SEARCH

SEARCH BY CITATION

Keywords:

  • Snail;
  • Epithelial–Mesenchymal Transition;
  • Epidermal Growth Factor Receptor;
  • Alcohol;
  • Cancer

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References

Background:  Alcohol consumption is associated with the risk of progressive cancers including colon and breast cancer. The mechanisms for the alcohol-induced aggressive behavior of these epithelial cancer cells have not been fully identified. Epithelial–mesenchymal transition (EMT) is a developmental program recently shown to play a role in cancer progression and metastases. We hypothesized that alcohol might promote cancer progression by inducing EMT in cancer cells and tested this hypothesis by assessing alcohol-stimulated changes in phenotypic markers of EMT as well as the EMT transcription factor Snail and its related cell signaling.

Methods:  Colon and breast cancer cell lines and a normal intestinal epithelial cell line were tested as well as colonic mucosal biopsy samples from alcoholic subjects. Cells were treated with alcohol and assessed for EMT-related changes using immunofluorescent microscopy, western blotting, reporter assays, RT-PCR, and knockdown of Snail with siRNA.

Results:  We show alcohol upregulated the signature EMT phenotypic marker vimentin as well as matrix metalloprotease (MMP)-2, MMP-7, and MMP-9 and cell migration in colon and breast cancer cells—all characteristics of EMT. Alcohol also stimulated nuclear localization of Snail phosphorylated at Ser246, transcription from a Snail reporter plasmid, and Snail mRNA expression by RT-PCR. Snail siRNA knockdown prevented alcohol-stimulated vimentin expression. In vivo, Snail expression was significantly elevated in colonic mucosal biopsies from alcoholics. Also, we found alcohol stimulated activation of epidermal growth factor receptor (EGFR) signaling and an EGFR inhibitor blocked alcohol-induced cell migration and Snail mRNA expression.

Conclusions:  Collectively, our data support a novel mechanism for alcohol promoting cancer progression through stimulating the EMT program in cancer cells via an EGFR-Snail mediated pathway. This study reveals new pathways for alcohol-mediated promotion of cancer that could be targeted for therapy or prevention of alcohol-related cancers.

Alcohol (ethanol) consumption is a risk factor for development as well as progression of many cancers (Boffetta and Hashibe, 2006; Seitz and Stickel, 2007). Chronic alcohol consumption is estimated to be directly responsible for about 4% of all cancers worldwide (Boffetta and Hashibe, 2006). Cancers promoted by alcohol include cancers of the upper aerodigestive tract (oropharynx, larynx, esophagus), liver, breast, colon, and rectum (Boffetta and Hashibe, 2006; Seitz and Stickel, 2007). The mechanisms underlying alcohol promotion of these cancers are not certain. However, several distinct alcohol-induced cancer promoting mechanisms have been postulated. These include alcohol mutagenic effects, changes in retinoic acid, folate, or estrogen metabolism, and oxidative stress (Seitz and Stickel, 2007). But, none has been shown to fully explain the mechanism of alcohol-induced aggressive behavior of cancer cells that ultimately results in metastasis, cancer occurrence, and poor outcome.

One cellular pathway involved in cancer progression that has received much recent attention is epithelial–mesenchymal transition (EMT) (Hugo et al., 2007; Thiery, 2002). EMT is a cellular program important during normal development (Nieto, 2002). However, recent studies have shown its importance in cancer progression and metastases (Hugo et al., 2007; Thiery and Sleeman, 2006; Turley et al., 2008). During EMT, epithelial cells lose tight junctions and acquire a more migratory and invasive (mesenchymal) phenotype. These changes include downregulation of junctional proteins including E-cadherin and expression of mesenchymal proteins like vimentin. In addition cells express matrix metalloprotease (MMP) enzymes that dissolve the extracellular matrix and promote cell migration and cancer cell invasion and metastases (Egeblad and Werb, 2002; Thiery and Sleeman, 2006). One key transcription factor regulating EMT is Snail (Nieto, 2002). Others include Snail2, ZEB1/2, and Twist with WNT/β-catenin signaling also being a critical component of EMT (Peinado et al., 2007; Thiery and Sleeman, 2006). Signaling through several growth factor receptors including the epidermal growth factor receptor (EGFR) can induce EMT and Snail expression (Ackland et al., 2003; Peinado et al., 2007).

