Plastimos: a live cell imaging-based framework to study dynamics of EMT-mediated cellular plasticity in breast cancer

Abstract

Cancer cell plasticity, primarily mediated by the process of epithelial-mesenchymal transition (EMT), plays a critical role in promoting therapeutic resistance, tumor heterogeneity, and metastasis. EMT-mediated plasticity enables cancer cells to undergo molecular and phenotypic changes, which alter their invasiveness and resistance to treatments. This study aims to develop a quantitative, high-throughput system to assess EMT-mediated plasticity, which can inform therapeutic strategies for metastatic cancers. To accomplish this, we developed Plastimos, a semi-automated imaging analysis pipeline. This framework utilizes time-series images to track live cells through deep learning-based segmentation and employs a greedy algorithm to map cell trajectories, enabling the extraction of cellular phenotypic features. These features are used to study the EMT-mediated plasticity state of cells in response to the well-known EMT-inducing factors EGF and TGF-{beta}1. We selected two breast cancer cell lines, MCF7 and MDA-MB-231, representing classical epithelial-like and mesenchymal-like cell. The pipeline assigns a Plasticity Index based on various parameters, including motility, morphology, and proliferation, thus providing a quantitative estimate for the plasticity of deviating from the epithelial state. Our results indicate that epithelial-like cells respond to EMT-inducing factors at both molecular and phenotypic levels, while mesenchymal-like cells responses are only seen phenotypically. The Plasticity Index classifies cells along the EMT spectrum, converting to a Plasticity Score that reflects mesenchymal proportions. Availability and implementationThe pipeline implementation and the source code of Plastimos can be found at: https://github.com/Durdam/Plastimos.git