14-3-3ε Overexpression Contributes to Epithelial-Mesenchymal Transition of Hepatocellular Carcinoma

Background 14-3-3ε is implicated in regulating tumor progression, including hepatocellular carcinoma (HCC). Our earlier study indicated that elevated 14-3-3ε expression is significantly associated with higher risk of metastasis and lower survival rates of HCC patients. However, the molecular mechanisms of how 14-3-3ε regulates HCC tumor metastasis are still unclear. Methodology and Principal Findings In this study, we show that increased 14-3-3ε expression induces HCC cell migration and promotes epithelial-mesenchymal transition (EMT), which is determined by the reduction of E-cadherin expression and induction of N-cadherin and vimentin expression. Knockdown with specific siRNA abolished 14-3-3ε-induced cell migration and EMT. Furthermore, 14-3-3ε selectively induced Zeb-1 and Snail expression, and 14-3-3ε-induced cell migration was abrogated by Zeb-1 or Snail siRNA. In addition, the effect of 14-3-3ε-reduced E-cadherin was specifically restored by Zeb-1 siRNA. Positive 14-3-3ε expression was significantly correlated with negative E-cadherin expression, as determined by immunohistochemistry analysis in HCC tumors. Analysis of 14-3-3ε/E-cadherin expression associated with clinicopathological characteristics revealed that the combination of positive 14-3-3ε and negative E-cadherin expression is significantly correlated with higher incidence of HCC metastasis and poor 5-year overall survival. In contrast, patients with positive 14-3-3ε and positive E-cadherin expression had better prognostic outcomes than did those with negative E-cadherin expression. Significance Our findings show for the first time that E-cadherin is one of the downstream targets of 14-3-3ε in modulating HCC tumor progression. Thus, 14-3-3ε may act as an important regulator in modulating tumor metastasis by promoting EMT as well as cell migration, and it may serve as a novel prognostic biomarker or therapeutic target for HCC.


Introduction
Epithelial-mesenchymal transition (EMT), a developmental process by which epithelial cells reduce cell-cell adhesion and lose apical-basal cell polarity, plays a critical role in the embryogenesis and conversion of early stage tumors into aggressive malignancies [1][2][3]. EMT promotes multiple physiological processes that increase the invasiveness and metastasis potentials of human tumors [1][2][3]. EMT is typically characterized by loss of Ecadherin, gain of N-cadherin and vimentin, and translocation of bcatenin from membrane to the nuclear compartment [4,5]. The impairment of E-cadherin is a hallmark of EMT, and E-cadherin expression is often inversely correlated with tumor malignancy and patient survival [4,5]. The E-cadherin expression level is downregulated by gene silencing with CpG methylation on promoter in hepatocellular carcinoma (HCC) [6,7] and may be associated with tumor grade and poor prognosis of HCC [8]. Several transcrip-tional regulators that act as E-cadherin repressors are mediated by recognizing the E-box motif on the E-cadherin promoter region [9][10][11]. Factors of Snail zinc finger, Zeb and bHLH families are known to suppress E-cadherin, thereby promoting the EMT process and tumor metastasis [9][10][11]. In addition, increased Zeb-1, Snail, SIP1, and Twist expressions are reportedly associated with the clinicopathological significances of HCC malignant progression, including cancer invasion and poor patient survival [12][13][14][15][16].

Transient Transfection
Huh-7 and HepG2 cells were transiently transfected with control and 14-3-3e by use of Polyjet TM transfection reagent (Signa-Gen Laboratories, Ijamsville, MD). Cells were transfected with control or 0.5 to 1.5 mg of 14-3-3e vectors per 6-well plate followed by incubation with Polyjet TM /DNA complex-containing medium and replaced with complete medium for 24 hours. Transfected cells were incubated for additional 24 hours before performing cell migration assays or protein expression analysis.

Immunofluorescence Staining
Immunofluorescence staining was performed as described previously [25]. Briefly, 14-3-3e and control cells were fixed with 2% paraformaldehyde for 15 minutes at 4uC. After washing, cells were permeabilized with 0.1% Triton X-100 in PBS for 5 minutes and blocked with PBS containing 10% FBS at room temperature for 1 hour. For the immunofluorescence staining, cells were incubated with the primary antibodies of anti-E-cadherin and anti-N-cadherin (BD Biosciences, San Jose, CA), and anti-vimentin (Millipore, Temecula, CA) in PBS containing 1% FBS at 4uC overnight, followed by incubation with Alexa FluorH 488 secondary antibody (Invitrogen, Grand Island, NY) in PBS containing 5% bovine serum albumin at room temperature for 2 hours. Samples were mounted and images were analyzed by use of the Leica TCS SP5 Confocal Imaging System (Leica, Germany).