Because of the important role of EMT in cancer progression, we hypothesized that alcohol abuse may worsen the outcome of epithelial-derived cancers such as colon and breast cancer, by stimulating EMT in cancer cells and thus promoting cancer progression and metastases. The aim of our study was to test our hypothesis by determining if alcohol, at relevant in vivo levels, stimulates key phenotypic and functional signatures of EMT associated with cancer progression in colon and breast cancer cell lines as well as in a nontransformed (normal) intestinal epithelial cell line. We also sought to investigate the signaling mechanism of alcohol-induced promotion of EMT. To this end, we determined the effects of alcohol on activation and mRNA expression of the key EMT Snail transcription factor as well as upstream signaling pathways related to EGFR signaling that might be turning on Snail and EMT. Also, in order to determine the potential relevance of our in vitro findings, we measured Snail mRNA expression in sigmoid colon mucosal samples from alcoholic subjects.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References

Reagents

Alcohol (ethanol) solutions were made daily (AAPER Alcohol and Chemical Co., Shelbyville, KY). TACE inhibitor: TAPI-2 (TNF-α Protease Inhibitor-2) (Chemicon, Temecula, CA). EGFR inhibitor: AG1478 (Calbiochem, San Diego, CA). All EGFR, MAPK (ERK1 and ERK2), and Histone H3 Abs were from Cell Signaling Technology (CST) (Danvers, MA). ERK1/2 (MEK1) inhibitor: PD98059 (Calbiochem). β-actin, MMP-2, MMP-7, and MMP-9 Abs: Sigma-Aldrich (St Louis, MO) and MMP-7 positive control (Chemicon). Phospho-Snail Ser246 Ab: Abcam (Cambridge, MA), and total Snail Ab: CST. TNF-α: R&D Systems (Minneapolis, MN). Vimentin mouse Ab: Santa Cruz Biotechnology (Santa Cruz, CA) and secondary Abs were anti-mouse Alexa fluor 488 and anti-rabbit Alexa fluor 594 (Invitrogen, Carlsbad, CA).

Cell Culture

Caco-2 cells (ATCC #CRL2101, human colorectal adenocarcinoma); MCF-7 (ATCC #HTB-22) human breast adenocarcinoma; MDA-MB-231 (ATCC #HTB-26) human breast adenocarcinoma; and IEC-6 (ATCC #CRL-1592) rat normal intestine epithelium were grown in 6-well plates (Corning, Corning, NY). Cells were cultured at 37°C/5% CO2 in DMEM/10% fetal bovine serum media with 5 mM penicillin-streptomycin. Caco-2 only: 0.01 mg/ml human transferrin. MCF-7 only: 0.1 mg/ml insulin (Invitrogen).

Treatment of Cells With Alcohol

Cells were stimulated once with alcohol (at the indicated concentrations) at the start of each time course experiment or once each day (with fresh media) for the 4-day studies. Alcohol concentrations in media were determined with the alcohol testing kit (Pointe Scientific, Canton, MI). For example in the long-term (4-day) studies, we found that after addition of alcohol each day alcohol concentration declined from 0.2% (43 mM) over 24 hours to reach undetectable (zero) concentration. Thus, in order to expose the cells to alcohol during the 4-day experiments, we change media each morning and added fresh media with alcohol every day during the 4-day experiments. This paradigm models cellular alcohol exposure through alcohol in the blood circulation during daily and regular drinking by alcoholics. Nuclear extracts were prepared using a nuclear extraction kit (Pierce Biotechnology, Rockford, IL). Signaling experiments were terminated by removal of media and addition of PBS for scrapping, SDS/RIPA buffer for whole cell lysates, or nuclear extract kit buffer.

Immunofluorescent Staining and Analysis of Cells

Cells grown on glass coverslips in complete medium in 6-well plates were treated with alcohol (0.2% [43 mM]) for the indicated times in 3 separate experiments. Cells were then washed with PBS, fixed/permeabilized with paraformaldhyde/triton x-100, and stained with primary and fluorescent secondary Abs as described (Forsyth et al., 2007). For the TACE/ZO-1 images (Fig. 6C) (20 × 1 μM z-stacks for 3D) were taken. All images were taken using a Zeiss Axiovert 100 microscope with Axiovision software (Carl Zeiss Inc., Thornwood, NY) using an oil immersion 40× objective as described (Forsyth et al., 2007). The plane of focus selected was the one that revealed the greatest overall clarity for any given image. Quantitative analysis of immunofluorescent staining for vimentin (Fig. 1) was performed with Image J software (NIH) using blinded selection and analysis to prevent bias. Three separate images from 3 separate experiments were analyzed by determining immunofluorescent intensity in 5 areas from each of the 3 images (thus N = 15) for each condition. Background values were subtracted using images for cells treated only with secondary antibodies. Means were compared for control and alcohol-treated cells for each cell type by t-test using SPSS (SPSS, Inc., Chicago, IL) with p ≤ 0.05 being significant.

