EGFR axis contributes to the progression of cholangiocarcinoma through the induction of an epithelial-mesenchymal transition

Accepted Manuscript EGF/EGFR axis contributes to the progression of cholangiocarcinoma through the induction of an epithelial-mesenchymal transition Audrey Clapéron, Martine Mergey, Thanh Huong Nguyen Ho-Bouldoires, Danijela Vignjevic, Dominique Wendum, Yves Chrétien, Fatiha Merabtene, Alexandra Frazao, Valérie Paradis, Chantal Housset, Nathalie Guedj, Laura Fouassier PII: DOI: Reference:

S0168-8278(14)00214-1 http://dx.doi.org/10.1016/j.jhep.2014.03.033 JHEPAT 5098

To appear in:

Journal of Hepatology

Received Date: Revised Date: Accepted Date:

15 October 2013 12 March 2014 24 March 2014

Please cite this article as: Clapéron, A., Mergey, M., Ho-Bouldoires, T.H.N., Vignjevic, D., Wendum, D., Chrétien, Y., Merabtene, F., Frazao, A., Paradis, V., Housset, C., Guedj, N., Fouassier, L., EGF/EGFR axis contributes to the progression of cholangiocarcinoma through the induction of an epithelial-mesenchymal transition, Journal of Hepatology (2014), doi: http://dx.doi.org/10.1016/j.jhep.2014.03.033

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 EGF/EGFR axis contributes to the progression of cholangiocarcinoma through the induction of an epithelial-mesenchymal transition Audrey Clapéron1,2, Martine Mergey1,2, Thanh Huong Nguyen Ho-Bouldoires1,2, Danijela Vignjevic3, Dominique Wendum1,2,4, Yves Chrétien1,2, Fatiha Merabtene5, Alexandra Frazao1,2, Valérie Paradis6, Chantal Housset1,2, Nathalie Guedj6 and Laura Fouassier1,2 1

INSERM, UMR_S 938, Centre de Recherche Saint-Antoine, F-75012 Paris, France;

2

UPMC, Univ Paris 06, UMR_S 938, Centre de Recherche Saint-Antoine, F-75012 Paris,

France; 3

CNRS UMR 144, Institut Curie, F-75005 Paris, France;

4

AP-HP, Hôpital Saint-Antoine, Service d’Anatomie et Cytologie Pathologiques, F-75012

Paris, France; 5

INSERM, UMR_S938, Centre de Recherche Saint-Antoine, Plateforme Morphologie du

Petit Animal, F-75012 Paris, France; 6

INSERM, UMRS_U773 & AP-HP, Hôpital Beaujon, Service de Pathologie, F-92100

Clichy, France Corresponding author Laura Fouassier, Ph.D. INSERM UMR_S 938, Centre de Recherche Saint-Antoine Faculté de Médecine Pierre et Marie Curie, site Saint-Antoine 27 rue Chaligny 75571 Paris cedex 12, France 33 6 98 77 40 01 (phone) 33 1 40 01 13 52 (fax) [email protected]

2

4900 words 4 figures 1 table List of abbreviations EMT, epithelial-mesenchymal transition, CCA, cholangiocarcinoma; EGFR, epidermal growth factor receptor; EGF, epidermal growth factor; AJs, adherens junctions; IHC, immunohistochemistry; HPS, hematoxylin phloxine safran; SEM, standard error of the mean. Conflict of interest The authors declare no conflict of interest. Financial support This work was supported by the Fondation de France (to A.C.), GEFLUC (to L.F.) and Fonds CSP (to L.F. and C.H.).

3

ABSTRACT Background & aims. Epithelial-mesenchymal transition (EMT) is a cellular process involved in cancer progression. The first step of EMT consists in the disruption of E-cadherin-mediated adherens junctions. Cholangiocarcinoma (CCA), a cancer with a poor prognosis due to local invasion and metastasis, displays EMT features. EGFR, a receptor tyrosine kinase, plays a major role in CCA progression. The aim of the study was to determine if EMT is induced by EGFR in CCA cells. Methods. In vivo, the expression of E-cadherin was analyzed in CCA tumors of 100 patients and correlated with pathological features and EGFR expression, and in a xenograft model in mice treated with gefitinib, an inhibitor of EGFR. In vitro, the regulation of EMT by EGFR was investigated in CCA cell lines. Results. In human CCA, a cytoplasmic localization of E-cadherin occurring in 50% of the tumors and was associated with the peripheral type of CCA, tumor size, the presence of satellite nodules and EGFR overexpression. In xenografted tumors, E-cadherin displayed a cytoplasmic pattern whereas the treatment of mice with gefitinib restored the membranous expression of E-cadherin. In vitro, EGF induced scattering of CCA cells that resulted from the disruption of adherens junctions. Internalization and decreased expression of E-cadherin, as well as nuclear translocation of β-catenin, were observed in EGF-treated CCA cells. In these cells, EMT-transcription factors (i.e. Slug and Zeb-1) and mesenchymal markers (i.e. Ncadherin and α-SMA) were induced, favoring cell invasiveness through cytoskeleton remodeling. All these effects were inhibited by gefitinib. Conclusions. The EGF/EGFR axis triggers EMT in CCA cells highlighting the key role of this pathway in CCA progression.

