DLC1 Induces Expression of E-Cadherin in Prostate Cancer Cells Through Rho Pathway and Suppresses Invasion

ORIGINAL ARTICLE DLC1 induces expression of E-cadherin in prostate cancer cells through Rho pathway and suppresses invasion

V Tripathi, NC Popescu and DB Zimonjic

E-cadherin is a cell–cell adhesion molecule that acts as a suppressor of cancer cell invasion and its expression is downregulated in many advanced, poorly differentiated, human cancers. In this study, we found that the expression of DLC1 (deleted in liver cancer 1) tumor-suppressor gene in metastatic prostate carcinoma (PCA) cells increased the expression of E-cadherin and resulted in an elevated rate of cell–cell aggregation as measured by aggregation assay. DLC1-mediated increase in E-cadherin expression was not dependent on a-catenin, a DLC1-binding protein associated with E-cadherin, and/or cellular density. The increase of E-cadherin expression occurred at mRNA level and relied on DLC1 RhoGAP function, leading to suppression of high level of RhoA-GTP and RhoC-GTP activity in metastatic PCA cells. Application of Rho/ROCK inhibitors produced the same effect as introduction of DLC1. Knocking down of RhoA produced a moderate increase in E-cadherin whereas knocking down of RhoC resulted in a signiﬁcant increase of E-cadherin. Downregulation of E-cadherin caused by constitutively active RhoAV14 and RhoCV14 could not be reversed by expression of DLC1 in DLC1-negative cell line. DLC1-mediated suppression of metastatic PCA cells invasion was comparable with the one associated with ectopic E-cadherin expression, or caused by suppression of Rho pathway either by Rho/ROCK inhibitors, or by shRNA repression. This study demonstrates that DLC1 expression positively regulates E-cadherin and suppresses highly metastatic PCA cell invasion by modulating Rho pathway, which appears as a feasible therapeutic target in cancers with high activity of RhoGTPases.

Oncogene (2014) 33, 724–733; doi:10.1038/onc.2013.7; published online 4 February 2013 Keywords: E-cadherin; DLC1; Rho; invasion; prostate cancer

INTRODUCTION functionally interact with Rac1, RhoA and Cdc42, all of which have 15,16 Deleted in liver cancer 1 (DLC1) is a potent tumor-suppressor gene role in coupling cytoskeletal structure to cell–cell contacts. in several cancers and is increasingly recognized by experts in the Several studies have reported a strong correlation between field as a metastasis-suppressor gene.1–5 DLC1 is a multidomain reduced or lost expression of E-cadherin and tumor progression; protein that contains SAM (sterile alpha motif), GAP (GTPase- E-cadherin downregulation is generally a result of trans- 17–19 20,21 activating protein) and START (steroidogenic acute regulatory criptional repression, post-translational processing or its 22 protein-related lipid transfer) domains. It suppresses proliferation, endocytosis. A direct role of E-cadherin expression in migration and invasion and induces apoptosis of human liver, suppression of invasive phenotype was explored in transgenic 23 breast and prostate cancer cells,1,6,7 by acting mainly as a GAP for mouse model of pancreatic b-cell carcinogenesis, and was RhoA, RhoC and, to lesser extent, for Cdc42.8 shown that loss of E-cadherin facilitated a transition from well- DLC1 is critically involved in the control of Rho signaling and, differentiated adenoma to carcinoma and, in the case of forced therefore, affects actin cytoskeletal remodeling and cell–cell expression of a truncated dominant negative gene, increased contacts. DLC1 protein co localizes and binds with a-catenin9 incidence of carcinomas and of metastasis. Conversely, and interacts with tensin, talin and focal adhesion kinase,10–13 thus reintroduction of E-cadherin expression resulted in significant influencing molecular machineries of adherens junctions (AJs) and reduction in frequency of carcinomas vs well-differentiated focal adhesions, both of which are highly relevant in cancer cell adenomas, thus indicating that the loss of E-cadherin-mediated migration and invasion. High level of RhoGTPases has been cell adhesion is required for invasion in vivo. Recently, it was implicated in various steps during cellular transformation and reported that E-cadherin-mediated suppression of anchorage- tumor cell invasion.14 During tumor progression, loss of epithelial independent growth and cell migration is associated with reduced characteristics involves morphological changes that affect activity of RhoA or cdc42, apparently consequential to DLC1 and cytoskeletal network, cell–cell adhesion and cell–matrix adhesion. p190RhoGAP action.24 Inappropriate activation of Rho proteins, reported in highly In this study, we show that DLC1expression increases the invasive human cancer cells, could be responsible for destabiliza- expression of E-cadherin and, consequently, abrogate the inva- tion of cell–cell contacts and facilitation of invasion. siveness of a highly malignant prostate carcinoma (PCA) cell line. E-cadherin, on the other side, is a transmembrane protein that is The DLC1-mediated process apparently results from its RhoGAP a key component of cell–cell adhesion machinery. It forms a action; attenuation of Rho proteins activity by specific Rho/Rock complex with other cytoplasmic proteins like b-catenin, a-catenin inhibitors, or intervention with shRNA, restores E-cadherin and p120 catenin and stabilizes the cell–cell junctions. Cadherins expression. Expression and localization of E-cadherin, disrupted