image

Figure 6.  (A,B) Alcohol stimulates transactivation of the EGFR and ERK1/2 signaling. Caco-2 and MDA-MB-231 cells made serum free for 24 hours were stimulated with alcohol (0.2%, 43 mM) for 10 minutes for p-EGFR and 30 minutes for the p-ERK1/2 analysis. Some cells were preincubated with an inhibitor of the EGFR (AG1478, 500 μM) or TACE (TAPI-2, 20 μM). Western blots of cell lysates were developed with antibody to phospho-EGFRTyr1068 or phospho-ERK1/2 for Caco-2 cells (A) and for MDA-MB-231 cells (B). Blots were stripped and reprobed with Ab to total EGFR or ERK1/2 to show equal loading. Blots are representative of 3 experiments. (C) Alcohol stimulates translocation of TACE to the cell surface in colon and breast cancer cells. Caco-2 and MDA-MB-231 cells grown on glass coverslips were stimulated with alcohol (0.2%, 43 mM) for 30 minutes. Cells were then fixed and permeabilized and stained with Ab to TACE or ZO-1 and then fluorescent secondary Abs to TACE (red) or ZO-1 (green). 20 × 1 μm z-stack images were taken (40× oil immersion) and deconvoluted with Zeiss Axiovision software to yield the 3D images shown. Images are representative from 3 experiments. Bars = 20 μm. (D) Alcohol stimulates EGFR-dependent migration of breast cancer cells. MDA-MB-231 breast cancer cells in serum free medium (5 × 105 cells/well) were added to the upper well of 12-well/8 μm format PET transwells. Then 100 ng/ml TNF-α (T) was placed in the bottom well and migration measured after 24 hours. For some wells ethanol (E, 0.2%, 43 mM) was added before adding cells ± specific inhibitors of the EGFR (AG1478, 500 μM); TACE (TAPI-2, 20 μM); or ERK1/2 (MEK) (PD98059, 30 μM). Data are means of cells per hpf (40×) from duplicate wells of a representative of 3 experiments (*p ≤ 0.05 for lane 2 vs. lane1, lane 3 vs. lane 2, and lanes 4, 5, 6 vs. lane 3).

Download figure to PowerPoint

image

Figure 1.  Alcohol stimulates upregulation of the signature EMT mesenchymal marker protein vimentin in breast and colon cancer cells and in normal intestinal epithelial cells. The human cell lines Caco-2 (colon cancer) (A) and MDA-MB-231 (B), and MCF-7 (C) (both breast cancer) as well as the rat normal intestinal epithelial cell line IEC-6 (D) were grown on glass coverslips in 6-well plates in complete medium. Treated cells were stimulated with alcohol (0.2%, 43 mM) for 4 days and then Control (A, no alcohol) and alcohol-treated (B) cells were fixed and stained with Ab to vimentin and Alexa fluor 488 (green) fluorescent secondary Ab. Images (40× oil immersion) are representative from duplicate wells in 3 experiments. Arrows note increased cytoplasmic vimentin staining. Scale bar is 20 μm.

Download figure to PowerPoint

Cell Migration Assay

Cell migration assays were performed for 24 hours as previously described (Forsyth et al., 2001). 1 × 105 cells (MDA-MB-231) in serum-free media (± alcohol ± inhibitors) were added to the upper well of PET membrane transwells (12 well/8 μM pore; Falcon, Becton Dickinson, Franklin Lakes, NJ). TNF-α (100 ng/ml) was then added to the lower well. For some wells ethanol (E, 0.2%, 43 mM) was added before adding cells ± specific inhibitors of the EGFR (AG1478, 500 μM); TACE (TAPI-2, 20 μM); or ERK1/2 (MEK) (PD98059, 30 μM). Assay was stopped by removing cells from the upper well with a cotton swab and inserts fixed and stained with crystal violet to visualize cells. Data are mean ± SE cells per high power field (40× objective) from 5 fields per insert for duplicate wells in a representative from 3 experiments.

Human Subjects and Sigmoid Mucosal Samples

Endoscopic Pinch sigmoid mucosal biopsies were obtained from 4 healthy and 4 alcoholic male subjects. These samples were randomly selected from our IRB approved tissue repository. Alcoholic subjects were recruited as part of our ongoing project studying the effects of chronic alcohol consumption on intestinal function. All alcoholic subjects fulfilled NIAAA criteria for at-risk drinking and alcohol abuse or dependence (O’connor and Schottenfeld, 1998) and DSM-IV criteria for alcoholism (Ball et al., 1997). This included a recent drinking history (regular and heavy EtOH consumption for a minimum of 3 months prior to enrollment). These 4 alcoholics were actively drinking for the last 10 years and their last drink was 7 to 10 days prior to the endoscopic procedure. However, blood alcohol level was undetectable at the time of the procedure. A history of alcohol consumption was assessed by a validated NIAAA-endorsed assessment instrument—The Lifetime Drinking History (Skinner and Sheu, 1982). Control subjects were otherwise healthy individuals with no known liver disease, who did not fulfill any of the exclusion criteria, and who were willing to participate in the study. None of the controls had ever been a daily or heavy drinker.

