Malignant astrocytomas are highly invasive neoplasms infiltrating diffusely into regions of normal brain. Whereas the molecular and cellular mechanisms governing astrocytoma invasion remain poorly understood, evidence in other cell systems has implicated a role for the Rho-GTPases in cell motility and invasion. Here, we examine how the inhibition or activation of Rho-kinase (ROCK) affects astrocytoma morphology, motility, and invasion. ROCK was inhibited in astrocytoma cells by using 5 to 100 μmol/L of Y27632 or by expressing the dominant-negative ROCK mutant, RB/PH TT. ROCK activation was achieved by expressing a constitutively active mutant, CAT. ROCK inhibition led to morphologic and cytoskeletal alterations characterized by an increase in the number and length of cell processes, increased membrane ruffling, and collapse of actin stress fibers. Using two-dimensional radial migration and Boyden chamber assays, we show that astrocytoma migration and invasion were increased at least 2-fold by ROCK inhibition. On the contrary, ROCK activation significantly inhibited migration and invasion of astrocytoma cells. Furthermore, using a Rac-GTP pull-down assay, we show that Rac1 is activated as a consequence of ROCK inhibition. Finally, we show that treatment of astrocytoma cells with small interfering RNA duplexes specific for Rac1-reversed stellation, prevented membrane ruffling formation and abrogated the increased motility observed following treatment with Y27632. Our data show that Rac1 plays a major role in astrocytoma morphology, motility, and invasion. These findings warrant further investigation to determine precisely how the modulation of Rac1 and ROCK can be exploited to inhibit glioma invasion.
- cell motility
- astrocytoma invasion
Astrocytomas are the most common primary human brain tumors. The majority of astrocytomas are histopathologically malignant lesions associated with a poor prognosis. Despite aggressive surgical resection, radiation, and chemotherapy, most patients with glioblastoma multiforme will die within 12 months of their diagnosis. Although malignant astrocytomas rarely metastasize systemically, death results from inexorable local tumor growth and brain invasion ( 1). The infiltration of contiguous and distant regions of the brain by astrocytoma cells from the primary tumor is a major cause of treatment failure and tumor recurrence ( 2). Invasion is a complex cellular phenomenon involving cell/cell and cell/extracellular matrix adhesion, enzyme degradation of the extracellular matrix, and cell migration ( 1– 6). The intracellular mechanisms of astrocytoma migration are poorly understood but ultimately require a balance of environmental cues and responsive intracellular signals that lead to dynamic regulation of the interactions between actin microfilaments, microtubules, and intermediate filaments ( 7). The driving force for cell movement, however, is normally provided by dynamic reorganization of the actin cytoskeleton, directing protrusion at the front of the cell and retraction at the rear ( 7– 9). In this regard, the key mediators of actin cytoskeleton reorganization are the Rho family of GTPases.
Members of the Rho family of GTPases include Rho, Rac, and Cdc42 ( 7– 9). These molecules have been previously implicated in cell motility and invasive phenotypes ( 8– 10). In the process of cell migration, Rac is required for the formation of actin-rich membrane ruffles, called lamellipodia, at the leading edge of the migrating cells and is thought to be the driving force for cell movement ( 8– 11). Together with Cdc42, Rac also controls the turnover of focal adhesion complexes ( 11). Cdc42, although not directly required for cell migration, has been shown to be important in the regulation of cell polarity and filopodia formation, thereby controlling the direction of cell movement ( 11). Rho regulates the formation of contractile actin-myosin filaments, which form stress fibers, and maintains focal adhesions at the rear of the cell ( 8– 11).
One important Rho target involved in stimulating actin-myosin contractility and focal adhesion assembly is the Ser/Thr kinase, p160 Rho-kinase (ROCK; ref. 12). Expression of a dominant-negative form of this kinase or the addition of a specific inhibitor, Y27632, inhibits stress fibers and focal adhesions induced by lysophosphatidic acid (LPA) or activated Rho in several cell lines ( 12– 14). As with Rho, ROCK has been implicated in altering cell migration during tumor cell invasion and metastasis ( 12). To enhance our understanding of the molecular mechanisms underlying astrocytoma invasiveness, we undertook the present study to determine the effects of ROCK inhibition on astrocytoma morphology, migration, and signaling.
