A minimal Markov state model to quantify epithelial-mesenchymal transition dynamics

Abstract

Epithelial-to-Mesenchymal transitions in the context of cancer are considered drivers of tumoral plasticity and hallmarks of cancer metastasis, resistance to therapy, and relapse. Although classically seen as processes induced by specific extracellular inputs, recent findings pointed to the question of whether and to what extent transitions could happen spontaneously within the tumor population. Here, we address this issue by comparing EMT dynamics in lung cancer cells under normal growth conditions and in the presence of TGF-β, a known EMT inducer. Combining time-course flow cytometry measurements with stochastic modeling, we found that (i) untreated cells undergo density-independent EMT. In particular, (ii) the long-term populations present a combination of epithelial, hybrid, and mesenchymal states, where (iii) the hybrid state acts as a transient phenotype across all experimental conditions. Instead, the spontaneous dynamics differ from those observed after the withdrawal of the treatment, suggesting (iv) the presence of hysteric effects due to EMT induction. Finally, we discuss how our minimal quantitative framework may provide insights into the structure of the Waddington landscape associated with the epithelial-mesenchymal plasticity.