Subjects recruited were excluded if they reported gastrointestinal diseases (except for hiatal hernia or hemorrhoids), clinically detectable ascites or severe edema, evidence of ongoing infection, major depression or anxiety requiring therapy, clinically significant lung or heart disease, or regular use of NSAIDs. All subjects had normal liver function tests and had no clinical evidence of liver disease.

Sigmoid biopsy samples were taken during an unprepped sigmoidoscopy and mucosal samples were snap frozen in liquid nitrogen, transferred to the laboratory and stored in −80°C freezer. The procedure was performed after a Rush IRB approved informed, written consent.

Western Blotting Analysis

Western blotting and densitometry analysis with Image J software (NIH) was performed with cell lysates equalized for protein or cell culture supernatants (conditioned media) equalized for total protein and cell number as previously described (Forsyth et al., 2002, 2007). To our knowledge there is no loading control marker for serum-free cell culture conditioned media for Fig. 2, all other western blotting data was normalized to densitometry values for the loading controls.

image

Figure 2.  Alcohol stimulates production of MMP-2, MMP-7, and MMP-9 characteristic for EMT. (AC) The cell lines Caco-2, MDA-MB-231, and IEC-6 (all 1 × 106 cells/well) were cultured in 6-well plates and made serum free for 24 hours before stimulation overnight (24 hours) with alcohol (0.2%, 43 mM) in 1 ml of serum free media. Media was removed and 50 μl/lane from control (no alcohol) and alcohol-treated cells was loaded onto SDS–PAGE gels and MMP proteins were determined by western blot with Abs to MMP-2, MMP-9, and MMP-7 (lower). (D) Caco-2 cells were treated as above for 24 hours with differing doses of alcohol (0.1% [22 mM], 2% [434 mM]) and MMP-7 measured by western blot as above. Far right “+C” lane is positive control for MMP-7 (densitometry histogram: ProMMP-7 = black, activated MMP-7 = grey). All blots are representative of 3 experiments. *p ≤ 0.05 for all figures versus control.

Download figure to PowerPoint

Isolation of RNA and RT-PCR

Isolation of cell total RNA, reverse transcription, and real time PCR was carried out as previously described (Forsyth et al., 2007). Primers for PCR were: Beta actin (157 bp): F:GCCAGGTCCAGACGCAGG;R:TGCTATCCAGGCTGTGCTA; Snail F:(142 bp):ACCCCACATCCTTCTCACTG; R:TACAAAAACCCACGCAGACA; MMP-7(138 bp) F:GAGTGCCAGATGTTGCAGAA, R:AAATGCAGGGGGATCTCTTT. Snail and MMP-7 Ct data were normalized to beta-actin and then to Snail or MMP-7 mRNA expression in the control to arrive at the fold expression value.

SiRNA Snail Inhibition Studies

Smartpool short inhibitory RNA (siRNA) for human Snail (Snail1, SNAI1) was purchased from Dharmacon (Dharmacon Inc., Lafayette, CO). Caco-2 cells grown to 60% confluence in 6-well plates were transfected with 100 nM Snail siRNA or control siRNA (Santa Cruz Biotechnology), using Lipofectamine (Invitrogen) and Optimem (Invitrogen) media (Forsyth et al., 2007). At 2 days post siRNA transfection cells were treated with alcohol (0.2%, 43 mM) for 4 additional days and then tested for expression of Snail and vimentin proteins by western blotting. Blot shown is representative from 3 experiments while densitometry data are means ± SE from all 3 experiments.

Snail and E-cadherin Reporter Assays

Caco-2 colon cancer cells grown in 96-well plates were transfected with 100 ng/well of Snail reporter plasmid (gift of A.G. de Herreros) (Barbera et al., 2004) or CDH1::LUC E-cadherin reporter (gift of L. Larue) (Julien et al., 2007). Values were read 8 hours after initial stimulation with 0.2% (43 mM) alcohol for the Snail reporter and after 24 and 48 hours for the E-cadherin reporter. Cells were co-transfected with 1 ng SV40 renilla luciferase plasmid (Promega, Madison, WI) and developed using the Stop and Glow kit (Promega), read on a Lumioskan Ascent plate reader (Thermo Labsystems, Waltham, MA), and data normalized to renilla values to control for transfection efficiency.

