Cadherin switching and activation of p120 catenin signaling ... - Nature

3 downloads 74 Views 1MB Size Report
Feb 1, 2010 - ... Hong Kong, PR China and 2Department of Obstetrics and Gynecology,. University of British Columbia, Vancouver, British Columbia, Canada.
Oncogene (2010) 29, 2427–2440

& 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 $32.00 www.nature.com/onc

ORIGINAL ARTICLE

Cadherin switching and activation of p120 catenin signaling are mediators of gonadotropin-releasing hormone to promote tumor cell migration and invasion in ovarian cancer LWT Cheung1, PCK Leung2 and AST Wong1 1

School of Biological Sciences, University of Hong Kong, Hong Kong, PR China and 2Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, British Columbia, Canada

Gonadotropin-releasing hormone (GnRH) receptor expression is often elevated in ovarian cancer, but its potential role in ovarian cancer metastasis has just begun to be revealed. Cadherin switching is a crucial step during tumorigenesis, particularly in metastasis. Here, we showed that GnRH is an inducer of E- to P-cadherin switching, which is reminiscent of that seen during ovarian tumor progression. Overexpression of P-cadherin significantly enhanced, whereas knockdown of P-cadherin reduced migration and invasion regardless of E-cadherin expression, suggesting that inappropriate expression of P-cadherin contributes to the invasive phenotype. These effects of P-cadherin were mediated by activation of the Rho GTPases, Rac1, and Cdc42, through accumulation of p120 catenin (p120ctn) in the cytoplasm. The use of p120ctn small interfering RNA or chimeric cadherin construct to inhibit p120ctn expression and cytoplasmic localization, respectively, resulted in significant inhibition of cell migration and invasion, with a concomitant reduction in Rac1 and Cdc42 activation, confirming that the effect was p120ctn specific. Similarly, the migratory/invasive phenotype could be reversed by expression of dominant-negative Rac1 and Cdc42. These results identify for the first time cadherin switching and p120ctn signaling as important targets of GnRH function and as novel mediators of invasiveness and tumor progression in ovarian cancer. Oncogene (2010) 29, 2427–2440; doi:10.1038/onc.2009.523; published online 1 February 2010 Keywords: ovarian cancer; GnRH; cadherin; p120ctn; motility

Introduction Ovarian cancer is a highly metastatic disease and shows the worst mortality among all gynecological malignancies. As most (B70%) ovarian cancer patients are diagnosed at an advanced stage with widespread Correspondence: Dr AST Wong, School of Biological Sciences, University of Hong Kong, 4S-14 Kadoorie Biological Sciences Building, Pokfulam Road, Hong Kong, PR China. E-mail: [email protected] Received 11 June 2009; revised 10 December 2009; accepted 23 December 2009; published online 1 February 2010

peritoneal metastasis, current therapies are ineffective. The 5-year survival for these women is low (o25%) (Jemal et al., 2008), necessitating new therapeutic targets and a better understanding of the mechanisms that regulate ovarian cancer metastasis. The hypothalamic decapeptide gonadotropin-releasing hormone (GnRH), beyond its activities in the control of pituitary gonadotropin secretion, has a key role in extrapituitary tissues and tumor cells, including ovarian cancer. The widespread presence (480%) of GnRH receptor in ovarian carcinomas underlies the importance of understanding the function of GnRH in ovarian cancer (Emons et al., 1989; Irmer et al., 1995). In addition to its well-established role in regulating proliferation and cell survival, our recent findings show for the first time that GnRH may be involved in other aspects of ovarian cancer progression, such as invasion and metastasis (Cheung et al., 2006). However, the molecular mechanism whereby GnRH contributes to ovarian cancer metastatic potential is poorly understood. A hallmark of metastatic progression is the dynamic regulation of cadherins. Cadherins are a family of transmembrane glycoproteins that mediate Ca2 þ -dependent homophilic cell–cell adhesion. Cadherin subtype switching in which other cadherins replace or are co-expressed with E-cadherin during embryogenesis has important functional roles in cell segregation (Wheelock et al., 2008). Tumor cells often recapitulate this activity during cancer progression in which tumor cells acquire the capacity to migrate, invade, and metastasize. Although E-cadherin is present in primary ovarian carcinomas, it often is scanty or absent in ovarian carcinoma metastases (Sundfeldt, 2003). This decrease in E-cadherin seems to be associated with concomitant increase in P-cadherin, suggesting that cadherin switching from E- to P-cadherin may have a role in ovarian cancer progression (Patel et al., 2003). In addition to their adhesive functions, cadherins can modulate signal transduction by interacting with b-catenin and p120 catenin (p120ctn) (Cavallaro and Christofori, 2004). For example, b-catenin mediates Wnt signaling pathway through its interaction with LEF/TCF transcription factors for target gene activation (Clevers, 2006). p120ctn, originally identified as an Src substrate, was subsequently shown to signal through

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2428

its activity on the Rho family GTPases (for example, Rac1, Cdc42, RhoA), which mediate cytoskeletal dynamics in cell migration (Nobes and Hall, 1999; Noren et al., 2000). In this study, we show for the first time a role for GnRH in the E- to P-cadherin switching that is reminiscent of metastasis from ovarian carcinomas. We also define the metastatic signaling cascade activated by P-cadherin, which alters trafficking of p120ctn to the cytoplasm, and thereby enhancing activities of Rho GTPases, Rac1, and Cdc42.

