Inhibition of the Gtpase Rac1 Mediates the Antimigratory Effects

Published OnlineFirst December 19, 2014; DOI: 10.1158/1535-7163.MCT-14-0102

Cancer Biology and Signal Transduction Molecular Cancer Therapeutics Inhibition of the GTPase Rac1 Mediates the Antimigratory Effects of Metformin in Prostate Cancer Cells Beatrice Dirat1,2, Isabelle Ader3,4, Muriel Golzio3,4, Fabienne Massa1,2, Amel Mettouchi2,5, Kathiane Laurent1,2,Fred eric Larbret6, Bernard Malavaud3,4,7, Mireille Cormont1,2, Emmanuel Lemichez2,5, Olivier Cuvillier3,4, Jean Francois¸ Tanti1,2, and Fred eric Bost1,2

Abstract

Cell migration is a critical step in the progression of prostate migration. We report that metformin leads to a major inhi- cancer to the metastatic state, the lethal form of the disease. bition of Rac1 GTPase activity by interfering with some of its The antidiabetic drug metformin has been shown to display multiple upstream signaling pathways, namely P-Rex1 (a Gua- antitumoral properties in prostate cancer cell and animal nine nucleotide exchange factor and activator of Rac1), cAMP, models; however, its role in the formation of metastases and CXCL12/CXCR4, resulting in decreased migration of pros- remains poorly documented. Here, we show that metformin tate cancer cells. Importantly, overexpression of a constitutive- reduces the formation of metastases to fewer solid organs in an ly active form of Rac1, or P-Rex, as well as the inhibition of orthotopic metastatic prostate cancer cell model established in the adenylate cyclase, was able to reverse the antimigratory nude mice. As predicted, metformin hampers cell motility in effects of metformin. These results establish a novel mecha- PC3 and DU145 prostate cancer cells and triggers a radical nism of action for metformin and highlight its potential reorganization of the cell cytoskeleton. The small GTPase Rac1 antimetastatic properties in prostate cancer. Mol Cancer Ther; is a master regulator of cytoskeleton organization and cell 14(2); 586–96. 2014 AACR.

Introduction effect on cancer cells; however, its mechanism of action remains largely unknown (8–13). In addition, how metformin interferes Metformin is an antidiabetic drug used by more than 120 with the small GTPase Rac1, one of a master regulator of cell million people worldwide. In agreement with retrospective epi- migration, is not known. demiologic studies in which diabetic patients on metformin Rac1 belongs to the family of the Rho GTPases that play a display decreased cancer incidence and cancer-related mortality central role in the control of cytoskeleton organization and cell (1–3), metformin has been shown to inhibit cancer cell prolifer- motility. The best characterized family members are: Rho, ation and decrease tumor growth in many animal models (4–7). involved in stress fibers and focal adhesion formation, together Prostate cancer is the second leading cause of death by cancer in with Rac and Cdc42, respectively, involved in lamellipodia and men, and most prostate cancer-related deaths are due to metas- filipodia formation (14). Rho GTPases switch from a GTP-bound tasis, a process that requires cancer cell migration. This migration active form to a GDP-bound inactive form. The exchange of is a complex biologic process orchestrated by environmental GDP to GTP is regulated by guanine nucleotide exchange factors factors, signal transduction, and cytoskeletal rearrangement. Sev- (GEF), and the inactivation of Rho GTPases is controlled by eral studies demonstrated that metformin exerts an antimigratory GTPase-activating enzymes. The Rac subclass (or subfamily) of RhoGTPases includes Rac1, Rac2, and Rac3. Rac1 is required for 1INSERM, C3M, U1065, Team Cellular and Molecular Physiopathology lamellipodium extension induced by growth factors, cytokines, of Obesity and Diabetes, Nice, France. 2Univ. Nice Sophia Antipolis, and extracellular matrix (ECM) components (15). Rac1 is over- C3M, U1065, Nice, France. 3CNRS, Institut de Pharmacologie et de 4 expressed in cancers, including prostate cancer, in which its Biologie Structurale, Toulouse, France. Universite de Toulouse, UPS, fi IPBS,Toulouse,France. 5INSERM, C3M, U1065, Equipe Labellisee Ligue expression is signi cantly increased in aggressive tumors (16). Contre le Cancer, Team Microtoxins in Host Pathogens Interactions, The PIP3 phosphatidylinositol (3,4,5)-triphosphate–dependent Nice, France. 6University of Nice Sophia Antipolis, EA6302, Flow Rac exchanger 1 (P-Rex1), a Rac-selective GEF, plays an important ^ 7 ^ Cytometry Facility, Hopital l'Archet 1, Nice, France. Hopital Rangueil, role in actin remodeling and cell migration. Importantly, upre- Service d'Urologie et de Transplantation Renale, Toulouse, France. gulation of P-Rex1 promotes metastasis whereas its downregula- Note: Supplementary data for this article are available at Molecular Cancer tion inhibits cell migration in prostate cancer cells (17). Therapeutics Online (http://mct.aacrjournals.org/). Rac1 activity is regulated by numerous biologic signals, such as Corresponding Author: Fred eric Bost, C3M, Batiment ARCHIMED, 151 route de cAMP and cytokines. Recent studies have highlighted an impor- St Antoine de Ginestiere, BP2 3194, Nice 06204 Cedex 3, France. Phone: 33-48- tant role for cAMP metabolism in the migration of carcinoma cells 906-4229; Fax: 33-48-906-4221; E-mail: [email protected] (18) and the regulation of Rac1 activity (19). For example, cAMP- doi: 10.1158/1535-7163.MCT-14-0102 specific phosphodiesterases facilitate cell migration as well as 2014 American Association for Cancer Research. lamellae formation by lowering cAMP levels. In addition, Chen

