Figure 1.
Activin B dependent invasion of ccRCC cells is inhibited by serum.
(A) Relative Activin B expression of 786.0 parental tumor cells and individual clones carrying empty vector (neo#1, #3) and two different shRNAs targeting Activin B mRNA (si1-βB#1 and #2, si2-βB#1 and #3), respectively, determined by quantitative realtime RT-PCR. Results were normalized on β-actin expression. (B) Invasion of parental 786.0 tumor cells, neo control clones and Activin B knockdown clones plated on collagen I gels in the presence of 2% (black bars) and 10% FCS (grey bars). Number of invaded cells was quantified after 3 days. (C) Phase contrast micrographs corresponding to B depicting the morphology of the indicated cell lines on the surface of the collagen I gel. (D) Relative Activin B expression of RCC10 parental tumor cells and individual clones carrying scrambled shRNA (scr#3, #4, #5) and shRNA targeting Activin B mRNA (si2-βB#4, #6, #10), respectively, determined by quantitative realtime RT-PCR. Results were normalized on β-actin expression. (E) Invasion of parental RCC10 tumor cells, scrambled control clones and Activin B knockdown clones plated on collagen I gels in the presence of 2% (black bars) and 10% FCS (grey bars). Number of invaded cells was quantified after 3 days. (F) Phase contrast micrographs corresponding to E depicting the morphology of the indicated cell lines on the surface of the collagen I gel. Bars in A, B, D and E represent the mean of three independent experiments, error bars indicate standard deviation. Statistical significance was determined by Student’s t-test and denoted by asterisks: **P<0.01; ***P<0.001; n.s. not significant.
Figure 2.
Knockdown of Activin B stabilizes actin stress fibers in ccRCC cells.
(A) Subconfluent parental 786.0 and RCC10 cells, control clones (786.0 neo#1, RCC10 scr#3) and Activin B knockdown clones (786.0 si1-βB#2, RCC10 si2-βB#4), respectively, were maintained in the presence of 10% FCS or serum starved (0% FCS) for 3 hours. Micrographs show TRITC-labeled Phalloidin staining of the actin cytoskeleton. (B) and (C) Quantification of the indicated 786.0 (B) and RCC10 cells (C) with actin stress fibers upon serum starvation for 3 hours. Phalloidin stained cells were classified by microscopic analysis and at least 200 cells were counted per experiment. Bars represent the mean of four independent experiments, error bars indicate standard deviation. Statistical significance was determined by ANOVA analysis and denoted by asterisks: ***P<0.001.
Figure 3.
Knockdown of Activin B activates the RhoA pathway.
(A) Endogenous RhoA activity of the indicated cell lines determined by G-Lisa assay. Bars represent the mean of three independent experiments, error bars indicate standard deviation. Statistical significance was determined by ANOVA analysis and denoted by asterisks; ***P<0.001. (B) Western Blot analysis of Myosin Light Chain 2 (MLC2) phosphorylated at Serin 19 in the indicated cell lines upon serum starvation (0% FCS) for 4 hours. pMLC2 was detected by a phospho specific antibody. Total MLC2 and β-actin served as loading controls. Numbers below the lanes reflect relative levels of phosphorylated MLC2 normalized to total MLC2 as determined by densitometry.
Figure 4.
The RhoA pathway is required for actin stress fiber formation induced by Activin B knockdown.
(A) TRITC-labeled Phalloidin staining of serum starved (0% FCS) neo#1 control cells and si1-βB#2 Activin B knockdown cells either untreated or treated with a Rho-Kinase inhibitor (Y-27632) or a Rho inhibitor (C3-Exoenzyme). (B) Quantification of untreated (grey bars) and Y-27632 treated (black bars) cells with actin stress fibers. (C) Generation of pools of 786.0 neo#1 and Activin B knockdown clones (si1-βB#2, si2-βB#1) stably transfected with EGFP empty vector, expression constructs for GFP-tagged wildtype RhoA, dominant active RhoA (G14V) or dominant negative RhoA (T19N). Expression of EGFP and GFP-tagged RhoA proteins was confirmed by Western Blotting. (D) TRITC-labeled Phalloidin staining of the indicated pools after serum starvation (0% FCS) for 3 hours. (E) Quantification of cells with actin stress fibers upon serum starvation for 3 hours. Bars represent the mean of three (B) or five (E) independent experiments, error bars indicate standard deviation. Statistical significance was determined by Student’s t-test (4B) and ANOVA analysis (4E) and denoted by asterisks: **P<0.01; ***P<0.001; n.s. not significant.
Figure 5.
The RhoA pathway interferes with invasiveness of ccRCC cells.
(A) Parental 786.0, neo control clones and Activin B knockdown clones were plated in the presence of 2% FCS on collagen I gels and treated with the Rho-Kinase inhibitor Y-27632 (black bars) or left untreated (grey bars). Number of invaded cells was quantified after 3 days. (B) Stable pools expressing either EGFP, wildtype, dominant active (G14V) and dominant negative (T19N) RhoA, respectively, were plated in the presence of 2% FCS (grey bars) or 10% FCS (black bars) on collagen I gels. Invaded cells were quantified after 3 days. Bars represent the mean of three independent experiments, error bars indicate standard deviation. Statistical significance was determined by Student’s t-test (5A) and ANOVA analysis (5B) and denoted by asterisks: *P<0.05; **P<0.01; ***P<0.001; n.s. not significant.
Figure 6.
Opposite roles for Rac1 and RhoA in Activin B dependent invasion and actin rearrangement.
(A) Generation of pools of the neo#1 clone and of the si2-βB#1 Activin B knockdown clone stably transfected with EGFP empty vector, expression constructs for GFP-tagged wildtype Rac1, dominant active Rac1 (G12V) or dominant negative Rac1 (T17N). Expression of EGFP and GFP-tagged Rac1 proteins was confirmed by Western Blotting. (B) TRITC-labeled Phalloidin staining of the indicated pools after serum starvation (0% FCS) for 3 hours. (C) Quantification of cells with actin stress fibers upon serum starvation for 3 hours. (D) and (E) The indicated pools were plated in the presence of 2% FCS (D) or 10% FCS (E) on collagen I gels and the number of invaded cells was quantified after 3 days. (F) Phosphorylation of MLC2 (Myosin light chain 2) at Serin 19 in 786.0 pools expressing either EGFP or the indicated Rac1 proteins was determined by Western Blotting. Cells were serum starved (0% FCS) for 4 hours. pMLC2 was detected by a phospho specific antibody. Total MLC2 and β-actin served as loading controls. Numbers below the lanes reflect relative levels of phosphorylated MLC2 normalized to total MLC2 as determined by densitometry. Bars in C, D and E represent the mean of three (C) and four (D; E) independent experiments, error bars indicate standard deviation. Statistical significance was determined by ANOVA analysis and denoted by asterisks: *P<0.05; **P<0.01.
Figure 7.
Autocrine stimulation of VHL negative ccRCC tumor cells by Activin B reduces RhoA activity thereby reorganizing the actin cytoskeleton and inducing invasiveness of spindle-shaped cells. Serum, which promotes RhoA activity, counteracts Activin B mediated effects. Rac1 activation by an unknown mechanism plays an opposite role to RhoA and is required for the invasive phenotype.