4

INTRODUCTION Epithelial to mesenchymal transition (EMT) is a process by which epithelial cancer cells acquire mesenchymal traits and undergo deep cytoskeleton remodeling leading to cell scattering and metastatic potential. The first step of EMT is the disruption of adherens junctions (AJs) due to internalization and/or down-regulation of E-cadherin, a major component of AJs [1, 2]. During the acquisition of mesenchymal features, E-cadherin is transcriptionally regulated by EMT-transcriptional factors such as Snail, Slug, Twist and ZEB1/2 [3, 4]. Cholangiocarcinoma (CCA), the second most common primary liver cancer after hepatocellular carcinoma, displays rapid progression and poor outcome [5]. Several studies have demonstrated a correlation between the presence of EMT features and pathological parameters suggesting a role of EMT during CCA progression and metastasis [6-11]. Loss of epithelial markers (i.e. E-cadherin) and acquisition of mesenchymal markers (i.e. Slug, S100A4, vimentin or N-cadherin) have been associated with aggressive tumor behavior, i.e. lymph node metastasis, venous and neural invasion, intrahepatic metastasis, advanced tumor stage, undifferentiated phenotype and poor outcome [6-8, 10, 12]. Epidermal growth factor receptor (EGFR) belongs to the HER/ErbB family of tyrosine kinase receptors. In different types of cancer, EGFR activation disturbs cell-cell adhesion by destabilizing E-cadherin/β-catenin complex, promotes EMT and contributes to the acquisition of a motile phenotype [13-15]. In CCA, an overexpression of EGFR has been associated with tumor progression [16-22]. However, whether EMT features observed in CCA depend on the EGF/EGFR axis activation has not been established. In the present study, we show that cytosolic expression of E-cadherin is positively correlated with CCA type, tumor size and the presence of satellite nodules in human CCA samples. In

5 addition, cytoplasmic localization of E-cadherin is associated with EGFR expression in human CCA. In xenografted nude mice, inhibition of EGFR by a specific inhibitor prevents the expression of E-cadherin in the cytoplasm of tumor cells. In vitro, EGFR activation by EGF, a ligand expressed in human CCA by tumor cells, induces an EMT program that enables CCA cells to display a significant reorganization of cytoskeleton and develop invasive properties. Overall, these findings indicate that the EGF/EGFR axis contributes to EMT in CCA tumor cells, favoring the progression of this type of cancer.

6

MATERIALS AND METHODS Cell culture and treatments Human CCA cell lines (Mz-ChA-1 and SK-ChA-1) were provided by Dr. A. Knuth (Zurich University, Switzerland). Both cell lines were cultured in DMEM supplemented with 1 g/L glucose, 10 mmol/L HEPES and 10% fetal bovine serum (Invitrogen, Carlsbad, CA). MzChA-1 and SK-ChA-1 cells were incubated in serum-free and 0.5% fetal bovine serum medium, respectively, for 24 h before starting experiments. Immunohistochemical analyses of human cholangiocarcinoma Tumor specimens were collected from 100 patients who underwent liver resection for cholangiocarcinoma at Beaujon Hospital (Clichy sous Bois, France); clinicopathological characteristics of the patients are described in the table S1. Normal liver tissue was obtained from patients who underwent liver surgery for focal benign lesions. Paraffin-embedded tissue blocks were processed to prepare tissue microarrays (five cores/CCA case), on which staining for E-cadherin and EGFR was performed using the primary antibodies listed in table S2 and an automated immunohistochemical stainer (streptavidin-peroxidase protocol; Ventana Medical Systems, Tucson, AZ). In all cases studied, E-cadherin was expressed in all tumor cells and displayed either a membranous or a cytoplamsic distribution. Potential correlations between this pattern of expression and the different histopronostic factors were analyzed. EGFR expression was detected in 68% of the tumors and was predominantly located at the plasma membrane, as previously shown [18]. The percentage of EGFR stained cells was measured on the basis of the median value (30%), two subgroups of tumors were defined. Tumors with a percentage of stained cells below the median defined a negative/weak subgroup. Tumor with a percentage of stained cells above median defined a moderate/strong

