecting metastasis and drug resistance. While circulating tumor cells (CTCs) from liquid biopsies offer prognostic information on disease progression, metastasis, and therapy response, their heterogeneity limits detection by immunofluorescence imaging, especially for so-called hybrid CTCs (H-CTCs) that express epithelial and mesenchymal markers to varying levels. Since CTCs with increasing mesenchymal expression exhibit progressively higher deformability levels than those with epithelial expression, this project seeks to advance single-cell deformability as a complementary metric to immunofluorescence for quantifying heterogeneous H-CTCs over a progression of epithelial to mesenchymal transition (EMT) states. For this purpose, a perfusable system for co-culture with cancer associated fibroblasts will be developed to generate PDAC cells expressing a progression of hybrid mesenchymal and epithelial states over varying levels through modulating secretory factors under shear stress and hypoxia conditions. The EMT states of PDAC cells validated by immunofluorescence will be classified on single-cell biophysical metrics obtained under microfluidic viscoelastic extensional flows to assess single-cell deformation and recovery dynamics. Using microgels of known stiffness levels as co-flowing standards, multilayer perceptron-based neural network training approaches will be used to learn the non-linear relationships between deformability metrics of H-CTCs over a progression of EMT states. In this manner, heterogeneous samples of H-CTCs will be generated at concentration levels and in sample matrices that resemble their levels in portal vein samples for their classification on deformability metrics and downstream validation against immunofluorescence for utilization in disease prognosis. This will lay the foundation for future work on microfluidic enrichment of CTCs from liquid biopsies to detect their heterogeneity, using portal vein samples obtained from a splenic injection mouse model of liver metastasis with implanted GFP- expressing cancer cells, so that the shed CTC numbers from bioluminescence at various disease stages under circulation, micro metastases and tumor progression can be compared to the heterogeneity of their EMT states.

Public health relevance public domain (NIH RePORTER)

Prognostic information on the progression of pancreatic ductal adenocarcinoma (PDAC), its metastasis, and its response to therapies can be obtained based on the shed circulating tumor cells (CTCs) in liquid biopsies from the portal vein. However, CTC detection by immunofluorescence imaging is limited by the heterogeneity of CTCs that can express epithelial and mesenchymal markers to varying levels. To advance the consideration of single- cell deformability as a complementary metric to immunofluorescence for quantifying CTCs over a progression of epithelial to mesenchymal transition (EMT) states, this project will develop systems to generate pancreatic cancer cells presenting this hybrid phenotype for multimodal deformability and immunofluorescence validation.