Figure 1.
Visualization of cell cycle-dependent cancer cell mobilization and invasion.
(A) Establishment and analyses of HCT116 colon cancer cells stably expressing Fucci. (Upper) The Fucci system enables monitoring of the cell cycle in live cells in real time. The nuclei of cells in the G1/G0, early S, and S/G2/M phases are labeled red, yellow, and green, respectively. (Lower) Snapshots of Fucci-expressing HCT116 cells. Scale bars represent 20 μm. (B) Intravital multiphoton imaging of Fucci-positive HCT116 cells inoculated into NOD/SCID mice. (Left) A representative image of Fucci-expressing HCT116 cells implanted in the cecum (green: Fucci-green (mAG), S/G2/M; red: Fucci-red (mKO2), G1; blue: collagen fibers (second harmonic generation (SHG) imaging)). Scale bars represent 75 μm. (Right) Quantification of the numbers of Fucci-green and -red HCT116 cells in different areas of inoculated tumors. Central and marginal zones were defined as areas further or closer than 75 μm from the border between the tumor and normal tissues, respectively. (C) Representative image at the edge of a Fucci-expressing HCT116 tumor mass. The entire area (left) and a time series (right) of magnified images (one per 400 s) of cancer cells invading the interstitium (green: Fucci-green (mAG), S/G2/M; red: Fucci-red (mKO2), G1; blue: collagen fibers (SHG imaging) (see also Movie S1). Actual images (upper panels) and cell trajectories (lower panels) are shown. Scale bars represent 100 μm (left) and 10 μm (right). (D) Representative image of extravasating Fucci-expressing HeLa cells. The entire area (left) and a time series (right) of magnified images of cancer cells extravasating from blood vessels (one per 12 min) (green: Fucci-green (mAG), S/G2/M; red: Fucci-red (mKO2), G1; blue: collagen fibers (SHG imaging) (see also Movie S2). Actual images (upper panels) and cell trajectories (lower panels) are shown. Scale bars represent 100 μm (left) and 10 μm (right). (E) Cellular motility in Fucci-green- and -red-positive cells was measured for 4 h (see also Movie S3). (Left) Green and red spheres represent Fucci-green- and -red-positive cells, respectively, and yellow lines show the associated trajectories. Scale bars represent 100 μm. (Right) Mean tracking velocities of Fucci-green- and -red-positive cells. Data (n = 379 for Fucci green and n = 259 for Fucci red) were obtained from individual cells in three independent experiments. The velocities of the two groups were compared by Mann-Whitney U-test (p = 0.0191). The median and interquartile ranges for each group are overlaid on the dot plots.
Figure 2.
Identification of ARHGAP11A as a cell cycle-dependent mobility- controlling molecule.
(A) Fucci signal-based microarray analyses. Fucci-positive HCT116 cells were separated into Fucci-green (mAG)- and Fucci-red (mKO2)- positive cells by FACS. mRNA was extracted from these cells and compared by microarray analysis (two dye-swap experiments, giving four independent microarray analyses). (B) In total, 2,023 probes (1,656 genes) showed >twofold changes in expression (Table S1) (P>0.05). Of them, Arhgap11a was highly ranked, and all three probes for Arhgap11a were among the top-ranked probes for RhoGAPs. The three probes were specific for the indicated potions of the Arhgap11a mRNA (left). The three marked dots (#1, #2, #3) in the scatter plots represent fold changes (right). (C) Cell cycle-dependent expression of Arhgap11a mRNA was confirmed by qPCR. The expression data were normalized to Gapdh (n = 3). (D) Cell cycle-dependent expression of Arhgap11a proteins. (Right) Flow cytometric analyses of Fucci-expressing HCT116 cells. Cell cycle profiles were color-coded: G1, red; early S, yellow; and S/G2/M, green (upper right). DNA contents were measured by Hoechst33342 fluorescence of (lower right), confirming that Fucci signals accurately represent cell cycle levels (n = 3). (Left) Cell cycle-dependent expression of ARHGAP11A and cell cycle markers, as determined by Western blotting (n = 3). (E) Time-dependent expression of Arhgap11a during progression of the cell cycle from G1 to S/G2/M. Fucci-red (mKO2)-positive HCT116 cells were sorted using a FACSAria cell sorter and were cultured for the indicated periods of time. Flow cytometry analyses (upper) and ratios of Fucci colors (left) are shown for each time point. (Right) Relative expression of Arhgap11a was examined by qPCR (n = 6). Data were analyzed by one-way ANOVA (p = 0.0001) and Bonferroni's multiple comparison test (*** p<0.01). (F) Cell cycle-dependent Arhgap11a expression in various human colon cancer cell lines. Fucci was introduced into different human colon cancer cell lines (HCT116, DLD1, HT29, and KM12SM). In all of the cell lines, Arhgap11a expression was significant higher in S/G2/M (green) than in G1 cells (red). (G) A chromatin immunoprecipitation (ChIP) assay with an anti-E2F1 antibody showed that E2F1 bound to the putative E2F-binding site in the Arhgap11a promoter (n = 3). (H) Luciferase reporter assay of the Arhgap11a promoter region (−500 bp), including the E2F-binding site (GTTTCGCGC) at −20 bp from the transcription starting point. Co-transfection with E2F1 enhanced transcriptional activity, whereas simultaneous expression of Rb blocked it. Values for luciferase activity were normalized across each experiment and, to control for differences in transfection efficiency, to β-galactosidase.