7 subgroup. Immunohistochemical analysis of EGF was performed on a small series of CCA (n=10) according the following procedure. The sections were pre-incubated with H2O2 for 5 min, washed and incubated overnight with EGF antibody (Table S2). Post-primary block (Novolink Polymer Detection System; Novocastra Laboratories, Ltd.) was applied for 15 min. Specimens were then washed and incubated with Novolink Polymer for 15 min. The Autostainer Plus (Dako) was used to perform immunostainings. The color was developed using amino-ethyl-carbazole (AEC peroxidase substrate kit; Vector Laboratories, Ltd.). Mouse xenografts Animal experiments were performed in accordance with the French Animal Research Committee guidelines. Mz-Ch-A1 cells (1 × 10 6 cells) suspended in 0.2 mL of phosphate buffer (PBS) were implanted subcutaneously into the flank of 5-week-old immunodeficient female NMRI-nu (nu/nu) mice (Janvier, Le Genest Saint Isle, France). Tumor growth was followed with a caliper, and tumor volume (V) was calculated as follows: V = ab2π/6, where a is the longest and b, the shortest of two perpendicular diameters. Approximately 45 days post-cell inoculation, when tumor volume reached 150-200 mm3, mice were treated by gavage with gefitinib (100 mg/kg; daily) and dissolved in PBS + 0.1% Tween-80 (vehicle). Twentyfour days after gefitinib treatment was started, tumor was removed, fixed in 10 % formalin and in embedded-paraffin or frozen in liquid nitrogen. For immunohistochemical analyses, paraffin-embedded tumor samples were cut in 4-µm sections, and the sections were processed as human samples. Statistics For human tumor samples, qualitative variables were analyzed using the two-sided Fisher exact test. Quantitative data were compared using the non-parametric Mann-Whitney test.

8 Univariate and multivariate regression analyses were performed using IBM SPSS Statistics 17.0 or Expo32 software (Beckman-Coulter, Brea, CA). Supplemental Methods Materials,

real-time

PCR

analysis,

immunoprecipitation,

immunoblot

and

immunofluorescence experiments, luciferase assays, cell migration and invasion assays are described in the Supporting Information.

9

RESULTS The mislocalization of E-cadherin is associated with features of progression and EGFR expression in human CCA In benign human liver, E-cadherin is localized at the plasma membrane of bile duct epithelial cells (Figure 1A). Among CCA tumors from 100 patients, half of the cases displayed a marked intracellular staining of E-cadherin indicating ectopic localization of E-cadherin in tumor cells (Figure 1B). In cell cytoplasm, E-cadherin immunostaining was diffuse and homogeneous. The remaining tumors displayed a membranous localization of E-cadherin (Figure 1C). In univariate analysis, the cytoplasmic pattern of E-cadherin expression was significantly correlated with a tumor size above median (> 4 cm) and the presence of satellite nodules (Table 1). There was a trend towards an association with tumor grade and no correlation with perineural, node or vascular invasion and TNM stage (data not shown). In multivariate analysis, cytoplasmic E-cadherin was correlated with the peripheral type of CCA (p<0.0001). A significant association was also found between the cytoplasmic localization of E-cadherin and the amount of EGFR expressing cells (Table 1). These results suggested a link between the ectopic localization of E-cadherin, a hallmark of EMT, and the EGFR-dependent signaling pathway. Abrogation of EGFR activity restores membranous expression of E-cadherin in vivo and in vitro To test in vivo whether EGFR is involved in ectopic localization of E-cadherin in tumor cells, CCA cells (i.e. Mz-ChA-1) that express EGFR [23] were injected subcutaneously into nude mice, which were treated or not with a specific inhibitor of EGFR tyrosine kinase activity, gefitinib. Upon treatment with gefitinib, tumor volume and weight were significantly reduced

10 by an average of 5-fold and 3.5-fold, respectively (Figure 2A and B). In vehicle treated mice, E-cadherin was predominantly localized in the cytoplasm of tumor cells. By contrast, in gefitinib-treated mice, a strong membranous staining was observed in tumor cells (Figure 2C). In vitro, EGF, a ligand that is expressed in CCA tumor cells (Figure S1A), was employed to stimulate EGFR-dependent cell signaling (i.e. STAT3, AKT and ERK1/2) (Figure S1B). Upon EGF treatment, both CCA cell lines (i.e. Mz-ChA-1 and SK-ChA-1) were dispersed and displayed a fibroblastic-like phenotype (Figure S1C). This event was abrogated by two inhibitors of EGFR (gefitinib and cetuximab C225), and by MEK1/2 and STAT3 inhibitors whereas an AKT inhibitor had no effect (Figure S1C). Immunofluorescence experiments showed that in unstimulated CCA cells, E-cadherin was concentrated at the plasma membrane (Figure 2D and S2A). EGF-mediated EGFR activation triggered the internalization of Ecadherin in CCA cell cytoplasm, and consistent with in vivo data, this effect was abolished by gefitinib (Figure 2D and S2A). In addition, the promoter activity of E-cadherin was repressed (Figure 2E) and E-cadherin protein level was decreased upon EGF stimulation (Figure 2F and S2B). We next investigated the association between E-cadherin and β-catenin, another component of the adherens junctions. An association between E-cadherin and β-catenin was evidenced in unstimulated conditions, and diminished upon EGF stimulation. Gefitinib restored the association between both proteins as in control (Figure 3A). Immunofluorescence analyses indicated an internalization of β-catenin in CCA cell cytoplasm and nucleus following EGF treatment that was inhibited by gefitinib (Figure 3B and S2C). Consistently, transcriptional activity of β-catenin was significantly increased in EGF-stimulated CCA cells (Figure 3C). Collectively, these data demonstrated that EGF-mediated EGFR activation induced a down-regulation of E-cadherin leading to disruption of junctional complexes in CCA cells.