Results GnRH induces E- to P-cadherin switching in ovarian cancer cells To test the possibility that GnRH acts through cadherins to promote invasiveness, we first examined whether GnRH could affect the expression of the main cadherins, that is E-, N-, and P-cadherin, that have key roles in ovarian tumor differentiation and progression (Wong et al., 1999; Patel et al., 2003; Sundfeldt, 2003). We chose to use Caov-3 and OVCAR-3, which express high levels of GnRH receptor (Cheung et al., 2006). Interestingly, in line with the dose-dependent response of cellular migration and invasion as shown earlier (Cheung et al., 2006), cells treated with GnRHa at doses of 0.1–10 nM for 24 h caused a significant decrease in Ecadherin and increase in P-cadherin proteins, whereas the response to higher doses of GnRHa (100 nM and 1 mM) was insignificant (Figure 1a). These results were confirmed at mRNA levels by real-time PCR (Figure 1b). Maximum changes in cadherin expression were observed at 0.1 nM of GnRHa (Figures 1a and b), which corroborates with the maximal effective dose of GnRHa on migration and invasion (Cheung et al., 2006). This concentration was used in all subsequent experiments. There was no significant difference in Ncadherin level (Figures 1a and b), consistent with its reported constitutive expression in ovarian cancer (Wong et al., 1999). These results suggest that a switch to P-cadherin may have a unique role in GnRHenhanced cell migration/invasion in ovarian cancer cells. Expression of GnRH receptor small interfering RNA (siRNA), but not nonspecific siRNA, revealed a significant decrease in GnRHa-induced P-cadherin expression (Figure 1c). In addition, cotreatment of cells with the GnRH antagonist antide, which is known to inhibit the functional activity of GnRH receptor (Li et al., 1994) and its downstream signaling (Kim et al., 2006), also prevented the upregulation of P-cadherin (Figure 1d), confirming a direct involvement of GnRH receptor in regulating P-cadherin expression. The increase in P-cadherin was not the result of changes in mRNA stability, as GnRHa did not affect the turnover rate of the P-cadherin mRNA (Figure 2a). Instead, GnRHa stimulated activation of the P-cadherin promoter (Figure 2b; Po0.05), suggesting a transcriptional mechanism of regulation. Oncogene

P-cadherin contributes to the invasive phenotype It has been suggested that expression of an inappropriate cadherin may directly contribute to the invasive phenotype (Nieman et al., 1999; Feltes et al., 2002; Suyama et al., 2002; Maeda et al., 2006). To test whether P-cadherin may promote migration/invasion, we stably transfected Caov-3 and OVCAR-3 cells with P-cadherin and analyzed their effects on cell migration/invasion. Three independent clones of each cell line were isolated. We confirmed overexpression of P-cadherin by western blot (Figures 3a and b, left panels) and also showed that levels of P-cadherin in these stable cell lines were similar to those detected on GnRHa treatment. It is worth noting that P-cadherin overexpression did not affect endogenous expression of either E- or N-cadherin (Figures 3a and b, left panels). Immunofluorescence microscopy revealed P-cadherin staining was enriched at the intercellular boundaries of P-cadherin transfectants (Figures 3a and b, right panels). Using transwell assays, we show that P-cadherin-overexpressing cells were significantly more motile and invasive (Figure 3c) than parental and empty vector control cells (B3-fold for migration and 3.5-fold for invasion; Po0.005). To determine whether P-cadherin was directly involved in regulating GnRH-mediated cellular migration and invasion, we used siRNA approach. The increases in migration/invasion in GnRHa-activated cells were completely abolished by GnRH receptor siRNA but not the nonspecific siRNA (Figure 4b). Importantly, siRNA-mediated depletion of P-cadherin, in which reduced the levels of P-cadherin but not the levels of E- or N-cadherin (Figure 4a), inhibited the ability of GnRHa to stimulate migration to near basal levels (97 and 94% inhibition in Caov-3 and OVCAR-3, respectively; Figure 4b, upper panel). Similarly, we showed that depletion of P-cadherin inhibited invasion by 91% in Caov-3 and 93% in OVCAR-3 (Figure 4b, lower panel). Under these conditions, and consistent with previous observation (Kang et al., 2000), GnRHa did not cause significant alterations of cell proliferation as determined by MTT assays (Supplementary Figure 1). We also showed no effect of P-cadherin on cell proliferation (Supplementary Figure 1), indicating that the increase in cell migration/invasion was not because of a proliferative effect. These results suggest that P-cadherin may have a dominant effect over E-cadherin on the invasive phenotype of ovarian cancer cells. p120ctn subcellular localization is regulated by P-cadherin and is critical for GnRH-induced migration/invasion of ovarian cancer cells Catenins, which associate with the intracellular domain of cadherins, have a dual role in adhesion and signaling. An important aspect of catenins is their ability to shuttle between cadherin-bound (membrane) and cytoplasmic pools, and only cytoplasmic catenins can exert signaling functions (Perez-Moreno and Fuchs, 2006). This prompted us to investigate whether the increase in P-cadherin induces changes in the subcellular localization of the catenins: a-catenin, b-catenin, and p120ctn. In

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2429

Figure 1 GnRH induces E- to P-cadherin switching in ovarian cancer cells. Caov-3 and OVCAR-3 cells were treated with increasing concentrations of GnRHa (0.1 nM–1 mM) for 24 h. The expression of E-, N-, and P-cadherin was evaluated by (a) western blot using specific antibodies (upper panel) and normalized against b-actin (lower panel) or by (b) real-time PCR with gene-specific primers. (c) Caov-3 and OVCAR-3 cells were transiently transfected with 20 nM GnRH receptor (GnRHR) or nonspecific (NS) siRNA for 24 h, or (d) pretreated with antide (100 nM) for 30 min before treatment with GnRHa (0.1 nM) for 24 h. Total protein was collected for western blot using specific antibodies to P-cadherin and b-actin (left panel). The levels of P-cadherin were normalized to b-actin as determined by densitometry (right panel). *Po0.05, difference with control.

