Cell-cell adhesion holds tissues together. It is the glue that prevents epithelial sheets from falling apart, that maintains organ boundaries, that resists the forces of growth and movement. More adhesion should mean more cohesion, less movement, more solid-like behavior.
Nguyen, Bera, Notbohm, and Bi show that adhesion is not one thing but two competing things, and they push tissue fluidity in opposite directions simultaneously.
The energetic component of adhesion — the rate-independent part, the thermodynamic preference for cell-cell contact — promotes fluidity. It modifies cell shape by favoring configurations with more shared interface, and in vertex models, cell shape is the order parameter for the solid-fluid transition. More shared interface means cells closer to the fluid threshold. The glue, through its effect on geometry, makes the tissue more liquid.
The dissipative component — the rate-dependent part, the viscous resistance to separating bonded surfaces — suppresses fluidity. It acts as a drag on cell rearrangement. Even if a cell's geometry permits movement, the kinetic cost of breaking and reforming adhesive contacts slows or prevents the rearrangement. The glue, through its effect on dynamics, makes the tissue more solid.
These are not two effects of two different molecules. They are two effects of the same molecular adhesion, operating through different physical channels. The energetic channel modifies the landscape. The dissipative channel modifies the rate of traversal. A tissue's fluidity depends on which channel dominates, and that depends on the timescale of observation and the rate of forcing.
Standard vertex models treat adhesion solely as interfacial tension — the energetic component. They predict that increased adhesion promotes fluidity, which is experimentally observed in some contexts and confusing in others. The dissipative channel was invisible to the prevailing theoretical framework because the framework modeled statics, not dynamics. Adding rate-dependent adhesion resolves contradictions that looked like failures of the model but were actually failures of scope.
The same molecule. Two mechanisms. Opposite effects. Which one you see depends on how fast you look.
Note (2026-11-11): checked against the full text of the paper (Nguyen, Bera, Notbohm and Bi, arXiv 2603.04170, 'Cell-cell adhesion as a double-edged sword in tissue fluidity'). The two components of adhesion (an energetic, rate-independent one that raises fluidity through cell shape, and a dissipative, rate-dependent one that causes jamming) and the paper's statement that its framework resolves previously contradictory experimental findings are as in the paper. Two points differ. (1) The essay says which channel dominates 'depends on the timescale of observation and the rate of forcing' and ends 'which one you see depends on how fast you look'. In the paper the two mechanisms are set by two model parameters, the shape index and the junctional viscosity, and its phase diagram shows the regions of parameter space where each dominates; I found no statement that it depends on the observation time or the rate of forcing. (2) The work is computational: an extended two-dimensional vertex model (the simulations use 400 cells), with a linear rheology analysis that finds power-law viscoelastic behavior in the unjammed regime, with adhesion modulating the exponent, which the essay does not mention. The experimental observations it cites (for example bronchial epithelial cells slowing as junctions mature; asthmatic airway epithelium unjamming) are from other studies.