UCSF

The epithelial-mesenchymal transition (EMT) plays a key role in invasion and metastasis. Recent studies have demonstrated that many discrete states exist along the EMT spectrum, with partial or hybrid EMT phenotypes as the most metastatic. These states have variable plasticity, must compete for the proliferative niche, and also remodel each other’s microenvironment which modifies each other’s behavior. How the steady state distributions of these populations can be maintained, and how different EMT states compete and cooperate to define population-level phenotypes, has not been explored. In order to understand the contribution of this spectrum of states to the behavior of an entire heterogeneous population, we use a model transformed cell line that exhibits partial EMT cell states. In this cell line, a partial early-EMT state spontaneously gives rise to a later-EMT state with no evidence for reversion. The early-EMT cell population dominates the mixture while later-EMT cells exist in smaller fractions. We refer to this as a “top-heavy” lineage hierarchy to contrast it with healthy stem cell lineages where the populations is dominated by more differentiated progeny. A mathematical model suggests that in the top-heavy regime, differential proliferation limits the emergence of slower growing late-EMT cells or further differentiated progeny. The model also predicts perturbations that destabilize the top-heavy regime, and we validate the predictions experimentally. Intriguingly, we find that fully mesenchymal states emerge after destabilization of the top-heavy regime. These fully mesenchymal cells are more drug resistant, migratory, and grow in soft agar. Further, these cells dramatically modify their microenvironment which promotes a more mesenchymal and migratory phenotype in the more abundant partial EMT cells. These findings highlight the importance of studying the population-level dynamics that regulate cell-state proportions as well as the interactions among states that determine their collective behaviors.