Figure 5a, we showed that, in contrast to a-catenin and b-catenin, p120ctn was localized predominantly in the membrane in control cells, but accumulated extensively

in the cytoplasm of GnRHa-treated cells. There was no change in total p120ctn, suggesting there was a redistribution of p120ctn from the plasma membrane to the Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2430

Figure 2 GnRH activates P-cadherin promoter activity. (a) Caov-3 cells were pretreated with 0.1 nM GnRHa for 3 h and then treated with actinomycin D (4 mg/ml) for the indicated time periods. Real-time PCR was performed with P-cadherin sequence-specific primers, and results are expressed as fold change relative to time zero. (b) Caov-3 cells were transfected with P-cadherin promoter-reporter gene construct and pSV-b-gal for normalization. Eight hours after transfection, cells were incubated with 0.1 nM GnRHa for further 24 h and then collected for luciferase and b-galactosidase assays. *Po0.05, difference with control.

Figure 3 P-cadherin expression contributes to the invasive phenotype. (a) Caov-3 and (b) OVCAR-3 cells were stably transfected with either the empty vector pcDNA3.1 (Neo) or full-length P-cadherin (P-cad) construct. Total cell lysates were analyzed for P-, E-, and N-cadherin expressions by western blot (left panel). Cells were grown on glass coverslips and immunostained with anti-P-cadherin antibody and then detected with Texas red-conjugated secondary antibody (right panel). (c) Cells were allowed to migrate or invade through transwell chambers. Results are mean±s.d. of migrated/invaded cells of five fields of triplicate wells from three independent experiments. **Po0.005, difference with control.

cytoplasm after GnRHa stimulation. Moreover, inhibition of P-cadherin expression with siRNA abolished p120ctn cytoplasmic accumulation (Figure 5b). These Oncogene

changes were consistent with a prominent membrane staining in transfected cells observed by immunofluorescence microscopy (Figure 5c), indicating that p120ctn

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2431

membrane proximal chimera significantly increased the amount of p120ctn recruited to the membrane, whereas the level of cytoplasmic p120ctn was largely reduced. Importantly, the IL2R/E-cadherin membrane proximal chimera potently inhibited the migratory and invasive phenotypes of P-cadherin-overexpressing cells (Figure 7c). In line with this, there was also a strong inhibition of GnRHa-induced cell migration and invasion (Figure 7d). These results indicate that p120ctn contributes to the migratory/invasive behavior and suggest that the effect was mediated through the cytoplasmic localization of p120ctn.

Figure 4 P-cadherin is required for GnRH-induced cell migration and invasion. (a) Cells were transiently transfected with P-cadherin (P-cad) or nonspecific (NS) siRNA for 24 h. Whole-cell lysates were analyzed for the levels of P-, E-, and N-cadherin by western blot. (b) GnRH receptor (GnRHR), P-cad, or NS siRNA transfected cells were seeded on transwell inserts for cell migration (upper panel) and invasion (lower panel) in the presence or absence of 0.1 nM GnRHa for 24 h. Results are mean±s.d. of migrated/ invaded cells of five fields of triplicate wells from three independent experiments. **Po0.005, difference with control.

relocalization in response to GnRH was induced by P-cadherin. To determine whether p120ctn was required for the migratory/invasive phenotype, we depleted p120ctn using siRNAs. Specific inhibition of p120ctn was confirmed by western blot analysis (Figure 6a). Both the migratory and invasive phenotypes of P-cadherin-overexpressing Caov-3 and OVCAR-3 cells were clearly reduced in p120ctn siRNA transfectants (Figure 6b). Treatment of cells with p120ctn siRNA also suppressed GnRHainduced cell migration and invasion, but no inhibition was observed for nonspecific siRNA (Figure 6c). To further determine whether the observed cell migration and invasion occurred through accumulation of p120ctn in the cytoplasm, we asked whether expression of the IL2R/E-cadherin membrane proximal chimera to sequester p120ctn to the plasma membrane (away from the cytoplasm) could inhibit cell migration and invasion in these cell lines. This construct contains the membrane proximal region (corresponding to p120ctn binding site) of the E-cadherin cytoplasmic tail fused to the extracellular domain of the IL2R (interleukin-2 receptor) a subunit (Figure 7a), and thereby binds p120ctn but does not mediate adhesion (Miranda et al., 2001). As shown in Figure 7b, expression of the IL2R/E-cadherin

p120ctn mediates GnRH-induced cell migration and invasion through Rac1 and Cdc42 Given the ability of p120ctn to regulate the activity of Rho family GTPase and the potential contribution of these GTPases to the invasive/metastatic phenotype (Nobes and Hall, 1999; Noren et al., 2000), we examined the activities of Rac1, Cdc42, and RhoA using pull-down assays. A significant increase in the levels of GTP-bound (active) Rac1 (3.4-fold) and Cdc42 (2.5fold) was detected in P-cadherin-overexpressing cells when compared with control cells (Figure 8a; Po0.05). Furthermore, knocking down of p120ctn by siRNA (#1) was able to decrease the P-cadherin effect on Rac1 and Cdc42 activities (Figure 8a). In contrast, we detected no change of active RhoA (Figure 8a). Similar results were obtained with a different siRNA (#2) shown to deplete p120ctn (Figure 6a), indicating that the effect was p120ctn specific (Supplementary Figure 2). In accordance with above results on p120ctn cytoplasmic localization, we also found that Rac1 and Cdc42 activities could be completely abolished by sequestering p120ctn with the IL2R/E-cadherin membrane proximal construct (Figure 8b). To further establish the role that GnRH has in Rac1 and Cdc42 activation, we repeated the pull-down assays using GnRHa-stimulated cells. As shown, GnRHa increased activities of Rac1 and Cdc42 as early as 4 h after treatment (Figure 9a) and this activation coincided with the initial induction of P-cadherin (Supplementary Figure 3). Consistent with the above observation, GnRHa did not have any effect on RhoA activation (Figure 9a). Knockdown of P-cadherin by siRNA significantly inhibited GnRHa-induced activation of Rac1 and Cdc42 (Figure 9b), suggesting the involvement of P-cadherin. Consistently, there was a clear reduction of GnRHa-induced activation of Rac1 and Cdc42 by p120ctn downregulation with siRNA or IL2R/E-cadherin membrane proximal chimera (Figures 9b and c). The expression of dominant-negative Rac1 (N17) and Cdc42 (N17) also dramatically decreased cell migration/ invasion of P-cadherin-overexpressing cells (Figure 10a) and GnRHa-stimulated cells (Figure 10b). In line with the above results showing no effect of RhoA, dominantnegative RhoA (N19) did not reduce the number of migrated/invaded cells (Figures 10a and b), despite efficient expression of the mutant detected by the expression of myc protein (Figure 10a, inset of upper Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2432