Migration Assay
Bio-coat cell migration Boyden chambers were used for cell migration assay (Becton Dickinson, Pont-de-Claix, France). Briefly, cells were trypsinized and suspended in 0.1% BSA-DMEM and cells (1610 4 for SK-Hep1, 6610 4 for Huh-7 and 2610 5 for HepG2) were added to the upper wells with 8-mm pores. Cells were allowed to migrate toward the bottom wells containing 100 mg/ml fibronectin (Becton Dickinson, Pont-de-Claix, France), epithelial growth factor (EGF, 20 ng/ml, Sigma-Aldrich, St. Louis, MO) and 10% BSA-DMEM for 20 hours. Cells remaining on the upper side were removed, and migrated cells on the bottom side were fixed and stained with 0.1% crystal violet containing 20% ethanol and 1% formadehyde for 20 minutes. Cell migration was quantified by counting the total number of migrated cells.

Quantitative Real-time PCR
As described previously [25,36], total RNA was extracted by use of the RNAspin Mini Kit (GE Healthcare, Freiburg, Germany). cDNA was synthesized from 2-5 mg RNA by use of the oligo(dT) 18 primers and RevertAid TM First Strand cDNA Synthesis Kit (Fermentas, Thermo Fisher Scientific, Waltham, MA). Quantitative real-time PCR involved use of SYBR Green (Kapa biosystem, Woburn, MA) with specific oligonucleotide primers (Table S2) from the AB 7900HT system (Applied Biosystems, USA). Applied Biosystems Relative Quantification (RQ) Manager Software version 1.2 was used to analyze the relative gene expression in each sample by the comparative Ct method. Gene expression was normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

Clinical Specimens
Tissue samples were obtained from 113 HCC patients who had undergone surgery for tumor resection or biopsy at Taichung Veterans General Hospital from January 1999 to December 2001. The mean follow-up time was 51.5628.7 months. Thirty-three patients (29.2%) developed tissue-proved metastasis in 3 to 87 months after the resection of primary HCC. Slides from paraffinembedded surgical specimens of primary tumors with surrounding non-cancerous liver parenchyma were subjected to immunohistochemical (IHC) staining. The pathological features, IHC staining results, clinical parameters, including Barcelona-Clinic Liver Cancer (BCLC) staging [37], and disease outcomes were collected for analysis. This study was approved by the Institutional Review Board of Taichung Veterans General Hospital. The policy that no informed consents are required for using these de-linked samples for retrospective analysis was also approved by the Institutional Review Board.

Statistical Analysis
The Student's t-test was used to analyze differences between 2 groups. Kaplan-Meier curves were plotted and the log rank test was used to analyze time-related variables of probabilities for metastasis and overall survival. A P value ,0.05 was considered statistically significant.

14-3-3e Promotes Epithelial-mesenchymal Transition of HCC
To investigate whether 14-3-3e expression regulates EMT of HCC cells, we determined the expression of EMT markers, Ecadherin, N-cadherin and vimentin, by Western blotting analysis. We found that 14-3-3e overexpression significantly reduced Ecadherin expression, but it induced N-cadherin and vimentin expression (Figure 2A). The expression levels and subcellular localizations of E-cadherin, N-cadherin, and vimentin were further examined by immunofluorescent confocal microscopy. E-cadherin expression was detected at the cell-cell contacts in control cells, while it was dramatically reduced in 14-3-3e overexpression cells ( Figure 2B, left panel). Slight expression of N-cadherin and vimentin was detected in control cells, but such expression was significantly induced by 14-3-3e overexpression ( Figure 2B, middle and right panels). Furthermore, reduction of E-cadherin expression and the induction of N-cadherin and vimentin expression in 14-3-3e overexpression cells were abrogated by transfection with 14-3-3e siRNA, as determined by Western blotting analysis ( Figure 2C). The regulation of these expressions of E-cadherin, N-cadherin and vimentin by 14-3-3e knockdown were further confirmed by confocal microscopy. 14-3-3e siRNA restored Ecadherin expression, which localized at the cell junctions ( Figure 2D). In addition, knockdown with siRNA suppressed the N-cadherin and vimentin expression induced by 14-3-3e ( Figure 2D). These results indicate that 14-3-3e overexpression promotes EMT of HCC.

