Metastasis involves cycles of forward and backward transitions between epithelial and mesenchymal phenotypes (EMT and MET). Cells undergoing these transitions often attain a hybrid epithelial/mesenchymal (E/M) or partial EMT phenotype that allows collective cell migration and presumably is responsible for observed clusters of Circulating Tumor Cells (CTCs). These clusters possess heightened tumor-initiating potential, maximum phenotypic plasticity, and thereby facilitate dissemination, thus acting as primary ‘bad actors’ of metastasis.

Using a detailed theoretical framework to mathematically model microRNA-mediated translational regulation, we show that the core EMT network of miR-200/ZEB feedback loop can act as a three-way switch for transitions between the three phenotypes – E (high miR-200, low ZEB), M (low miR-200, high ZEB) and E/M (medium miR-200, medium ZEB). Next, we couple miR-200/ZEB with LIN28/let-7, a key part of the stemness decision-making circuit. We find that the hybrid E/M phenotype can be highly likely, sometimes even more so than complete EMT, to gain ‘stemness’; generically, the ‘stemness window’ lies somewhere midway on the EMT axis.

Our results offer a unifying mechanistic explanation for the emerging notion that partial EMT is a likely path to metastatic disease.