Figure 5 P-cadherin regulates cytoplasmic accumulation of p120ctn. (a) Caov-3 cells were treated with GnRHa (0.1 nM) for 24 h and fractionated as described in ‘Materials and methods.’ Equal amounts of cytoplasmic, membrane, nuclear proteins (left panel) and total cell lysates (right panel) were analyzed by western blot using anti-a-catanin, anti-b-catenin, and anti-p120ctn antibodies. Equal loading and fraction purity were confirmed by reprobing for the cytoplasmic marker protein a-tubulin, membrane marker protein N-cadherin, and nuclear marker protein histone H1. (b) Cells were transfected with P-cadherin (P-cad) or nonspecific (NS) siRNA. Twenty-four hours after transfection, the cells were treated with GnRHa (0.1 nM). Cytoplasmic and membrane fractions were prepared and analyzed by western blot with anti-p120ctn antibody. (c) Cells were grown on glass coverslips and immunostained with anti-p120ctn antibody and then detected with Texas red-conjugated secondary antibody.

panel). These results suggest that the migratory/invasive effect of GnRH is Rac1/Cdc42 dependent.

Discussion Considering the extremely high rate of metastasis of ovarian tumors, we are interested in mechanisms leading to the progression of ovarian carcinoma. Alteration in cell adhesion is a hallmark feature of metastasis (Wheelock et al., 2008). Earlier studies have documented Oncogene

the switching of cadherin profiles during ovarian tumor progression (Patel et al., 2003). Here, we provide evidence that GnRH is a critical inducer of the E- to P-cadherin switch, which is reminiscent of that observed in the malignant progression of ovarian cancer. We also show that P-cadherin activates a metastatic signaling cascade that results in p120ctn trafficking to the cytoplasm, Rho GTPase activation, and consequently, cell motility and invasion in ovarian cancer cells (Figure 11). One intriguing question is how GnRH may have a biphasic action: whereas low concentrations of GnRH

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2433

Figure 6 Knockdown of p120ctn by siRNA inhibited GnRH-mediated cell migration and invasion. (a) Caov-3 and OVCAR-3 cells were transfected with p120ctn or nonspecific (NS) siRNA for 24 h. Knockdown of p120ctn was confirmed by western blot. (b) Caov-3 (upper panel) and OVCAR-3 (lower panel) control cells and cells stably overexpressing P-cadherin (P-cad-29 for Caov-3 and P-cad-24 for OVCAR-3, respectively) were transfected with p120ctn or NS siRNA. Twenty-four hours after transfection, the cells were collected for migration and invasion assays. (c) p120ctn or NS siRNA transfected cells were seeded on transwell inserts for cell migration and invasion in the presence or absence of 0.1 nM GnRHa for 24 h. Results are mean±s.d. of migrated/invaded cells of five fields of triplicate wells from three independent experiments. **Po0.005, difference with control.

stimulate cell migration/invasion, high concentrations inhibit these functions (Cheung et al., 2006; Chen et al., 2007). The reasons are unknown, but the receptor expression level is known to be a determinant of different signal outcomes. We and others have shown earlier that low doses of GnRH upregulate its receptor, whereas high doses decrease it (Cheung et al., 2006). This difference in regulation suggests that high levels of GnRH receptor may enhance cellular response to GnRH stimulation, presumably because of more

efficient signal amplification or altered signaling through coupling to different G proteins (Cheung and Wong, 2008). In support, we found that GnRH induces cell motility and invasive response selectively in Caov-3 and OVCAR-3, which express high levels of GnRH receptor (Cheung et al., 2006; Chen et al., 2007). In addition, GnRH has been shown to regulate its own synthesis, suggesting that GnRH may be involved in autocrine/ paracrine regulation (Kang et al., 2000). Although our results and other studies failed to show an effect by Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2434

Figure 7 Cytoplasmic localization of p120ctn is important for GnRH-mediated cell migration and invasion. (a) Schematic representation of the IL2R/E-cadherin membrane proximal chimera (IL2R/Ecad MP). This chimera was designed to bind p120ctn without mediating adhesion. Wild-type E-cadherin is shown at the top. (b) Caov-3 cells were transiently transfected with IL2R/Ecad MP or empty vector (pcDNA3) for 24 h. Whole-cell lysates were analyzed by western blotting using E-cadherin and b-actin antibodies (left panel) or subjected to subcellular fractionation (right panel). The cytoplasmic, membrane, and nuclear proteins were immunoblotted with anti-p120ctn antibody. Equal loading and fraction purity were confirmed by reprobing for the cytoplasmic marker protein a-tubulin, membrane marker protein N-cadherin, and nuclear marker protein histone H1. (c) Caov-3 (left panel) and OVCAR-3 (right panel) control cells and cells stably overexpressing P-cadherin (P-cad-29 for Caov-3 and P-cad-24 for OVCAR-3, respectively) were transfected with IL2R/Ecad MP or pcDNA3. Twenty-four hours after transfection, the cells were collected for migration and invasion assays. (d) IL2R/EcadMP or pcDNA3 transfected cells were seeded on transwell inserts for cell migration and invasion in the presence or absence of 0.1 nM GnRHa for 24 h. Results are mean±s.d. of migrated/invaded cells of five fields of triplicate wells from three independent experiments. **Po0.005, difference with control.