14-3-3e Promotes HCC Cell Migration via Upregulation of Zeb-1 and Snail
To understand the molecular regulation of how 14-3-3e induces EMT and reduces E-cadherin expression in HCC, we examined the expression levels of distinct E-box transcriptional suppressors. We found that 14-3-3e overexpression selectively induced Zeb-1 and Snail expression but had no significant effect on Zeb-2, Twist or Slug ( Figure 3A). The induced expression of Zeb-1 and Snail by 14-3-3e was further confirmed by quantitative real-time PCR analysis ( Figure 3B). In addition, results of transient transfection indicated that overexpression of 14-3-3e dose-dependently induced Zeb-1/Snail and reduced E-cadherin expression in Huh-7 ( Figure 3C, left panel) and HepG2 cells ( Figure 3C, right panel).  Furthermore, 14-3-3e-induced expression of Zeb-1 and Snail was abrogated by knockdown of 14-3-3e with siRNA ( Figure 3D). These findings suggest that Zeb-1 and Snail may be involved in 14-3-3e-induced HCC cell migration and EMT. We next determined the role of 14-3-3e-induced Zeb-1 and Snail on cell migration. We found that knockdown of either Zeb-1 or Snail expression by siRNA significantly abolished 14-3-3e induced cell migration ( Figure 3E). These results indicate that Zeb-1 and Snail play important roles in regulating 14-3-3e-induced HCC cell migration.

14-3-3e Suppresses E-cadherin Expression Selectively Mediated by Induction of Zeb-1
To further explore the role of Zeb-1 and Snail on 14-3-3einduced cell migration or EMT, we knockdown Zeb-1 and Snail with siRNAs and examined E-cadherin expression by Western blotting analysis. Interestingly, 14-3-3e-reduced E-cadherin expression was specifically restored by Zeb-1 siRNA, but not by Snail siRNA (Figure 4A). This specific effect of Zeb-1 on regulating 14-3-3e-reduced E-cadherin was validated by quantitative real-time PCR ( Figure 4B). In addition, the selective effect of Zeb-1 knockdown restoring E-cadherin expression in 14-3-3e overexpression cells was further confirmed by confocal microscopy ( Figure 4C). SK-Hep1 cells expressed higher levels of 14-3-3e and lower levels of E-cadherin than other HCC cell lines. We next transfected SK-Hep1 cells with 14-3-3e siRNA and determined the expression of Zeb-1, Snail and E-cadherin by Western blotting analysis. Knockdown of 14-3-3e reduced the expression of Zeb-1/ Snail and induced that of E-cadherin in SK-Hep1 cells with a concentration-dependent manner ( Figure 4D). These results demonstrate that the reduction of E-cadherin expression by 14-3-3e is selectively mediated by regulation of Zeb-1. Thus, Zeb-1/Ecadherin expression is a downstream factor of 14-3-3e for promoting EMT in HCC.

Correlation and Impact of Positive 14-3-3e with Negative E-cadherin Expression in HCC
To further support the likelihood that 14-3-3e suppresses Ecadherin and regulates EMT as well as tumor progression, we examined the expression of 14-3-3e and E-cadherin by immunohistochemical analysis in HCC tumors. Expression of 14-3-3e was higher in HCC primary tumors than in the surrounding noncancerous liver tissues ( Figure 5A, left panel). We next determined the expression of E-cadherin and found that E-cadherin expression was reduced in HCC tumors ( Figure 5A, right panel). Positive 14-3-3e expression was significantly correlated with negative Ecadherin in HCC tumors (p = 0.043) ( Figure 5B). In addition, expression of 14-3-3e was correlated with Zeb-1 in HCC tumors ( Figure S1).

Association of 14-3-3e/E-cadherin Expression with Extrahepatic Metastasis and Patient Survival of HCC
We have previously shown that 14-3-3e overexpression in HCC primary tumors was significantly associated with subsequent extrahepatic metastasis and reduced 5-year overall survival [30]. To evaluate whether E-cadherin plays an important role as a downstream effector of 14-3-3e in promoting tumor progression, the associations of E-cadherin with clinicopathological characteristics and with 14-3-3e expression were compared. In addition to 14-3-3e positivity, expression of E-cadherin is significantly correlated with gender (p = 0.011), histology grade (p = 0.001), BCLC staging (p = 0.030), tumor size (p = 0.003), and subsequent extrahepatic metastasis (p = 0.004) (Table S3). Patients with positive E-cadherin expression exhibit a lower risk of metastasis ( Figure 6A, p = 0.013) and better overall survival rate ( Figure 6B, p = 0.047) than do those with negative E-cadherin expression in 14-3-3e positive HCC tumors. These results provide clinical evidence to support the hypothesis that E-cadherin is one of the crucial downstream regulators of 14-3-3e that modulate HCC tumor progression.

