β-adrenergic signaling modulates breast cancer cell mechanical behaviors through a RhoA-ROCK-myosin II axis

Abstract

The ability of cancer cells to deform and generate force is implicated in metastasis. We previously showed that β-adrenergic agonists increase cancer cell stiffness, which was associated with enhanced motility and invasion. Here, we investigate how β-adrenoceptor (βAR) activation alters the mechanical behaviors of triple-negative breast cancer cells. We find that βAR activation increases traction forces in metastatic MDA-MB-231^HM and MDA-MB-468 cells, but not in non-tumorigenic MCF10A cells. Using computational modeling, we show that βAR activation increases the number of active myosin motors via myosin light chain phosphorylation. To identify molecular regulators, we use a deformability assay to screen for pharmacologic and genetic perturbations. Our results define a βAR-RhoA-ROCK-non-muscle myosin II (NMII) signaling axis that modulates the mechanical behaviors of MDA-MB-231^HM and MDA-MB-468 cells. These findings provide insight into how stress signaling regulates cancer cell mechanics and suggest potential targets to block metastasis in triple-negative breast cancer.

Keywords: Cell biology; Functional aspects of cell biology; Mechanobiology.

Graphical abstract

Figure 1

βAR activation results in increased cellular force generation (A) Schematic illustration showing the traction force microscopy (TFM) assays. (i) TFM-Pillars Assay. Gold disks (yellow) are embedded on top of polydimethylsiloxane (PDMS) pillars to facilitate imaging lateral displacements due to cellular traction forces. (ii) TFM-Beads Assay. Gold nanoparticles (yellow) are embedded in a PDMS matrix to facilitate imaging lateral displacements due to cellular traction forces. (B) Representative images of MDA-MB-231^HM cells on micropillars with superimposed vector force map. Color scale indicates the force per pillar. Scale, 4 μm. (C) Traction forces of MDA-MB-231^HM cells using TFM-Pillars assay after treatment for 24 h with: vehicle, Veh; βAR agonist isoproterenol, Iso (100 nM); βAR antagonist propranolol, Pro (10 μM); or 100 nM Iso and 10 μM Pro. Each dot represents an individual pillar from at least 16 single cells across 3 independent experiments. Bars show the median; error bars represent standard error. (D) Representative images from traction force measurements of TNBC and MCF10A cells treated with isoproterenol (Iso) for 24 h using the TFM-Beads assay. Each arrow indicates the direction of force and the color gradient corresponds to the magnitude of stress. Scale: 10 μm. (E) Quantification of traction forces. Each data point represents the average traction force per bead for an individual cell, which is averaged over multiple beads that are within the boundary of each individual cell. Data shown here represent at least 7 individual cells across 3 independent experiments. Bars show the median; error bars represent standard error. (F) Western blotting against mono-phosphorylated myosin light chain 2 (pMLC2), non-phosphorylated MLC2, and GAPDH from MDA-MB-231^HM cells treated with 100 nM of isoproterenol for up to 48 h. Quantification of band intensity for pMLC2/MLC2 was normalized to 0 h sample. (G) Western blotting against di-phosphorylated MLC2 (ppMLC2), non-phosphorylated MLC2, and GAPDH after MDA-MB-231^HM cells were treated with increasing concentrations of isoproterenol for 2 h. The ratio of ppMLC2 to MLC2 was normalized to vehicle control. (H) Expression levels of ADRB2 transcripts in TNBC, MCF10A, and MCF7 cells measured by qRT-PCR. Images in (A) are adapted from Servier Medical Art by Servier and are published a Creative Commons BY license (https://creativecommons.org/licenses/by-nc/3.0/). Unless otherwise stated, all error bars represent mean ± s.e.m (N = 3). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test (F–H) and statistical significance in (C and E) is determined using a permutation test to evaluate the difference in medians between control and treatment conditions. ∗p < 0.05].

Figure 2

βAR activation results in decreased cancer cell deformability (A) Schematic illustration showing the PMF assay. Suspensions of cells were loaded into the top well separated with porous membranes from the bottom well. (B) Filtration measurements by PMF reveal cellular deformability of non-transformed epithelial cells (MCF10A) versus triple-negative breast cancer cells (MDA-MB-231) and the highly metastatic variant of MDA-MB-231 cells (MDA-MB-231^HM) after treatment for 24 h with: vehicle, Veh; βAR agonist isoproterenol, Iso (100 nM); βAR antagonist propranolol, Pro (10 μM); or 100 nM Iso and 10 μM Pro. All error bars represent mean ± s.e.m (N = 3). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test]. Images in A are adapted from Servier Medical Art by Servier and are published under a Creative Commons BY license (https://creativecommons.org/licenses/by-nc/3.0/).