inhibiting the GnRH receptor in the absence of GnRHa in these cells (Kang et al., 2000; Cheung et al., 2006; Kim et al., 2006), an autocrine system might exist Oncogene

in vivo. It is known that the GnRH precursor must be cleaved to become active (Cheng and Leung, 2005), and the lack of response with GnRH receptor inhibition in

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2435

Figure 8 p120ctn mediates P-cadherin activation of Rac1 and Cdc42. Caov-3 control cells and cells stably overexpressing P-cadherin (P-cad-29) were transfected with (a) p120ctn (#1) or nonspecific (NS) siRNA or (b) IL2R/E-cadherin membrane proximal (IL2R/Ecad MP) or empty vector (pcDNA3) and subjected to the pull-down assays. GTP-bound (active) Rac1, Cdc42, and RhoA were analyzed by western blot. GTPgS served as a positive control. GDP was a negative control. The signal intensity was determined by densitometry, and the levels of active Rac1, Cdc42, and RhoA were normalized to the total amounts of the corresponding proteins. *Po0.05, difference with control.

culture may have been influenced by the presence of an inactive precursor form of GnRH, which requires in vivo factors for activation. The results of this study have implications regarding the pathogenesis of ovarian cancer. Our finding that P-cadherin as a critical mediator of the invasive phenotype further supports the view that the incorrect expression of another adhesion molecule, rather than a decrease in E-cadherin, is associated with an increased invasive potential (Nieman et al., 1999; Feltes et al., 2002; Suyama et al., 2002; Maeda et al., 2006). Unlike E-cadherin, P-cadherin can promote motility and invasion in carcinoma cells. For example, expression of P-cadherin correlates with metastatic disease and poor prognosis of breast cancer and pancreatic carci-

nomas (Taniuchi et al., 2005; Paredes et al., 2007). Likewise, P-cadherin is frequently upregulated in ovarian cancer cells and increased with tumor progression (Patel et al., 2003). In this study, we showed that P-cadherin caused significant migration and invasion even in the presence of E-cadherin. Conversely, inhibition of P-cadherin did not affect the expression of E-cadherin, yet still resulted in reduced migratory and invasive capacities. These results suggest that P-cadherin may have a dominant effect over E-cadherin in contributing to the invasive behavior of ovarian carcinoma cells. In support of our results, others have shown that transfection of E-cadherin into the invasive and P-cadherin-expressing BT549 breast cancer cells did not revert their invasive phenotype (Sommers et al., Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2436

Figure 9 GnRH-induced activation of Rac1 and Cdc42 depends on p120ctn. (a) Caov-3 cells were incubated with GnRHa (0.1 nM) for the indicated time points and then collected for the pull-down assays. (b) Cells were transfected with P-cadherin (P-cad), p120ctn, or nonspecific (NS) siRNA or (c) IL2R/E-cadherin membrane proximal (IL2R/Ecad MP) or empty vector (pcDNA3) in the presence or absence of 0.1 nM GnRHa and then collected for the pull-down assays. GTP-bound (active) Rac1, Cdc42, and RhoA were analyzed by western blot. GTPgS served as a positive control. GDP was a negative control. The signal intensity was determined by densitometry, and the levels of active Rac1, Cdc42, and RhoA were normalized to the total amounts of the corresponding proteins. *Po0.05, difference with control.

1994). Moreover, E-cadherin is diminished in some, but not all, highly invasive and ascitic ovarian carcinomas (Sundfeldt, 2003). We have shown that P-cadherin can promote invasiveness of ovarian cancer cells by a mechanism involving more than a simple change in cellular adhesion. A cascade of signaling events activated by P-cadherin and dependent on p120ctn result in activation of the Rho family GTPases, Rac1, and Cdc42, which are known modulators of actin dynamics essential for cell migration and invasion. Although the mechanism by which p120ctn regulates Rac1/Cdc42 activation is still obscure, we found that p120ctn translocation to the Oncogene

cytoplasm seems to be a crucial factor. In support of our finding, others have shown that p120ctn in the cytoplasmic pool is important for the complex formation between p120ctn and the guanine nucleotide exchange factor, Vav2, in the regulation of Rac1 and Cdc42 (Noren et al., 2000). In addition to cytoskeletal changes, cytoplasmic p120ctn is also thought to be important for regulating invasion-related genes through binding to the transcriptional repressor Kaiso (Soubry et al., 2005). Indeed, cytoplasmic localization of p120ctn is observed in ovarian tumors, and a correlation has been noted between cytoplasmic p120ctn and distant metastases (Soubry et al., 2005).

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2437

Figure 11 Schematic representation of a P-cadherin-mediated metastatic signaling cascade activated by GnRH in ovarian cancer cells. GnRH induces E- to P-cadherin switching. The induction of P-cadherin results in the translocation of p120ctn from plasma membrane to the cytoplasm, which in turn activates Rac1 and Cdc42 and cellular invasiveness.

Figure 10 Rac1 and Cdc42 are required for GnRH-induced cell migration and invasion. (a) Caov-3 (upper panel) and OVCAR-3 (lower panel) control cells and cells stably overexpressing Pcadherin (P-cad-29 for Caov-3 and P-cad-24 for OVCAR-3, respectively) were transfected with myc-tagged dominant-negative (DN) Rac1, Cdc42, or RhoA. Twenty-four hours after transfection, whole-cell lysates were analyzed by western blot with anti-myc-tag antibody (inset). Transfected cells were collected for migration and invasion assays. (b) DN-Rac1, Cdc42, or RhoA transfected cells were seeded on transwell inserts for cell migration and invasion in the presence or absence of 0.1 nM GnRHa for 24 h. Results are mean±s.d. of migrated/ invaded cells of five fields of triplicate wells from three independent experiments. **Po0.005, difference with control.