Laboratory of Experimental Carcinogenesis, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. Correspondence: Dr NC Popescu, Laboratory of Experimental Carcinogenesis, National Cancer Institute, National Institutes of Health, Building 37, Room 4128B, 37 Convent Drive, MSC 4262, Bethesda, MD 20892-4262, USA. E-mail: [email protected] Received 21 August 2012; revised 29 November 2012; accepted 17 December 2012; published online 4 February 2013 DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 725 by constitutively active, GTP-bound, RhoAV14 and RhoCV14, shown was measured at 0 and 60 min in a calcium-containing buffer 2 þ to be resistant to GAP action, is not reversed by DLC1 expression. (2 mM Ca ). At the 60-min endpoint, DLC1-transduced cells with Reduction in invasiveness of highly malignant PCA cells mimics increased levels of E-cadherin expression showed increased the one observed in cases of E-cadherin overexpression or aggregation compared with their control counterparts (Figures downregulation of Rho activity. 1c and d). Treatment with either anti-E-cadherin antibody (HEDC-1) or with calcium chelating agent EGTA (ethylene glycol tetraacetic acid), both used to disrupt E-cadherin activity, abolished DLC1/E-cadherin-induced aggregation (Figure 1d). RESULTS DLC1 increases expression of E-cadherin in cancer cells and elevates the rate of cell–cell aggregation DLC1-mediated increase in E-cadherin expression is independent Androgen independent and metastatic prostate cancer cells C4-2- of a-catenin and culture density, but E-cadherin localization is B2 and PC-3 express lower levels of E-cadherin as compared with dependent on a-catenin RWPE-1, a normal immortalized prostate epithelial cell (Figure 1a). Alpha-catenin acts as a molecular link between DLC1 and Transduction of DLC1 in these cells and its expression to the E-cadherin. To check whether DLC-mediated increase in E-cad- physiological levels was associated with noticeable increase in the herin expression was dependent on a-catenin, shRNA was used to expression of E-cadherin as shown by western blot and knock down a-catenin in C4-2-B2 cells, and stable clone exhibiting immunoﬂuorescence staining (Figures 1a and b). To examine the maximum effect was selected for the study.9 Transduction of effects of DLC1-mediated increased expression of E-cadherin on these cells with DLC1 resulted in increased expression of cell aggregation ability in vitro, Ad-LacZ- and DLC1-transduced E-cadherin compared with controls and proved that such C4-2-B2 and PC-3 cells were cultured, and their aggregation index increase was independent of a-catenin (Figure 2a). Yet, as shown

Figure 1. DLC1 induces expression of E-cadherin. (a) E-cadherin expression in RWPE-1, C4-2-B2 and PC-3 cells (left) and of C4-2-B2 and PC-3 cells transduced with either Ad-Lac Z or DLC1 (right). (b) Immunoﬂuorescence analysis of E-cadherin in Ad-Lac Z or DLC1 transduced C4-2-B2 and PC-3 cells. Bar ¼ 20 mm. (c) Aggregation of Ad-LacZ- or Ad-DLC1-transduced C4-2-B2 and PC-3 cells at time 0 and 60 mins. Bar ¼ 50 mm. (d) Effect of DLC1 expression on cell aggregation at various time points. C4-2-B2 and PC-3 cells were grown in calcium-negative (in the presence of EGTA) and calcium-positive medium. Each batch of cells was transduced with DLC1 or LacZ in the presence or absence of HECD-1.

& 2014 Macmillan Publishers Limited Oncogene (2014) 724 – 733 DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 726

Figure 2. DLC1-mediated E-cadherin induction is independent of a-catenin and cell density. (a) E-cadherin expression in C4-2-B2 and PC-3 cells in the presence or absence of DLC1 and a-catenin. (b, c) Immunoﬂuorescence analysis of E-cadherin in the above-mentioned cell lines. Bars ¼ 10 and 20 mm. (d) Expression of E-cadherin in Ad-LacZ- or Ad- DLC1-transduced C4-2-B2 cells in sparse, medium and conﬂuent culture (right). Densitometric analysis by image quat of western blotted bands compared with GAPDH (glyceraldehyde 3-phosphate dehydrogenase; left). (e) E-cadherin expression in TritonX-100 soluble and insoluble fractions from Ad-LacZ- and Ad-DLC1-transduced C4-2-B2 cells. Bar ¼ 20 mm. DAPI, 40,6-diamidino-2-phenylindole.

by Immunofluorescence analysis, localization of E-cadherin at AJs of cell membranes into TritonX-100-soluble and insoluble fractions required both DLC1 and a-catenin (Figure 2b). This observation and their subsequent probing for E-cadherin confirmed irrele- was confirmed in PC-3 cells negative for both DLC1 and a-catenin vance of cell density and showed that DLC1-mediated increase in (Figures 2a and c). To examine whether culture density E-cadherin expression was confined to insoluble skeletal fraction affects DLC1-mediated increase in E-cadherin expression, of membranes (Figure 2e). DLC1-positive and -negative C4-2-B2 cells were grown for 2 days in sparse (5 Â 103 cells/cm2), medium (2 Â 104 cells/cm2) or dense (5 Â 104 cells/ cm2) culture and checked for expression of DLC1 increases E-cadherin expression in a GAP-dependent E-cadherin (Figure 2d). Densitometric analysis showed that manner at transcriptional level transduction with DLC1 caused approximately twofold increase To determine whether DLC1 RhoGAP activity is required for the in E-cadherin expression compared with transduction with LacZ restoration of E-cadherin expression, C4-2-B2 and PC-3 cells irrespective of culture density (Figure 2d). Differential partitioning were transduced with V5-tagged full-length DLC1 and stably

Oncogene (2014) 724 – 733 & 2014 Macmillan Publishers Limited DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 727 expression (Figures 4a and b). The comparison between two PCA cell lines (PC-3 and C4-2-B2) and one normal prostate immortalized cell line (RWPE-1) revealed that in cancer cells expression of both RhoA and RhoC was significantly higher than in normal immortal one (Figure 4c). Interestingly, the expression of E-cadherin was inversely correlated to expression of Rho proteins; it was high in RWPE-1 and significantly lower in C4-2-B2 and PC-3 as confirmed by both blotting (Figure 4c) and immunofluorescent detection (Figure 4d). To understand the relation between DLC1, E-cadherin and Rho proteins, E-cadherin was stably knocked down in C4-2-B2 cells (Figure 4e), and stable clone (D3) was then transduced with either Ad-Lac Z or DLC1 and examined for RhoA and RhoC activity (Figures 4f and g). In the absence of E-cadherin, the efficiency of DLC1-mediated suppression of RhoA and RhoC was markedly reduced, thus indicating that E-cadherin is required for effective inhibition of those two proteins by DLC1.

Rho/ROCK inhibitors affect E-cadherin expression similar to DLC1 To gain insights into the molecular mechanisms underlying DLC1-mediated increase in E-cadherin expression, C4-2-B2 cells were treated with Rho inhibitor C-3 or ROCK inhibitor Y-27632 (Figure 4h). As expected, Y-27632 showed no effect in suppressing Rho, whereas C3 reduced RhoA and RhoC activity as much as DLC1 did. A similar, but reversed, pattern was observed with E-cadherin expression (Figure 4i), which appears to have been equally increased with either introduction of DLC1 or by application of Rho or ROCK inhibitors. Immunoﬂuorescent labeling (Figure 4j) conﬁrmed the boost in E-cadherin expression. Effects of Rho or ROCK inhibitors’ phenocopied the effect caused by introduction of DLC1.

DLC1 restores the Rho protein-mediated disruption of AJs To differentiate the roles of RhoA and RhoC in DLC1-mediated increase of E-cadherin expression, different shRNA plasmids were tested for their ability to effectively suppress the expression of either of them (Figure 5a). Stable clones with maximum knock down of RhoA (D2) and RhoC (D3) were selected, and E-cadherin expression was analyzed in the presence or absence of DLC1. The Figure 3. DLC1-mediated E-cadherin induction is GAP dependent reduction in each respective Rho’s expression positively correlated and at mRNA level. (a) E-cadherin and DLC1 (V-5) expression in C4-2-B2 and PC-3 cells transduced with Ad-LacZ, Ad-DLC1 and with suppression of its activity (Figure 5b). Analysis showed that stably transfected with DLC1 GAP-deleted or DLC1 GAP-mutated inhibition of RhoC had more profound effect on E-cadherin constructs. (b) Immunofluorescence analysis of E-cadherin in the expression than the inhibition of RhoA, as use of shRhoC resulted above-mentioned cells. Bar ¼ 20 mm. (c). Real-time PCR analysis of in E-cadherin’s increase comparable with that caused by DLC1 E-cadherin of C4-2-B2 and PC-3 cells transduced with Ad-LacZ, alone (Figures 5c and d). Additional transduction of shRhoA- or Ad-DLC1 and stably transfected with DLC1R718E. Bar ¼ 20 mm. shRhoC-transfected cells with Ad-DLC1 leveled the field by GAPDH, glyceraldehyde 3-phosphate dehydrogenase; DAPI, showing somewhat stronger effect in case of shRhoA-transfected 4,6-diamidino-2-phenylindole. cells, thus raising the possibility that active RhoC may have stronger role in suppressing E-cadherin expression and that DLC1 restores E-cadherin by negatively modulating RhoC activity more transfected with DLC1 GAP mutant (R718E) or DLC1 GAP-deleted than that of RhoA. construct (1–630), and E-cadherin expression was examined by Transfection of C4-2-B2 cells with myc-tagged constitutively western blot and immunofluorescence (Figures 3a and b). Because active RhoAV14 and RhoCV14, refractory to DLC1 action, showed E-cadherin expression was positively affected only by the full- significant reduction in E-cadherin presence on membranes of length DLC1 but not by either GAP mutant or the GAP-deleted cells containing the respective Rho transgene(s) and, particularly, construct, the experiment confirmed that GAP domain is required disruption of E-cadherin at AJs (Figure 6a). Although DLC1 was for DLC1-mediated E-cadherin expression. Real-time PCR analysis able to increase the expression of E-cadherin in all Rho transgene- of C4-2-B2 and PC-3 cells transduced and/or transfected with LacZ, negative, it was not able to correct the expression and localization DLC1 and DLC1R718E showed that DLC1-mediated increase in of E-cadherin in RhoAV14- or RhoCV14-positive cells. E-cadherin occurred at the mRNA level (Figure 3c). Only full-length DLC1 was able induce E-cadherin expression, whereas DLC1GAP- Invasion potential of C4-2-B2 cell is positively correlated with dead R718E was ineffective (Figure 3c). expression of Rho proteins but negatively correlated with DLC1 and E-cadherin expression DLC1 suppresses RhoA and RhoC activation and requires To evaluate the role of DLC1, E-cadherin and Rho proteins on E-cadherin for optimum suppression invasion capacity, C4-2-B2 cells were transduced/transfected with Introduction of DLC1 in DLC1-negative cells not only increases the different combinations of transgenes and suppressive RNAs and/ expression of E-cadherin but also suppresses both RhoA and RhoC or treated with chemical inhibitors, grown in invasion chambers

& 2014 Macmillan Publishers Limited Oncogene (2014) 724 – 733 DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 728

Figure 4. DLC1 reduces the level of active RhoA and RhoC. (a, b) Levels of RhoA-GTP and RhoC-GTP level in C4-2-B2 cells transduced with Ad-LacZ or Ad-DLC1, assessed by Rhotekin assay. (c) Expression of RhoA, RhoC, E-Cadherin and DLC1 in RWPE1, C4-2B2 and PC-3 cells. (d) Immunoﬂuorescence analysis of E-cadherin expression in the above-mentioned cells. Bar ¼ 20 mm. (e) Western blot of E-Cadherin stable knock downs in C4-2-B2 cells. (f, g) Rhotekin assay of active RhoA and RhoC in lysate from C4-2-B2 cell with stable E-cadherin knock down in the presence of DLC1. (h) Activity of RhoA and RhoC in Ad-LacZ- or Ad-DLC1-transduced and Rho- or ROCK inhibitor-treated C4-2-B2 cells. (i) E-cadherin and DLC1 expression in C4-2-B2 cells transduced with Ad-LacZ or Ad-DLC1 and treated with Rho inhibitor C3 or ROCK inhibitor Y-27632. (j) Immunoﬂuorescence analysis of E-cadherin in above-mentioned cells. Bar ¼ 20 mm. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; DAPI, 4,6-diamidino-2-phenylindole.

