Shape-Independent Fluidization in Epithelial Cell Monolayers

TL;DR

A generalized model quantitatively captures the experimental data, demonstrating that the interplay between adhesive energetics and dissipative friction is essential for a complete understanding of epithelial fluidity.

Abstract

Tissue fluidity regulates many critical biological processes, including embryonic development, wound healing, and cancer metastasis. In confluent epithelia, where cell packing fraction is effectively fixed, the prevailing paradigm postulates that transitions between solid-like jammed and fluid-like unjammed states are governed by a geometric cell shape index determined by the balance of cortical tension and intercellular adhesion. Here, we challenge this geometric framework by reporting a mode of fluidization in epithelial monolayers that is entirely shape-independent. We observe that reducing cell-cell adhesion triggers a substantial increase in fluidity, yet this occurs without any corresponding change in cell shape, cell density, substrate traction, or junctional line tension. This decoupling of shape and fluidity reveals that current vertex models, which treat adhesion solely as a contribution to interfacial tension, are incomplete. To reconcile these findings, we extend the theoretical framework to account for the dual nature of adhesion -- its thermodynamic role in setting interfacial adhesion energy at the cell-cell junctions and its kinetic role in generating viscous drag as cells slide past their neighbors. This generalized model quantitatively captures the experimental data, demonstrating that the interplay between adhesive energetics and dissipative friction is essential for a complete understanding of epithelial fluidity.

Figures (11)