586 Mol Cancer Ther; 14(2) February 2015

Metformin Inhibits Rac1 GTPase and Cancer Cell Migration

and colleagues (20) have shown that increased cAMP levels expression vectors: RacQ61L and RacV12 expression vector using correlated with the inhibition of cell migration in both Mouse Lipofectamine 2000 (Invitrogen). Embryonic Fibroblasts and 4T1a breast tumors cells by interfering with the formation of lamellipodia. Chemicals Chemokines are also important regulators of Rac1, one of Metformin, the adenylate cyclase inhibitor (SQ22536), dibu- them, CXCL12 (also known as SDF-1a) activates Rac1, decreases tyryl-cAMP (dbcAMP), and fibronectin were purchased from – cAMP levels, and favors prostate cancer cells migration (21 23). Sigma-Aldrich. The Rac inhibitor and AMD3100 were from Merck In addition, CXCR4, the CXCL12 receptor, is frequently over- Chemicals. CXCL12 was purchased from Peprotech. expressed in malignant epithelial cells, and the CXCL12/CXCR4 axis plays a pivotal role in directing the metastasis of CXCR4- positive tumor cells to organs that express CXCL12, such as lungs, Boyden chamber assay fi fi liver, and bones (24, 25). Here, we investigated the effects of Boyden chambers with lter inserts coated with bronectin (10 metformin on Rac1 GTPase activity and determined whether it mg/mL) and 8-mm pores (BD Biosciences) were used to quantify interferes with some of Rac1 multiple upstream signaling path- cell migration. To respond better to the chemoattractant cells were 4 ways, namely P-Rex1, cAMP, and CXCL12/CXCR4. serum starved overnight, 12 10 cells were seeded in the upper We demonstrate that metformin inhibits the migration of chamber in serum-free DMEM medium in the presence or absence prostate cancer cells and limits the formation of metastasis to of 5 mmol/L metformin. The lower chamber contained complete fewer solid organs in an orthotopic xenograft model using PC3 DMEM, 10% FBS, or DMEM with CXCL12 at the indicated cells. In addition, we show that metformin strongly modifies concentration. Cell migration was determined after 4 hours by fi actin cytoskeletal organization. Reversal of the decreased counting all cells in ve randomly selected counting areas at the fi Rac1GTPase activity through the expression of constitutively lower surface of the lter. Cells on the upper surface were removed fi fi active Rac1GTP or P-Rex1, overturned the antimigratory effects with a cotton swab; lters were xed and stained with blue of metformin. Similarly, blocking the metformin-induced toluidin. Each experiment was repeated at least three times. For cAMP increase with an adenylate cyclase inhibitor hampered invasion experiments, the inserts were coated with 25 mg/mLof the effects of metformin on migration. We also show that Matrigel (Beckton Dickinson), and invading cells were counted metformin inhibits CXCL12 chemotactism and counteracts the after a 24-hour incubation with metformin. increase of Rac1GTP by CXCL12. Our study reveals a novel mechanism of action for metformin, in which it targets Spheroid migration assays in three-dimensional Matrigel Rac1GTPase and cytoskeletal organization. matrices Prostate spheroids were generated using the liquid overlay Materials and Methods technique. Briefly, 24-well culture plates were coated with 1.5% Orthotopic implantation of PC3–GFP prostate cancer cells and agarose prepared in sterile water. Cells from a single-cell suspen- analysis of metastasis sion were added at 10,000 cells per well. The plates were gently Intraprostatic human prostate cancer xenografts were estab- swirled and incubated at 37 Cin5%CO2 atmosphere until lished in nude mice by surgical orthotopic implantation as spheroid aggregates were formed. Then, spheroids were included originally described (26). Briefly, mice were anesthetized by in a Matrigel matrix and images of invasion were obtained 24 isoflurane inhalation and placed in the supine position. A lower hours later. midline abdominal incision was made and a tumor cell suspen- 6 sion (1 10 cells/20 mL) was injected into the dorsal lobe of the Cell migration observation with video microscopy prostate using a 30-gauge needle and glass syringe (Hamilton). Cell migration was monitored in duplicate experiments by After implantation, the surgical wound was closed in two layers time-lapse digital microscopy. Cells were seeded on a 6-well plate with 4-0 Dexon-interrupted sutures. All procedures were per- at low density. Computer-assisted cell tracking of 20 to 30 formed with a dissecting microscope. Autopsy and in vivo fluo- randomly selected cells was performed. Briefly, the x and y rescence imaging were conducted as previously detailed. The coordinates were collected from the center of the cell with a step measurements were performed blinded. Animal use and care was interval of 5 minutes and reconstructed either as path at ortho- approved by the local Animal Care committee according to the topic position or as migration speed over time. European Legislation. fl fl Cell culture and transfection Immuno uorescence and uorescence microscopy fi The human PC3 and DU145 cancer cell lines were obtained Cells grown on coverslips were xed in 3.7% paraformalde- from the ATCC and authenticated by the ATCC, the experiments hyde, permeabilized in 0.2% Triton X-100 for 20 minutes, performed in this work were performed during the year and half blocked with 2% BSA for 1 hour (all reagents were diluted in following the reception of the cells. Cells were grown in DMEM PBS), and then incubated with Texas red Phalloidin and anti-HA supplemented with 10% FCS, 100 U/mL penicillin, and 50 mg/mL antibodies (Covance). Cells were simultaneously stained with streptomycin. GFP-expressing PC3 cells were generated as Hoescht. Images were recorded with a Leica scanning microscopy described previously (26). Cells were maintained at 37Cina system DM5500B. Image acquisition and image analysis were performed on the C3M (or MicorBio) Cell Imaging Facility. humidified atmosphere of 5% CO2, and the media were replaced every 2 to 3 days. In all experiments, cells were treated for 4 hours with 5 mmol/L metformin. Cells were transiently transfected with Western blot analysis HA–P-rex1–expressing vector (kind gift of Dr. C. Mitchell, Mon- A total of 40 mg cell lysate protein was separated by SDSPAGE, ash University, Victoria, Australia; ref. 27), or the Rac1-mutant transferred on a polyvinylidene difluoride membrane (Millipore),

www.aacrjournals.org Mol Cancer Ther; 14(2) February 2015 587

Dirat et al.

and incubated with the antibodies against Rac1 (BD Biosciences); AMP–responsive element (CRE) coupled to the luciferase gene ERK and HSP90 (Santa Cruz Biotechnology); HA (Covance). (CRE–Luc), and 1 mg of pRL Renilla Luciferase Control vector. Two days after transfection, the culture medium was discarded and cells Pull-down assay for the measurement of RAC1 GTP activity were treated with DMEM supplemented with 10% FCS 5mmol/L The assay was performed as previously described (28). DU145 metformin. After a 4-hour incubation at 37 C, the stimulation and PC3 were seeded in DMEM medium with 10% FBS. After a 4- medium was discarded and the Luciferase activity was determined hour treatment with 5 mmol/L metformin, cells were washed once using the Dual Luciferase reporter Assay System (Promega). with ice-cold PBS and immediately lysed in 25 mmol/L Tris buffer, pH 7.5, 150 mmol/L NaCl, 5 mmol/L MgCl2, 0.5% Triton X100, Statistical analysis 4% glycerol, 10 mmol/L sodium fluoride, 2 mmol/L sodium The statistical significance of differences between the means of orthovanadate, 5 mmol/L DTT, 1 mmol/L phenylmethylsulfonyl- two groups was evaluated using the Student t test. fluoride. Cleared extracts were mixed with 20 mgofGST–PAK in the presence of glutathione–agarose beads (Sigma-Aldrich). After a 40- Results minute incubation at 4C, beads were pelleted by centrifugation Metformin inhibits tumor growth and reduces metastasis in an and washed three times in lysis buffer, and the proteins were eluted orthotopic model of PC3 cells in SDSPAGE sample buffer for analysis by Western blot analysis We first investigated the effects of metformin on the formation using a monoclonal antibody to Rac1 (BD Biosciences). of metastases using an orthotopic model of PC3 cells overexpres- sing GFP. In these experimental conditions, cells grow in their cAMP concentration native environment, and the primary tumor forms distant metas- cAMP levels were assessed using a commercially available tasis (26, 29). Tumor growth and metastasis dissemination were fl uorimetric kit (Arbor Assays). In brief, DU145 or PC3 were analyzed 5 weeks after the injection of PC3–GFP cells into the seeded in 6-well plates and half the wells were treated with 5 prostate. Metformin (100 mg/kg/d) was given in drinking water mmol/L of metformin for 4 hours before cAMP measurement for 5 weeks (Met, 5 wk) starting 3 days after cell injection or only 2 performed according to the manufacturer's protocol. final weeks (Met, 2 wk). A group was injected i.p with docetaxel (20 mg/kg) for the last 2 weeks. Metformin had no toxic effect on Flow cytometry mice, and it did not affect animal weight and insulinemia (Sup- Cells were harvested after 4 hours of metformin treatment (5 plementary Fig.S1). A whole-body open imaging of the animals mmol/L). Cells were labeled with anti–CRCR4-APC–conjugated revealed the fluorescence of primary tumors and metastases, antibody (R&D Systems) and fixed in PAF 3.7% for 10 minutes. including periaortic and periadrenal lymph nodes, liver, pancreas, Labeling was carried out in ice for 2 hours. Cells were then washed lungs, and mesentery, indicating a disseminating disease as in PBS 0.5% BSA at 1,100 rpm for 5 minutes and resuspended in described previously (29). A representative picture of the GFP- 400 mL of PBS. For each tube, 10,000 events were acquired. Samples positive tumors is shown in Fig. 1A. As expected, the tumors were were analyzed using FACSCanto II cytometer (Beckton Dickinson). significantly smaller in the docetaxel-treated group and metformin induced a strong antitumoral effect. Indeed, it significantly CRE–luciferase reporter gene assay reduced by more than 50% the growth of the primary tumor when PC3 and DU145 cells were transiently transfected using lipo- given for 5 weeks (Fig. 1 A and B). However, when administrated fectamine 2000 with 1 mg of a plasmid encoding for the cyclic only during the last 2 weeks like docetaxel, metformin did not