Discussion
We previously demonstrated that 14-3-3e expression is increased in primary and metastatic HCC. Elevated 14-3-3e expression is correlated with higher risk of extrahepatic metastasis and lower survival rates of HCC patients [30]. In this study, we investigated the molecular mechanism to determine how 14-3-3e regulates tumor progression. Attenuated expression of E-cadherin has been recognized as an important determinant and biomarker of tumor progression, one especially indicative of EMT in various tumors. In addition, gene silencing and loss of E-cadherin expression in the malignant progression of HCC have been demonstrated [6,7], and it is suggested that E-cadherin is associated with reduced survival of HCC patients [8]. Our current investigation indicates that 14-3-3e promotes HCC EMT and cell migration and also suppresses E-cadherin expression via upregulation of Zeb-1. We found that the expression of Zeb-1 was increased (14 of 113) in HCC primary tissues ( Figure S1), although the increase is not as significant as in a previous report [14]. This difference may due to the sensitivity of reagents, sample size or differences in the cohort. It has been suggested that TGF-b and downstream signals of Smad2/3 activation regulate Zeb expression and EMT [41]. Although we performed the experiments of Smad2 knockdown in 14-3-3e overexpression cells, we did not observe significant restoration of E-cadherin ( Figure S2). In addition, increased 14-3-3f expression has been shown to promote EMT via associating with TGF-b receptor signaling and PI3-K subunit p85 in breast cancer cells [42,43]. However, treating cells with TGF-b receptor or PI3-K inhibitors (SB-431542 or LY-294002) did not abolish E-cadherin-suppression induced by 14-3-3e ( Figure S3). These results suggest the effect of 14-3-3esuppressed E-cadherin expression may not be regulated through TGF-b/Smad2/3 or PI3-K signal pathways. Thus, 14-3-3e contributes to EMT via induction of Zeb-1 may be mediated by a novel mechanism. Further work is currently ongoing to investigate how 14-3-3e regulates Zeb-1 expression.
Expression of transcriptional repressors for E-cadherin, including Zeb-1, Snail, SIP1 and Twist, is associated with cell invasion, EMT, metastasis and poor patient survival of HCC [12-   ]. However, no previous studies have shown that 14-3-3e promotes HCC tumor progression via modulating E-cadherin transcriptional repressors. Our study shows for the first time that 14-3-3e induces Zeb-1 expression, thereby repressing E-cadherin expression and promoting EMT. The 14-3-3e regulation of Ecadherin reduction occurs through Zeb-1, and not through Snail or other E-cadherin repressors, as supported by Figure 3A. To further clarify the regulation of 14-3-3e-reduced E-cadherin expression by Zeb-1, 14-3-3e overexpression cells were transfected with Zeb-1 siRNA or control scramble siRNA, and the gene expression profile was analyzed by use of microarray analysis. Altered gene expression (fold change .2) was identified of 557 transcripts in 14-3-3e overexpression vs. the control cells and 160 transcripts in Zeb-1 siRNA vs. scramble siRNA cells. Among them, CDH1 (E-cadherin), SMAD2, and PLA2G2A were regulated in 14-3-3e overexpression cells but had a reversed expression pattern in Zeb-1 knockdown cells (Data not shown). These results provide additional evidence to support our findings.
In addition to Zeb-1, our results indicated that 14-3-3e induces Snail expression and promotes HCC cell migration ( Figure 3A and 3E). However, knockdown of Snail did not restore 14-3-3ereduced E-cadherin expression ( Figure 4A and 4B). Interestingly, partially increased of Snail expression was found by treatment with Zeb-1 siRNA ( Figure 4A). As Snail and Zeb-1 regulate EMT of HCC may be mediated by separate and complicated pathways, a compensative effect is possibly involved. Further investigation is needed to elucidate this finding. Additionally, our results indicated that 14-3-3e overexpression-induced EMT (increase of Ncadherin, Vimentin, Zeb-1 and Snail as well as decrease of Ecadherin expression) was impaired by 14-3-3e siRNA ( Figure 2C and Figure 3D). However, knockdown of 14-3-3e has no significant effect on affecting EMT markers in control cells ( Figure 2C and Figure 3D). We therefore postulate that other endogenous house-keeping regulators may be involved in maintaining basal level of Snail/Zeb-1 expression. Endogenous level of Snail/Zeb-1 modulates expression of EMT markers which is independent of 14-3-3e expression in HCC. Moreover, 14-3-3e may upregulate FAK expression via activation of NFkB to enhance HCC cell migration [35]. These results reveal the complicated signal mechanisms that are involved in 14-3-3e induced HCC cell migration, EMT, and metastasis. Uncovering the complex role of 14-3-3e in tumor progression could contribute to the development of therapeutic strategies for treatment of aggressive and advanced HCC.