Materials and Methods
Chemical reagents. Y27632 [(+)-(R)-trans-4(1-aminoethyl)-N-(4-pyridyl)cyclohexanecarboxamide dihydrochloride] was purchased from Calbiochem (San Diego, CA) and reconstituted in H2O. The sodium salt of LPA (1-oleoyl-2-hydroxy-sn-glycero-3-phosphate) was purchased from Sigma (Oakville, Canada) and dissolved in PBS.
Plasmids. Expression vectors for mutant ROCK, pEF BOS myc RB/PH TT (dominant-negative ROCK, amino acids 941-1,388 with a NK to TT mutation of the Rho-binding domain at amino acids 1,036 and 1,037, respectively) and pEF BOS myc CAT (constitutively active ROCK, amino acids 6-553), were provided by Dr. Kaibuchi (Nara Institute of Science and Technology, Nara, Japan; ref. 15). These plasmids were digested with EcoRI and the myc RB/PH TT and myc CAT fragments were subcloned into pTRE (obtained from Dr. Ab Guha, University of Toronto, Toronto, Ontario, Canada) for regulation by tetracycline or its derivative, doxycycline.
Cell culture and transfections. The permanent human glioma cell lines U251, U87, SF539, SF126, and SF188 (obtained from the University of California at San Francisco; ref. 16) were maintained in DMEM containing 10% fetal bovine serum (FBS) at 37°C, 5% CO2 in a humidified chamber. Normal human astrocytes were purchased from Clonetics (East Rutherford, NJ) and maintained in AGM media containing 3% fetal bovine serum, 0.1% ascorbic acid, 0.5% recombinant human epidermal growth factor, 0.1% GA-1000, 0.25% insulin, and 1% l-glutamine (all from Clonetics). Transient transfections were done in U251 and U87 glioma cell lines by Effectene (Qiagen, Valencia, CA) following the protocol of the manufacturer. Cells transfected with empty plasmid vector were used as controls. Transfected cells were cultured for an additional 48 hours before use.
Inducible and stable expression of dominant-negative (myc-RB/PH TT) and constitutively active (myc-CAT) ROCK in U251 cells was accomplished using the tetracycline-repressor gene expression system as described previously ( 17). U251 cells were initially transfected with the Tet-off plasmid, pUHD-15-1neo, encoding the tet-responsive transcriptional activator downstream of a human cytomegalovirus promoter. Transfections were done using FuGENE 6 transfection reagent (Roche Diagnostics, Basel, Switzerland). Stable transfectants were selected with geneticin (G418 sulfate, Life Technologies, Inc., Gaithersburg, MD) and the expression of the tet-responsive transcriptional activator was determined by detecting the herpes simplex virus protein VP16 by Western blot analysis using a specific antibody (Clontech, San Jose, CA; ref. 18). A high VP16-expressing cell clone was selected and used to transfect pTRE RB/PH TT myc or pTRE CAT with pTK puro for expression of the puromycin resistance gene. Stably transfected lines were maintained in αMEM containing 10% tetracycline-free FBS, 300 μg/mL geneticin, and 1 μg/mL puromycin (Sigma). The inducibility of the RB/PH TT myc or CAT myc proteins was determined by detection of the myc tag with an anti-myc monoclonal antibody (mAb; 9E10, UBI, Lake Placid, NY) after treatment of cells with 2 μg/mL doxycycline for 24 hours. The effects of mutant ROCK expression were determined by confocal microscopy (see below).