A B

400 )

3 350 300 250 C Met 5w 200 * 150 100 * 50

Tumor volume (mm 0 C

Met 2w Docetaxel Met 5 wk Met 2 wk Docetaxel

Figure 1. Effects of metformin on growth of established human fluorescent PC-3 orthotopic xenografts in nude mice. The day of the orthotopic implantations with 1 106 PC-3 cells, mice were randomized into four groups. Animals were given drinking water (control, C) or 100 mg/kg metformin (Met, 5 wk) in drinking water for 5 weeks. Alternatively, 3 weeks after implantation, animals were treated for the final 2 weeks with metformin in drinking water (Met, 2 wk) or weekly i.p. with 20 mg/kg docetaxel A, representative photos of the primary tumors (magnification 0.8). The tumors are in green and the bladder appears in orange (autofluorescence) on the picture. B, tumor volume calculated as described in Materials and Methods. Columns, mean from 6 to 8 animals; bars, SEM. The statistical analysis was performed using the Student t test; , P < 0.05.

588 Mol Cancer Ther; 14(2) February 2015 Molecular Cancer Therapeutics

Metformin Inhibits Rac1 GTPase and Cancer Cell Migration

Table 1. Pattern of metastatic dissemination Treatments Sham Metformin 5 weeks Metformin 2 weeks Docetaxel 2 weeks Number of mice with metastases/total number of mice 7/7 (100%) 6/8 (75%) 6/6 (100%) 7/7 (100%) Retroperitoneal lymph nodes Periaortic lymph nodes 2, 2, 2, 2, 2, 2, 1 2, 0, 2, 2, 2, 2, 2, 0 2, 2, 2, 2, 2, 2 2, 1, 1, 2, 1, 2, 2 Periadrenal 2, 2, 1, 2, 2, 1, 1 0, 0, 2, 2, 2, 2, 1, 0 1, 2, 2, 1, 2, 1 2, 0, 0, 2, 0, 0, 0 Number of metastases 24/28 (3.5/animal) 21/32 (2.6/animal)NS 21/24 (3.5/animal)NS 15/28 (2.1/animal)a

Liver 1, 1, 1, 1, 1, 1, 1 0, 0, 1, 1, 0, 1, 1, 0 0, 1, 1, 1, 0, 1 0, 0, 0, 1, 0, 0, 1 Pancreas 1, 1, 1, 1, 1, 1, 1 0, 0, 1, 1, 0, 1, 1, 0 1, 1, 1, 1, 0, 1 0, 0, 0, 1, 0, 1, 1 Lung 1, 1, 1, 1, 1, 1, 0 0, 0, 1, 1, 1, 1, 0, 0 1, 1, 1, 1, 0, 1 1, 0, 0, 1, 0, 1, 0 Mesentery 1, 1, 0, 1, 1, 1, 1 0, 0, 1, 1, 0, 1, 1, 0 1, 1, 1, 1, 0, 1 0, 0, 0, 1, 0, 0, 1 Number of metastases 26/28 (3.7/animal) 16/32 (2.0/animal)b 19/24 (3.2/animal)NS 10/28 (1.4/animal)c NOTE: For retroperitoneal lymph nodes, the numbers 0, 1, or 2 represent the quantity of invaded lymph nodes. For other organs, 0 means no presence of metastasis; 1 means a metastatic organ (regardless of the intensity of metastasis dissemination in this organ). Abbreviation: NS, not signiﬁcant. aP ¼ 0.049, compared with sham-treated animals (t test). bP ¼ 0.040, compared with sham-treated animals (t test). cP ¼ 0.002, compared with sham-treated animals (t test). have any impact on tumor growth (Fig. 1 A and B). Our ﬁndings matrix in a collective migration/invasion pattern. Spheroids trea- show that metformin has a preventative effect on primary tumor ted with metformin remained compact with almost no cells growth, but yet does not manifest a curative effect when the tumor migrating out (Supplementary Fig. S4). We then tracked individ- is already established. Interestingly, the dissemination pattern of ual cell migration over a period of 12 hours using time-lapse video metastases showed that all mice had metastases regardless of the microscopy. Untreated PC3 cells moved in several directions over treatment except 2 mice in the "metformin 5 weeks" group (Table an extended area compared with those treated with metformin 1). Among the 3 mice without solid metastasis 2 of them had (Supplementary Fig. S5). The total accumulated distance covered primary tumors bigger than the average tumor volume of the by the untreated cells was 1,329.1 369.2 mm versus 9.30 7.74 metformin 5 weeks group (162.51 and 295.64 vs. 135.53 mm3 for mm for those treated with metformin, and the mean euclidean the average tumor volume), suggesting that the absence of metas- distance (shortest linear distance between points A and B) tasis is not associated with small tumors. Only docetaxel signif- was 89.02 56.73 mm versus 6.47 4.22 mm. Metformin also icantly reduced the formation of retroperitoneal lymph nodes as affected cell velocity because untreated cells migrated at a 1.84 well as liver, pancreas, lung, and mesentary metastases (Table 1). 0.51 mm/min versus 0.012 0.01 mm/min for metformin-treated Nevertheless, mice from the "metformin 5 weeks" arm exhibited cells. Our results establish that metformin inhibits all movement statistically less metastasis (P ¼ 0.04), suggesting that metformin parameters of prostate cancer cells with a major inhibitory may hinder the metastatic dissemination. impact on their invasive properties.

Metformin inhibits the migration and the invasive properties of Metformin induces the reorganization of actin cytoskeleton PC3 and DU145 prostate cancer cell lines Because cells coordinate their migration through the regulation Because metastasis requires cancer cell migration, we investi- of actin dynamics (31), we studied the effect of metformin on gated the effects of metformin on human prostate cancer cell a-actin, b-actin, and fascin expression, three important proteins migration using Boyden chamber assay. According to our previ- implicated in cell migration. We did not observe any change in the ous studies (4, 30) and a dose response experiment (data not expression these proteins except a slight decrease of fascin expres- shown), we treated the cells with 5 mmol/L metformin. To sion in PC3 cells only (Supplementary Fig. S6). More importantly, exclude any action of metformin on cell proliferation, PC3 and we analyzed actin cytoskeleton organization, PC3 and DU145 DU145 were treated with metformin for 4 hours during the cells were seeded on fibronectin-coated wells and fluorescence migration toward culture medium supplemented with FBS (che- microscopy was used to analyze F-actin. In the control (untreated) moattractant medium). We monitored cell viability and apopto- conditions, elongated cells forming lamellipodia extensions rich sis in the same conditions. As expected, viabilities in all cell in F-actin and stress fibers and ruffle formations were visible (Fig. cultures treated for 4 hours with metformin exceeded 95% (data 2C). Four hours of treatment with 5 mmol/L metformin induced a not shown) and markers of apoptosis were negative (Supplemen- drastic change of cell morphology, with cells reorganizing their tary Fig. S2). Interestingly, a significant inhibitory effect of met- actin cytoskeleton, becoming circular, displaying less lamellipo- formin of 50% on the migration of PC3 and DU145 cells was dia (Fig. 2C). The shape of the PC3 and DU145 cells confirmed revealed (Fig. 2A). In contrary, metformin did not alter the that metformin treatment significantly decreased invasive mor- migration of normal epithelial prostate cells (P69 cells; Supple- phology (Fig. 2C). mentary Fig. S3). Next, we determined the impact of metformin on invasion. Cells were treated 4 hours with metformin before Metformin decreases Rac1 GTPase activity assessing 2D-invasion in Matrigel using Boyden chambers, as The known role of the small GTPase Rac1 as a major driver of described in Materials and Methods section. As shown in Fig. cell motility (32, 33) prompted us to assess Rac1 activity, using 2B, metformin strongly inhibited the invasive properties of PC3 GST-Pak pull-down assay, as described previously (34). Interest- and DU145 cells. To further explore whether metformin reduces ingly, this series of measurements revealed a significant decrease in invasion, we performed a spheroid assay with DU145 cells. Rac1–GTP levels in PC3 and DU145 cells treated with 5 mmol/L Untreated DU145 cells were able to invade the adjacent Matrigel metformin for 4 hours (Fig. 3A). Rho activity was not affected by

www.aacrjournals.org Mol Cancer Ther; 14(2) February 2015 589

Dirat et al.