11 Epithelial to mesenchymal transition is induced by the EGF/EGFR axis in CCA cells The disruption of adherens junctions is a major event during EMT. The expression of mesenchymal markers that repress E-cadherin expression (i.e. Snail, ZEB1 and Slug) was further investigated. EGF-mediated activation of EGFR triggered a significant increase in the mRNA levels of Slug and ZEB1 that were inhibited by gefitinib in both CCA cell lines (Figure 3D and, S2D). No modification of Snail or Twist mRNA levels was observed upon EGF stimulation (data not shown). Upregulation of Slug mediates the switch from E-cadherin to N-cadherin expression, a hallmark of EMT [24]. In untreated condition, CCA cells expressed no or low levels of N-cadherin whereas EGF increased its expression notably (Figure 3E and S2E). The expression of another marker of EMT, α-SMA, was also increased in CCA cells exposed to EGF (Figure 3F and S2F). Finally, the expression of Zonula Occludens (ZO-1), a marker of tight junction involved in the apico-basal polarity of epithelial cells, was markedly decreased in CCA cells treated with EGF (Figure 3G and S2G). These data demonstrated that EGFR activation triggered the loss of apico-basal polarity in CCA cells. Migration and invasion of CCA cells are induced by the EGF/EGFR axis EMT phenotype is characterized by cell scattering and enhanced motility. To assess if EGFR activation results in CCA cell motility, time-lapse microscopy was used to capture the motility response of individual CCA cells treated or not with EGF. In untreated condition, CCA cells grew in clusters (Figure 4A, upper panel and Movie S1). Upon EGF, cells started to scatter 7 hours after treatment and were fully individualized after 24 hours with the appearance of membrane protrusions, a feature of migrating cells (Figure 4A, lower panel and Movie S2). Representative tracks of five cells in each condition are shown in Figure 4B. The activation of EGFR in CCA cells increased the speed of migration by 3-fold (0.047±0.005 µm/min in control cells vs. 0.159±0.025 µm/min in EGF-treated cells, p<0.001) and the

12 persistence by 4-fold (0.113±0.011 in control cells vs. 0.408±0.058 in EGF-treated cells, p<0.001). Cell motility requires a strong reorganization of the actin cytoskeleton, including dynamic formation and disassembly of focal adhesions (FAs) [25]. Phosphorylation of FAK (Focal Adhesion Kinase) on Tyr925 increases FAs turnover at the leading edges and enhances cell migration [26]. Upon EGF treatment, FAK was phosphorylated and concentrated at the leading edge of Mz-ChA-1 cells (Figure 4C and 4D, respectively) and SK-ChA-1 cells (Figure S3A and S3B, respectively). The actin-modulating protein VASP is concentrated at the tips of migratory protrusions (i.e. lamellipodia or filopodia) located at the leading edge of migrating cells [27]. Marked staining of VASP was observed at the leading edge of CCA cells upon EGFR activation by EGF compared to control cells (Figure 4E and S3C). These protrusions were positive for the lamellipodia marker, cortactin (Figure 4F and S3D). Cells undergoing EMT ultimately become invasive. The activation of EGFR by EGF increased by 50% the capacity of the cell to invade. Invasion induced by EGFR activation was inhibited by gefitinib (Figure 4G). Accordingly, treatment of CCA cells with EGF increased mRNA levels of two matrix metalloproteinases involved in extracellular matrix degradation, MMP1 and MMP9 by 7-fold and 11-fold, respectively. The induction of both MMPs induced by EGF was abolished in the presence of gefitinib (Figure 4H).

13 DISCUSSION EGFR can be highly expressed in tumor cells of CCA [18, 23, 28, 29]. Previous reports have shown that expression of EGFR was associated with CCA progression and poor clinical outcome [16, 29]. To date, underlying cellular mechanisms by which EGFR contributes to CCA progression has been poorly investigated. Here, we demonstrate that the EGF/EGFR axis is a potent inducer of EMT, the molecular signature of which is associated with CCA progression. One of the major events in the process of EMT is the loss of cell-cell contacts maintained by E-cadherin between cancer cells. In this report, we described an ectopic expression of Ecadherin in the cytoplasm of tumor cells in resected human CCA. Cytoplasmic expression of E-cadherin was correlated with aggressive phenotypic features of the CCA tumors, including CCA type, tumor size and satellite nodules supporting previous data [8]. Interestingly, we also showed a correlation between cytoplasmic localization of E-cadherin and the abundance of EGFR expression. This result was reinforced by our in vivo experimental findings using a xenografted model of CCA tumors, in which the inhibition of EGFR activity restored proper localization of E-cadherin at the plasma membrane of tumor cells. We found an expression of EGF, an EGFR ligand, in tumor cells of human CCA samples. Another source of EGF in CCA is the tumor environment including stromal myofibroblasts (data not shown). These data suggest the existence of autocrine and paracrine activating loops of EGFR. While in normal liver, EGF is barely detected in the biliary epithelium [30], an overexpression of EGF was previously reported both in proliferative biliary ductules of cirrhotic liver [30] and in CCA cells [21]. We demonstrated that EGF induced the disruption of cell-cell contacts along with an internalization of E-cadherin, a mechanism previously reported to occur through caveolin-dependent endocytosis [31]. Several EMT-inducing transcriptional factors have been shown to bind E-cadherin promoter

14 and directly repress its activity [4]. Internalization of E-cadherin by EGF in CCA cells was accompanied by a down-regulation of transcription and protein levels of E-cadherin. We found an upregulation of two major EMT-inducing transcriptional factors by the EGF/EGFR axis in CCA cells, Slug and ZEB1. Both factors bind to E-box sequences in the E-cadherin promoter and thereby down-regulate E-cadherin expression [32]. Slug is not expressed in intrahepatic bile ducts of normal liver but an expression has been detected in the cytoplasm and nuclei of tumor cells in CCA [10]. An overexpression of ZEB1 has been reported in gallbladder carcinoma at sites of cancer invasion and in CCA [33, 34]. Beside repression of the epithelial marker E-cadherin, ZEB1 displays other functions such as the activation of the mesenchymal markers expression, vimentin and N-cadherin [35]. Because a switch from Ecadherin to N-cadherin expression by EGF was observed in CCA cells, we assume that ZEB1 may contribute to the repression of E-cadherin and activation of N-cadherin expression. A rearrangement of the actin cytoskeleton is necessary for cell to migrate and invade [2, 36]. We showed that activation of the EGFR by EGF elicited CCA cell motility by regulating proteins involved in actin dynamics such as ezrin (data not shown), VASP, cortactin and FAK. CCA cells exposed to EGF were characterized by the formation of lamellipodia with a strong expression of VASP and cortactin at the leading edge of the cells. VASP regulates assembly of the actin-filament network and modulates the morphology and behavior of membrane protrusions named lamellipodia, a migratory structure involved in cancer cell migration and invasion during tumor metastasis [27, 36]. Cortactin, another actin-binding protein, is a regulator of branched actin assembly strongly expressed in lamellipodia [37]. The focal adhesion kinase, FAK, acts as a switch that coordinates focal adhesion disassembly and cell protrusion. To perform its functions, FAK must be phosphorylated in particular on Tyr925 [26]. Here, we report a phosphoregulation of FAK on Tyr-925 by EGF in CCA cells. Previous studies indicated that FAK in CCA cells was phosphorylated by HGF, a ligand of

15 the receptor tyrosine kinase c-Met [38]. FAK has been also implicated in MMP-9 synthesis by TNF-α in CCA cells [39]. In this report, we show that MMP-9 expression is increased in CCA cells exposed to EGF, which one can attribute to FAK phosphorylation. In conclusion, we established a link between EGFR and EMT in the CCA. Thus, the present study reinforces the major implication of EGFR in CCA progression.

16

ACKNOWLEDGMENT We thank “Tumeur-Est tissue bank” for cholangiocarcinoma human samples and the Nikon Imaging Center, Institut Curie-CNRS Paris, for time-lapse microscopy.

17

REFERENCES [1]

Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymal transitions in

development and disease. Cell 2009;139:871-890. [2]

Guarino M, Rubino B, Ballabio G. The role of epithelial-mesenchymal transition in

cancer pathology. Pathology 2007;39:305-318. [3]

Firrincieli D, Boissan M, Chignard N. Epithelial-mesenchymal transition in the liver.

Gastroenterol Clin Biol 2010;34:523-528. [4]

De Craene B, Berx G. Regulatory networks defining EMT during cancer initiation and

progression. Nat Rev Cancer 2013;13:97-110. [5]

Razumilava N, Gores GJ. Cholangiocarcinoma. Lancet 2014.

[6]

Ryu HS, Chung JH, Lee K, Shin E, Jing J, Choe G, et al. Overexpression of epithelial-

mesenchymal transition-related markers according to cell dedifferentiation: clinical implications as an independent predictor of poor prognosis in cholangiocarcinoma. Hum Pathol 2012;43:2360-2370. [7]

Fabris L, Cadamuro M, Moserle L, Dziura J, Cong X, Sambado L, et al. Nuclear

expression of S100A4 calcium binding protein increases cholangiocarcinoma invasiveness and metastasization. Hepatology 2011. [8]

Yao X, Wang X, Wang Z, Dai L, Zhang G, Yan Q, et al. Clinicopathological and

prognostic significance of epithelial mesenchymal transition-related protein expression in intrahepatic cholangiocarcinoma. Onco Targets Ther 2012;5:255-261. [9]

Korita PV, Wakai T, Ajioka Y, Inoue M, Takamura M, Shirai Y, et al. Aberrant

expression of vimentin correlates with dedifferentiation and poor prognosis in patients with intrahepatic cholangiocarcinoma. Anticancer Res 2010;30:2279-2285. [10]

Zhang KJ, Zhang BY, Zhang KP, Tang LM, Liu SS, Zhu DM, et al. Clinicopathologic

significance of slug expression in human intrahepatic cholangiocarcinoma. World J

18 Gastroenterol 2010;16:2554-2557. [11]

Dos Santos A, Court M, Thiers V, Sar S, Guettier C, Samuel D, et al. Identification of

cellular targets in human intrahepatic cholangiocarcinoma using laser microdissection and accurate mass and time tag proteomics. Mol Cell Proteomics 2010;9:1991-2004. [12]

Mosnier JF, Kandel C, Cazals-Hatem D, Bou-Hanna C, Gournay J, Jarry A, et al. N-

cadherin serves as diagnostic biomarker in intrahepatic and perihilar cholangiocarcinomas. Mod Pathol 2009;22:182-190. [13]

Henson ES, Gibson SB. Surviving cell death through epidermal growth factor (EGF)

signal transduction pathways: implications for cancer therapy. Cell Signal 2006;18:20892097. [14]

Sebastian S, Settleman J, Reshkin SJ, Azzariti A, Bellizzi A, Paradiso A. The

complexity of targeting EGFR signalling in cancer: from expression to turnover. Biochim Biophys Acta 2006;1766:120-139. [15]

Barr S, Thomson S, Buck E, Russo S, Petti F, Sujka-Kwok I, et al. Bypassing cellular

EGF receptor dependence through epithelial-to-mesenchymal-like transitions. Clin Exp Metastasis 2008;25:685-693. [16]

Yoshikawa D, Ojima H, Iwasaki M, Hiraoka N, Kosuge T, Kasai S, et al.

Clinicopathological and prognostic significance of EGFR, VEGF, and HER2 expression in cholangiocarcinoma. Br J Cancer 2008;98:418-425. [17]

Yoon JH, Gwak GY, Lee HS, Bronk SF, Werneburg NW, Gores GJ. Enhanced

epidermal growth factor receptor activation in human cholangiocarcinoma cells. J Hepatol 2004;41:808-814. [18]

Claperon A, Guedj N, Mergey M, Vignjevic D, Desbois-Mouthon C, Boissan M, et al.

Loss of EBP50 stimulates EGFR activity to induce EMT phenotypic features in biliary cancer cells. Oncogene 2012;31:1376-1388.

19 [19]

Ito Y, Takeda T, Sasaki Y, Sakon M, Yamada T, Ishiguro S, et al. Expression and

clinical significance of the erbB family in intrahepatic cholangiocellular carcinoma. Pathol Res Pract 2001;197:95-100. [20]

Harder J, Waiz O, Otto F, Geissler M, Olschewski M, Weinhold B, et al. EGFR and

HER2 expression in advanced biliary tract cancer. World J Gastroenterol 2009;15:4511-4517. [21]

Nonomura A, Ohta G, Nakanuma Y, Izumi R, Mizukami Y, Matsubara F, et al.

Simultaneous detection of epidermal growth factor receptor (EGF-R), epidermal growth factor (EGF) and ras p21 in cholangiocarcinoma by an immunocytochemical method. Liver 1988;8:157-166. [22]

Sia D, Hoshida Y, Villanueva A, Roayaie S, Ferrer J, Tabak B, et al. Integrative

molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes. Gastroenterology 2013;144:829-840. [23]

Claperon A, Mergey M, Aoudjehane L, Ho-Bouldoires TH, Wendum D, Prignon A, et

al. Hepatic myofibroblasts promote the progression of human cholangiocarcinoma through activation of epidermal growth factor receptor. Hepatology 2013. [24]

Cheng JC, Auersperg N, Leung PC. EGF-induced EMT and invasiveness in serous

borderline ovarian tumor cells: a possible step in the transition to low-grade serous carcinoma cells? PLoS One 2012;7:e34071. [25]

Webb DJ, Donais K, Whitmore LA, Thomas SM, Turner CE, Parsons JT, et al. FAK-

Src signalling through paxillin, ERK and MLCK regulates adhesion disassembly. Nat Cell Biol 2004;6:154-161. [26]

Deramaudt TB, Dujardin D, Hamadi A, Noulet F, Kolli K, De Mey J, et al. FAK

phosphorylation at Tyr-925 regulates cross-talk between focal adhesion turnover and cell protrusion. Mol Biol Cell 2011;22:964-975. [27]

Bear JE, Svitkina TM, Krause M, Schafer DA, Loureiro JJ, Strasser GA, et al.

20 Antagonism between Ena/VASP proteins and actin filament capping regulates fibroblast motility. Cell 2002;109:509-521. [28]

Andersen JB, Spee B, Blechacz BR, Avital I, Komuta M, Barbour A, et al. Genomic

and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology 2012;142:1021-1031 e1015. [29]

Sia D, Tovar V, Moeini A, Llovet JM. Intrahepatic cholangiocarcinoma: pathogenesis

and rationale for molecular therapies. Oncogene 2013. [30]

Komuves LG, Feren A, Jones AL, Fodor E. Expression of epidermal growth factor

and its receptor in cirrhotic liver disease. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society 2000;48:821-830. [31]

Lu Z, Ghosh S, Wang Z, Hunter T. Downregulation of caveolin-1 function by EGF

leads to the loss of E-cadherin, increased transcriptional activity of beta-catenin, and enhanced tumor cell invasion. Cancer Cell 2003;4:499-515. [32]

Sanchez-Tillo E, Liu Y, de Barrios O, Siles L, Fanlo L, Cuatrecasas M, et al. EMT-

activating transcription factors in cancer: beyond EMT and tumor invasiveness. Cellular and molecular life sciences : CMLS 2012;69:3429-3456. [33]

Adachi Y, Takeuchi T, Nagayama T, Ohtsuki Y, Furihata M. Zeb1-mediated T-

cadherin repression increases the invasive potential of gallbladder cancer. FEBS Lett 2009;583:430-436. [34]

Oishi N, Kumar MR, Roessler S, Ji J, Forgues M, Budhu A, et al. Transcriptomic

profiling reveals hepatic stem-like gene signatures and interplay of miR-200c and epithelialmesenchymal transition in intrahepatic cholangiocarcinoma. Hepatology 2012;56:1792-1803. [35]

Sanchez-Tillo E, Siles L, de Barrios O, Cuatrecasas M, Vaquero EC, Castells A, et al.

Expanding roles of ZEB factors in tumorigenesis and tumor progression. American journal of cancer research 2011;1:897-912.

21 [36]

Vignjevic D, Montagnac G. Reorganisation of the dendritic actin network during

cancer cell migration and invasion. Semin Cancer Biol 2008;18:12-22. [37]

Weaver AM. Cortactin in tumor invasiveness. Cancer Lett 2008;265:157-166.

[38]

Pongchairerk U, Guan JL, Leardkamolkarn V. Focal adhesion kinase and Src

phosphorylations in HGF-induced proliferation and invasion of human cholangiocarcinoma cell line, HuCCA-1. World J Gastroenterol 2005;11:5845-5852. [39]

Mon NN, Hasegawa H, Thant AA, Huang P, Tanimura Y, Senga T, et al. A role for

focal adhesion kinase signaling in tumor necrosis factor-alpha-dependent matrix metalloproteinase-9 production in a cholangiocarcinoma cell line, CCKS1. Cancer Res 2006;66:6778-6784.

22

FIGURE LEGENDS Figure 1: Expression of E-cadherin in human CCA Representative immunostaining of E-cadherin in normal bile duct epithelial cells (A) and in CCA cells (B and C). (A) In normal bile duct epithelial cells, E-cadherin was primarily localized at the cell-cell junctions. In CCA cells, E-cadherin displayed either a marked intracellular (B) or a membranous (C) expression. Original magnification, x200. Figure 2: Activation of EGFR induces mislocalization of E-cadherin in vitro and in vivo (A) Tumor volume of mice bearing Mz-ChA-1 cells treated with vehicle (white bars) or with gefitinib (black bars ; 100 mg/kg daily). Mean of tumor volume ± SEM. (B) Mean of tumor weight ± SEM at sacrifice. * p<0.05; *** p<0.001 ; CCA cells/vehicle vs. CCA cells/gefitinib. Each group comprised 10 mice. (C) Representative immunostaining of Ecadherin in CCA cells/vehicle (left panel) and CCA cells/gefitinib (right panel). Magnification x40. (D) Mz-ChA-1 cells were pretreated or not with gefitinib (1µM), 30 min prior to incubation with or without EGF (50 ng/ml) for 24h. Localization of E-cadherin was assessed by immunofluorescence using confocal microscopy. Scale bar, 20 µm. (E) Mz-ChA1 cells were transiently transfected with the pGL3-Control or pGL3-E-Cadherin luciferase reporter constructs. Forty-eight hours after transfection, cells were stimulated with EGF (50 ng/ml). After 16 h, luciferase activity was determined. Results obtained for pGL3-Control in presence of control medium were set to 1. Mean of luciferase activity ± SEM of four independent experiments, each performed in duplicate. *** p<0.001 vs. pGL3-Ctl with EGF stimulation; ### p<0.001 vs. pGL3-E-cadherin without EGF stimulation. (F) Mz-ChA-1 cells were stimulated with EGF (50 ng/ml) for 24 hours. E-cadherin expression was analyzed by western blot. Representative image of two independent experiments are shown. Figure 3: EGFR activation promotes loss of epithelial features and gain of mesenchymal

23 markers in CCA cells (A) Mz-ChA-1 cells were pretreated or not with gefitinib (1 µM) for 30 min prior stimulation by EGF (50 ng/ml) for 24 h. Cell lysates were immunoprecipitated with an anti β-catenin antibody or control rabbit immunoglobulin G. Similar amounts of immunoprecipitates were analyzed by western blotting using E-cadherin or β-catenin antibodies. Representative blots of three experiments are shown. (B) Mz-ChA-1 cells were pretreated with gefitinib (1 µM), 30 min prior to incubation with EGF (50 ng/ml) for 24 h. Localization of β-catenin was assessed by immunofluorescence using confocal microscopy. The blue labeling is nuclear DNA staining by TOPRO. Scale bar, 20 µm. (C) Mz-ChA-1 cells were transiently transfected with the pTOP-Flash or pFOP-Flash luciferase reporter constructs. Forty-eight hours after transfection, cells were stimulated with EGF (50 ng/ml). After 16 h, luciferase activity was determined. Results obtained for pFOP-Flash in presence of control medium were set to 1. Mean of luciferase activity ± SEM of four independent experiments, each performed in duplicate. * p<0.05 vs. pFOP-Flash with EGF stimulation; # p<0.05 vs. pTOP-Flash without EGF stimulation. (D) Mz-ChA-1 cells were pretreated with gefitinib (1µM), 30 min prior to incubation with EGF (100 ng/ml) for 6h. Slug (left panel) and ZEB-1 (right panel) mRNA levels were analyzed by RT-PCR. Quantitative data are means ± SEM of three experiments performed in duplicate; * p<0.05, *** p<0.001 vs. control medium; # p<0.05, ### p<0.001 vs. medium + EGF. (E) Mz-ChA-1 cells were stimulated with EGF (25 ng/ml) for 10 days and N-cadherin expression was analyzed by western blot. (F) Mz-ChA-1 cells were pretreated with gefitinib (1µM), 30 min prior to incubation with EGF (50 ng/ml) for 48 h. Cells were analyzed for α-SMA expression by western blot and immunofluorescence. β-actin was used as a loading control. (G) Mz-ChA-1 cells were stimulated with EGF (25 ng/ml) for 24h and ZO-1 localization was analyzed by immunofluorescence. Representative images of three experiments are shown. Scale bar, 20 µm.

24 Figure 4: EGF promotes CCA cell migration through EGFR activation (A-B) Mz-ChA-1 cells stimulated with EGF (50 ng/ml) were analyzed by time-lapse microscopy. (A) Frames were selected from a 48-h long time-lapse experiment. Time after exposure to EGF is indicated in hours in the lower right corner of each frame. (B) Tracks of five representative cells for each condition are shown. (C) Activated pFAK-Y925 and total FAK were examined by western blot after stimulation of Mz-ChA-1 cells with EGF (100 ng/ml) for 5 and 30 min. (D-F) Mz-ChA-1 cells were pretreated with gefitinib (1µM), 30 min prior to incubation with EGF (100 ng/ml) for 1h. Localization of pFAK-Y925 (D), VASP (E) and cortactin (F) was assessed by immunofluorescence. Arrows indicate the leading edges of the cells. Scale bar, 4 µm. (G) Cell invasion towards a chemoattractant (2.5% serum) was measured by Transwell chamber assay coated with matrigel. Data are means ± SEM of four experiments in 3 random fields. * p<0.05 vs. control medium, # p<0.05 vs. medium + EGF. (H) Mz-ChA-1 cells were pretreated with gefitinib (1µM), 30 min prior to incubation with EGF (100 ng/ml) for 6h. MMP1 (left panel) and MMP9 (right panel) mRNA levels were analyzed by RT-PCR. Quantitative data are means ± SEM of three experiments performed in duplicate; *** p<0.001 vs. control medium; ### p<0.001 vs. medium + EGF.

Cytoplasmic pattern (n=51)

Membranous pattern (n=49)

p

Type Hilar Peripheral Size > 4 cm (median) Satellite nodules Grade: poorly differentiated Vascular invasion Perineural invasion Lymph node metastasis

15 (29%) 36 (71%) 31 (61%) 21 (41%) 29 (57 %) 36 (71%) 29 (57%) 17 (34%)

36 (73%) 13 (27%) 14 (25%) 9 (16%) 21 (40%) 33 (61%) 37 (67%) 12 (22%)

<0.0001 0.001 0.033 0.076 0.51 0.50 0.33

EGFR > 30% (median)

30 (60%)

19 (39%)

0.037

Table 1: Pathological features and EGFR expression in cholangiocarcinoma tumors with cytoplasmic vs. membranous pattern of E-cadherin expression. Univariate analysis (Mann Whitney).