Figure 3.
GTPase-activating properties of ARHGAP11A.
(A) Halo-Tagged ARHGAP11A and Halo-Tag proteins expressed and purified from HEK293 cells. (B) Detection of GTPase activity by quantifying inorganic phosphate released by GTP hydrolysis by Rho family proteins. ARHGAP11A enhanced the GTPase activity of RhoA, but not of Rac1 or Cdc42. (C) Detection of active and total forms of various Rho-family proteins (RhoA, RhoB, RhoC, Rac1, and Cdc42) in HEK293 cells transfected with Halo-ARHGAP11A or its control. Expression of ARHGAP11A reduced the amounts of the active forms of RhoA, RhoB, and RhoC. (D) Assessment of mutant ARHGAP11A without the putative GAP domain (ΔGAP).
Figure 4.
Functional analyses of ARHGAP11A in RhoA-mediated cellular reactions in HCT116 colon cancer cells.
(A) Schematic illustration of RhoA-mediated cellular reactions. (B) Effect of overexpression of wild-type (WT) or constitutively active (Q63L) RhoA on the formation of F-actin stress fibers (visualized using Alexa 568-phalloidin) and focal adhesions (stained with anti-paxillin). GFP was co-transfected to identify the transfected cells. Scale bars represent 15 μm. (C) Effect of CT04, a potent RhoA inhibitor, on the formation of F-actin stress fibers (visualized using rhodamine-phalloidin) and focal adhesions (stained with anti-paxillin). Nuclei were stained with DAPI. Scale bars represent 15 μm. (D) Effect of overexpression of Halo-Tagged ARHGAP11A or its control on the formation of F-actin stress fibers (visualized using Alexa 568-phalloidin) and focal adhesions (stained with anti-paxillin) in HCT116 cells. Arrowheads identify Halo-Tag-expressing cells (labeled with Oregon green-conjugated Halo-Tag ligand). Scale bars represent 10 μm. (E) Quantification of mean intensities of F-actin in Halo-control (n = 80) and Halo-ARHGAP11A-expressing (n = 80) HCT116 cells. Data were compiled from three independent experiments. (F) Quantification of focal adhesions in Halo-control (n = 80) and Halo-ARHGAP11A-expressing (n = 80) HCT116 cells. Data were compiled from three independent experiments. (G) Effect of overexpression of Halo-Tagged ARHGAP11A or its control on the formation of F-actin stress fibers (visualized using Alexa 568-phalloidin) and focal adhesions (stained with anti-paxillin) in HeLa cells. Arrowheads identify Halo-Tag-expressing cells (labeled with Oregon green-conjugated Halo-Tag ligand). Scale bars represent 10 μm. (H) Quantification of mean intensities of F-actin in Halo-control (n = 80) and Halo-ARHGAP11A-expressing (n = 80) HeLa cells. Data were compiled from three independent experiments. (I) Quantification of the number of focal adhesions in Halo-control (n = 40) and Halo-ARHGAP11A-expressing (n = 46) HeLa cells. Data were compiled from three independent experiments.
Figure 5.
Relative augmentation of Rac1 activity and increased invasive migration in ARHGAP11A-expressing cancer cells.
(A) Schema representing the balance between RhoA and Rac1 for cell migration. (B) Analyses of Rac1 activity at the single-cell level in HCT116 cells expressing Halo-ARHGAP11A or its Halo control. Representative images of Raichu-Rac1-expressing HCT116 cells under Halo-control (left) or Halo-ARHGAP11A transfection (right) conditions. Rac1 activity was monitored by CFP/YFP FRET ratios derived from Raichu-Rac1. Expression of Halo-Tag was identified with TMR-conjugated Halo-Tag ligand. The scale bar represents 5 μm. (C) Quantification of FRET ratios in Halo-control (n = 30) and Halo-ARHGAP11A-expressing (n = 30) HCT116 cells. (D) Three-dimensional culture of HCT116 transfected with Halo-control or Halo-tagged ARHGAP11A, supplemented with Y27632 (for Halo-control only). The scale bar represents 50 μm. (E) Proportions of round-type HCT116 in 3D culture transfected with Halo-ARHGAP11A or Halo-control. Round-type cells were counted in three visual fields for each of three independent experiments. Columns represent the mean ± s.e.m. (F) In vitro invasion assay using 3D Matrigel plate. Migrated cells were visualized by staining culture membrane with Diff Quik stain (Dade Behring). HCT116 transfected with Halo-ARHGAP11A or Halo-control, and wild-type HCT116 treated with Y27632 were used in the assay. (G) Quantification of invasion indices from 3D Matrigel plate assays. Invasion indexes were calculated according to the equation shown in the Method section. Columns represent the means ± s.e.m.
