Stimulation of cortical myosin phosphorylation by p114RhoGEF drives cell migration and tumor cell invasion

Abstract

Actinomyosin activity is an important driver of cell locomotion and has been shown to promote collective cell migration of epithelial sheets as well as single cell migration and tumor cell invasion. However, the molecular mechanisms underlying activation of cortical myosin to stimulate single cell movement, and the relationship between the mechanisms that drive single cell locomotion and those that mediate collective cell migration of epithelial sheets are incompletely understood. Here, we demonstrate that p114RhoGEF, an activator of RhoA that associates with non-muscle myosin IIA, regulates collective cell migration of epithelial sheets and tumor cell invasion. Depletion of p114RhoGEF resulted in specific spatial inhibition of myosin activation at cell-cell contacts in migrating epithelial sheets and the cortex of migrating single cells, but only affected double and not single phosphorylation of myosin light chain. In agreement, overall elasticity and contractility of the cells, processes that rely on persistent and more constant forces, were not affected, suggesting that p114RhoGEF mediates process-specific myosin activation. Locomotion was p114RhoGEF-dependent on Matrigel, which favors more roundish cells and amoeboid-like actinomyosin-driven movement, but not on fibronectin, which stimulates flatter cells and lamellipodia-driven, mesenchymal-like migration. Accordingly, depletion of p114RhoGEF led to reduced RhoA, but increased Rac activity. Invasion of 3D matrices was p114RhoGEF-dependent under conditions that do not require metalloproteinase activity, supporting a role of p114RhoGEF in myosin-dependent, amoeboid-like locomotion. Our data demonstrate that p114RhoGEF drives cortical myosin activation by stimulating myosin light chain double phosphorylation and, thereby, collective cell migration of epithelial sheets and amoeboid-like motility of tumor cells.

Figure 1. p114RhoGEF regulates myosin activation and migration during wound repair.

(A) HCE cells, grown to confluence in dishes with two chamber culture inserts, were induced to migrate by removing the insert. Cell were fixed after different periods of migration and stained for double phosphorylated MLC and f-actin. Bar, 10 µm. (B,C) HCE cells were transfected with siRNAs and depletion of p114RhoGEF (B) and MLC phosphorylation (C) were analyzed as indicated. Note, MLC phosphorylation is only reduced at cell junctions, not at the leading edge, upon p114RhoGEF depletion. (D,E) Collective migration of HCE cells transfected with the indicated siRNAs was analyzed by measuring wound closure after different periods of time. (F) HCE cells were grown to confluence and left to stabilize for several days. The monolayers were then either directly extracted, or wounded with multiple scratches using a needle and re-incubated for 45 minutes prior to extraction. p114RhoGEF was then immunoprecipitated, and the precipitates were then analyzed by immunoblotting for the GEF and myosin IIA. Bar, 250 µm (D). Panel E shows means ±1SD, n = 3.

Figure 2. Regulation of single cell migration by p114RhoGEF.

MDA-MB-231 cells were transfected with siRNAs as indicated (p114RG siRNA-1 and siRNA-2 refer to distinct individual siRNAs; siRNA-p refers to a pool of the two siRNAs). Expression of indicated proteins was analyzed by immunoblotting (A), effect on total levels of active RhoA by G-LISA assay (B; shown are means ± 1SD, n = 4), and migration by time-lapse microscopy over 5 hours (C,D). Migration distances were quantified by single cell tracking. Panel D shows means ± 1SD of four different fields (20 cells were analyzed for each field).

Figure 3. Rac activation in response to p114RhoGEF depletion in MDA-MB-231 cells.

(A) MDA-MB-231 cells were transfected with the indicated siRNAs and the levels of active Rac were determined using a G-LISA assay 72 hours after transfection. Shown are means ± 1SD, n = 6. The numbers were normalized to control siRNA transfections. (B,C) Cells that had been transfected with siRNAs as in panel A were fixed and processed for immunofluorescence using an antibody specific for active Rac1 and fluorescent phalloidin. The intensity of the active Rac1 staining was quantified using image J. Panel C shows means ± 1SD of three experiments (in each experiment, at least 5 fields were analyzed).

Figure 4. Matrix-dependence of p114RhoGEF-regulated migration.

MDA-MB-231 cells were plated on the indicated 2D matrices and migration was analyzed by time-lapse microscopy as in figure 2 for 5 hours. In panels B and C, the cells had been transfected with the indicated siRNAs. All quantifications show means ± 1SD of four different fields (12 cells were tracked for each field). Note, only migration on collagen and Matrigel is p114RhoGEF-dependent. Bar, 30 µm.

Figure 5. Regulation of tumor cell invasion by p114RhoGEF.

(A,B) MDA-MB-231 cells were transfected as indicated and invasion across Matrigel covered filter inserts was then analyzed. Panel A shows images of the matrix after crystal violet staining. Panel B shows quantifications of matrix associated cells (shown are means ± 1SD, n = 7). (C-E) Cells transfected as indicated were plated into 3D matrices and migration as analyzed by time-lapse microscopy for 5 hours. Panel C shows still images and panel D quantification of migration distances of cells plated in collagen or Matrigel, and panel E migration in the presence of the metalloproteinase inhibitor GM6001. Quantifications show means ± 1SD of four different fields (12 cells were analyzed for each field).

Figure 6. Regulation of cortical myosin phosphorylation by p114RhoGEF.

(A) MDA-MB-231 cells were extracted and p114RhoGEF was immunoprecipated. The precipitates were then analyzed by immunoblotting for the GEF, myosin IIA, and ROCKII. (B) MDA-MB-231 cells that had transfected with the siRNAs indicated were lyzed and subjected to immunoblotting using the indicated antibodies. (C-G). Cells, transfected and plated as indicated, were fixed and then processed for immunofluorescence microscopy using antibodies against the proteins indicated. Panel C2 shows a larger magnification of the control and p114RhoGEF siRNA transfected cells plated on Matrigel shown in panel C1. Panel D shows a quantification of the percentage of cells with cortical staining for double phosphorylated myosin (shown are means ± 1SD). (C1, G) Bars, 10 µm; (C2, E, F) Bars, 5 µm.

Figure 7. Analysis of collagen gel contraction and cell stiffness.

(A) MDA-MB-231 cells were transfected with control or p114RhoGEF-specific siRNAs and were then embedded into collagen gels. Contraction of the gels was then followed for 6 days. Gel contraction was recorded daily using digital photography and the gel area was measured using image J. Contraction is expressed as a percentage decrease compared to the original gel area. Shown are means ± 1SD of 5 experiments. (B) Stiffness of MDA-MB-231 cells was measured by atomic force microscopy 72 hours after siRNA transfection. As a positive control, cells were incubated for 30 minutes with the ROCK inhibitor Y27632 prior to the start of the measurements. Shown are means ± 1SD of 3 experiments; in each experiment at least 30 cells were analyzed per condition.