Although cadherins do not exhibit any enzymatic activities, several recent studies have shown that they can interact with receptor tyrosine kinases to modulate

catenin signaling downstream of adherens junctions. For example, E-cadherin is associated with epidermal growth factor receptor, thus activating the mitogenactivated protein kinase pathway, and N-cadherin has also been found to interact with fibroblast growth factor receptor (Pece and Gutkind, 2000; Suyama et al., 2002; Qian et al., 2004). In this context, we have found that P-cadherin was able to transactivate the insulin-like growth factor-1 receptor in regulation of p120ctn signaling (our unpublished observations). This is also relevant to the clinical situation as insulin-like growth factor-1 receptor is frequently overexpressed in ovarian cancer and confers adverse prognosis (Spentzos et al., 2007). In addition to promoting motile and invasive signals, P-cadherin may also facilitate the adhesion of ovarian tumor cells to the peritoneum, a critical event in the metastatic progression of ovarian cancer, through homophilic and heterotypic interactions. In contrast to most solid tumors that metastasize through the lymphatics or hematogenously, ovarian cancer metastasizes by peritoneal dissemination. Successful adhesion to the peritoneum is an important first step in ovarian cancer metastasis. P-cadherin has been shown to be the predominant form of cadherin expressed by normal peritoneum (Chen et al., 2002), suggesting that it can actively participate in this process. Moreover, intimate cell–cell contacts may also facilitate juxtacrine ligand and receptor interactions, for example Jagged and Notch, which have shown to focally support the growth and invasion of cancer cells within the peritoneal cavity (Fagotto and Gumbiner, 1996; Choi et al., 2008). In summary, we report a novel role for GnRH in regulating the E- to P-cadherin switching that is seen during ovarian cancer progression. We also provide the signaling mechanism by which P-cadherin-expressing tumor cells can acquire metastatic properties. Metastasis is a complex, multistep process, which involves cell Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2438

adhesion, migration, and matrix degradation. On the basis of results from our previous (Cheung et al., 2006) and present studies, it seems that GnRH has developed different mechanisms for the invasive growth of ovarian tumor cells. These data highlight the importance of GnRH in ovarian cancer metastasis. Targeting GnRH receptor or its associated signaling may thus be a useful approach to molecular therapy for metastatic ovarian cancer.

Materials and methods Cell culture and treatments The human ovarian cancer cell lines Caov-3 and OVCAR-3 (kindly provided by Dr N Auersperg, University of British Columbia, Vancouver, BC, Canada) were grown in medium 199:105 (Sigma, St Louis, MO, USA) supplemented with 5% fetal bovine serum and 100 units/ml penicillin–streptomycin (Invitrogen, San Diego, CA, USA). Cells were maintained at 37 1C in humidified atmosphere of 5% CO2. To measure mRNA stability, cells were pretreated with GnRH agonist (GnRHa; D-Ala6) (Sigma) for 4 h before transcription was blocked with actinomycin D (Calbiochem, San Diego, CA, USA). Constructs and transfections To express cDNA constructs, cells were transiently transfected with 1.5 mg of plasmid DNA per well in six-well plates for 24 h using Lipofectamine (Invitrogen). The myc-tagged dominantnegative Rac1 (N17) was a gift from Dr A Hall (SloanKettering Institute, New York, NY, USA) (Olson et al., 1995), and dominant-negative Cdc42 (N17) and RhoA (N19) (Olson et al., 1995) were obtained from UMR cDNA Resource Center (Rolla, MO, USA). The IL2R/E-cadherin membrane proximal chimera construct was provided by Dr C Gottardi (Northwestern University, Chicago, IL, USA) (Miranda et al., 2001). The cDNA for human P-cadherin in the vector pcDNA3.1, amplified by PCR using primers, sense, 50 -GGATCCATGGG GCTCCCTCGTGGACCTCT-30 and antisense, 50 -GAATTC CTAGTCGTCCTCCCCGCCACCGTA-30 , was transfected using Lipofectamine (Invitrogen). Cells transfected with an empty vector serve as a control. The transfected clones were selected with 400 mg/ml of G418, and subsequent subcloning was performed by limiting dilution. siRNA transfection siRNA duplex specific to GnRH receptor (50 -CAAGAACAA UAUACCAAGA-30 ), P-cadherin (#1, 50 -GAGGGUGUCUU CGCUGUAG-30 ; #2, 50 -UAUCAGUGCUAAACAGAGCU30 ), or p120ctn (#1, 50 -GGACCUUACUGAAGUUAUU-30 ; #2, 50 -UAGCUGACCUCCUGACUAA-30 ) were purchased from Dharmacon (Lafayette, CO, USA). Nontargeting siRNA duplex (Dharmacon) was used as nonspecific control. Transfection was performed using siLentFect (Bio-Rad, Hercules, CA, USA). Reverse-transcription PCR Total RNA was isolated using TRIzol reagent and cDNA was synthesized using SuperScript Reverse Transcriptase (Invitrogen). Real-time PCR was performed on an ABI Prism 7300 Sequence Detector System using the SYBR green PCR master mix. The primers used were E-cadherin, sense, 50 -ACAGCC CCGCCTTATGATT-30 and antisense, 50 -TCGGAACCGCT Oncogene

