Substrate stiffness regulates cadherin-dependent collective migration through myosin-II contractility

Abstract

The mechanical microenvironment is known to influence single-cell migration; however, the extent to which mechanical cues affect collective migration of adherent cells is not well understood. We measured the effects of varying substrate compliance on individual cell migratory properties in an epithelial wound-healing assay. Increasing substrate stiffness increased collective cell migration speed, persistence, and directionality as well as the coordination of cell movements. Dynamic analysis revealed that wounding initiated a wave of motion coordination from the wound edge into the sheet. This was accompanied by a front-to-back gradient of myosin-II activation and establishment of cell polarity. The propagation was faster and farther reaching on stiff substrates, indicating that substrate stiffness affects the transmission of directional cues. Manipulation of myosin-II activity and cadherin-catenin complexes revealed that this transmission is mediated by coupling of contractile forces between neighboring cells. Thus, our findings suggest that the mechanical environment integrates in a feedback with cell contractility and cell-cell adhesion to regulate collective migration.

Figure 1.

Collective migration of MCF10A epithelial cell sheets on PAA gel substrates of various compliances. (A) Progression of wound healing on soft (3 kPa) and stiff (65 kPa) substrates. Red lines trace the wound edge; blue line on 12-h images reflects the wound edge at 0 h after wounding. (B) The distance advanced by cell sheets over 14 h of wound healing on various substrate stiffnesses. (C) Tracks of cell movements overlaid with images of H2B-mCherry–labeled nuclei in cell sheets at 12 h after wounding on soft (3 kPa) and stiff (65 kPa) substrates. (D and E) Mean speed (D) and mean wound-directed velocity (E) of individual cells within the collectively migrating cell sheet at 12 h after wounding, plotted as a function of cell distance from the wound edge. (F) Mean cell migration persistence for cells positioned at various distances from the wound edge after 5 h of migration starting at 8 h after wounding. N = number of experiments; n = total number of cells measured from N experiments. Error bars show 95% confidence interval of the mean (95% SEM); all nonoverlapping error bars are statistically significant with P < 0.05. Bars, 100 µm.

Figure 2.

Coordination of cell movements on soft and stiff substrates. (A) Depiction of normalized cell pair separation distance measurement. Measurements were made starting 8 h after wounding (t₀). See Materials and methods for details. (B) Mean normalized cell pair separation distance over a time course (t) of 5 h for all neighboring cell pairs within 160 µm from the wound edge. (inset) Mean cell pair separation distance at the end of the 5-h observation window for different substrate stiffness. (C) Mean normalized cell pair separation distance as a function of cell position from the wound edge for different substrate stiffness. (D) Depiction of velocity correlation measurement. The migrating cell sheet was divided into bands of 160 µm starting from the wound edge. (inset) For each band, the velocities (arrows) of all cells were correlated with those of other cells falling in a ring of radius R and width ΔR and then averaged. See Materials and methods for details. (E) Velocity correlation 12 h after wounding for cells 160 µm from the wound edge for different substrate stiffness. (F) Velocity correlation as a function of cell–cell distance R, for cells in different distance bands from the wound edge and at different time points after wounding. N = number of experiments. Each plot displays means of 150–3,000 cell measurements pooled from N experiments. All nonoverlapping error bars (95% SEM) are statistically significant with P < 0.05.

Figure 3.

Analysis of coordination of cell movements in unwounded monolayers and between sparse cells. (A and B) Mean normalized cell pair separation distance (A) and mean velocity correlation (B) for confluent, unwounded MCF10A monolayers on soft (3 kPa) and stiff (65 kPa) substrates compared with those for cells within 160 µm of wound edge undergoing collective migration at 12 h after wounding. (C) Mean velocity correlation for sparsely seeded MCF10A cells on various substrate stiffness. N = number of experiments. All nonoverlapping error bars (95% SEM) are statistically significant with P < 0.05.

Figure 4.

Analysis of cell polarization orientations on stiff and soft substrates. (A) MCF10As transfected with Golgi-GFP undergoing wound healing on soft (3 kPa) versus stiff (65 kPa) substrates. Bar, 50 µm. (B) The orientations of Golgi relative to the direction of wound healing, as measured from centroids of cell nuclei, were plotted in rose plots at various time points after wounding for cells at the wound edge. The direction of each bar in the rose plots indicates the angular Golgi orientations, whereas the magnitude of each bar shows the fraction of cells with the indicated Golgi orientations. (C) Rose plots of Golgi orientations at 12 h after wounding for cells at various rows behind the wound edge cells. (D) The primary direction of lamellipodial protrusions in cells on soft (3 kPa) and stiff (65 kPa) substrates was quantified as in the direction of wound healing, opposite the direction of the wound, and toward other directions. Measurements were obtained for >30 cells at various distances from the wound edge from N experiments. Error bars show SEM.

Figure 5.