• Figure 1: Fluidity of MDCK cell layers treated with the anti-E-cadherin antibody DECMA-1. (a) Phase-contrast image of a cell island, with the yellow boxed region magnified in later panels. (b) Representative trajectories overlaid on the nuclei image and color-coded to distinguish between different trajectories. (c) Mean-square displacement (MSD) calculated from steady-state trajectories ($t$ = 4 h to 24 h) for each DECMA-1 concentration. Blue lines show fits to the 0 µ g ml$^{-1}$ DECMA-1 condition over the $\Delta t$ (lag time) ranges 0.25--1.0 h and 2.5--6.25 h, yielding MSD exponents $\gamma_1$ and $\gamma_2$, respectively. Fits were extrapolated to estimate the superdiffusive-subdiffusive crossover time $\Delta t_c \approx$ 1.5 h, and length scale, $L_c \approx$ 17 µ m. (d--f) Scatter plots for individual islands at different DECMA-1 concentrations showing (d) long-time MSD exponent, $\gamma_2$, with dashed line indicating the subdiffusive boundary ($p$ = 8.5 $\times 10^{-7}$), (e) self-diffusivity, $D_s$ in semi-log scale ($p$ = 0.014), (f) structural relaxation time (uncaging time), $\tau_\alpha$, extracted using DDM analysis ($p$ = 3.3 $\times 10^{-3}$). (g) Image of cell nuclei overlaid with Voronoi tessellations, color-coded by the cell shape index, $q$. (h) Average $q$ over time. Inset: steady-state shape index $q_\text{steady}$, defined as the average over $t$ = 23--24 h ($p$ = 1.1 $\times 10^{-3}$). (i) Probability distributions of steady-state shape index, $P(q)$. Error bars are standard deviations across different independent cell monolayers. In all scatter plots, markers indicate data from independent cell monolayers, and horizontal lines denote means of respective groups.
• Figure 1: Number density during the DECMA-1 experiment (Fig. 1). Number density was computed at each frame by counting nuclei within each island and dividing by island area. (a) Average number density $n$ for each DECMA-1 concentration, with error bars indicating standard deviations across different cell monolayers. (b) Plotted $n_\text{steady}$, the steady state density defined as the average over $t$ = 23--24 h ($p$ = 5.8 $\times 10^{-5}$). Horizontal lines denote means of respective groups. Although the number density was statistically different across the different DECMA-1 concentrations, the relative values are in a small range of $\approx$ 2500--3200 mm$^{-2}$. Prior studies that have shown effects of density on collective migration [main text Ref. saraswathibhatla2020tractions] have used a much larger range, e.g., 1200--4200 mm$^{-1}$, meaning that any effects of the small range of density in these experiments are likely to be modest. Furthermore, additional experimental data in this manuscript (Fig. 3 and Supplementary Fig. 4) show no correlation between density and kinematics, which rules out density as the cause of the fluidization reported here.
• Figure 2: Effects of DECMA-1 on cell-cell interfacial adhesion energy and cell-substrate traction. (a) Representative confocal image of MDCK monolayers with a live stain for F-actin; the yellow arrow indicates the junction segment 1 s, before ablation. (b-e) The same region after ablation at successive time points. Overlaid velocity vectors were obtained from digital image correlation, and yellow boxes denote regions of interest at opposite ends of the ablation site. The scale bar shown in (a) is the same for all panels (a)-(e). (f) End-to-end recoil displacement $d_\mathrm{IR}$ as a function of time, averaged over 50 ablation sites for each DECMA-1 concentration. (g) Initial recoil velocity $v_\mathrm{IR}$, a proxy for junctional line tension, computed over the first 8 s after ablation; horizontal lines denote means ($p$ = 0.55; extended statistical test in Supplementary Table 1). (h) Root-mean-square (RMS) traction $T$ as a function of time; inset: steady-state traction $T_\text{steady}$, defined as the average over $t$ = 23--24 h ($p$ = 0.56; extended statistical test in Supplementary Table 1). Error bars indicate standard deviations. Horizontal lines denote means of respective groups.
• Figure 2: Verification that DECMA-1 binds to E-cadherin. Top row, left to right: Confocal images of E-cadherin (red) and DECMA-1 (green) in cells fixed after 17 h, for a represented region in a cell monolayer treated with 5 µ g ml$^{-1}$ DECMA-1; merging color channels appears yellow, indicating co-localization of E-cadherin and DECMA-1, which verifies that DECMA-1 bound to E-cadherin. Bottom row, left to right: similar plots for 0 µ g ml$^{-1}$ show no green fluorescence, indicating that the antibody used for imaging DECMA-1 bound specifically to DECMA-1 in the 5 µ g ml$^{-1}$ condition.
• Figure 3: Effects of EGTA on tissue fluidization, cell shape, junctional line tension, and traction. (a) MSD calculated from steady-state trajectories ($t$ = 7--24 h), for all EGTA conditions. Blue lines show fits to the 0 mM EGTA data in the $\Delta t$ intervals 0.25--1.0 h and 2--6.25 h, with MSD exponents $\gamma_1$ = 1.3 and $\gamma_2$ = 0.6, respectively, yielding $\Delta t_c \approx$ 1.1 h and $L_c \approx$ 10 µ m. (b--d) Scatter plots for each islands with EGTA showing (b) long-time MSD exponent, $\gamma_2$ ($p$ = 4.2 $\times 10^{-4}$), (c) self-diffusivity, $D_s$ ($p$ = 8.2 $\times 10^{-3}$), and (d) uncaging time, $\tau_\alpha$ ($p$ = 0.034). (e) Average shape index, $q$, Inset: steady-state shape index $q_\text{steady}$, defined as the average over $t$ = 22--23 h ($p$ = 0.86; extended statistical test in Supplementary Table 2). (f) Steady-state distributions of $q$. (g) Initial recoil velocity, $v_\mathrm{IR}$, measured after laser ablation; horizontal lines indicate mean ($p$ = 0.63; extended statistical test in Supplementary Table 2). (h) RMS traction, $T$, reaching steady state after $t \approx 7$ h. (i) Steady-state traction, $T_\text{steady}$ ($p$ = 2.1 $\times 10^{-4}$). Error bars are standard deviations. In the scatter plots, markers indicate data from independent cell monolayers, and horizontal lines denote means of respective groups.