Data Analysis

Results are expressed as means ± SE. Differences among multiple means compared to control were determined by ANOVA followed by Dunnett post hoc test. Differences between 2 groups [control vs. one experimental group] were assessed using Student’s t-test. Significance for all tests was defined as p ≤ 0.05. All statistical analyses were performed using SPSS.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References

Alcohol Stimulates Expression of the EMT Signature Mesenchymal Marker Protein Vimentin in Breast and Colon Cancer Cells and Normal Intestinal Epithelial Cells

We hypothesized that alcohol is promoting cancer progression by stimulating EMT in cancer cells. The most widely validated signature biomarker of EMT is the mesenchymal protein vimentin (De Wever et al., 2008). Caco-2 (colon cancer), MCF-7 and MDA-MB-231 (breast cancer), and IEC-6 (nontransformed, normal intestinal) cells grown on glass coverslips were stimulated with 0.2% (43 mM) alcohol containing media or media alone for 4 days. As seen in Fig. 1A–D, alcohol stimulated the expression of vimentin in normal intestinal IEC-6 cells as well as both the colon cancer and breast cancer cells. Mean Image J densitometry fluorescence values for the 4 cell types were: Caco-2 control (64 ± 9) and alcohol treated (119 ± 14); MCF-7 control (62 ± 13) and alcohol treated (211 ± 15); MDA-MB-231 control (101 ± 12) and alcohol treated (196 ± 12); and IEC-6 control (81 ± 4) and alcohol treated (157 ± 6). Thus, the untreated Caco-2, MCF-7, and IEC-6 cells showed low vimentin expression. However, the MDA-MB-231 cells did show higher basal expression of vimentin. High basal expression of vimentin in the MDA-MB-231 cells was also noted by others (Nagaraja et al., 2006). Nonetheless, alcohol stimulated a significant increase in vimentin staining in all 4 cell types (Fig. 1A–D; p ≤ 0.05; control vs. alcohol treated t-tests).

Alcohol Stimulates Production of EMT Characteristic MMP-2, MMP-7, and MMP-9

Three MMPs especially characteristic of the EMT proteome are MMP-2, MMP-7, and MMP-9 (De Wever et al., 2008; Thiery, 2002; Turley et al., 2008). MMPs are secreted as inactive proenzymes that require activation or cleavage by other MMPs to become active (Egeblad and Werb, 2002). The lower molecular weight form is the activated form for each MMP shown (MMP-9:92/88 kd; MMP-7:28/18 kd; MMP-2:72/63 kd). Equal numbers of Caco-2, MDA-MB-231, and IEC-6 cells (1 × 106/well) were stimulated with alcohol in serum free media (24 hours, 1 ml) and the levels of MMP-2 and MMP-9 (upper sections) and MMP-7 (lower sections) were determined in 50 μl aliquots of conditioned media by western blotting as described (Forsyth et al., 2002). As seen in Fig. 2A–C, alcohol robustly and significantly stimulated the production of all 3 MMPs over 24 hours by both transformed, cancer cell lines (colon cancer and breast cancer cells) and the nontransformed (“normal”) intestinal cell line IEC-6. Fold increases for Caco-2, MDA-MB-231, and IEC-6 (respectively) were: MMP-2: 3,6,2; MMP-7: 10,8,6; and MMP-9: 8,8, and 5 fold (p ≤ 0.05 for all MMPs vs. control, t-tests). We also assessed Caco-2 cell MMP-7 expression with ethanol at varying concentrations (0.01% [2 mM]; 0.05% [11 mM]; 0.075% [16 mM]; 0.1% [22 mM]; 0.2% [43 mM]; 1.0% [217 mM]; 2.0% [434 mM]) for 24 hours (Fig. 2D). As shown in Fig. 2D, ethanol stimulated a dose dependent significant increase in MMP-7 secretion compared to control (p ≤ 0.05, AVOVA with post hoc Dunnett test). Remarkably, the peak MMP-7 detected in conditioned media was with alcohol at physiological concentrations (0.1% [22 mM] and 0.2% [43 mM]). RT-PCR analysis of MMP-7 mRNA expression revealed an 8-fold increase for 0.1% (22 mM) ethanol (not shown).

Alcohol Stimulates Phosphorylation of Snail at Ser246 and Snail Localization to the Nucleus

We next sought to determine a mechanism whereby alcohol might be turning on the EMT program. The most studied transcription factor regulating EMT is Snail. One recent study showed that PAK1 phosphorylates (activates) Snail at Ser246 resulting in Snail localization to the nucleus and is required for Snail-mediated promotion of EMT (Yang et al., 2005). To determine whether alcohol promotes EMT through Snail, we performed a time course experiment in which cells were stimulated with alcohol (30 minutes, 1, 2, 4, 6 hours) and nuclear extracts were tested by western blotting with an Ab specific for Snail Ser246. Cytoplasmic extracts were negative (not shown). As seen in Fig. 3A–D, alcohol stimulated significant (p ≤ 0.05, ANOVA with post hoc Dunnett test) time-dependent nuclear localization of Snail phosphorylated at Ser246 in Caco-2 cells, MDA-MB-231, MCF-7, and IEC-6 cells.

image

Figure 3.  Alcohol stimulates phosphorylation of the EMT transcription factor Snail at Ser246 and its localization to the nucleus. (AD) Equal cell numbers of Caco-2, MDA-MB-231, MCF-7, and IEC-6 cells were stimulated with alcohol (0.2% [43 mM]) and nuclear extracts prepared after 30 minutes, 1, 2, 4, and 6 hours. Aliquots equalized for protein were analyzed by western blot with Ab to Snail p-Ser246 and Histone H3 to measure loading. Blots are representative from 4 experiments while the densitometry data shown represents the means of all 4 experiments for each cell type, so some bars are slightly different from the blots shown.

Download figure to PowerPoint

Alcohol Stimulates Increased mRNA Expression From a Snail Reporter and Decreased mRNA Expression From an E-cadherin Reporter

We sought to determine if alcohol also stimulates mRNA expression of Snail. We first assessed the ability of alcohol to stimulate mRNA expression from a Snail reporter plasmid. Caco-2 cells were transfected with a PGL3 luciferase reporter containing the Snail promoter (gift of A.G. Herreros) (Barbera et al., 2004). As seen in Fig. 4A, after 8 hours alcohol stimulated a 2-fold increase in Snail promoter activity compared to control (no alcohol) cells (p ≤ 0.05, t-test). Because Snail is known to repress E-cadherin mRNA expression we also determined the effect of alcohol on an E-cadherin reporter (gift of L. Larue) (Julien et al., 2007). Recent studies show E-cadherin mRNA expression is suppressed at 24 to 48 hours during EMT (Billottet et al., 2008), so we assessed the effects of alcohol on E-cadherin reporter activity at 24 and 48 hours after alcohol stimulation. Figure 4A shows that alcohol stimulation resulted in significant inhibition of E-cadherin reporter activity at 24/48 hours (50%/48%; p ≤ 0.05, t-test) consistent with activation of Snail nuclear activity by alcohol.

image

Figure 4.  (A) Alcohol stimulates increased mRNA expression from a Snail reporter plasmid and repressed mRNA expression from an E-cadherin reporter. Caco-2 cells in complete medium in 96-well plates were transfected with 100 ng/well of the Snail-PGL3 reporter plasmid containing the human Snail promoter (−869/+59) or the CDH1::LUC E-cadherin reporter (−308/+21) as well as a control renilla luciferase plasmid (1 ng/well). Cells made serum free for 24 hours were stimulated with alcohol (0.2% [43 mM]) for 8 hours for the Snail reporter data or for 24 and 48 hours for the E-cadherin data. Firefly luciferase values were determined and normalized to renilla luciferase values to control for transfection efficiency. Data (% Control) are means of duplicate wells from 3 experiments (N = 6). (B,C) Alcohol stimulates mRNA expression of Snail and Snail protein expression in colon and breast cancer cells. Caco-2 colon cancer and MDA-MB-231 breast cancer cells grown in 6-well plates in complete medium were made serum free for 24 hours and then stimulated for 24 hours with alcohol (0.1% [22 mM], 0.2% [43 mM], or 1% [217 mM]). RNA was prepared and Snail mRNA expression normalized to β-actin by qRT-PCR analysis. Data are means from a representative of 3 experiments (upper sections). In addition cell lysates from matched wells were analyzed by western blot for total Snail protein with a representative blot shown from 3 experiments (lower sections). (D) Snail mRNA expression is significantly elevated in the colonic mucosa of heavy drinkers. Pinch sigmoid mucosal biopsies were obtained at 20 to 25 cm from the anus during sigmoidoscopy procedure (N = 4/group). Controls were healthy subjects who had never been a daily or heavy drinker. Alcoholic subjects were selected based on NIAAA and DSM-IV criteria. RNA was prepared and Snail mRNA expression determined by qRT-PCR and data are expressed relative to β-actin.

Download figure to PowerPoint

Alcohol Induces mRNA and Protein Expression of Snail in Colon and Breast Cancer Cells

We used qRT-PCR to assess Snail mRNA expression directly in Caco-2 colon cancer and MDA-MB-231 breast cancer cells stimulated for 24 hours with alcohol. As seen in Fig. 4B,C (upper sections), for Caco-2 cells the values for fold increase in Snail mRNA expression were 0.1–7.5%, 0.2–8.5%, and 1–11.6% fold. To identify possible mechanisms stimulating Snail mRNA expression Caco-2 cells were also treated with an EGFR inhibitor AG1478 + 1% alcohol which blocked stimulation by 75% (2.8 fold vs. 11.6 fold). For MDA-MB-231 cells alcohol stimulated Snail mRNA expression by 0.2–5.4% and 1–8.2%. Total Snail protein in cells stimulated under the same conditions (lower sections 4B and 4C) were: Caco-2: 0.1–4.8%, 0.2–3.5%, 1.0–4.1%, and 1.0%+ EGFR inhibitor 0.8 (inhibited 80%) and for MDA-MB-231: 0.2–7.8%, 1.0–8.5% (p ≤ 0.05 for all increases and the EGFR inhibition, ANOVA with Dunnett post hoc test).

Snail mRNA Expression is Significantly Elevated in the Colonic Mucosa of Heavy Drinkers

To investigate the potential relevance of our in vitro findings in vivo, we sought to determine Snail mRNA expression in the colonic mucosa of chronic (active) heavy alcohol drinkers. We used qRT-PCR to assess Snail mRNA expression using RNA prepared from intestinal biopsies from normal vs. alcoholic subjects. As shown in Fig. 4D, Snail mRNA expression was significantly higher (83% higher) in colonic biopsies from chronic heavy drinkers (NIAAA criteria) as compared to controls (N = 4/group, p ≤ 0.05, t-test).

Snail Knockdown with siRNA Inhibits Alcohol-Stimulated Vimentin Expression

We investigated whether knockdown of Snail expression with siRNA would prevent alcohol-stimulated EMT (vimentin expression). Caco-2 cells were treated with alcohol (0.2%, 43 mM) for 4 days. Some cells were pretreated (48 hours) with an irrelevant siRNA control or siRNA (100 nM final) specific for Snail as described (Forsyth et al., 2007). Western blots were used to assess Snail and vimentin protein expression. As seen in Fig. 5, treatment with alcohol caused an increase in vimentin expression in media treated as well as control siRNA treated cells (lanes 2&3). However, Snail knockdown (48% mean knockdown) with specific siRNA prevented alcohol-induced vimentin expression (lane 4; p ≤ 0.05 vs. siRNA control, t-test). Note that these data are for total Snail only to assess knockdown. Our previous data (Fig. 3A–D) show alcohol stimulates nuclear localization of Ser246-phosphorylated Snail.

image

Figure 5.  SiRNA knockdown of Snail prevents alcohol-induced vimentin expression. Caco-2 cells were treated with either media alone (lane 1), EtOH alone (lane 2), EtOH with control (scrambled) siRNA (lane 3), or EtOH with siRNA specific for Snail (lane 4). After 48 hours, cells were then treated with alcohol for 4 days as in Fig. 1 and then vimentin and total Snail proteins were determined by western blot. Blot is a representative from 3 experiments; histogram shows summarized data from all 3 experiments.

Download figure to PowerPoint

Alcohol Stimulates Transactivation of EGFR Signaling

We attempted to identify upstream signaling mechanisms for alcohol activating PAK1 phosphorylation of Snail and Snail mRNA expression. We have previously shown oxidative stress activates EGFR signaling in Caco-2 cells (Forsyth et al., 2007). EGFR signaling can regulate EMT and cancer as well as PAK1 mediated phosphorylation of Snail and stimulation of Snail mRNA expression via ERK1/2 signaling (Barbera et al., 2004; Yang et al., 2005). In Fig. 6, the data show alcohol (0.2%, 43 mM) stimulates phosphorylation 5- to 10-fold of the EGFR and downstream ERK1/2 in both colon cancer (Fig. 6A) and breast cancer (Fig. 6B) cells. Alcohol-induced EGFR and ERK1/2 activation was significantly blocked by an EGFR inhibitor, AG1478 (500 nM, 85% inhibition), or inhibition of the metalloprotease tumor necrosis factor converting enzyme-TACE (ADAM17) (TAPI-2, 20 μM; 85% inhibition). This supports the requirement for TACE-EGFR transactivation by alcohol for EGFR and ERK1/2 activation as we have shown for oxidative stress (Forsyth et al., 2007).

To further investigate a role for TACE in alcohol-mediated EGFR transactivation, we used 3D microscopy of deconvoluted z-stack images (20 × 1 μm) of Caco-2 or MDA-MB-231 cells to identify changes in TACE location in the cells (Fig. 6Ca–d). Translocation of TACE to the apical surface from a perinuclear location is associated with EGFR transactivation and cleavage of EGFR proligands including TGF-α (Blobel, 2002; Forsyth et al., 2007). Although TACE inhibitors block EGFR activation at the 10-minute timepoint, we observed maximum translocation (surface staining) of TACE after 30 minutes. Therefore, we show data for 30 minutes in Fig. 6C in which alcohol clearly stimulates translocation of TACE to the apical (Caco-2) or outer (MDA-MB-231) cell surface further supporting a specific role for TACE in alcohol transactivation of the EGFR.

