Live Cell Imaging of Epithelial-Mesenchymal Plasticity during Collective Migration and Spatial Patterning in Monolayer Culture

Introduction: Epithelial cells dynamically reprogram their adhesive phenotypes to facilitate collective migration during tissue development and wound healing. For example, the classical epithelial-mesenchymal transition (EMT) is associated with the loss of cell-cell adhesion and apicobasal polarity, which can promote individual cell dissemination. More recently, partial or hybrid EMT states have been suggested that retain some cell-cell adhesions and coordinate multicellular motility as “leader cells.” However, profiling the functional phenotypes of partial EMT states remains challenging since they are transient and likely depend on local cell-cell interactions. Here, we use live cell imaging of EMT fluorescent reporters and traction force microscopy to elucidate the emergence of leader-like phenotypes and collective migration.

Materials and

Methods: Mammary epithelial cells (MCF10A) and breast cancer adenocarcinoma cells (MDA-MB-231) with a genetically encoded EMT fluorescence reporter (Z-cad) were a generous gift from MJ Toneff and JM Rosen (UT Southwestern). These cells express red fluorescent protein (RFP) driven by an E-cadherin (CDH1) promoter, which switches to green fluorescent protein (GFP) regulated by ZEB1 3’ UTR after EMT induction. Cells were further labeled with CellTracker FarRed to display partial states in the absence of RFP or GFP expression. Epithelial cells were treated with TGF-beta1 to induce partial or complete EMT at varying concentrations and durations (1-5 ng/mL for 7 days). Epithelial cells were then cultured in varying ratios with cells that had undergone partial or full EMT induction, then imaged using live cell microscopy. Cell morphologies were analyzed using Cellpose in each fluorescence channel, while collective migration was analyzed using digital image correlation. Cells were also plated on compliant polyacrylamide substrates coated with collagen I for traction force microscopy.

Results, Conclusions, and Discussions: We mixed epithelial, partial and full EMT cells at varying ratios and densities, which were imaged over time with live cell microscopy (Figure 1A). First, we analyzed cell morphology after various TGF-beta1 treatments to profile differences associated with fluorescent reporter expression and EMT state. We hypothesized that epithelial cells with RFP expression would be clustered with compact, cobblestone-like morphologies while mesenchymal cells with GFP expression would be dispersed with elongated, spindle-like morphologies. Although morphology and fluorescent protein expression were somewhat correlated, there were numerous “mismatched” examples where cells with mesenchymal morphology exhibited RFP expression and vice versa. Instead, we observed that mixtures of epithelial, partial, and full cells could spatially “sort” by type, which was modulated by overall density (Figure 1B). Thus, we comprehensively tracked cells in each fluorescence channel and classified trajectories into regions of epithelial; partial and full EMT. Interestingly, mixtures of epithelial and partial EMT exhibited trajectories with some intermixing, indicating homophilic contact behaviors between epithelial cells and partial cells. In contrast, mixtures of epithelial and full EMT cells exhibited trajectories that were hindered at interfacial boundaries due to heterotypic contact, resulting in segregated patterns (Figure 1C). This collective migration also exhibits differences in spatial patterns and temporal fluctuation of cell-generated tractions.
We find that collective cell migration is shaped by cell-cell adhesions, resulting in spatial patterning of distinct regions of epithelial; partial, and full EMT state. We envision that additional mechanical or biochemical stimuli can be used to dynamically reprogram EMT state with increased spatial and temporal resolution. This approach can yield new fundamental insights into epithelial-mesenchymal plasticity in microenvironmental context, as well as dynamic control of collective cell migration.

Acknowledgements (Optional): We acknowledge funding from NIGMS (R01GM140108) and the Brown University Hibbitt Postdoctoral Fellowship.