Modelling variability and heterogeneity of EMT scenarios highlights nuclear positioning and protrusions as main drivers of extrusion

Epithelial‐Mesenchymal Transition (EMT) is a key process in physiological and pathological settings (i.e.
development, fibrosis, cancer). EMT is often presented as a linear sequence of events including (i) disassembly of
cell‐cell junctions, (ii) loss of epithelial polarity and (iii) reorganization of the cytoskeleton leading to basal extrusion
from the epithelium. Once out, cells can adopt a migratory phenotype with a front‐rear polarity and may additionally
become invasive. While this stereotyped sequence can occur, many in vivo observations have challenged this notion.
It is now accepted that there are multiple EMT scenarios and that cell populations implementing EMT are often
heterogeneous. However, the relative importance of each EMT step towards extrusion is unclear. Similarly, the
overall impact of variability and heterogeneity on the efficiency and directionality of cell extrusion has not been
assessed. Here we used computational modelling of a pseudostratified epithelium to model multiple EMT‐like
scenarios. We confronted these in silico data to the EMT occurring during neural crest delamination. Overall, our
simulated and biological data point to a key role of nuclear positioning and protrusive activity to generate timely
basal extrusion of cells and suggest a non‐linear model of EMT allowing multiple scenarios to co‐exist.