Immunofluorescence and confocal microscopy. Cells were washed thrice with PBS and fixed with 4% paraformaldehyde for 10 minutes at room temperature. Following three washes with PBS, cells were permeabilized for 5 minutes with 0.5% Triton X-100 in PBS. Nonspecific binding was blocked by incubating cells with 1% bovine serum albumin (BSA) in PBS for 1 hour at room temperature. Primary antibodies were diluted in blocking buffer and incubated with an anti-myc mAb (1:200; UBI). The primary antibody was washed off with PBS and the cells were incubated with Alexa Fluor 594–conjugated goat anti-mouse immunoglobulin G secondary antibody for about 30 minutes. Filamentous actin was always visualized with TRITC Phalloidin at 1:50 dilution (Molecular Probes, Eugene, OR). A Zeiss Axiovert 100M confocal microscope (Carl Zeiss Inc., Thornwood, NY) or a Leica DMLB fluorescent microscope (Leica Microsystems, Inc, Bannockburn, IL) was used to visualize fluorescence.
Cell motility assay. Migration assays were done using the microliter scale radial monolayer assay as described ( 19, 20). Briefly, 10-well Teflon-coated slides (CSM Inc., Phoenix, AZ) were coated with 10 or 0.1 μg/mL laminin and 0.1% BSA. Cells were seeded through a cell sedimentation manifold (CSM) at 3,000 cells/well to establish a circular confluent monolayer at the center of the substrate-coated well. Several hours postseeding, a circle of best-fit that circumscribed the cells was drawn. The cells were allowed to migrate out over a 24-hour period and another circle circumscribing the cells was drawn. Migration results are reported as the change in the diameter of the circle circumscribing the cell population over a 24-hour period (microns per hour). Measurements are taken using a Leica DM IRE2 inverted microscope (Leica Microsystems), digitized using a Spot camera, and image analysis was done (Scion Image).
For treatment of cells, cells were seeded through the cell sedimentation manifold and allowed to adhere. Media was then exchanged with media containing either Y27632 (5-100 μmol/L) or LPA (1-100 μmol/L) and cells were followed for their migration. U251 and U87 cells transfected with empty plasmid vector, RB/PH TT, or CAT were also applied to the migration assay. Treatments or transfections were repeated in five replicates and experiments were repeated thrice, and the average migration rate of representative five replicates was calculated.
Cell invasion assay. Cell invasion assays were carried out using modified Boyden chambers consisting of Transwell chambers with precoated Matrigel membrane filter inserts in 24-well tissue culture plates (BD Biosciences Discovery Labware, Bedford, MA) as described previously ( 1). Serum-deprived cells (2 × 105) suspended in 100 μL of DMEM containing 1 mg/mL BSA and 0.5% serum were added to each Transwell. After 16 hours, noninvading cells were removed by wiping the upper side of the membrane, and the invaded cells were fixed with methanol and stained using crystal violet (Sigma). The number of invaded cells was quantified by counting them in at least six random fields (total magnification, ×200) per filter. In certain experiments, Y27632 (5 μmol/L) was applied to the upper chamber. U251 and U87 cells transfected with empty plasmid vector, RB/PH TT, or CAT were also applied to the invasion assay.
Myosin binding subunit phosphorylation. For the determination of phosphorylation of myosin binding subunit (MBS) of myosin light chain phosphatase, the CycLex Rho-kinase Assay Kit (MBL International, Nagano, Japan) was used. Briefly, 96-well plates were precoated with a substrate corresponding to the recombinant COOH terminus of MBS, which contains a threonine residue that is phosphorylated by Rho-kinase family members. Cells were harvested and lysed as described above (see Rac-GTP pull down assay). To perform the kinase assay, clarified lysates were diluted in kinase buffer, treated with 5, 25, or 100 μmol/L Y27632, pipetted into the wells, and allowed to phosphorylate the bound substrate in the presence of Mg2+ and ATP for 1 hour at 30°C. The amount of phosphorylated MBS was calculated in a binding reaction with a horseradish peroxidase conjugate of AF20, an anti–phospho-MBS threonine-696 specific antibody, which catalyzes the conversion of the chromogenic substrate tetra-methylbenzidine from a clear to a blue-colored solution. The color reaction product was quantified by measuring the absorbance at 450 nm using a VERSAmax microtiter plate reader (Molecular Devices Corp., Sunnyvale, CA). Each treatment was repeated in triplicate and results were plotted using the mean values and the SE.