TCCTTCA-30 ; N-cadherin, sense, 50 -TGGGAATCCGACGA ATGG-30 and antisense, 50 -GCAGATCGGACCGGATAC TG-30 ; and P-cadherin, sense, 50 -AGTGGAGGACCCCATG AACA-30 and antisense, 50 -TTGGGCTTGTGGTCATTCT G-30 . PCR specificity was verified by melting curve analysis and gel electrophoresis. The mean Ct of E-, N-, and P-cadherin was calculated and normalized with that of the internal control glyceraldehyde 3-phosphate dehydrogenase. Luciferase reporter assay A 531 bp fragment of the P-cadherin promoter was inserted into the pGL3-basic plasmid upstream of the luciferase reporter gene. Cells were transiently transfected with 1 mg of the P-cadherin reporter plasmid. Twenty-four hours post transfection, luciferase reporter activity was measured by using the Luciferase assay system (Promega, Madison, WI, USA). b-galactosidase (pSV-b-gal) was used to normalize data. Western blot and cell fractionation Equal amounts of protein (40 mg) were subjected to SDS–PAGE and then transferred into nitrocellulose membrane. Antibodies specific for E-cadherin (1:2000), N-cadherin (1:1000), P-cadherin (1:1000), a-catenin (1:1000), b-catenin (1:2000), p120ctn (1:2000) (Transduction Laboratories, San Diego, CA, USA), myc (1:1000) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and b-actin (1:2000) (Sigma) were used. Bound proteins were detected by using appropriate horseradish peroxidase-conjugated secondary antibodies (Bio-Rad) in the enhanced chemiluminescence (Amersham Biosciences, Little Chalfont, UK). For cell fractionation analysis, cells were fractionated into membrane, cytoplasmic, or nuclear fractions using the Qproteome Cell Compartment kit (Qiagen, Chatsworth, CA, USA). The purity of subcellular fractions was confirmed by using antibodies against N-cadherin (a membrane marker) (Transduction Laboratories), a-tubulin (a cytoplasmic marker) (Lab Vision, Fremont, CA, USA), and histone H1 (a nuclear marker) (Upstate Biotechnology, Lake Placid, NY, USA). Immunofluorescence microscopy Cells were cultured on glass coverslips until subconfluence. Coverslips were fixed in 4% paraformaldehyde, permeabilized with 0.05% Triton X-100, and blocked with 5% bovine serum albumin. The coverslips were then incubated with antiP-cadherin (1:100) or anti-p120ctn (1:200) at room temperature for 1 h followed by Texas red-conjugated secondary antibodies for 1 h. For negative controls, nonspecific IgG was used instead of primary antibody. The cells were mounted in aqueous antifade medium (Vector laboratories, Burlingame, CA, USA) and analyzed with a Nikon Eclipse E600 fluorescence microscope. Rho GTPases activity assay Rho GTPases activation was determined using Rac1/Cdc42 and RhoA activation assay kits (Upstate). Briefly, cell lysates were immunoprecipitated with a glutathione S-transferase fusion protein corresponding to the binding domain of human PAK (for Rac1 and Cdc42) or rhotekin (for RhoA) bound to glutathione-sepharose, run on SDS–PAGE, and western blotted using anti-Rac1, anti-Cdc42, and anti-RhoA antibodies (1:1000). Whole-cell lysates were also blotted to measure total Rac1, Cdc42, and RhoA. Transwell migration and invasion assay Cells (1  105) were seeded into the upper chamber of a Transwell insert (BD Biosciences, Palo Alto, CA, USA) in

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2439 serum-free medium for 24 h, and the ability of cells to migrate through filters was assayed as reported earlier by us (Cheung et al., 2006). To assess cellular invasion, the same protocol was used, except that inserts were precoated with Matrigel. Results were presented as the mean cell number of five fields (±s.d.) of triplicate wells from three independent experiments.

Conflict of interest The authors declare no conflict of interest.

Acknowledgements Statistical analysis All experiments were repeated at least three times, and data were presented as mean±s.d. Statistical analysis was carried out using ANOVA followed by Tukey’s post hoc test (GraphPad Software, San Diego, CA, USA). Po0.05 was considered significant.

We thank Drs N Auersperg, A Hall, and C Gottardi for cell lines and plasmids. This work was supported by Canadian Institutes of Health Research grant (PCK Leung), and by Hong Kong Research Grant Council Grant 778108 and HKU Outstanding Young Researcher Award (AST Wong).

References Cavallaro U, Christofori G. (2004). Multitasking in tumor progression: signaling functions of cell adhesion molecules. Ann N Y Acad Sci 1014: 58–66. Chen CL, Cheung LW, Lau MT, Choi JH, Auersperg N, Wang HS et al. (2007). Differential role of gonadotropin-releasing hormone on human ovarian epithelial cancer cell invasion. Endocrine 31: 311–320. Chen GT, Tai CT, Yeh LS, Yang TC, Tsai HD. (2002). Identification of the cadherin subtypes present in the human peritoneum and endometriotic lesions: potential role for P-cadherin in the development of endometriosis. Mol Reprod Dev 62: 289–294. Cheng CK, Leung PC. (2005). Molecular biology of gonadotropinreleasing hormone (GnRH)-I, GnRH-II, and their receptors in humans. Endocr Rev 26: 283–306. Cheung LW, Leung PC, Wong AS. (2006). Gonadotropin-releasing hormone promotes ovarian cancer cell invasiveness through c-Jun NH2-terminal kinase-mediated activation of matrix metalloproteinase (MMP)-2 and MMP-9. Cancer Res 66: 10902–10910. Cheung LW, Wong AS. (2008). Gonadotropin-releasing hormone: GnRH receptor signaling in extrapituitary tissues. FEBS J 275: 5479–5495. Choi JH, Park JT, Davidson B, Morin PJ, Shih Ie M, Wang TL. (2008). Jagged-1 and Notch3 juxtacrine loop regulates ovarian tumor growth and adhesion. Cancer Res 68: 5716–5723. Clevers H. (2006). Wnt/beta-catenin signaling in development and disease. Cell 127: 469–480. Emons G, Pahwa GS, Brack C, Sturm R, Oberheuser F, Knuppen R. (1989). Gonadotropin releasing hormone binding sites in human epithelial ovarian carcinomata. Eur J Cancer Clin Oncol 25: 215–221. Fagotto F, Gumbiner BM. (1996). Cell contact-dependent signaling. Dev Biol 180: 445–454. Feltes CM, Kudo A, Blaschuk O, Byers SW. (2002). An alternatively spliced cadherin-11 enhances human breast cancer cell invasion. Cancer Res 62: 6688–6697. Irmer G, Burger C, Muller R, Ortmann O, Peter U, Kakar SS et al. (1995). Expression of the messenger RNAs for luteinizing hormonereleasing hormone (LHRH) and its receptor in human ovarian epithelial carcinoma. Cancer Res 55: 817–822. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T et al. (2008). Cancer statistics, 2008. CA Cancer J Clin 58: 71–96. Kang SK, Cheng KW, Nathwani PS, Choi KC, Leung PC. (2000). Autocrine role of gonadotropin-releasing hormone and its receptor in ovarian cancer cell growth. Endocrine 13: 297–304. Kim KY, Choi KC, Auersperg N, Leung PC. (2006). Mechanism of gonadotropin-releasing hormone (GnRH)-I and -II-induced cell growth inhibition in ovarian cancer cells: role of the GnRH-I receptor and protein kinase C pathway. Endocr Relat Cancer 13: 211–220. Li SL, Vuagnat B, Gruaz NM, Eshkol A, Sizonenko PC, Aubert ML. (1994). Binding kinetics of the long-acting gonadotropin-releasing