and compared for their invasion competence. Introduction of through RhoC and that such action is responsible for suppression DLC1 or activation of E-cadherin expression equally repressed of invasion of C4-2-B2 cells. invasion; conversely, suppression of E-cadherin or absence of DLC1 both resulted in significantly increased invasion capacity (Figures 6b and d). Suppression of RhoA or RhoC by their DISCUSSION corresponding shRNAs reduced invasion as well, but the The crucial step in progression of tumor cell toward a more effects was significantly more pronounced when RhoC was aggressive phenotype depends on deregulation of cytoskeleton suppressed than RhoA, thus confirming an earlier observation structure and adhesion machinery, which allows separation of on their respective roles in E-cadherin induction. Use of Rho individual cells from the bulk of tumor and, consequently, their inhibitor C3 and of ROCK inhibitor Y27632 also significantly invasion of surrounding tissues and/or distant sites. Such reduced invasion capacity of C4-2-B2 cells, and the extent of separation is frequently associated with local depletion of suppression mimicked the one caused by RhoC shRNA (Figure 6c). E-cadherin, whose extracellular domains at the surface of adjacent Expression of DLC1 or suppression of RhoC produced comparable cells interact with one another, while their cytoplasmatic domains attenuation of invasion, strongly indicating that DLC1 RhoGAP connect to the respective cell’s actin cytoskeletons via catenins. activity is inducing E-cadherin expression by acting mostly In this study, we report that expression of DLC1, a tumor-

Oncogene (2014) 724 – 733 & 2014 Macmillan Publishers Limited DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 729

Figure 5. DLC1 restores E-cadherin activity in similar fashion as RhoA and RhoC knock down. (a) Knock down of RhoA and RhoC. Clones labeled as D1, D2 and D3. (b) Activity of RhoA and RhoC in knock down cells. (c) E-cadherin expression in C4-2-B2 cells transduced with Ad-Lac Z or DLC1 and stably transfected with ShRNA for RhoA and RhoC. (d) Immunofluorescence analysis of E-cadherin in the above-mentioned varieties of C4-2-B2 cells. Bar ¼ 20 mm. GAPDH, glyceraldehyde 3-phosphate dehydrogenase; DAPI, 40,6-diamidino-2-phenylindole. suppressor gene, induces E-cadherin expression in androgen- E-cadherin, which has paracrine effect and inhibits E-cadherin independent and metastatic PCA cells and that such an effect is function.29 The loss or downregulation of E-cadherin expression independent of a-catenin and/or of cell culture density. Down- in epithelial cells leads to abrogation of cell–cell contacts and regulation of E-cadherin in these cells coincided with high level of increased motility, whereas forced expression of E-cadherin expression of active Rho proteins. As either expression of DLC1 or protein in metastatic tumor cells is sufficient for reversal attenuation of RhoA and RhoC by other means—chemical of metastatic phenotype.23,30 Even cytoplasmic expression of inhibitors or shRNAs—both resulted in restored E-cadherin E-cadherin was shown as sufficient to suppress invasion of expression, mechanism of DLC1 action was clearly attributable MDA-MB-231 breast cancer cells and TSU-Pr1 prostate cancer to this protein’s RhoGAP domain and function. DLC1-mediated cells.31 Immunohistochemical studies demonstrated that increase in E-cadherin reduced invasion capacity of metastatic aggressive PCA exhibits decreased or otherwise abnormal PCA cells and suppressed their invasion in vitro. expression of E-cadherin.21,32,33 Metastatic PCA cell lines C4-2-B2 and PC-3 chosen for this study In a majority of human cancers, E-cadherin remains genetically have been extensively examined and well characterized in our intact but is frequently silenced by transcriptional and transla- laboratory.7,25 They both are deficient in DLC1 expression due to tional mechanisms during the progression to invasive phenotype. either promoter hypermethylation (C4-2-B2) or genomic deletion Repression of E-cadherin transcripts via E-box-binding proteins (PC-3). DLC1 transduction in these cells suppresses tumorigenicity, (for example, Snail and Slug) has been described in detail and is invasion and proliferation. In addition, both the cell lines express also associated with tumor metastasis. In that context, culture lower level of E-cadherin, as compared with the normal prostate density appears to affect the E-cadherin expression due to epithelial cell, which makes them a suitable model for exploring transcriptional repression of E-cadherin in sparse cultures by the interactions of those two proteins. activated extracellular signal–regulated kinase- or b-catenin- The fact that DLC1-mediated E-cadherin expression affected signaling through transcription factor Slug.34 Our results invasion is not surprising because the role of E-cadherin as a obtained on sparse, intermediate and confluent cell cultures rule repressor of metastasis is well established.26,27 Migration and out the possibility that transcriptional repression by Slug—on the invasion of tumor cells are also promoted by the loss of interaction E-box element in E-cadherin promoter—is the mechanism between cytoskeleton and AJs, whose major component is underlying the E-cadherin variations in this study. E-cadherin, subsequent to changes in the activities of Rho family They, however, clearly point to GAP function of DLC1 molecule small GTPases and the concomitant reorganization of the actin in triggering signaling cascade leading to E-cadherin re-expression cytoskeleton.14,28 On the other side, matrix metalloproteinase in C4-2-B2 and PC-3 cells. Real-time PCR data revealed that (MMP) molecules MMP-7 and MMP-13 were shown to cleave DLC1-mediated expression occurred at mRNA level. Knocking cell surface E-cadherin, thus forming a soluble ectodomain of down of RhoA and RhoC lowered the levels of active RhoA and

& 2014 Macmillan Publishers Limited Oncogene (2014) 724 – 733 DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 730

Figure 6. E-cadherin, DLC1 and Rho interactions affect invasion. (a) Immunoﬂuorescence analysis of E-cadherin in C4-2-B2 cells transfected with RhoAV14 and RhoCV14 and transduced with Ad-LacZ or Ad-DLC1. Bar ¼ 20 mm. (b) Introduction of DLC1 or activation of E-cadherin expression represses invasion. (c) Suppression of Rho activity by knocking down or by Rho and ROCK inhibitors suppresses invasion equal to transduction with Ad-DLC1. (d) Suppression of E-cadherin or the absence of DLC1 result in increased invasion capacity. DAPI, 40,6-diamidino-2- phenylindole.