A Migration PC3 DU145

120 120 100 100 80 80 * 60 ** 60 40 40 (% of control) 20 (% of control) 20 Relative migration 0 Relative migration 0 C M C M

B PC3 Invasion DU145 Figure 2. Metformin inhibits prostate cancer cell migration and invasion. A, PC3 and DU145 cells were seeded in Boyden chambers, and metformin (5 mmol/L) was added during the migration for 4 120 120 hours. Graphs are expressed as a percentage of 100 100 cells migrating across the Boyden chamber relative to the control conditions (100%), and the 80 80 * ** insets represent picture of the counted fields. B, 60 60 quantification of the invasion assay performed in 40 40 Boyden chambers during 24 hours in presence of 5 mmol/L metformin. C, immunofluorescence

(% of control) 20 20 (% of control) performed with Texas red Phalloidin in PC3 and Relative invasion 0 Relative invasion 0 DU145 treated with 5 mmol/L metformin for 4 C M C M hours. The statistical analysis was performed using the Student t test; , P < 0.05; , P < 0.01. C PC3

C M

15 µm 15 µm

DU145 C M

µ 15 µm 15 m

metformin, thereby pointing to a specific decrease in Rac1 activity forms of Rac1 (HA–Rac1–Q61L or HA–Rac1–V12) were over- (Supplementary Fig. S7). To establish the link between the inhi- expressed in PC3 and DU145 cells. In the presence of metformin, bition of migration and Rac1GTPase activity triggered by metfor- cells expressing the active form of Rac1 no longer displayed the min treatment, we used a Rac1 inhibitor that specifically and "rounded shape" phenotype that could be observed in nontrans- reversibly inhibits Rac1 GDP/GTP exchange activity, while exhi- fected cells (Fig. 3C). Furthermore, we found that the expression of biting no effect on Cdc42 or RhoA (35). We found that treatment the constitutive forms of Rac1 slightly but significantly inhibits of PC3 and DU145 cells phenocopied the effects of metformin on basal cell migration (Fig 3D). Importantly, the inhibitory effect of cell migration (Fig. 3B) and induced a circular cell morphology metformin on control PC3 and DU145 cell migration was abol- (Supplementary Fig. S8). To further gain insight in the relation- ished in cells expressing the constitutive forms of Rac1: Rac1– ship between metformin and Rac1GTPase, constitutively active Q61L or Rac1–V12 (Fig. 3D). This reveals that constitutive

590 Mol Cancer Ther; 14(2) February 2015 Molecular Cancer Therapeutics

Metformin Inhibits Rac1 GTPase and Cancer Cell Migration

A PC3 DU145 - + Metformin - + Rac1GTP

Total Rac1

Tubulin

Figure 3. 2 Metformin inhibits Rac 1 activity and * 2 * constitutively active Rac 1 restores cell 1 migration. A, immunoblot analysis of 1 Rac1 performed with DU145 and PC3 (arbitrary unit) (arbitrary unit) 0 prostate cancer cells treated with 5 Rac1GTP/total Rac1 0 Rac1GTP/total Rac1 mmol/L metformin for 4 hours as C M C M described in Materials and Methods. B PC3 120 DU145 The graph represents the 120 ﬁ quanti cation of the ratio of Rac1 100 100 * ﬁ GTP/Total Rac1. B, quanti cation of 80 the migration assay in Boyden 80 ** chambers of cells treated for 60 60 4 hours with the Rac inhibitor (50 40 40 mmol/L). C, DU145 transfected with 20 20 (% of control) HA-RacQ61L and treated with 5 (% of control) Relative migration

Relative migration 0 mmol/L metformin was analyzed by 0 C Rac Inh immunoﬂuorescence using Texas red C Rac Inh Phalloidin (red) and HA (green). D, C PC3 and DU145 were transfected with empty vector, active forms of Rac1: RacQ61L or RacV12 and treated with 5 mmol/L metformin for 4 hours during the migration assay. Columns are the mean of ﬁve independent experiments; bars are SEM. The 20 µm F-actin 20 µm HA-RacQ61L 20 µm Merge statistical analysis was performed using the Student t test; D 120 120 DU145 , P < 0.05; , P < 0.01. PC3 Control Metformin 100 100

80 80 ** ** 60 60

40 40 (% of control) (% of control) Relative migration Relative migration 20 20

0 0 Empty Rac61L RacV12 Empty Rac61L RacV12

activation of Rac1 overrides the effects of metformin on actin idea that the forced activation of Rac1 alleviates the metformin- cytoskeleton reorganization and cancer cell migration. mediated inhibition of cancer cell migration.

P-Rex1 overexpression reverses the antimigratory action of Metformin increases cAMP levels in prostate cancer cells metformin Because cyclic AMP inhibits Rac1 activity (37), we investigated P-Rex1 is a GEF that modulates cellular Rac1–GTP levels. It is whether cAMP acts as a potential mediator by which metformin implicated in cytoskeleton remodeling (36) and facilitates pros- modulates migration of prostate tumor cells. Accordingly, we tate cancer metastasis (17). We asked whether P-Rex1 overexpres- measured cAMP content in cells treated with metformin. We did sion (HA–P-Rex1 wt) reversed metformin effects on cell migra- not detect any change in cAMP concentration after 4 hours of tion. HA–P-Rex1 expression did not affect basal Rac1 GTP levels, treatment with metformin in PC3 cells (Supplementary Fig. S9). but restored Rac1 GTP levels in cancer cells treated with metfor- On contrary, metformin induced a slight but signiﬁcant increase min (Fig. 4A). Accordingly, the overexpression of wild-type P- in cAMP levels in DU145 (Fig. 5A), which was associated with the Rex1 does not affect cell migration but reversed the antimigratory augmentation of luciferase activity in cells transfected with the effects of metformin (Fig. 4B). Altogether, our results support the CRE–Luc construct (to monitor cAMP increase through the

www.aacrjournals.org Mol Cancer Ther; 14(2) February 2015 591

Dirat et al.