Cell viability, apoptosis, and cell cycle assays. Cells were seeded in 96-well dishes at a density of 3,000 cells/well and allowed to adhere overnight before treatment. Cells were treated with 1, 5, 25, 50, and 100 μmol/L concentrations of Y27632 for 3 and 24 hours. Each concentration was tested in triplicate with MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (Promega, Madison, WI)] to assess cell viability. The MTS assay was done according to the instructions of the manufacturer. MTS coupled with an electron coupling reagent (phenazine ethosulfate) in a one-solution reagent is added to cells following treatment and incubated at 37°C for 1 hour. Absorbance of the reaction is measured at 520 nm with a VERSAmax microtiter plate reader (Molecular Devices). The quantity of formazan produced, as measured by absorbance, was directly proportional to the number of living cells in culture. Untreated cells were used as controls. Absorbances were plotted as a percentage of the control value. All experiments were repeated in triplicate.
For flow cytometric analysis, cells treated with 5, 25, or 100 μmol/L Y27632 for 24 hours were trypsinized, washed in PBS, and resuspended in ice-cold 80% ethanol. Briefly, ∼2.5 × 105 fixed cells were incubated in 250 μL propidium iodide solution (500 mg/mL propidium iodide in 3.8 mol/L sodium citrate at pH 7.0) and 250 μL RNase A (10 mg/mL prepared in 10 mmol/L Tris-HCl at pH 7.5) for 30 minutes at 37°C in the dark. The stained cells were filtered through the cell strainer caps of Falcon polystyrene round-bottomed tubes. DNA content was analyzed on a Becton Dickinson FACScan (Becton Dickinson, San Jose, CA). Percent cell cycle phase was determined using Cell Fit software (Becton Dickinson). Data were collected from at least 10,000 cells as described previously ( 21).
Rac-GTP pull-down assay. U251 cells were plated on 10 cm dishes to 80% confluency and serum starved for 24 hours. Cells were treated for 90 minutes with 5 μmol/L Y27632 and/or 5 μmol/L LPA for 2 minutes following serum starvation for 24 hours. Cells were washed twice with cold TBS and lysed in 50 mmol/L Tris (pH 7.2), 1% Triton X-100, 10 mmol/L MgCl2 in a cocktail of protease inhibitors (Roche Diagnostics). Lysates were clarified by centrifugation, and equal concentrations of lysates were incubated with 20 μg of purified glutathione S-transferase (GST)-CRIB (for Rac/Cdc42 activation; Cytoskeleton, Denver, CO) immobilized on glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech, Baie d'Urfé, Canada) for 1-hour rocking at 4°C. The beads were washed and bound Rac-GTP was detected using a mAb against Rac (1:500; Cell Signaling, Beverly, MA).
Rac1 small interfering RNA preparation and transfections. Small interfering RNA (siRNA) duplexes specific for Rac1 were designed according to Elbashir et al. ( 22). RNAs of 21 nucleotides were purchased from Dharmacon (Lafayette, CO) in deprotected and desalted forms. Two different Rac siRNA sequences were used: Rac1-1, corresponding to bp 439 to 459 after the start codon of the Rac1 gene (5′-AAGGAGATTGGTGGTGCTGTAAAA), and Rac1-2, corresponding to bp 754 to 774 after the start codon of the Rac1 gene (5′-AACCTTTGTACGCTTTGCTCA). The sequences for the siRNA duplexes to Rac were provided by Dr. Marc Symons (North Shore-Long Island Jewish Research Institute, Bronx, NY). As a control for off-target effects caused by RNA interference, nontargeting siRNAs, bioinformatically designed to have at least four mismatches with human and mouse sequences, were used (Dharmacon). Transient transfections of siRNA were carried out using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Cells were plated to 80% confluency in DMEM containing 10% serum without antibiotics. Transfections were carried out as soon as cells were fully spread, usually <24 hours later. The siRNA was diluted in serum-free DMEM. Lipofectamine 2000 was diluted in serum-free DMEM for 5 minutes. The two mixtures were combined and incubated for 20 minutes at room temperature to enable complex formation. Cells were assayed 3 to 4 days after transfection. Lipofectamine 2000-only transfections served as controls. Rac expression was determined by Western blot analysis using a mAb against Rac (Cell Signaling). A mAb against the human transferrin receptor was used on the same blot and served as a loading control (Zymed, San Francisco, CA).
Statistical analyses. Statistical analyses were done using the χ2 test and the two-tailed Mann-Whitney U test. P < 0.05 was considered significant.
Results and Discussion
Y27632 treatment leads to astrocytoma stellation and membrane ruffling. All astrocytoma cells treated with Y27632 showed a loss of actin stress fibers, adopting more of a stellate morphology with an increase both in the number and length of actin-rich cell processes, sometimes several times the length of the cell body ( Fig. 1 ). These changes occurred in a time- and dose-dependent manner ( Fig. 1B). At higher concentrations of Y27632 (e.g., 50-100 μmol/L), actin-rich microspikes were observed on the newly formed cell processes ( Fig. 1B, bottom left). Similar phenotypic changes have been reported in several other cell types previously studied ( 11, 12, 23– 29).
Another striking feature in response to treatment with Y27632 was the induction of actin-positive membrane ruffles ( Fig. 1A and B). The extent of membrane ruffling seen varied between the different cell lines. For example, SF188 and SF539 showed less peripheral actin staining ( Fig. 1A) than did SF126 and U251 ( Fig. 1A and B). At least four other studies have shown that ROCK inhibition induces membrane ruffling in different cell lines including PC3 prostate cancer cells, Swiss 3T3 cells, human umbilical vascular endothelial cells, and BE colon cancer cells ( 23, 27, 29, 30).
To determine if the effects of Y27632 on astrocytoma morphology were reversible, we did recovery experiments in which Y27632 was washed off with PBS and astrocytoma cells were incubated for an additional 24 hours before examination by confocal microscopy. Astrocytoma cells treated at all doses of Y27632 completely regained their original polygonal morphology, as seen in the untreated control cells, and regained multiple stress fibers spanning the length of the cell ( Fig. 1B, bottom right). These results indicate that the effects of Y27632 are reversible and that Y27632 does not have a toxic effect on astrocytoma cells.
Interestingly, normal human astrocytes were not as responsive to Y27632 as were the astrocytoma cell lines. Normal human astrocytes retained their polygonal shape following Y27632 treatment. Although the induction of thin actin-containing cytoplasmic processes was sometimes seen. Normal human astrocytes rarely lost stress fibers, or formed membrane ruffles in response to Y27632 ( Fig. 1C).
Molecular mimicry of Y27632 using a dominant-negative Rho-kinase mutant. ROCK is composed of catalytic, coiled-coil, Rho-binding (RB) and pleckstrin homology (PH) domains ( 31– 33). A constitutively active ROCK is created by expressing the NH2-terminal catalytic domain ( 34). Point mutations introduced in the COOH-terminal portion of Rho-kinase containing RB and PH domains abolish Rho-binding and inhibit Rho-kinase ( 34). Expression of this COOH-terminal portion (RB/PH TT) serves as a dominant-negative mutant ( 34). The expression of these constructs in U251 astrocytoma cells was confirmed by Western blot analysis and immunofluorescence for the myc-tagged antigen. Here, 60 and 49 kDa species corresponding to myc-CAT and myc-RB/PH TT, respectively, were detected. Expression of these ROCK-mutant proteins was eliminated with 2 μg/mL doxycycline for 24 hours ( Fig. 2A ).
In previous studies, immunofluorescence for ROCK has shown its distribution mainly in the cytoplasm ( 12), but some ROCK can translocate from the cell cytosol to the plasma membrane or cell periphery on stimulation by RhoA or by a constitutively active ROCK mutant ( 35). No detection of the myc antigen was seen in the doxycycline-treated controls, indicating that gene expression was successfully turned off ( Fig. 2B). The morphology of U251 astrocytoma cells expressing the dominant-negative mutant resembled that of cells treated with Y27632 in that they showed an increase in the length and number of actin-rich cell processes, a loss of actin stress fibers, and induced membrane ruffling ( Fig. 2B and C, top). Interestingly, many cells adopted a bipolar phenotype. On the other hand, U251 cells expressing the constitutively active variant of ROCK were pyramidal in shape. Pseudopodia formation became evident in some cells ( Fig. 2B and C, bottom). Addition of doxycycline to the media abolished these phenotypic changes and cells reestablished a morphology resembling parental U251 cells ( Fig. 2B and C).
Expression of the constitutively active ROCK mutant in U251 led to the formation of an aberrant form of centrally contracted and stellate actin bundles, not seen in cells treated with doxycycline ( Fig. 2C, bottom). One possible explanation for these abnormally situated actin fibers comes from the study by Tsuji et al. in which mDia and ROCK were shown to be antagonistic molecules, both contributing to the proper alignment of actin microfilaments and microtubules ( 12, 29). Accordingly, the overexpression of a constitutively active ROCK mutant, as seen in our experiments, could potentially disturb the required balance between ROCK and mDia to maintain a well-aligned actin microfilamentous system.
Rho-kinase inhibition stimulates astrocytoma motility. A microscale radial monolayer migration assay was used to asses the effects of Y27632 on the migration rate of astrocytoma and normal human astrocyte cells ( 24). Cells were either plated onto BSA, to serve as a control nonpermissive substrate, or onto laminin, a migration stimulating substrate. For cells seeded on laminin, a densely packed core of cells with a rim of actively migrating cells was observed. Morphologic changes in U251 and U87 cells treated with varying doses of Y27632 were also clearly seen in the migrating rim cells ( Fig. 3A ). When concentrations of Y27632 were increased, the astrocytoma cells became more stellate and extended longer cell processes. Interestingly, with laminin as a substrate, normal human astrocytes were sensitized to morphologic changes. Normal human astrocytes became elongated, less spread out, and many cells adopted a bipolar shape ( Fig. 3, bottom). With increasing doses of Y27632, normal human astrocytes extended longer, thinner cell processes and became more stellate in shape.
An important observation in our study was the change in astrocytoma migration after ROCK inhibition with Y27632 and dominant-negative ROCK (RB/PH TT) or activation with constitutively active ROCK (CAT). Previous studies have shown either an increase or a decrease in migration following ROCK inhibition, depending on the cell line studied or the model system used ( 11, 24– 28, 30, 36– 39). In our study, Y27632 increased the motility of all astrocytoma cells to varying degrees ( Fig. 3B; data not shown for all cell lines). Y27632 (5 μmol/L) stimulated the migration of U251 astrocytoma cells by at least 2-fold ( Fig. 3B). U87 and normal human astrocytes were moderately sensitive to Y27632-induced migration ( Fig. 3B; data not shown for normal human astrocytes). Inhibition of ROCK by RB/PH TT also induced an increase in the migration rate in U251 cells relative to mock plasmid–transfected cells (2.2-fold). However, the migration of U251 was significantly suppressed by the transfection with CAT (0.37-fold). Furthermore, Y27632 stimulated the motility of CAT-expressing cells, confirming the validity of our findings. Similar outcomes were obtained in the migration of U87, but to a lesser degree ( Fig. 3B).
To confirm that ROCK inhibition–induced motility is not a laminin-specific effect, astrocytoma cells were seeded onto BSA as a control and the migration assay was done in a similar fashion. Under these conditions, the same motility-stimulating effects of Y27632 and RB/PH TT were observed. As well, in this setting, CAT was able to inhibit migration ( Fig. 3B). The relative fold increases in migration between cells seeded on laminin and BSA were approximately the same, but the actual migration rates on BSA were somewhat lower. These data indicate that although laminin is capable of enhancing the migration of astrocytoma cells and other cell types ( 40), its effects on migration are likely separate from those induced by ROCK inhibition.
To complement the two-dimensional radial cell migration assay, we tested the effects of ROCK inhibition and activation in a Boyden chamber cell invasion assay. As shown in Fig. 3C, cells increased their invasion of Matrigel-coated membranes when treated with Y27632 (5 μmol/L) by ∼3.8-fold in U251 and by 1.9-fold in U87 compared with untreated cells. RB/PH TT also effectively stimulated the invasion of U251 and U87 glioma cells (2.3- and 1.5-fold, respectively) compared with mock-transfected controls. However, transfection with CAT decreased invasion (0.3- and 0.43-fold for U251 and U87 astrocytoma cells, respectively; Fig. 3C). Treatment of astrocytoma cells with Y27632 consistently overcame the invasive phenotype of CAT-expressing cells by severalfold in both U251 and U87 cells. Taken together, these data indicate that inhibition of ROCK enhances both astrocytoma cell migration and invasion, and activation of ROCK is clearly inhibitory to the migratory and invasive phenotype of astrocytoma cells.
To show that the molecular effects of Y27632 were intact in our system, we tested the phosphorylation status of MBS at serine-696 of myosin light chain phosphatase, a downstream effector known to be phosphorylated by ROCK ( 12). Phosphorylation of residues in the COOH-terminal half of MBS results in the inhibition of myosin light chain phosphatase activity and in a concomitant increase in phosphorylated myosin light chain and cell contractility. In our study, Y27632 (5, 25, and 100 μmol/L) effectively inhibited the phosphorylation of MBS in U251 astrocytoma cells ( Fig. 3D).
Y27632 does not affect astrocytoma cell viability or proliferation. To rule out an increase or a decrease in cell number following Y27632 treatment, the viability of astrocytoma cells treated with Y27632 was assessed by the MTS assay. The viability of U251 cells was unaffected at all concentrations of Y27632 tested and compared favorably with control, untreated cells ( Fig. 4A ). Furthermore, by fluorescence-activated cell sorting (FACS) analysis, Y27632 treatment at 5 and 25 μmol/L had no effects on the cell cycle progression ( Fig. 4B, M2 and M3 peaks). There was also no evidence of DNA degradation following Y27632 treatment ( Fig. 4B, M1 peak).
Astrocytoma motility stimulated by Rho-kinase inhibition is associated with Rac1 activation. The observation that ROCK inhibition causes increased membrane ruffling suggested that Rac activation was a possible mechanism for Y27632-induced motility of astrocytoma cells. Rac-GTPases are thought to regulate membrane ruffling formation and cell migration largely by stimulating actin polymerization. Figure 5A illustrates the formation of membrane ruffles in Y27632-treated cells and their disappearance in Rac1-depleted cells even in the presence of Y27632. In addition, these cells also seem to regain a polygonal shape and lose the stellation induced by Y27632, suggesting that membrane ruffling and stellation in astrocytoma cells are Rac1-dependent phenomena.
To confirm that Rac activation is an outcome of ROCK inhibition, we have used Rac-GTP pull-down assays in astrocytoma cells. U251 cells treated with 5 μmol/L Y27632 or transfected with RB/PH TT showed increased Rac activity compared with mock-transfected controls ( Fig. 5B). Furthermore, cells transfected with CAT inhibited Rac activity ( Fig. 5B). In addition, LPA alone (5 μmol/L) or in combination with Y27632 also increased the levels of Rac-GTP when compared with the untreated control ( Fig. 5B). In Swiss 3T3 cells, Tsuji et al. ( 29) showed that ROCK inhibition increased Rac activity only in LPA-treated cells. Although the activation of Rac following ROCK inhibition results from the effects of Rho on Rac, as was originally observed in neuronal cells or Swiss 3T3 cells ( 29, 41, 42), other mechanisms may also be involved.
The mechanism by which LPA activates RhoA to induce cytoskeletal contraction is through the Gα12/13 subunits of G-protein–coupled receptors, which bind directly to at least three distinct Rho-specific guanine nucleotide exchange factors (RhoGEFs) that promote RhoA-GTP accumulation ( 43, 44): ROCK, which is a downstream effector, is activated and leads ultimately to actomyosin-based contractile events ( 43, 44). Other authors have shown that LPA, acting via LPA1 receptors, can activate Rac and thereby promote cell spreading, lamellipodia formation, and cell migration ( 45, 46). Through diverse complementary approaches, it was established that the Rac-specific guanine nucleotide exchange factor Tiam1 is responsible for Rac activation by LPA ( 45, 46). Taken together, we suggest that Rac activation may follow from two separate pathways: one involving LPA and the other involving the inhibition of ROCK. Indeed, the exact role of Rho in the context of these findings is currently being examined in our laboratory.
To further examine the role of Rac in astrocytomas, we did Rac1-directed siRNA experiments. Figure 5C confirms the reduction of Rac protein expression in Rac1-depleted U251 cells compared with untransfected and nontargeting siRNA controls. Cells transfected with Rac1 siRNA alone were then placed in the radial cell migration assay. Rac1 depletion with either the Rac1-1 or Rac1-2 siRNA oligo inhibited the migration of U251 cells ∼2-fold ( Fig. 5D). Whereas the effects of Rac in glioma migration have been previously reported ( 47, 48), here we show that Rac1-directed siRNA treatment effectively abrogated the Y27632-induced motility of astrocytoma cells by nearly one half ( Fig. 5D).
Whereas several studies have shown that the mechanism that leads to reduced migration following ROCK inhibition is the reduction of myosin light chain phosphorylation and loss of stress fiber and focal adhesions, fewer studies have examined the mechanisms by which ROCK inhibition can lead to increased cell motility ( 12). Interestingly, Lepley et al. ( 49) have recently shown that ROCK is required for sphinogosine-1-phosphate–mediated inhibition of migration in U118 astrocytoma cells. When U118 cells are pretreated with Y27632 and then exposed to sphinogosine-1-phosphate, migration is enhanced ( 49).
Lysophosphatidic acid stimulates astrocytoma motility. LPA is a potent inducer of cell proliferation, survival, and migration in many cell lines ( 43, 44). Manning et al. ( 50) have previously shown that LPA stimulates motility of normal rat astrocytes and astrocytoma cell lines in a transwell migration assay. LPA is also well known for its contractile responses in nonmuscle cells, as was first observed in LPA-treated neuronal cells undergoing rapid neurite retraction and cell rounding ( 43, 44). LPA not only signals through classic second messenger pathways but also activates the Rho-family of GTPases to control migration and morphogenesis ( 43, 44).
In our study, 5 and 10 μmol/L LPA stimulated the motility of U251 cells about 3-fold ( Fig. 6 ). Given the fact that LPA activates Rac, we wanted to test the dependency of Rac on LPA-induced motility. Cells treated with LPA and depleted for Rac1 by siRNA lose their ability to migrate compared with untreated and nontargeting siRNA controls ( Fig. 6). In contrast to U251 cells, U87 motility was not as responsive to LPA stimulation (data not shown). The reasons for the differences observed in migration rates in the different astrocytoma cell lines following treatment with Y27632 or LPA are not clear. Presumably, differences in cell line–specific expression of various extracellular matrix proteins might modulate the degree by which ROCK inhibition could induce migration ( 51). Ongoing work in our laboratory is currently investigating the role of different extracellular matrix components as substrates for migration in the context of ROCK inhibition.
In summary, our data show that ROCK plays a major role in regulating astrocytoma cell morphology, the actin microfilamentous system, cell migration, and invasion. These cell biological alterations are associated with the activation of Rac1 through ROCK inhibition or LPA stimulation. Future studies will be directed towards a better elucidation of the cross-talk between the different Rho-GTPase family members and their effector proteins in astrocytomas. These studies could then serve as the basis for the inhibition of astrocytoma invasion using the Rho-GTPase family members as targets.
Grant support: Canadian Institutes of Health Research (MOP-74610), Brainchild, the Wiley Fund, the Laurie Berman Fund for Brain Tumor Research, Canadian Institutes of Health Research (J.T. Rutka), Restracomp fellowship from The Hospital for Sick Children, and Canada Graduate Scholarship Award from the Canadian Institutes of Health Research (B. Salhia).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received January 18, 2005.
- Revision received June 30, 2005.
- Accepted August 1, 2005.
- ©2005 American Association for Cancer Research.