Picking winners in cell-cell collisions: wetting, speed, and contact

Abstract

Groups of eukaryotic cells can coordinate their crawling motion to follow cues more effectively, stay together, or invade new areas. This collective cell migration depends on cell-cell interactions, which are often studied by colliding pairs of cells together. Can the outcome of these collisions be predicted? Recent experiments on trains of colliding epithelial cells suggest that cells with a smaller contact angle to the surface or larger speeds are more likely to maintain their direction (“win”) upon collision. When should we expect shape or speed to correlate with the outcome of a collision? We build a model for two-cell collisions within the phase field approach, which treats cells as deformable objects. We can reproduce the observation that cells with high speed and small contact angles are more likely to win with two different assumptions for how cells interact: (1) velocity-aligning, in which we hypothesize that cells sense their own velocity and align to it over a finite timescale, and (2) front-front contact repolarization, where cells polarize away from cell-cell contact, akin to contact inhibition of locomotion. Surprisingly, though we simulate collisions between cells with widely varying properties, in each case, the probability of a cell winning is completely captured by a single summary variable: its relative speed (in the velocity-aligning model) or its relative contact angle (in the contact repolarization model). Both models are currently consistent with reported experimental results, but they can be distinguished by varying cell contact angle and speed through orthogonal perturbations.

Migration and division in cell monolayers on substrates with topological defects

“…Fibroblast polarity occasionally flips direction [4], which we model by stochastically reversing p with average flip period of ∼ 2.5 hours. We also assume cells tend to align to their past displacement [18][19][20], leading to some coordination between cell velocities, as in Fig. 1g.…”

“…Our approach also avoids issues with orientational anisotropy associated with lattice models like the Cellular Potts Model [45]. In principle, phase field approaches [20,[55][56][57][58][59] would also avoid lattice artifacts, but our scale of ∼ 3000 cells is an order of magnitude larger than typical applications of even simplified phase-field models [60][61][62]. Earlier papers have modeled elongated self-propelled objects with particle-field and/or Gay-Berne approaches [63,64], though without explicitly describing deformability.…”