hormone (GnRH) antagonist antide to rat pituitary GnRH receptors. Endocrinology 135: 45–52. Maeda M, Johnson E, Mandal SH, Lawson KR, Keim SA, Svoboda RA et al. (2006). Expression of inappropriate cadherins by epithelial tumor cells promotes endocytosis and degradation of E-cadherin via competition for p120(ctn). Oncogene 25: 4595–4604. Miranda KC, Khromykh T, Christy P, Le TL, Gottardi CJ, Yap AS et al. (2001). A dileucine motif targets E-cadherin to the basolateral cell surface in Madin-Darby canine kidney and LLC-PK1 epithelial cells. J Biol Chem 276: 22565–22572. Nieman MT, Prudoff RS, Johnson KR, Wheelock MJ. (1999). Ncadherin promotes motility in human breast cancer cells regardless of their E-cadherin expression. J Cell Biol 147: 631–644. Nobes CD, Hall A. (1999). Rho GTPases control polarity, protrusion, and adhesion during cell movement. J Cell Biol 144: 1235–1244. Noren NK, Liu BP, Burridge K, Kreft B. (2000). p120 catenin regulates the actin cytoskeleton via Rho family GTPases. J Cell Biol 150: 567–580. Olson MF, Ashworth A, Hall A. (1995). An essential role for Rho, Rac, and Cdc42 GTPases in cell cycle progression through G1. Science 269: 1270–1272. Paredes J, Correia AL, Ribeiro AS, Albergaria A, Milanezi F, Schmitt FC. (2007). P-cadherin expression in breast cancer: a review. Breast Cancer Res 9: 214. Patel IS, Madan P, Getsios S, Bertrand MA, MacCalman CD. (2003). Cadherin switching in ovarian cancer progression. Int J Cancer 106: 172–177. Pece S, Gutkind JS. (2000). Signaling from E-cadherins to the MAPK pathway by the recruitment and activation of epidermal growth factor receptors upon cell-cell contact formation. J Biol Chem 275: 41227–41233. Perez-Moreno M, Fuchs E. (2006). Catenins: keeping cells from getting their signals crossed. Dev Cell 11: 601–612. Qian X, Karpova T, Sheppard AM, McNally J, Lowy DR. (2004). E-cadherin-mediated adhesion inhibits ligand-dependent activation of diverse receptor tyrosine kinases. EMBO J 23: 1739–1748. Sommers CL, Gelmann EP, Kemler R, Cowin P, Byers SW. (1994). Alterations in beta-catenin phosphorylation and plakoglobin expression in human breast cancer cells. Cancer Res 54: 3544–3552. Soubry A, van Hengel J, Parthoens E, Colpaert C, Van Marck E, Waltregny D et al. (2005). Expression and nuclear location of the transcriptional repressor Kaiso is regulated by the tumor microenvironment. Cancer Res 65: 2224–2233. Spentzos D, Cannistra SA, Grall F, Levine DA, Pillay K, Libermann TA et al. (2007). IGF axis gene expression patterns are prognostic of survival in epithelial ovarian cancer. Endocr Relat Cancer 14: 781–790. Sundfeldt K. (2003). Cell-cell adhesion in the normal ovary and ovarian tumors of epithelial origin; an exception to the rule. Mol Cell Endocrinol 202: 89–96. Oncogene

GnRH promotes cadherin switching in ovarian cancer LWT Cheung et al

2440 Suyama K, Shapiro I, Guttman M, Hazan RB. (2002). A signaling pathway leading to metastasis is controlled by N-cadherin and the FGF receptor. Cancer Cell 2: 301–314. Taniuchi K, Nakagawa H, Hosokawa M, Nakamura T, Eguchi H, Ohigashi H et al. (2005). Overexpressed P-cadherin/CDH3 promotes motility of pancreatic cancer cells by interacting with p120ctn and activating rho-family GTPases. Cancer Res 65: 3092–3099.

Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. (2008). Cadherin switching. J Cell Sci 121: 727–735. Wong AS, Maines-Bandiera SL, Rosen B, Wheelock MJ, Johnson KR, Leung PC et al. (1999). Constitutive and conditional cadherin expression in cultured human ovarian surface epithelium: influence of family history of ovarian cancer. Int J Cancer 81: 180–188.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene