This work has been carried out in accordance with the Declaration of Helsinki (2000) of the World Medical Association. This study was approved ethically by University Hospital Birmingham Trust (LREC 2002/166). All patients provided informed written consent.

Esophageal adenocarcinoma resection specimens: Samples of normal squamous esophagus (n = 40) and esophageal adenocarcinoma (n = 40) were collected during surgery. Half of the esophageal adenocarcinoma resection specimens (n = 20) collected also had associated intestinal Barrett’s metaplasia. In addition, samples (n = 20) of long segment (≥ 3 cm) Barrett’s metaplasia, defined as columnar mucosa with intestinal type goblet cells were also collected during endoscopy. All specimens were divided in two, half for RNA extraction and half for immunohistochemistry.

Immunohistochemistry: Immunohistochemistry for E-cadherin was performed using microwave antigen retrieval as previously described with an E-cadherin monoclonal antibody (clone 36, BD Biosciences, Oxford, UK) used at a concentration of 1:300[21]. Immunohistochemistry for Snail, Slug and Twist was performed as follows: Slides were immersed in W-cap buffer (Bio-Optica, Milan, Italy) and cycled in a Pixel antigen retriever (CellPath, Newtown, UK) for 60 min, washed in running water and placed in methanol:hydrogen peroxide (10:1) for 5 min. Sections were then incubated in a primary antibody to either Snail, 1:10, (SNAI1 clone E18, Autogen Bioclear, Wiltshire, UK), Slug, 1:20 (SNAI2 clone G18 Autogen Bioclear, UK) or Twist, 1:50 (clone C17 Autogen Bioclear, UK) in TBS 7.5 × buffer (Bios Europe Ltd, Skelmersdale, UK) at 4°C overnight, washed with TBS and reacted with peroxidase-linked rabbit anti-sheep antibody (Dako, Ely, UK) at a 1:100 dilution in TBS for 1 h. The immunoreactivity was then revealed as above using DAB. Slides were then dipped in hematoxylin, dehydrated and mounted. The slides were scored by a previously described method for (1) intensity of staining (0 = negative, 1 = weak, 2 = moderate, 3 = intense) and (2) percentage of epithelial cells staining (0 = 0%-5%; 1 = 6%-25%; 2 = 26%-50%; 3 = 51%-75%; 4 = 76%-100%); these two scores were multiplied to yield a final staining score[39]. In addition, cellular localization (nuclear, cytoplasmic, cell surface) was assessed. All sections were scored independently by two observers (PJ and CT).

In the series of immunofluorescent experiments following primary antibody incubations, sections were washed extensively and then incubated with either FITC goat anti-mouse or goat anti-rabbit (Jackson Immunoresearch, USA, 1:500) for 1 h. Sections were then washed and incubated in 4’, 6-Diamidino-3-phenylindole dihydrochloride hydrate (DAPI) (1:10 000) for 1 min prior to visualisation. Omission of primary antibody was employed as a negative control. Images were visualized using an Olympus BX40 microscope and digital images taken using a Sensys Photometrics camera (Middlesex, UK). Desksoft SmartCapture 2 software was used for image acquisition (Desksoft, USA).

Real time RT-PCR reactions were performed as previously described using 18S ribosomal RNA as an internal standard (PE Biosystems, Roche, USA)[21]. Each reaction was performed in triplicate and contained one of the following sets of probes and primers: (1) E-cadherin (probe 5’FAM AAATTCACTCTGCCCAGGACGCGGTAMRA3’), forward primer (5’-GGCGCCACCTCGAGAGA-3’) and reverse primer (5’-TGTCGACCGGTGCAATCTT3’); (2) Slug (probe 5'FAM'CACATACAGTGATTATTTCCC CGTATCTCTATGAGAGTTAMRA3'), forward primer (5’-AAAAGCCAAACTACAGCGAACTG-3’) and reverse primer (5’-AGAATCTCTGCTTGTGGT ATG ACA-3’); (3) Snail (probe 5’FAMTGCAGGACTCTA ATCCAGAGTTTACCTTCCAGCTAMRA3’), forward primer (5’-CCCAATCGGAAGCCTAACTACAG-3’) and reverse primer (5’-CAGGTGGGCCTGGTCGTA-3’), or (4) Twist (probe 5’-FAMACAATGACATCTAGG TCTC-3’), forward primer (5’-GCTCCAGAGTCTCTA GACTGTCCATT-3’) and reverse primer(5’-GGGGCCTGGTCCATGTC-3’).

Western blotting was performed as previously described[21] with antibodies to (1) E-cadherin (1:2 000, clone 36, Autogen Bioclear UK); (2) Fibronectin (1:1000, Clone P1F11 Autogen Bioclear, UK); (3) Vimentin (1:1000, clone C-20, Autogen Bioclear, UK); (4) Cytokeratin 19 (CK-19) (1:2000; Oncogene Research Products USA) and (5) GFP (1:2000, Clone AB290 Abcam, UK). Immunoreactive bands were then subject to densitometry using NIH Image 1.62 software.

Cell lines derived from esophageal adenocarcinoma (OE33[22], SEG-1[23] and oesophageal squamous carcinoma TE-7[24]) were all routinely cultured in Dulbecco’s modified Eagles medium (Gibco, USA) with 10% fetal calf serum supplemented with 1 × 10⁵ Units/L penicillin and 1 g/L streptomycin. Upon reaching 70% confluence cells were lysed into Trizol reagent (Gibco, UK) for mRNA extraction and evaluation of Slug mRNA expression by Real Time PCR.

Cell line OE33 was transiently transfected with either control empty vector, pMSCV-GFP or full-length human Slug-GFP tagged construct, pMSCV-Slug-GFP[25] or beta-galactosidase plasmid (beta-GAL) using Lipofectamine Plus according to manufacturers instruction. 48 h post transfection cells were either assayed for transfection efficiency using beta-galactosidase assay to determine transfection efficiency, lysed into RIPA buffer for Western blotting or lysed in Trizol for real time RT-PCR studies. Data analysis was only performed in experiments where a minimum of 40% transfection efficiency was achieved as determined by beta-galactosidase assay.

To determine the effect of Slug on the human E-cadherin promoter, an E-cadherin promoter assay was performed as previously described[21]. Briefly, OE33 cells were transiently transfected with full length human slug (pCDNA3-Slug[25]), Renilla luciferase plasmid pRL TK and the wild-type human E-cadherin promoter cloned into the pGL3basic luciferase reporter plasmid (EproWT) a kind gift from Professor Frans van Roy). Controls included empty pCDNA3 and empty pGL3basic vectors. 48 h post transfection cells were assayed for firefly and Renilla luciferase activities using the dual-reporter assay kit Stop ‘N’ Glow (Promega) according to manufacturer’s instruction. Firefly luciferase activity was normalized to Renilla luciferase activity as a transfection control. Promoter activity was expressed as a fold change in relative luciferase units (RLU) compared to that obtained in pGL3basic control transfected cells. Results shown represent the means ± SE of three independent experiments.

All data are presented as means ± SE. Statistical significance was calculated using unpaired Student’s t test. To assess the association between Slug and E-cadherin, linear regression analysis was performed on log-transformed mRNA fold change values using log Slug expression as the independent variable and log E-cadherin expression as the dependent variable. The R² value was determined to demonstrate the extent to which expression of Slug was correlated to E-cadherin. A Bonferroni adjustment was applied to P values attained from multiple comparisons during the analysis of prognostic factors. In univariate analyses significance was accepted at P≤ 0.05, following Bonferroni adjustment, P≤ 0.0125 was considered significant. All analyses were performed using SPSS version 10.0 (SPSS Inc, USA).

In normal stratified squamous esophagus, there was no detectable expression of Snail and only weak diffuse cytoplasmic immunoreactivity for Twist, while there was strong nuclear immunoreactivity for Slug in basal and supra-basal layers (Figure 1). In Barrett’s metaplasia, Snail and Slug were generally localised to the cytoplasm with staining being patchy and diffuse in nature. In contrast, Twist was largely nuclear in localization (Figure 1). In esophageal adenocarcinoma, all three proteins were abundantly expressed with Snail and Twist being mostly localised in the cytoplasm whilst Slug was almost exclusively nuclear in all adenocarcinoma specimens (n = 20) (Figure 1).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Immunolocalization of Slug, Snail and Twist in normal esophagus, intestinal type Barrett’s metaplasia and oesophageal adenocarcinoma. Normal squamous esophagus (S), intestinal type Barrett’s metaplasia (BM) and esophageal adenocarcinoma (ADC) were subjected to immunohistochemistry using antibodies to Slug, Snail and Twist. Arrows denote areas of positivity (× 20, × 40).

In addition, all slides were scored and data are shown in Table 1. To determine if Snail, Slug and Twist were over-expressed in adenocarcinoma specimens relative to non-matched Barrett’s metaplastic specimens (BM-ve), statistical analysis was performed and results showed that both Slug and Snail were significantly over-expressed (P < 0.05) in esophageal adenocarcinoma. Twist, however, was not significantly induced. To validate this data, we further performed immunohistochemistry on a further 20 sections of adenocarcinoma with matched Barrett’s metaplastic tissue (BM+ve). When comparing the staining scores of matched Barrett’s metaplastic specimens with adenocarcinoma, only Slug was significantly over-expressed (P < 0.05).

Since there is a growing body of evidence implicating Slug as a repressor of E-cadherin, we further sought to address if Slug overexpression was associated with depressed E-cadherin expression in sections of both Barrett’s metaplasia and adenocarcinoma using dual immunofluorescence with anti-Slug (FITC) and E-cadherin antibodies (Texas Red) (Figure 2). Sections of Barrett’s metaplasia mostly demonstrated preserved cell border E-cadherin immunoreactivity consistent with its role in cell-cell adhesion whilst there was little detectable expression of Slug. Conversely in sections of adenocarcinoma in areas of abundant Slug immunoreactivity, there was almost a complete loss of E-cadherin immunoreactivity with the expression mostly localised to the cytoplasm.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Co-immunofluoresence of Slug and E-cadherin in Barrett’s metaplasia and esophageal adenocarcinoma. Sections of Barrett's metaplasia (BM) mostly demonstrate preserved cell border E-cadherin (Texas Red) immunoreactivity consistent with its role in cell-cell adhesion. Slug (FITC) immunoreactivity in BM however was minimal and appeared non specific. Conversely in sections of adenocarcinoma (ADC) in areas of abundant Slug immunoreactivity, there was negligible E-cadherin immunoreactivity. DAPI staining was used to highlight nuclei in sections (× 40).

To determine if Slug, Snail, Twist and E-cadherin were modulated at the transcriptional level in esophageal adenocarcinoma, 10 specimens of esophageal adenocarcinoma each with matched intestinal Barrett’s metaplasia were analysed by real-time PCR (Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 mRNA expression of E-cadherin, Slug, Snail and Twist in esophageal adenocarcinoma. Real time RT-PCR was employed to examine the expression of E-cadherin, Slug, Snail and Twist in ten esophageal adenocarcinoma specimens (ADC1-ADC10) each compared to Barrett’s metaplastic tissue collected from the same resection specimen. Barrett’s metaplastic tissue control (BM), normalized to 1.0.

E-cadherin mRNA was repressed in 70% of adenocarcinoma specimens whilst Slug, Snail and Twist was over-expressed in 70%, 40% and 50%, respectively. The mean fold decrease for E-cadherin in the ten specimens examined was 0.73 whilst Slug, Snail and Twist had mean fold increases of 2.07, 1.47 and 1.50 respectively. We then sought to determine if the mean fold changes for Slug, Snail and Twist across the ten specimens were significantly different from E-cadherin values. This revealed that only Slug was significantly elevated compared to E-cadherin expression (P < 0.005). Additionally we assessed the association between E-cadherin and these EMT regulators using linear regression analysis. Results of this showed that again only Slug was negatively correlated with E-cadherin expression (R² 0.677, P < 0.03) (Figure 4).

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Correlation of E-cadherin and Slug mRNA expression in esophageal adenocarcinoma specimens. Using linear regression analysis, Slug mRNA expression was demonstrated to be negatively correlated with E-cadherin expression in the ten adenocarcinoma specimens examined (ADC1-10) (R² 0.677, P < 0.03).

Slug mRNA expression was examined in a panel of three esophageal cell lines OE33, SEG1 and TE-7 by real-time PCR and results showed that the cell line OE33 had the lowest expression of Slug mRNA whilst the highest Slug expressing cell line was TE-7. In this regard, the cell line OE33 was chosen for Slug overexpression studies.

The cell line OE33 was transiently transfected with either full length human Slug-GFP vector or the control empty GFP vector. Forty-eight h after transfection, cells were either lysed and processed for mRNA and protein analysis or fixed for immunofluorescence.

mRNA analysis revealed an approximate 6.5 × induction in Slug mRNA in cells transfected with full length human Slug-GFP compared to cells transfected with empty-GFP vector (P < 0.05) (Figure 5A). Exogenous Slug expression was further verified indirectly (due to a lack of a commercially available Slug antibody suitable for Western blotting) by Western blotting for GFP (Figure 5B).

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Overexpression of Slug in the esophageal adenocarcinoma derived cell line OE33. Cell line OE33 was transiently transfected with either full length human Slug-GFP vector (OE33 Slug-GFP) or with the control empty GFP vector (OE33 empty). 24 h post transfection, (A) Slug mRNA and (B) GFP protein expression was determined. CK-19 was used as an epithelial control. ^aP < 0.05 using Student's t test.

To ascertain whether overexpression of Slug in OE33 cells mediated an EMT phenotype, Western blotting was performed to examine protein expression of three EMT markers E-cadherin, vimentin and fibronectin (Figure 6). Densitometric analysis of Western blots (n = 6) revealed Slug overexpression was associated with a repression in E-cadherin (mean fold decrease of 53% ± 15%, P < 0.05) and increased vimentin (41% ± 11%, P < 0.05) and fibronectin expression (32% ± 12%, P < 0.05). In addition, the cell adhesion and signalling molecule beta-catenin was also examined and shown to be significantly repressed (57% ± 19%, P < 0.05).

Open in New Tab   Full Size Figure   Download Figure

Figure 6 The effect of Slug overexpression on proteins implicated in EMT. Expression of E-cadherin, beta-catenin, fibronectin and vimentin were all examined in Slug over-expressing (OE33 Slug-GFP) and control cells (OE33 empty-GFP). Lysates were normalized using GFP expression.

To address whether Slug mediated E-cadherin protein repression was through a direct effect at the level of the E-cadherin promoter and consequently a transcriptional effect, a luciferase E-cadherin promoter assay and real-time PCR was employed respectively as previously described[2125]. Overexpression of Slug in OE33 resulted in a significant reduction (35% ± 10%, P < 0.05) in E-cadherin promoter activity (Figure 7A) which was mirrored by decreased E-cadherin mRNA expression (20% ± 5%, P < 0.05) (Figure 7B).

Open in New Tab   Full Size Figure   Download Figure

Figure 7 Slug expression causes a decrease in E-cadherin promoter activity and mRNA expression. A: OE33 cells transfected with full length human Slug (Slug) showed decreased E-cadherin promoter activity compared to sham transfected control cells. Promoter activity was expressed as a fold change in relative luciferase units (RLU) compared to control cells. Results represent the means ± SE of three independent experiments. ^aP < 0.05 using Student’s t test; B: Expression of E-cadherin mRNA in Slug over-expressing cells (Slug) was also repressed in comparison to control cells (control). Relative gene expression was normalized to 1.0 (100%) of controls. Error bars = 2 mean ± SE. ^cP < 0.05 using Student’s t test. These data are the mean of three independent experiments.

The incidence of esophageal adenocarcinoma is currently rising faster than any other cancer in the Western world though the cause of this increase is largely unknown. Progression of this disease is associated with a repression in the cell adhesion molecule E-cadherin; a crucial event in invasion and metastasis.

E-cadherin silencing is a common event in nearly all epithelial malignancies and in several cancers this is either due to mutation, gene deletion or promoter methylation. However, how E-cadherin is silenced in esophageal adenocarcinoma has not been unequivocally addressed. In this study, the authors demonstrate that the overexpression of Slug could be a potential mechanism for mediating E-cadherin repression.

Recent reports have highlighted the importance of epithelial mesenchymal regulators including Snail, Slug and Twist in gastrointestinal carcinogenesis. In particular in esophageal squamous cell cancers, Slug is over-expressed. This is the first study to report that Slug is also over-expressed in esophageal adenocarcinomas. Furthermore, our in vitro studies would suggest that this protein may be the cause of the repression in E-cadherin observed in this cancer.

By understanding how Slug is induced and by blocking its expression, this study may represent a future strategy for therapeutic intervention in the treatment of patients with esophageal adenocarcinoma.

Slug, Snail and Twist are all proteins involved in the process called epithelial mesenchymal transition. This is when epithelial cells lose cell-cell adhesion and behave more like fibroblasts. Such a mechanism is thought to be crucial in invasion and metastasis of cancer. Non-surprisingly, the cell adhesion molecule E-cadherin is repressed during this process.

The authors examined the expression of E-cadherin and its repressors; Snail, Slug and Twist in squamous esophagus, Barrett’s metaplasia and esophageal adenocarcinoma. It revealed that Slug was increased in adenocarcinoma and in OE33 cells, the increase of Slug expression was inversely correlated to E-cadherin expression (mRNA and promoter activity) and induced EMT. The results are interesting and may represent a molecular mechanism of esophageal carcinogenesis.