A B Control Metformin Empty vector HA-Prex1 WT 120 CM MC 100 Rac1 GTP 80 *

Total Rac1 60

HA (% of control) 20 Relative migration Ratio Rac1GTP HSP90 0 /Total Rac1: 0.701 1.06 0.96 Empty HA-Prex1 WT

Figure 4. Overexpression of P-Rex1 counteracts the antimigratory effects of metformin. A, immunoblot analysis of Rac1GTP and total Rac1 in DU145 cells transfected with the indicated vectors and treated for 4 hours with 5 mmol/L metformin. Expression of the HA-tagged proteins in the cell lysates was revealed by an anti-HA immunoblot. Similar results were obtained in three independent experiments. B, DU145 transfected with control vector (empty vector) or P-Rex1 expression vector (HA–P-rex1 wt) were assayed for migration during 4 hours in the presence or absence of 5 mmol/L metformin. The graph represents the average of three independent migration assays. The statistical analysis was performed using the Student t test; , P < 0.05.

activation of CREB, the cAMP response element–binding through the regulation of CXCR4 and Rac1 to inhibit prostate protein; Fig. 5B) and increased CREB phosphorylation (Supple- cancer cell migration. mentary Fig. S10). To firmly establish that increased cAMP is directly implicated in the antimigratory effects of metformin, we treated DU145 Discussion cells with SQ22536, an inhibitor of adenylate cyclase. Treat- Prostate cancer can be very aggressive in advanced stages and ment with 100 mmol/L SQ22536 prevented the increase of commonly metastasizes to bone and lymph nodes, more rarely to cAMP (Fig. 5A) as well as the decrease in cell migration (Fig. the liver and lung and cell migration, which is required for 5C) induced by metformin, while leaving basal cAMP concen- metastasis, is a complex biologic process regulated by environ- tration and basal cell migration unaffected (Fig. 5A and C). To mental factors, signaling pathways and cytoskeletal rearrange- directly observe the effects of elevated cAMP on cell migration, ment. Here, we report that the antidiabetic drug metformin we treated DU145 cells with 500 mmol/L of dbcAMP, a cell- reduces the formation of metastasis to fewer solid organs in an permeable cAMP analogue. A 4 hours treatment with dbcAMP orthotopic mouse model and affects cell cytoskeleton organiza- inhibited the migration of DU145 cells (Fig. 5D) and decreased tion, which drastically inhibits prostate cancer cell migration Rac1 activity (Supplementary Fig. S11). Importantly, the over- through decreased Rac1 activity. Because our previous studies expression of a constitutively active Rac1 in DU145 cells showed that metformin inhibits cancer cell proliferation and overcame the antimigratory effects of dbcAMP (Fig. 5E). These blocks cell cycle in G0–G1 within 24 hours (4), all cell migration results suggest that the antimigratory effect of metformin assays were performed within 4 hours of treatment to exclude any requiresincreasedcAMPlevels. effects due to cell-cycle arrest. Metformin inhibits the migration of glioblastoma, ovarian, Metformin inhibits CXCL12 chemotactism in prostate cancer and pancreatic cancer cells (8, 12, 13). However, the cellular and cells molecular mechanisms responsible for this inhibition are poorly Regardless of its chemoattractive properties, CXCL12 was documented. In melanoma, metformin does not affect cell migra- recently shown to regulate Rac1 (38). Therefore, CXCL12 was tion, but inhibits invasion by reducing the activity of matrix used as a chemoattractant in a cell migration assay, in which it metalloproteinases (MMP; ref. 9). Similarly, two studies reported significantly promoted DU145 migration (Fig. 6A). Importantly, that metformin inhibits the activity of MMP-9, and therefore we found that addition of metformin prevented CXCL12 promi- blocks cancer cell invasion in endothelial and fibrosarcoma cells gratory effects (Fig. 6A). CXCL12 binds to the chemokine receptor (10, 40). Bao and colleagues (8) correlated the antimigratory 4 (CXCR4) to affect cell migration. To validate the role of CXCL- effects of metformin with the decreased expression of let-7b, miR- 12/CXCR4 signaling in prostate cancer cell migration, we treated 26a, and miR-200b. In glioma cell lines, metformin suppresses cells with AMD3100, a well-characterized and specific antagonist MMP-2 expression and affects cell adhesion through the dimi- of CXCR4, which inhibits the binding and function of CXCL12 nution of fibulin-3, a secreted glycoprotein that associates to the (39). In the presence of CXCL12, AMD3100 significantly inhib- ECM (41). Here, we show that metformin induces drastic changes ited DU145 cell migration showing that the CXCL12/CXCR4 axis in cell morphology with a marked reduction of lamellipodia. plays an important role in the migration of prostate cancer cells These modifications are not associated with changes in a-actin or (Fig. 6B). Flow-cytometry analysis to monitor expression of b-actin expressions (Supplementary Fig. S6). However, we CXCR4 at the cell surface revealed a decrease upon metformin observed a slight decrease of fascin upon metformin treatment. treatment (Fig. 6C and Supplementary Fig. S12). We measured Fascin downstream of Rac contributes to cancer cell migration and Rac1–GTP levels and found that CXCL12 increased Rac1 activity the formation of metastasis (42–44). Further investigations are in a metformin-sensitive manner (Fig. 6D). In conclusion, our required to determine how metformin interferes with lamellipo- results show that metformin interferes with CXCL12 signaling dia formation, and whether fascin is implicated in its

592 Mol Cancer Ther; 14(2) February 2015 Molecular Cancer Therapeutics

Metformin Inhibits Rac1 GTPase and Cancer Cell Migration

AB20 * ** 3 15 * Figure 5. 2 Metformin increases cAMP levels and 10 inhibition of adenylate cyclase reverses the antimigratory effects of 1 metformin. A, cAMP concentration in 5 DU145 cells treated with 5 mmol/L metformin for 4 hours in the presence cAMP pmol/µg prot 0 Fold activation (RLU) 0 or absence of 100 mmol/L SQ22536 C M SQ M SQ C M (Adenylate cyclase inhibitor). B, luciferase activity of the CRE promoter element in DU145 cells transfected C * D with CRE–Luc vector and treated with 120 5 mmol/L metformin for 4 hours. C, 100 quantiﬁcation of a migration assay * 100 performed in Boyden chambers with DU145 cells treated with both 5 mmol/ 80 ** L metformin and 100 mmol/L SQ22536 60 (Adenylate cyclase inhibitor) for 4 50 ** hours. D, quantiﬁcation of a migration 40 (% of control) assay performed in Boyden chambers (% of control) Relative migration 20

with DU145 cells treated with 5 mmol/ Relative migration L metformin (M) or 100 mmol/L of the 0 0 cell-permeant cAMP analogue dbcAMP for 4 hours. E, DU145 cells C SQ M MSQ C M dbcAMP were transfected with empty vector, active forms of Rac1:RacQ61L or E RacV12 and treated with dbcAMP for 4 Control dbcAMP hours during the migration assay. The 120 graphs represent the quantiﬁcation of 100 at least three experiments performed 80 independently. The statistical analysis * was performed using the Student t 60 test. The differences are signiﬁcant; 40 , P < 0.05; , P < 0.01.