Figure 6.
Functional analyses of ARHGAP11A-knockdown HCT116 human colon cancer cells.
(A) Establishment of ARHGAP11A-knockdown HCT116 cell lines (SH #1, #2). Decreased ARHGAP11A mRNA expression was confirmed by qPCR. (B) ARHGAP11A protein expression in control and sh-knockdown HCT116 cells was assessed by Western blotting. (C) BrdU proliferation assay of these HCT116 cell lines. Columns represent the means ± s.e.m. (D) In vitro invasion assay using 3D Matrigel culture plates. Columns represent the means ± s.e.m. (E) In vivo functional analyses of ARHGAP11A-knockdown HCT116 cells. Representative images of control (green) and ARHGAP-knockdown (red) HCT116 cells inoculated subcutaneously into NOD/SCID mice (upper). A raw ‘merged’ image and images extracted from the green and red channels are shown. Cell motility was measured for 7 h. Green and red circles represent control and ARHGAP11A-knockdown SH#1 HCT116 cells, respectively, and yellow lines show their trajectories (see also Movie S4). The scale bar represents 50 μm. (Lower) Mean tracking velocities of control and SH cells. Data (n = 440 for the control and n = 215 for SH#1) were obtained from individual cells in two independent experiments. The velocities of the two groups were compared by Mann-Whitney U-test (p = 0.0034). The median and interquartile ranges for each group are overlaid on the dot plots. (F) Extravasation of control (green) and SH#1 (red) HCT116 cells. (Left) Green cells were preferentially detected in extravascular spaces, suggesting a high potency for extravasation. (Right) Average numbers of extravasated cells per visual field. Data were extracted from 10 visual fields. (G) In vivo tumor expansion of HCT116 cells. Wild-type control HCT116 cells (black circles) and HCT116 cells treated with scrambled control shRNA (black squares), SH#1 (red circles), and SH#2 (red squares) are shown. Cancer cells (1.0×106/100 μl of PBS) were primarily inoculated into subcutaneous tissue. Tumor sizes were measured every week for 4 weeks after inoculation. Data represent the means ± s.e.m. of five independent experiments. Data were analyzed by two-way ANOVA (p = 0.0037). (H) Decreased expression of ARHGAP11A in HCT116 cells treated with siRNAs targeting ARHGAP11A (assessed by qPCR). (I) In vivo siRNA treatment of HCT116 tumors. Control HCT116 cells (5.0×106) were implanted into NOD/SCID mice, and 1 week later were treated with PBS (black filled circles), atelocollagen (black filled triangles), scrambled control siRNA plus atelocollagen (black filled squares), or two siRNAs (#1 and #2) against ARHGAP11A plus atelocollagen (red open circles and squares, respectively). Tumor sizes were measured weekly. Data represent the means ± s.e.m. of five independent experiments. Data were analyzed by two-way ANOVA (p = 0.0001). (J) Images of tumors excised on day 35 (left). The scale bar represents 10 mm. (Right) A representative images of SCID mice bearing HCT116 human colon cancer cells, 35 days after treatment with an siRNA (siRNA#1) or with a scrambled RNA duplex (control), together with atelocollagen. The scale bar represents 10 mm.
Figure 7.
Enhanced expression of ARHGAP11A in surgically resected human colorectal cancers.
(A) Clinicopathological data for human colon cancer patients (B) Microarray analysis of total RNA isolated from five normal tissue samples and 74 colorectal cancer specimens. The data points represent values from individual patients. The median and interquartile ranges are overlaid on each column. Relative expression of ARHGAP11A in normal colon tissues and colon cancer specimens (p = 0.0430 [Mann-Whitney U-test]). (C) Comparison of expression data based on TNM factors. There were significant increases in expression between T1 (tumor invades the submucosa) and T3 (tumor invades and penetrates the muscularis propria into the subserosa or non-peritonealized pericolic or perirectal tissue), and between T1 and T4 (tumor directly invades other organs or structures and/or perforates the visceral peritoneum). Data were analyzed by Kruskal–Wallis test and Dunnett's test (*** p<0.01, ** p<0.05, * p<0.1). The median and interquartile ranges for each group are overlaid on the dot plots.