Figure 3

βAR activation increases the number of NMII motors interacting with F-actin to yield increased traction force generation (A) A stochastic Monte Carlo model is used to simulate the state transitions and force generation in actin-myosin filaments that are transferred to the substrate via integrins. Schematic shows actomyosin complex attaching to integrin on the cytosolic side. Endogenous forces affect integrin bond lifetime, which is modeled by a spring with catch-slip dynamics. (B) Simulations predict the number of active NMII motors per actin filament (for a maximum of 60 NMII per actin filament and 120 actin filaments per μm²) with increasing concentrations of isoproterenol, [Iso]; 0 nM isoproterenol is vehicle control. (C) Traction forces per μm² predicted by simulation are normalized to vehicle control. (D) Kymographs show simulated forces at focal adhesions. Each row represents an individual actin-myosin filament and summation over all 120 filaments (rows) represents the total traction force at time (t). Color map gradient shows magnitude of force generated at focal adhesion.

Figure 4

βAR-induced changes in cellular deformability require ROCK activity (A) Cell filtration measurements by PMF with simultaneous βAR activation by isoproterenol (Iso) and suppression of myosin activity by pharmacological inhibitors, Bleb, blebbistatin (10 μM); Y27632 (10 μM); ML-7 (10 μM); and IPA-3 (10 μM). Cells were co-treated with inhibitors and isoproterenol for 24 h prior to filtration measurements. (B–D) Filtration measurements with increasing concentration of isoproterenol with or without NMII-regulating kinase inhibitors. (E) Western blotting against phosphorylated MLC2 (pMLC2), non-phosphorylated MLC2, and GAPDH. (F) Normalized ratio of pMLC2 to MLC2 levels after MDA-MB-231^HM cells were treated with isoproterenol with or without inhibition of ROCK (Y27632), MLCK (ML7), and PAK1 (IPA3). Data represents 3 independent experiments. n.s.: not significant, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test].

Figure 5

RhoA contributes to βAR regulation of cellular deformability (A) Confirmation of siRNA knockdown of RhoA by western blotting. GAPDH is loading control. (B) Ratio of RhoA to GAPDH normalized to control siRNA (siCon). (C) Cell filtration with PMF after 100 nM of isoproterenol treatment for 24 h in control and RhoA knockdown cells. (D) Cell filtration with PMF after increasing concentration of Rho activator, CN03, treatment for 4 h. Multiple comparisons from ANOVA test were made by comparing the mean of each treatment with the mean of Veh treatment. All experiments were performed 3 times (N = 3). n.s.: not significant, ∗p < 0.05; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test].

Figure 6

NMII, ROCK, and RhoA activity are required for the increased invasion of MDA-MB-231^HM cells due to βAR activation (A and B) Representative images from a 3D scratch wound invasion assay. The confluent cells appear gray; the scratch wound is teal; and the cells that enter the scratch wound are represented with purple. Drugs were added at time 0: isoproterenol (Iso, 100 nM), NMII inhibitor blebbistatin (Bleb, 10 μM), and ROCK inhibitor Y27632 (10 μM). Transfections were performed 72 h prior to time 0 h of the invasion assay. Scale: 300 μm. (C and D) Relative wound density as a function of time. Relative wound density is defined as the area of cells in the newly healed scratch wound region (purple) compared to the area of the initial scratch wound (teal at time 0 h). (E and F) Relative wound density at 48 h. Unless otherwise indicated, all comparisons were made to vehicle control (Veh). All experiments were performed 3 times (N = 3). n.s.: not significant, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test].

Figure 7

NMII and ROCK activity are required for invasion of MDA-MB-231^HM and MCF10A cells, while βAR activation increases invasion of MDA-MB-231^HM cells only 3D scratch wound invasion assays from MDA-MB-231^HM (A), and MCF10A (C) cells treated with isoproterenol (Iso, 100 nM), NMII inhibitor blebbistatin (Bleb, 10 μM), ROCK inhibitor Y27632 (10 μM), and Rho activator CN03 (1 μg/mL). (B and D) Relative wound density at 24 h. Unless otherwise indicated, all comparisons were made to vehicle control (Veh). All experiments were performed 3 times (N = 3). n.s.: not significant, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001 [one-way ANOVA with Tukey’s test].

Figure 8

Schematic illustration of the proposed mechanism for how a βAR-RhoA-ROCK-NMII axis regulates mechanical behaviors of TNBC cells that can contribute to metastasis Actin filaments are shown in yellow and myosins attached to these filaments are shown in black. Black arrows indicate traction forces generated by the cell. Images are adapted from Servier Medical Art by Servier (http://www.servier.com/Powerpoint-image-bank) and are published under a Creative Commons by license (https://creativecommons.org/licenses/by-nc/3.0/).