(% of control) 20 Relative migration 0 Empty Rac61L RacV12 antimigratory effect. The drastic change in prostate cancer cell cAMP plays an important and sometimes controversial role in morphology is associated with a decrease in the active form of apoptosis (52, 53), but cAMP is also a well-established inhibitor Rac1, a master regulator of actin polymerization (14). Expression of cell migration (20) and a regulator of cytoskeleton organization of a constitutively active form of Rac1 inhibited the antimigratory (54). It was demonstrated that cAMP inhibits the Rho family effects of metformin and restored the formation of lamellipodia small GTPases via PKA. For example, prostaglandin E2 inhibits in cancer cells. Conﬂicting reports have been published regarding insulin-like growth factor-I–induced cell migration and inhibited the role Rac1 in cell migration (45–49). For instance, the RacGEF Rac1 activity through a mechanism involving cAMP. In addition, Tiam1 inhibits cell migration of melanoma cells (48), in accor- cAMP has been shown to regulate Rac1 and breast cancer cell dance, we also observe a slight inhibition of cell migration when migration via PKA. We showed that metformin increases cAMP we express Rac61L and RacV12 in DU145. On the other hand, the levels in DU145, but not in PC3. The cellular cAMP level depends GEF P-Rex1 promotes cell migration and its downregulation with on the activity of two enzymes, the adenylyl cyclases that produce siRNA inhibits PC3 cell migration (17). In the present study, the cAMP and the phosphodiesterases that hydrolyze cAMP. This expression of P-Rex1 reversed the antimigratory effects of met- discrepancy between the cell lines may be related to a different formin, supporting the notion that metformin acts in a Rac1- action of metformin on adenylyl cyclase or/and phosphodiester- dependent manner. Metformin could, therefore, act as a GEF ase depending on the cell lines. Indeed, a study demonstrates that inhibitor. Indeed, P-Rex1 activity is enhanced by PIP3 and Gbg metformin decreases phopshodiesterase 3B mRNA levels in breast proteins, which is inhibited by cAMP through the phosphoryla- cancer biopsies after the treatment (55). More recently, a work tion of P-Rex1 by the protein kinase A (PKA; ref. 50). As a performed in primary hepatocytes showed that a pretreatment result, the increased cAMP levels induced by metformin could with metformin inhibited glucagon-induced accumulation of inhibit P-Rex1 through PKA and downregulate Rac1. In line with cAMP, but did not affect basal cAMP levels (56). Indeed, metfor- this hypothesis, increased intracellular cAMP levels and PKA min induced an increase in the AMP levels, possibly due to the activity following morphine treatment lead to inhibition of decreased ATP concentration, which inhibits the activity of ade- Rac1GTPase and p38 MAPK, cause attenuation of actin polymer- nylate cyclase. We and others have shown that metformin inhibits ization, and decrease bacterial phagocytosis (51). the activity of the mitochondrial complex 1, and decreases the

www.aacrjournals.org Mol Cancer Ther; 14(2) February 2015 593

Dirat et al.

A B Figure 6. 200 ** 120 Metformin inhibits the promigratory effects of CXCL12. A, quantiﬁcation of 160 100 a migration assay performed with DU145 cells treated with 250 ng/mL 120 80 ** of CXCL12 in the absence (C) or 60 ** presence of 5 mmol/L metformin for 80 40 4 hours. 0, the untreated condition. B, (% of control) quantiﬁcation of migration assay with (% of control)

Relative migration 40 20

Relative migration DU145 cells incubated with 250 ng/mL 0 CXCL12 for 4 hours in the presence or 0 C AMD3100 absence of 25 mg/mL AMD3100 (a 0 C M CXCR4 antagonist). C, relative CXCL12 CXCL12 expression of CXCR4, determined by flow-cytometry analysis, in DU145 C D cells treated with 5 mmol/L metformin CXCL12 for 4 hours. D, immunoblot analysis of Rac1GTP in DU145 cells treated with 5,000 MC0 CXCL12 in presence of 5 mmol/L metformin for 4 hours. The graphs 4,000 Rac1 GTP * represent the quantification of at least 3,000 three independent experiments Rac1 total performed independently. The 2,000 statistical analysis was performed using the Student t test. The 1,000 ERK differences are significant; , P < 0.05; , P < 0.01. CXCR4 relative expresion 0 C M

intracellular concentration of ATP, resulting in the increase of cisplatin-induced toxicity. To this regard, it would be interesting to AMP within 8 hours (30, 57). determine whether metformin can improve the efficiency of We established that CXCL12 increases Rac1 activity as previ- docetaxel, the standard treatment for patients with prostate cancer ously shown in endothelial cells (58), and that metformin inhi- who are refractory to hormonal manipulations. bits CXCL12-induced Rac1 activation. Our work suggests that Collectively, our results shed light on a new mechanism of metformin hampers the promigratory effects of CXCL12 by action of metformin and novel properties of this drug in prostate affecting Rac1 GTPase activity. Interestingly, the CXCL12/CXCR4 cancer. pathway was recently associated with Rac activation and metastasis (38). Therapeutic approaches target this pathway by either Disclosure of Potential Conflicts of Interest blocking CXCL12 with antibodies or acting on CXCR4 by pre- No potential conflicts of interest were disclosed. venting CXCL12 binding. We anticipate that metformin may represent a novel and alternative way of inhibiting this pathway known to play a major role in prostate cancer metastasis. Authors' Contributions Regarding prostate cancer therapy, we demonstrated in an Conception and design: B. Dirat, I. Ader, O. Cuvillier, F. Bost orthotopic metastatic model that metformin reduces the forma- Development of methodology: B. Dirat, I. Ader, M. Golzio, B. Malavaud, F. Bost tion of metastasis to fewer organs in addition to its inhibitory Acquisition of data (provided animals, acquired and managed patients, effect on the growth of primary tumors. Several studies have provided facilities, etc.): B. Dirat, I. Ader, M. Golzio, A. Mettouchi, F. Larbret, B. Malavaud, E. Lemichez, F. Bost shown in different mouse xenograft models and transgenic mice Analysis and interpretation of data (e.g., statistical analysis, biostatistics, that metformin inhibits tumor growth (reviewed in ref. 59), but computational analysis): B. Dirat, I. Ader, M. Golzio, K. Laurent, O. Cuvillier, few works analyzed metastasis dissemination. Our data are F. Bost encouraging for a potential use of metformin in the treatment Writing, review, and/or revision of the manuscript: B. Dirat, I. Ader, M. Golzio, of advanced metastatic prostate cancer. However, one of the F. Massa, B. Malavaud, M. Cormont, O. Cuvillier, J.F. Tanti, F. Bost limitation of our in vivo model is the injection of exogenous Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Dirat, F. Larbret, F. Bost cancer cells in the mouse prostate. Therefore, we are aware that we Study supervision: B. Dirat, O. Cuvillier, F. Bost need to confirm the effect of metformin on the formation of metastasis in another mouse model. Thus, it would be interesting to test the effects of metformin in the "RapidCaP" model recently Acknowledgments described by Cho and colleagues (60). In this new model, unlike The authors thank Anne Doye and Rachel Paul-Bellon for technical assis- – our study, mice develop metastasis from mouse prostate tumors. tance, Pr. Mitchell and Dr. Becanovic for The HA PREX plasmid, Issam Ben- Sahra, Stephane Ricoult, Jerome^ Gilleron, Sophie Giorgetti-Peraldi, and Yannick Rattan and colleagues (11) demonstrated that metformin signif- Le Marchand Brustel for their help and the critical reading of the article. The icantly reduces the growth of metastatic nodules of ovarian cancer authors greatly acknowledge Damien Alcor of the C3M (or MicorBio) Cell cells in nude mice. They also indicated that metformin potentiates Imaging Facility.

594 Mol Cancer Ther; 14(2) February 2015 Molecular Cancer Therapeutics

Metformin Inhibits Rac1 GTPase and Cancer Cell Migration

Grant Support The costs of publication of this article were defrayed in part by the This study was supported by The European Foundation for the Study of payment of page charges. This article must therefore be hereby marked advertisement Diabetes (EFSD), INCA (grants 2010-219 and 2010-214) and the "Fondation in accordance with 18 U.S.C. Section 1734 solely to indicate ARC." B. Dirat was supported by INCA grant 2010-219, the Cancerople PACA, this fact. and the Region PACA. F. Bost, J.F. Tanti, and A. Mettouchi are investigators of the Centre National de la Recherche Scientiﬁque (CNRS). F. Massa is supported by ITMO-Cancer. This work was supported by the French Government (National Received February 6, 2014; revised December 5, 2014; accepted December 9, Research Agency, ANR) through the "Investments for the Future" LABEX 2014; published OnlineFirst December 19, 2014. SIGNALIFE (grant ANR-11-0028-01).

References 1. Currie CJ, Poole CD, Jenkins-Jones S, Gale EA, Johnson JA, Morgan CL. 20. Chen L, Zhang JJ, Huang XY. cAMP inhibits cell migration by interfering Mortality after incident cancer in people with and without type 2 diabetes: with Rac-induced lamellipodium formation. J Biol Chem 2008;283: impact of metformin on survival. Diabetes Care 2012;35:299–304. 13799–805. 2. Evans JM, Donnelly LA, Emslie-Smith AM, Alessi DR, Morris AD. Metfor- 21. Dwinell MB, Ogawa H, Barrett KE, Kagnoff MF. SDF-1/CXCL12 reg- min and reduced risk of cancer in diabetic patients. BMJ 2005;330:1304–5. ulates cAMP production and ion transport in intestinal epithelial cells 3. Margel D, Urbach DR, Lipscombe LL, Bell CM, Kulkarni G, Austin PC, et al. via CXCR4. Am J Physiol Gastrointest Liver Physiol 2004;286: Metformin use and all-cause and prostate cancer-specific mortality among G844–50. men with diabetes. J Clin Oncol 2013;31:3069–75. 22. Ghosh MC, Makena PS, Gorantla V, Sinclair SE, Waters CM. CXCR4 4. Ben Sahra I, Laurent K, Loubat A, Giorgetti-Peraldi S, Colosetti P, Auberger regulates migration of lung alveolar epithelial cells through activation of P, et al. The antidiabetic drug metformin exerts an antitumoral effect in vitro Rac1 and matrix metalloproteinase-2. Am J Physiol Lung Cell Mol Physiol and in vivo through a decrease of cyclin D1 level. Oncogene 2008;27: 2012;302:L846–56. 3576–86. 23. Singh S, Singh UP, Grizzle WE, Lillard JWJr. CXCL12-CXCR4 interactions 5. Huang J, Manning BD. The TSC1–TSC2 complex: a molecular switchboard modulate prostate cancer cell migration, metalloproteinase expression, controlling cell growth. Biochem J 2008;412:179–90. and invasion. Lab Invest 2004;84:1666–76. 6. Memmott RM, Mercado JR, Maier CR, Kawabata S, Fox SD, Dennis PA. 24. Taichman RS, Cooper C, Keller ET, Pienta KJ, Taichman NS, McCauley LK. Metformin prevents tobacco carcinogen–induced lung tumorigenesis. Use of the stromal cell-derived factor-1/CXCR4 pathway in prostate cancer Cancer Prev Res 2010;3:1066–76. metastasis to bone. Cancer Res 2002;62:1832–7. 7. Zakikhani M, Dowling R, Fantus IG, Sonenberg N, Pollak M. Metformin is 25. Teicher BA, Fricker SP. CXCL12 (SDF-1)/CXCR4 pathway in cancer. Clin an AMP kinase-dependent growth inhibitor for breast cancer cells. Cancer Cancer Res 2010;16:2927–31. Res 2006;66:10269–73. 26. Pchejetski D, Golzio M, Bonhoure E, Calvet C, Doumerc N, Garcia V, et al. 8. Bao B, Wang Z, Ali S, Ahmad A, Azmi AS, Sarkar SH, et al. Metformin Sphingosine kinase-1 as a chemotherapy sensor in prostate adenocarci- inhibits cell proliferation, migration, and invasion by attenuating CSC noma cell and mouse models. Cancer Res 2005;65:11667–75. function mediated by deregulating miRNAs in pancreatic cancer cells. 27. Waters JE, Astle MV, Ooms LM, Balamatsias D, Gurung R, Mitchell CA. P- Cancer Prev Res 2012;5:355–64. Rex1—a multidomain protein that regulates neurite differentiation. J Cell 9. Cerezo M, Tichet M, Abbe P, Ohanna M, Lehraiki A, Rouaud F, et al. Sci 2008;121:2892–903. Metformin blocks melanoma invasion and metastasis development in a 28. Boyer L, Turchi L, Desnues B, Doye A, Ponzio G, Mege JL, et al. CNF1- AMPK/p53-dependent manner. Mol Cancer Ther 2013;12:1605–15. induced ubiquitylation and proteasome destruction of activated RhoA is 10. Hwang YP, Jeong HG. Metformin blocks migration and invasion of tumour impaired in Smurf1-/- cells. Mol Biol Cell 2006;17:2489–97. cells by inhibition of matrix metalloproteinase-9 activation through a 29. Brizuela L, Dayon A, Doumerc N, Ader I, Golzio M, Izard JC, et al. The calcium and protein kinase Calpha-dependent pathway: phorbol-12-myr- sphingosine kinase-1 survival pathway is a molecular target for the tumor- istate-13-acetate-induced/extracellular signal-regulated kinase/activator suppressive tea and wine polyphenols in prostate cancer. FASEB J protein-1. Br J Pharmacol 2010;160:1195–211. 2010;24:3882–94. 11. Rattan R, Graham RP, Maguire JL, Giri S, Shridhar V. Metformin suppresses 30. Ben Sahra I, Laurent K, Giuliano S, Larbret F, Ponzio G, Gounon P, et al. ovarian cancer growth and metastasis with enhancement of Cisplatin Targeting cancer cell metabolism: the combination of metformin and 2- cytotoxicity in vivo. Neoplasia 2010;13:483–91. deoxyglucose induces p53-dependent apoptosis in prostate cancer cells. 12. Tan BK, Adya R, Chen J, Lehnert H, Sant Cassia LJ, Randeva HS. Metformin Cancer Res 2010;70:2465–75. treatment exerts antiinvasive and antimetastatic effects in human endo- 31. Yamazaki D, Kurisu S, Takenawa T. Regulation of cancer cell motility metrial carcinoma cells. J Clin Endocrinol Metab 2011;96:808–16. through actin reorganization. Cancer Sci 2005;96:379–86. 13. Wu B, Li S, Sheng L, Zhu J, Gu L, Shen H, et al. Metformin inhibits the 32. Sanz-Moreno V, Gadea G, Ahn J, Paterson H, Marra P, Pinner S, et al. Rac development and metastasis of ovarian cancer. Oncol Rep 2012;28:903–8. activation and inactivation control plasticity of tumor cell movement. Cell 14. Parri M, Chiarugi P. Rac and Rho GTPases in cancer cell motility control. 2008;135:510–23. Cell Commun Signal 2010;8:23. 33. Schmitz AA, Govek EE, Bottner B, Van Aelst L. Rho GTPases: signaling, 15. Nobes CD, Hall A. Rho, rac, and cdc42 GTPases regulate the assembly of migration, and invasion. Exp Cell Res 2000;261:1–12. multimolecular focal complexes associated with actin stress fibers, lamel- 34. Manser E, Loo TH, Koh CG, Zhao ZS, Chen XQ, Tan L, et al. PAK kinases are lipodia, and filopodia. Cell 1995;81:53–62. directly coupled to the PIX family of nucleotide exchange factors. Mol Cell 16. Engers R, Ziegler S, Mueller M, Walter A, Willers R, Gabbert HE. Prognostic 1998;1:183–92. relevance of increased Rac GTPase expression in prostate carcinomas. 35. Dwivedi S, Pandey D, Khandoga AL, Brandl R, Siess W. Rac1-mediated Endocr Relat Cancer 2007;14:245–56. signaling plays a central role in secretion-dependent platelet aggregation in 17. Qin J, Xie Y, Wang B, Hoshino M, Wolff DW, Zhao J, et al. Upregulation of human blood stimulated by atherosclerotic plaque. J Transl Med PIP3-dependent Rac exchanger 1 (P-Rex1) promotes prostate cancer 2010;8:128. metastasis. Oncogene 2009;28:1853–63. 36. Balamatsias D, Kong AM, Waters JE, Sriratana A, Gurung R, Bailey CG, et al. 18. O'Connor KL, Shaw LM, Mercurio AM. Release of cAMP gating by the Identification of P-Rex1 as a novel Rac1-guanine nucleotide exchange alpha6beta4 integrin stimulates lamellae formation and the chemo- factor (GEF) that promotes actin remodeling and GLUT4 protein traffick- tactic migration of invasive carcinoma cells. J Cell Biol 1998; ing in adipocytes. J Biol Chem 2011;286:43229–40. 143:1749–60. 37. Nagasawa SY, Takuwa N, Sugimoto N, Mabuchi H, Takuwa Y. Inhibi- 19. Bond M, Wu YJ, Sala-Newby GB, Newby AC. Rho GTPase, Rac1, regulates tion of Rac activation as a mechanism for negative regulation of actin Skp2 levels, vascular smooth muscle cell proliferation, and intima forma- cytoskeletal reorganization and cell motility by cAMP. Biochem J 2005; tion in vitro and in vivo. Cardiovasc Res 2008;80:290–8. 385:737–44.

www.aacrjournals.org Mol Cancer Ther; 14(2) February 2015 595

Dirat et al.