The effects of myosin-II contractility on collective migration on soft and stiff substrates. (A) Maximum intensity projection of z-stack images of collectively migrating cells on soft and stiff substrates immunostained with pMLC (green), phalloidin as actin marker (red), and DAPI as the nuclei marker (blue). Bar, 50 µm. (B) Spatial variation of pMLC fluorescent intensity normalized by cell numbers. Error bars show SEM. (C) Focal adhesion lengths on stiff and soft substrates. Blue bars within box plots indicate means and SEM; data points outside the whiskers are outliers. (D) Effects of 25 µM blebbistatin treatment on the mean speed of individual cells at 12 h after wounding on soft and stiff substrates. (E and F) Effects of 25 µM blebbistatin on mean normalized cell pair separation distance (E; after 5 h of observation starting 8 h after wounding, see Fig. 2) and velocity correlation at 12 h after wounding (F) on soft and stiff substrates for cells 160 µm from the wound edge. N = number of experiments; n = total number of cells measured from N experiments. *, P < 0.05; **, P < 0.005; ***, P < 0.0005. Nonoverlapping error bars in D–F (95% SEM) are statistically significant with P < 0.05.

Figure 6.

The effects of P-cadherin knockdown (siCDH3) on migration properties during wound healing. (A) Effects of siCDH3 treatment on mean distance advanced by the cell sheet. (B–F) Migration properties of siCDH3-treated cells undergoing wound healing on soft (3 kPa) and stiff (65 kPa) substrates, compared with control cells, plotted at same time points after wounding and distances from wound edge as in Figs. 1 and 5. (G and H) Golgi orientation at 12 h after wounding on soft (3 kPa) or stiff (65 kPa) substrates. N = number of experiments; n = total number of cells measured from N experiments. All nonoverlapping error bars (95% SEM) are statistically significant with P < 0.05.

Figure 7.

The effects of siCDH3 and siCTNNA1 on actin organization and pMLC level in cells during wound healing. (A) Maximum projections of confocal immunofluorescence images of pMLC (green) costained with phalloidin (red) and DAPI (blue) in control and knockdown cells at the wound edge and within the migrating cell monolayer. Insets are numbered and magnified at the right with corresponding maximum projection E-cadherin immunofluorescence images. Bars, 50 µm. (B) Quantification of pMLC fluorescence intensity per cell, normalized to that of the control wound edge cells on stiff substrates. (C) pMLC fluorescence intensity per cell on stiff (65 kPa) substrates, normalized to that of wound edge cells for each condition. Error bars show SEM. N = number of independent experiments. *, P < 0.05; **, P < 0.005; ***, P < 0.0005.

Figure 8.

Effects of lowering myosin contractility in cells with reduced cell–cell adhesions. (A and B) Normalized cell pair separation distance (A) and velocity correlation (B) for control cells, cells treated with siCTNNA1, MIIA knocked down cells, and MIIA knocked down cells treated with siCTNNA1 on stiff (65 kPa) substrates. (C) Velocity correlation for control and siCDH3-treated cells on glass after blebbistatin treatment. Inset shows velocity correlation at R = 225 µm for siCDH3 cells. N = number of experiments. All nonoverlapping error bars (95% SEM) are statistically significant with P < 0.05. For the inset in C, statistical significance was specifically calculated for 0, 5, and 10 µM blebbistatin treatments to highlight the partial rescue in velocity correlation with low dose blebbistatin treatment. **, P < 0.005; ***, P < 0.0001.

Figure 9.

Side view model of mechanoresponsive collective migration of MCF10A cells. To achieve coordination in cell movements, cells at the wound edge respond to directional cue by front-rear polarization (P) in the direction of wound healing. The establishment and maintenance of cell polarity require actomyosin contractility (M), and they are influenced by cell–cell adhesions (C). Actomyosin contractility is governed by focal adhesions (FA) and cell–cell adhesion signaling. At the same time, actomyosin contractility promotes the formation and maturation of both focal adhesions and cell–cell adhesions (double-headed arrows). Actomyosin contractility can also be a negative regulator of cell–cell adhesions (flat-end arrow). The feedback coupling between actomyosin contractility and focal adhesions is stronger on stiff substrates and at the wound edge compared with >500 µm into the sheet, whereas within the sheet, the coupling between actomyosin contractility and cell–cell adhesions is stronger than between actomyosin contractility and focal adhesions. Forces generated by actomyosin contractility are transmitted to neighboring cells across cell–cell adhesions (blue dots connected by a line), where they activate actomyosin contractility and promote front-rear polarization in the follower cells. Propagation of front-rear polarization beyond the wound edge into the monolayer leads to cell–cell coordination. Various perturbations studied in this work affect focal adhesions, actomyosin contractility, and cell–cell adhesions, which in turn affect force transmission and front-rear polarization, leading to different levels of motion coordination. The sizes of the letters reflect relative magnitudes, and letters with asterisks indicate a parameter that is inferred and not directly measured in this study. All arrows are inferred, and the widths of the arrows reflect the relative strengths of interactions compared across the scenarios.