RhoC and restored the E-cadherin expression; however, the effect forms of Rho proteins in tumors. Small GTPases from Rho family was more pronounced when RhoC was suppressed and appeared affect complex cellular processes with repercussions on to be similar, if not the same, as the effect of DLC1 introduction. cytoskeletal dynamics, cell-cycle progression and cancer cell The Rho family of small GTPases, particularly subclasses Rho, Rac migration38 and augment prostate cancer cell invasion.39 Among and cdc42, have been increasingly recognized as key players in them, RhoA and RhoC are of particular interest because RhoA regulation of AJs, due to their effect on the actin cytoskele- expression correlate positively with the progression of breast ton.15,35–37 Rho proteins can be activated by guanine nucleotide cancer and testicular germ-cell tumors,40 disrupts AJs through exchange factor or inactivated by GTPase-activating protein (GAP). cyclooxygenase-2 signaling and increases cell motility in colorectal Unlike Ras, there are no reports of mutated, constitutively active cancer.41 On the other side, RhoC is overexpressed in pancreatic

Oncogene (2014) 724 – 733 & 2014 Macmillan Publishers Limited DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 731 ductal adenocarcinoma and inflammatory breast cancer42–44 and quantitative reverse transcriptase–PCR (Invitrogen). Quantitative PCR for its inactivation leads to morphological changes in PCA cell lines.45 E-cadherin (Hs00170423_a1CDH1) was performed using Taqman Gene RhoA and RhoC role in DLC1-mediated E-cadherin induction was Expression Assays (Applied Biosystem, Carlsbad, CA, USA). further clarified by using constitutively active RhoA and RhoC, whose action was not reversible by DLC1. Western blotting Our invasion data showed that DLC1 suppressed invasion Cells were lysed with NP-40 lysis buffer (BioSource, Camarillo, CA, USA) potential of C4-2-B2 cells to the same extent as ectopic expression supplemented with Protease Inhibitor Cocktail (Sigma, St Louis, MO, USA). of E-cadherin, or inhibition of Rho/ROCK pathway by pharmaco- Centrifugation was done at 12 000 g for 20 min at 4 1C. Supernatant was logical inhibitors or by shRNA knockdown. A ROCK pathway is one collected and protein level determined by BCA protein assay (Thermo of the two pathways downstream of Rho with effect on AJs; Dia1 Scientific, Rockford, IL, USA). Lysate was loaded on 4–12% SDS–PAGE stabilizes them, whereas ROCK disrupts cell–cell junctions through (sodium dodecyl sulfate-polyacrylamide gel electrophoresis) gels (Invitro- the generation of actin contractile force.46 DLC1-mediated gen), transferred to nitrocellulose membranes (Invitrogen). Antibodies suppression of ROCK has been shown to negatively affect cell used for western blot were: monoclonal anti-a-catenin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), monoclonal anti-DLC1 antibody migration47 and Rho/ROCK pathway has been strongly implicated 48,49 (BD Biosciences, San Jose, CA, USA), anti-E-cadherin, anti-RhoA and anti- in ovarian and bladder cancer cell migration and invasion. Pre- RhoC antibody (Abcam, Cambridge, MA, USA). treatment of cells with a pharmacological inhibitor of Rho/ROCK (Y-27632) or overexpression of dominant negative mutant of Rho N19 effectively inhibited invasion potential of ovarian cancer Preparation of Triton X-soluble and insoluble membrane fractions cells.50 As ROCK is more efficient substrate of RhoC, RhoC/ROCK Cells were harvested, lysed in NP-40 lysis buffer (BioSource, Camarillo, CA, could be the major pathway affecting E-cadherin in cell lines USA) and supplemented with protease inhibitor cocktail (Sigma-Aldrich, overexpressing RhoC. Our finding that ROCK inhibitor is effective St Louis, MO, USA). Lysate was ultracentrifuged at 100 000 g, for 30 min at 4 1C to pellet membrane fraction, which was then dissolved in phosphate- in restoring E-cadherin is, therefore, plausible in the context of the buffered saline containing 0.5% Triton X-100 and centrifuged for 20 min at above-mentioned results. 12 000 g at 4 1C. Supernatant was collected as Triton X-100-soluble fraction, Seemingly, restoration of E- cadherin function may have whereas the pellet was dissolved in phosphate-buffered saline with 1% 51,52 therapeutic implications in certain cases. For example, SDS and collected as Triton X-100-insoluble fraction. gamma-linolenic acidhas been used to modulate E-cadherin expression in several tumor cell lines.53 In addition, the Rhotekin assay treatment of breast tumor cells in 3-D culture with a -integrin antibody triggered increased E-cadherin expression54 and caused C4-2-B2 cells were grown to confluence and Rhotekin assay was used according to manufacturer’s (Cell Biolabs, San Diego, CA, USA) instruction. a reversion of the malignant phenotype. A combined treatment Briefly, Cells were lysed on ice for 10 mins in 1 Â lysis buffer supplemented with PPARg (peroxisome proliferator-activated receptor-g) agonist with protease inhibitors cocktail (Sigma-Aldrich) and 1 mM phenylmetha- and histone deacetylase inhibitor highly upregulated E-cadherin nesulfonyl fluoride (Sigma-Aldrich), and the lysate was centrifuged at and impaired the bone invasive potential of prostate cancer cells 13 000 g and 4 1C for 5 min. Small aliquot of lysate was analyzed for total in mice.55 Our data categorize DLC1 as an additional upregulatory protein content and total RhoA or RhoC content. The remainders of the cell factor of E-cadherin. Our study provides no evidences for lysates were used for the Rho activation assay (Cell Biolabs). Rhotekin transcriptional or translational modification of E-cadherin and beads were incubated with cell lysates for 4 h and washed thoroughly five demonstrates the DLC1’s GAP ability to modulate activity of small times in washing buffer before active Rho proteins attached to beads were Rho GTPases, upregulate E-cadherin, affect cell–cell adhesion and, separated by SDS–PAGE in 4–12% acrylamide gels. RhoA and RhoC were detected with anti-Rho primary antibodies (Abcam). consequently, migration and invasion. As previously dis- cussed,56,57 this study also strengthens the view of ROCK inhibitors as a promising anticancer agent. Immunofluorescence analysis Cells were grown on chamber slides, fixed with 4.0% paraformaldehyde for 20 min and permeabilization was performed in 0.5% Triton-X for 5 min. MATERIALS AND METHODS After blocking in 3% bovine serum albumin, the slides were incubated overnight with primary antibodies at 4 1C. The following primary Cell lines and culture conditions antibodies were used for detection: goat polyclonal anti-DLC1, mouse PC-3 and RWPE-1 cells lines were purchased from American Type Culture monoclonal anti-E cadherin, and anti-a-catenin, (Santa Cruz Biotechnol- Collection (Rockville, MD, USA). PC-3 was cultured in RPMI 1640 medium ogy). Secondary antibodies used were anti-mouse Alexa Fluor 488, anti- (Invitrogen, San Diego, CA, USA) supplemented with heat inactivated 10% goat Alexa Fluor 568 and anti-mouse Alexa Fluor 650 (Molecular Probe, fetal bovine serum. RWPE-1 was cultured in keratinocyte medium Carlsbad, CA, USA). Confocal imaging was performed with a Zeiss LSM 510 (Invitrogen) supplemented with epithelial growth factor (Invitrogen) and NLO confocal system (Carl Zeiss Inc., Thornwood, NY, USA) based on a bovine pituitary extract (Invitrogen). C4-2-B2 cell line was purchased from Axiovert 200 M inverted microscope and operating with a 30-mW argon ViroMed, Lab Inc. (Minneapolis, MN, USA) and was cultured in T-medium laser tuned to 488 nm, a 1-mW HeNe laser tuned to 543 nm and a 1-mW (Invitrogen) supplemented with 10% fetal bovine serum. All cell cultures HeNe laser tuned to 633 nm. Images were collected in Zeiss AIM software were grown in a humidified CO2 incubator at 37 1C. (v 4.0) using a Â 63 or Â 40 Plan-Apochromat 1.4 NA oil immersion objective and a multi-track configuration where different color signals were sequentially collected in separate photomultiplier tubes with a BP Plasmids, transfections and real-time PCR 500–530 nm, BP 565–615 nm and BP 650–710 filters after excitation with For stable silencing of a-catenin, RhoA, RhoC, and E-cadherin, four 488 nm, 543 and 633 nm laser, respectively. Images were 512 Â 512 pixels respective SureSilencing shRNA plasmid vectors (KH00305H, KH01089N, with line having average 4. KH00135P, SABiosciences, Frederick, MD, USA), containing different resistance gene, were transfected into cells using lipofectamine 2000, and the media containing antibiotic was changed after 48 h. Two weeks Invasion assay post transfection, single colonies were picked up and expanded. An Cell invasion assays were performed using 8-mm pore size BD Falcon- adenovirus encoding either DLC1 cDNA or LacZ was prepared and FluoroBlok (Becton Dickinson, Franklin Lakes, NJ, USA). Cells were transduced as previously described.7 Adenovirus encoding a-catenin was transduced with Ad-DLC1 or Ad-LacZ or stably transfected with ShRNA purchased from Vector Biolabs (Philadelphia, PA, USA). Myc tagged of E-cadherin and GFP-E-cadherin, and 1 Â 106 cells/well were plated on RhoAV14 and RhoC V14 were a kind gift from Christopher J Marshal (Cancer ECMatrix gel-coated layered cell culture inserts. After 24 h, the non- Research UK Center for Cell and Molecular Biology). For real-time PCR, RNA invading cells and the extracellular Matrix gel from the interior of the was prepared using RNAesy mini kit (Qiagen, Valencia, CA, USA) and cDNA inserts were removed using a cotton-tipped swab. Invaded cells were were prepared using SuperScript III First-Strand Synthesis SuperMix for stained or quantitated using calcein AM fluorescent dye (Becton

& 2014 Macmillan Publishers Limited Oncogene (2014) 724 – 733 DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 732 Dickinson), and fluorescence was measured at wavelength 494/517 15 Fukata M, Kaibuchi K. Rho-family GTPases in cadherin-mediated cell-cell adhesion. (absorbance/emission). Nat Rev Mol Cell Biol 2001; 2: 887–897. 16 Me`ge RM, Gavard J, Lambert M. Regulation of cell-cell junctions by the cytos- Cell aggregation assay keleton. Curr Opin Cell Biol 2006; 18: 541–548. 17 Batlle E, Sancho E, Francı´ C, Domıńguez D, Monfar M, Baulida J et al. The tran- Cells were transduced with Ad-LacZ or Ad-DLC1 in medium with 10% fetal scription factor snail is a repressor of E-cadherin gene expression in epithelial bovine serum. To dissociate cells, a solution 0.05% Trypsin in phosphate- tumour cells. Nat Cell Biol 2000; 2: 84–89. buffered saline with calcium was added, cells were collected and washed 18 Li LC, Zhao H, Nakajima K, Oh BR, Ribeiro Filho LA et al. Methylation of the with balanced salt solution containing 10 mM HEPES (4-(2-hydroxyethyl)-1- E-cadherin gene promoter correlates with progression of prostate cancer. J Urol piperazineethanesulfonic acid). Cells were dissolved in HBSS (Hank’s 2001; 166: 705–709. Balanced Salt Solution) buffer containing 0.01% trypsin and 10 mM HEPES for 30 min at 37 1C to have single-cell suspension followed by washing. 19 Hajra KM, Chen DY, Fearon ER. The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res 2002; 62: 1613–1618. Further, cells were resuspended in HBSS (with 10 mM HEPES) and concentration was adjusted to 1 Â 105 cells/ml. They were then added 20 Brouxhon S, Kyrkanides S, O’Banion MK, Johnson R, Pearce DA, Centola GM et al. into a 24-well tissue culture plate and placed on a shaker at 60 r.p.m. and Sequential down-regulation of E-cadherin with squamous cell carcinoma pro- incubated at various time points. In selected cultures, HECD-1 (Abcam) or gression: loss of E-cadherin via a prostaglandin E2-EP2 dependent posttransla- EGTA was included. Proportions of this cell suspension were removed at tional mechanism. Cancer Res 2007; 67: 7654–7664. 15, 30 and 45 min and placed in ice and then fixed in a one volume of 5% 21 Rashid MG, Sanda MG, Vallorosi CJ, Rios-Doria J, Rubin MA, Day ML. Post- paraformaldehyde. Total numbers of particles were counted, and the translational truncation and inactivation of human E-cadherin distinguishes aggregation index was calculated using formula N0 À Nt/N0, where N0 is prostate cancer from matched normal prostate. Cancer Res 2001; 61: 489–492. the number of particles at zero time, that is, just when the experiment 22 Mosesson Y, Mills GB, Yarden Y. Derailed endocytosis: an emerging feature of started, and Nt is the number at the particular time. cancer. Nat Rev Cancer 2008; 8: 835–850. 23 Perl AK, Wilgenbus P, Dahl U, Semb H, Christofori G. A causal role for E-cadherin in the transition from adenoma to carcinoma. Nature 1998; 392: 190–193. CONFLICT OF INTEREST 24 Asnaghi L, Vass WC, Quadri R, Day PM, Qian X, Braverman R et al. E-cadherin negatively regulates neoplastic growth in non-small cell lung cancer: role of Rho The authors declare no conflict of interest. GTPases. Oncogene 2010; 29: 2760–2771. 25 Guan M, Zhou X, Soulitzis N, Spandidos DA, Popescu NC. Aberrant methylation and deacetylation of deleted in liver cancer-1 gene in prostate cancer: potential ACKNOWLEDGEMENTS clinical applications. Clin Cancer Res 2006; 12: 1412–1419. This research was supported by the Intramural Research Program of the National 26 Vleminckx K, Vakaet Jr L, Mareel M, Fiers W, van Roy F. Genetic manipulation of Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA. E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 1991; 66: 107–119. 27 Christofori G, Semb H. The role of the cell-adhesion molecule E-cadherin as a REFERENCES tumour-suppressor gene. Trends Biochem Sci 1999; 24: 73–76. 28 Noren NK, Liu BP, Burridge K, Kreft B. p120 catenin regulates the actin cytoske- 1 Goodison S, Yuan J, Sloan D, Kim R, Li C, Popescu NC et al. The RhoGAP protein leton via Rho family GTPases. J Cell Biol 2000; 150: 567–580. DLC-1 functions as a metastasis suppressor in breast cancer cells. Cancer Res 2005; 29 Noe¨ V, Fingleton B, Jacobs K, Crawford HC, Vermeulen S, Steelant W et al. Release 65: 6042–6053. of an invasion promoter E-cadherin fragment by matrilysin and stromelysin-1. 2 Durkin ME, Yuan BZ, Zhou X, Zimonjic DB, Lowy DR, Thorgeirsson SS et al. DLC-1:a J Cell Sci 2001; 114: 111–118. Rho GTPase-activating protein and tumour suppressor. J Cell Mol Med 2007; 11: 30 Frixen UH, Behrens J, Sachs M, Eberle G, Voss B, Warda A et al. E-cadherin- 1185–1207. mediated cell-cell adhesion prevents invasiveness of human carcinoma cells. JCell 3 Xue W, Krasnitz A, Lucito R, Sordella R, Vanaelst L, Cordon-Cardo C et al. DLC1 is a Biol 1991; 113: 173–185. chromosome 8p tumor suppressor whose loss promotes hepatocellular carci- 31 Wong AS, Gumbiner BM. Adhesion-independent mechanism for suppression of noma. Genes Dev 2008; 22: 1439–1444. tumor cell invasion by E-cadherin. J Cell Biol 2003; 161: 1191–1203. 4 Vigil D, Cherfils J, Rossman KL, Der CJ. Ras superfamily GEFs and GAPs: validated 32 Ko¨ksal IT, Ozcan F, Kiliçaslan I, Tefekli A. Expression of E-cadherin in prostate and tractable targets for cancer therapy? Nat Rev Cancer 2010; 12: 842–857. 5 Thiolloy S, Rinker-Schaeffer CW. Thinking outside the box: using metastasis sup- cancer in formalin-fixed, paraffin-embedded tissues: correlation with pathological pressors as molecular tools. Semin Cancer Biol 2011; 21: 89–98. features. Pathology 2002; 34: 233–238. 6 Zhou X, Thorgeirsson SS, Popescu NC. Restoration of DLC-1 gene expression 33 Cheng L, Nagabhushan M, Pretlow TP, Amini SB, Pretlow TG. Expression of induces apoptosis and inhibits both cell growth and tumorigenicity in human E-cadherin in primary and metastatic prostate cancer. Am J Pathol 1996; 148: hepatocellular carcinoma cells. Oncogene 2004; 23: 1308–1313. 1375–1380. 7 Guan M, Tripathi V, Zhou X, Popescu NC. Adenovirus-mediated restoration of 34 Conacci-Sorrell M, Simcha I, Ben-Yedidia T, Blechman J, Savagner P, Ben-Ze’ev A. expression of the tumor suppressor gene DLC1 inhibits the proliferation and Autoregulation of E-cadherin expression by cadherin-cadherin interactions: tumorigenicity of aggressive, androgen-independent human prostate cancer cell the roles of beta-catenin signaling, Slug, and MAPK. J Cell Biol 2003; 163: lines: prospects for gene therapy. Cancer Gene Ther 2008; 15: 371–381. 847–857. 8 Healy KD, Hodgson L, Kim TY, Shutes A, Maddileti S, Juliano RL et al. DLC-1 35 Lozano E, Bretson M, Braga VM. Tumor progression: small GTPases and loss of cell- suppresses non-small cell lung cancer growth and invasion by RhoGAP-depen- cell adhesion. BioEssays 2003; 25: 452–463. dent and independent mechanisms. Mol Carcinog 2008; 47: 326–337. 36 Braga VM, Yap AS. The challenges of abundance: epithelial junctions and small 9 Tripathi V, Popescu NC, Zimonjic DB. DLC1 interaction with a-catenin stabilizes GTPase signalling. Curr Opin Cell Biol 2005; 17: 466–474. adherens junctions and enhances DLC1 antioncogenic activity. Mol Cell Biol 2012; 37 Etienne-Manneville S, Hall A. Rho GTPases in cell biology. Nature 2002; 420: 629–635. 32: 2145–2159. 38 Vega FM, Ridley AJ. Rho GTPases in cancer cell biology. FEBS Lett 2008; 582: 10 Yam JW, Ko FC, Chan CY, Jin DY, Ng IO. Interaction of deleted in liver cancer 1 2093–2101. with tensin2 in caveolae and implications in tumor suppression. Cancer Res 2006; 39 Hodge JC, Bub J, Kaul S, Kajdacsy-Balla A, Lindholm PF. Requirement of RhoA 66: 8367–8372. activity for increased nuclear factor kappaB activity and PC-3 human prostate 11 Liao YC, Si L, deVere White RW, Lo SH. The phosphotyrosine-independent inter- cancer cell invasion. Cancer Res 2003; 63: 1359–1364. action of DLC-1 and the SH2 domain of cten regulates focal adhesion localization 40 Burbelo P, Wellstein A, Pestell RG. Altered Rho GTPasesignaling pathways in and growth suppression activity of DLC-1. J Cell Biol 2007; 176: 43–49. breast cancer cells. Breast Cancer Res Treat 2004; 84: 43–48. 12 Qian X, Li G, Asmussen HK, Asnaghi L, Vass WC, Braverman R et al. Oncogenic 41 Chang YW, Marlin JW, Chance TW, Jakobi R. RhoA mediates cyclooxygenase-2 inhibition by a deleted in liver cancer gene requires cooperation between tensin signaling to disrupt the formation of adherens junctions and increase cell motility. binding and Rho-specific GTPase-activating protein activities. Proc Natl Acad Sci Cancer Res 2006; 66: 11700–11708. USA 2007; 104: 9012–9017. 42 Suwa H, Ohshio G, Imamura T, Watanabe G, Arii S, Imamura M et al. Over- 13 Li G, Du X, Vass WC, Papageorge AG, Lowy DR, Qian X. Full activity of the deleted in expression ofthe RHOC gene correlates with progression of ductal adenocarci- liver cancer 1 (DLC1) tumor suppressor depends on an LD-like motif that binds talin nomaof the pancreas. Br J Cancer 1998; 77: 147–152. and focal adhesion kinase (FAK). Proc Natl Acad Sci USA 2011; 108: 17129–17134. 43 Clark EA, Golub TR, Lander ES, Hynes RO. Genomic analysis of metastasis reveals 14 Sahai E, Marshall CJ. RHO-GTPases and cancer. Nat Rev Cancer 2002; 2: 133–142. an essential role for RHOC. Nature 2000; 406: 532–535.

Oncogene (2014) 724 – 733 & 2014 Macmillan Publishers Limited DLC1 modulates E-cadherin and suppresses invasion V Tripathi et al 733 44 Wu M, Wu ZF, Rosenthal DT, Rhee EM, Merajver SD. Characterization of the roles 51 Sawada K, Mitra AK, Radjabi AR, Bhaskar V, Kistner EO, Tretiakova M et al. Loss of of RHOC and RHOA GTPases in invasion, motility, and matrix adhesion in E-cadherin promotes ovarian cancer metastasis via alpha 5-integrin, which is a inflammatory and aggressive breast cancers. Cancer 2010; 116: 2768–2782. therapeutic target. Cancer Res 2008; 68: 2329–2339. 45 Sequeira L, Dubyk CW, Riesenberger TA, Cooper CR, van Golen KL. Rho GTPases in 52 Zhou Q, Yan B, Hu X, Li XB, Zhang J, Fang J. Luteolin inhibits invasion of PC-3 prostate cancer cell morphology, invasion and tumor cell diapedesis. Clin prostate cancer PC3 cells through E-cadherin. Mol Cancer Ther 2009; 8: Exp Metastasis 2008; 25: 569–579. 1684–1691. 46 Sahai E, Marshall CJ. ROCK and Dia have opposing effects on adherens junctions 53 Jiang WG, Hiscox S, Hallett MB, Horrobin DF, Mansel RE, Puntis MC. Regulation of downstream of Rho. Nat Cell Biol 2002; 4: 408–415. the expression of E-cadherin on human cancer cells by gamma-linolenic acid 47 Wong CC, Wong CM, Ko FC, Chan LK, Ching YP, Yam JW et al. Deleted in liver (GLA). Cancer Res 1995; 55: 5043–5048. cancer 1 (DLC1) negatively regulates Rho/ROCK/MLC pathway in hepatocellular 54 Weaver VM, Petersen OW, Wang F, Larabell CA, Briand P, Damsky C et al. carcinoma. PLoS One 2008; 3: e2779. Reversion of the malignant phenotype of human breast cells in three-dimensional 48 Kamai T, Tsujii T, Arai K, Takagi K, Asami H, Ito Y et al. Significant association of culture and in vivo by integrin blocking antibodies. J Cell Biol 1997; 137: 231–245. Rho/ROCK pathway with invasion and metastasis of bladder cancer. Clin Cancer 55 Annicotte JS, Iankova I, Miard S, Fritz V, Sarruf D, Abella A et al. Peroxisome Res 2003; 9: 2632–2641. proliferator-activated receptor gamma regulates E-cadherin expression and 49 Sawada K, Morishige K, Mabuchi S, Ogata S, Kawase C, Sakata M et al. In vitro and inhibits growth and invasion of prostate cancer. Mol Cell Biol 2006; 26: 7561–7574. in vivo assays to analyze the contribution of Rho kinase in angiogenesis. Methods 56 Lahoz A, Hall A. DLC1: a significant GAP in the cancer genome. Genes Dev 2008; Enzymol 2008; 439: 395–412. 22: 1724–1730. 50 Jeong KJ, Park SY, Cho KH, Sohn JS, Lee J, Kim YK et al. The Rho/ROCK pathway for 57 Zimonjic DB, Popescu NC. Role of DLC1 tumor suppressor gene and MYC onco- lysophosphatidic acid-induced proteolytic enzyme expression and ovarian cancer gene in pathogenesis of human hepatocellular carcinoma: Potential prospects for cell invasion. Oncogene 2012; 31: 4279–4289. combined targeted therapeutics. Int J Oncol 2012; 41: 393–406.

& 2014 Macmillan Publishers Limited Oncogene (2014) 724 – 733