38. Li H, Yang L, Fu H, Yan J, Wang Y, Guo H, et al. Association between 50. Mayeenuddin LH, Garrison JC. Phosphorylation of P-Rex1 by the cyclic Galphai2 and ELMO1/Dock180 connects chemokine signalling with Rac AMP-dependent protein kinase inhibits the phosphatidylinositiol (3,4,5)- activation and metastasis. Nat Commun 2013;4:1706. trisphosphate and Gbetagamma-mediated regulation of its activity. J Biol 39. Rosenkilde MM, Gerlach LO, Jakobsen JS, Skerlj RT, Bridger GJ, Schwartz Chem 2006;281:1921–8. TW. Molecular mechanism of AMD3100 antagonism in the CXCR4 recep- 51. Ninkovic J, Roy S. Morphine decreases bacterial phagocytosis by inhibiting tor: transfer of binding site to the CXCR3 receptor. J Biol Chem actin polymerization through cAMP-, Rac-1-, and p38 MAPK-dependent 2004;279:3033–41. mechanisms. Am J Pathol 2012;180:1068–79. 40. Esfahanian N, Shakiba Y, Nikbin B, Soraya H, Maleki-Dizaji N, Ghazi- 52. Kizaki H, Nakada S, Ohnishi Y, Azuma Y, Mizuno Y, Tadakuma T. Tumour Khansari M, et al. Effect of metformin on the proliferation, migration, and necrosis factor-alpha enhances cAMP-induced programmed cell death in MMP-2 and -9 expression of human umbilical vein endothelial cells. Mol mouse thymocytes. Cytokine 1993;5:342–7. Med Report 2012;5:1068–74. 53. Naderi EH, Ugland HK, Diep PP, Josefsen D, Ruud E, Naderi S, et al. 41. Gao LB, Tian S, Gao HH, Xu YY. Metformin inhibits glioma cell U251 Selective inhibition of cell death in malignant vs. normal B-cell precursors: invasion by downregulation of ﬁbulin-3. Neuroreport 2013;24:504–8. implications for cAMP in development and treatment of BCP-ALL. Blood 42. Hashimoto Y, Parsons M, Adams JC. Dual actin-bundling and protein 2013;121:1805–13. kinase C-binding activities of fascin regulate carcinoma cell migration 54. Han JD, Rubin CS. Regulation of cytoskeleton organization and paxillin downstream of Rac and contribute to metastasis. Mol Biol Cell dephosphorylation by cAMP. Studies on murine Y1 adrenal cells. J Biol 2007;18:4591–602. Chem 1996;271:29211–5. 43. Li A, Dawson JC, Forero-Vargas M, Spence HJ, Yu X, Konig I, et al. The actin- 55. Hadad S, Iwamoto T, Jordan L, Purdie C, Bray S, Baker L, et al. Evidence for bundling protein fascin stabilizes actin in invadopodia and potentiates biological effects of metformin in operable breast cancer: a pre-operative, protrusive invasion. Curr Biol 2010;20:339–45. window-of-opportunity, randomized trial. Breast Cancer Res Treat 44. Schoumacher M, Goldman RD, Louvard D, Vignjevic DM. Actin, micro- 2011;128:783–94. tubules, and vimentin intermediate ﬁlaments cooperate for elongation of 56. Miller RA, Chu Q, Xie J, Foretz M, Viollet B, Birnbaum MJ. Biguanides invadopodia. J Cell Biol 2010;189:541–56. suppress hepatic glucagon signalling by decreasing production of cyclic 45. Engers R, Springer E, Michiels F, Collard JG, Gabbert HE. Rac affects AMP. Nature 2013;494:256–60. invasion of human renal cell carcinomas by upregulating tissue inhibitor 57.El-MirMY,NogueiraV,FontaineE,AveretN,RigouletM,LeverveX. of metalloproteinases (TIMP)-1 and TIMP-2 expression. J Biol Chem Dimethylbiguanide inhibits cell respiration via an indirect effect 2001;276:41889–97. targeted on the respiratory chain complex I. J Biol Chem 2000;275: 46. Feng H, Hu B, Liu KW, Li Y, Lu X, Cheng T, et al. Activation of Rac1 by Src- 223–8. dependent phosphorylation of Dock180(Y1811) mediates PDGFRalpha- 58. Carretero-Ortega J, Walsh CT, Hernandez-Garcia R, Reyes-Cruz G, Brown stimulated glioma tumorigenesis in mice and humans. J Clin Invest JH, Vazquez-Prado J. Phosphatidylinositol 3,4,5-triphosphate-dependent 2011;121:4670–84. Rac exchanger 1 (P-Rex-1), a guanine nucleotide exchange factor for Rac, 47. Hwang SY, Jung JW, Jeong JS, Kim YJ, Oh ES, Kim TH, et al. Dominant- mediates angiogenic responses to stromal cell-derived factor-1/chemokine negative Rac increases both inherent and ionizing radiation-induced cell stromal cell derived factor-1 (SDF-1/CXCL-12) linked to Rac activation, migration in C6 rat glioma cells. Int J Cancer 2006;118:2056–63. endothelial cell migration, and in vitro angiogenesis. Mol Pharmacol 48. Uhlenbrock K, Eberth A, Herbrand U, Daryab N, Stege P, Meier F, et al. The 2010;77:435–42. RacGEF Tiam1 inhibits migration and invasion of metastatic melanoma 59. Bost F, Ben Sahra I, Le Marchand-Brustel Y, Tanti JF. Metformin and cancer via a novel adhesive mechanism. J Cell Sci 2004;117:4863–71. therapy. Curr Opin Oncol 2012;24:103–8. 49. Zhang JY, Zhang D, Wang EH. Overexpression of small GTPases directly 60. Cho H, Herzka T, Zheng W, Qi J, Wilkinson JE, Bradner JE, et al. RapidCaP, a correlates with expression of delta-catenin and their coexpression predicts a novel GEM model for metastatic prostate cancer analysis and therapy, poor clinical outcome in nonsmall cell lung cancer. Mol Carcinog reveals myc as a driver of Pten-mutant metastasis. Cancer Discov 2013;52:338–47. 2014;4:318–33.

596 Mol Cancer Ther; 14(2) February 2015 Molecular Cancer Therapeutics

Inhibition of the GTPase Rac1 Mediates the Antimigratory Effects of Metformin in Prostate Cancer Cells

Béatrice Dirat, Isabelle Ader, Muriel Golzio, et al.

Mol Cancer Ther 2015;14:586-596. Published OnlineFirst December 19, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-0102

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2014/12/20/1535-7163.MCT-14-0102.DC1

Cited articles This article cites 60 articles, 29 of which you can access for free at: http://mct.aacrjournals.org/content/14/2/586.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/14/2/586.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/14/2/586. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.