Alcohol Stimulates EGFR-Mediated Migration of Cancer Cells

One hallmark of Snail expression and EMT is increased cell migration/invasion (Nieto, 2002; Thiery, 2002). We wished to assess whether ethanol stimulated migration of cancer cells. We used MDA-MB-231 breast cancer cells because our data show alcohol stimulates MMPs and Snail activation and expression in these cells (see above). As shown in Fig. 6D, cells stimulated with TNF-α alone showed a significant 2.6-fold migration versus control (p ≤ 0.05, ANOVA with Dunnett post hoc test). Alcohol stimulated an additional 3-fold increase in migration through inserts together with TNF-α (100 ng/ml). This migration was blocked by inhibitors of either the EGFR (AG1478, 500 nM, 83%inhibition), or TACE (TAPI-2, 20 μM, 82% inhibition), or ERK1/2 (PD98059, 30 μM, 88% inhibition) (p ≤ 0.05, ANOVA with Dunnett post hoc test).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References

In the current study, we provide evidence that alcohol induces EMT in cancer cells and this alcohol effect is mediated through Snail activation. EMT is an active focus of current cancer research with a large body of in vitro, animal, and patient evidence strongly supporting a key role for EMT in cancer progression and metastases (Hugo et al., 2007; Thiery, 2002; Thiery and Sleeman, 2006; Turley et al., 2008; Weinberg, 2008). In addition to breast and colon cancer, evidence supports a role for EMT in oral, nasopharyngeal, esophageal, gastric, rectal, cervical, ovarian, thyroid, pancreatic, and prostate cancer to name only a partial list (Hugo et al., 2007; Natalwala et al., 2008; Turley et al., 2008). Recent animal experiments demonstrated direct in vivo evidence for EMT in breast cancer progression (Trimboli et al., 2008). In colon cancer, it has been shown that EMT (vimentin) is associated with early adenoma progression in the APCMin mouse model (Chen et al., 2008). We now, for the first time, provide compelling evidence to support our hypothesis that alcohol promotes breast and colon cancer migration by stimulating the EMT program in cancer cells.

Several prior studies with alcohol in different cell types are relevant to our findings. Two studies describe EMT changes by alcohol treatment in an immortalized keratinocyte cell line (9 weeks alcohol treatment) (Chamulitrat et al., 2003) or a breast epithelial cell line (7 days of alcohol treatment) (Robson et al., 2006). Specific mechanisms for EMT were not investigated in either study. In other studies alcohol was found to stimulate cell migration of breast cancer cells (Ke et al., 2006) and endothelial cells (Morrow et al., 2008). Alcohol has also been shown to both activate and inhibit TACE-mediated processing of TNF-α (Zhao et al., 2003, 2004) and to inhibit EGFR signaling or enhance ErbB2 cell migration (Ma et al., 2003, 2005).Thus, although limited in their investigation of possible mechanisms, these previous studies support our hypothesis for induction of EMT characteristics by alcohol stimulation.

In the present study, we investigated the effects of alcohol in breast and colon cancer cell lines because both of these cancers are promoted by alcohol use. Our data show that alcohol stimulates expression of the signature mesenchymal protein vimentin in breast and colon cancer cells. We also show that Snail knockdown inhibits this alcohol-induced vimentin expression in Caco-2 cells. Epithelial cells express cytokeratin intermediate filaments, while these filaments are composed of vimentin in mesenchymal cells. (De Wever et al., 2008). Vimentin is the most widely used and well validated signature phenotypic marker of EMT (De Wever et al., 2008; Turley et al., 2008). Vimentin expression has been identified in several studies of breast, colon, and other cancers as being associated with increased invasiveness and poor prognosis (De Wever et al., 2008).

Our data for both breast and colon cancer cells shows that alcohol stimulates nuclear localization and phosphorylation of Snail at Ser246, probably by PAK1(Yang et al., 2005). In addition we show alcohol stimulates mRNA expression from a Snail reporter plasmid and Snail mRNA expression in colon and breast cancer cell lines. Snail has been called a master regulator of EMT (Peinado et al., 2007). A hierarchical model of transcriptional regulation of EMT involves other key transcription factors that have been identified including Snail2, ZEB1/2, and Twist but places Snail at the top of this hierarchy (Peinado et al., 2007). Overexpression of Snail alone is sufficient to induce EMT in cell lines in vitro (Cano et al., 2000; Julien et al., 2007) and recently to generate possible cancer stem cells (Mani et al., 2008). In transgenic mice, chronic Snail activation results in multiple epithelial and mesenchymal tumors (Perez-Mancera et al., 2005). In another mouse model, Snail expression determined breast cancer invasiveness and recurrence (Moody et al., 2005). Overexpression of Snail has also been shown to correlate closely with invasiveness and recurrence of breast and colon cancer as well as other cancers in patient studies (Blanco et al., 2002; Natalwala et al., 2008; Roy et al., 2005). We now show, for the first time, that alcohol also promotes EMT in cancer cells through activation of Snail. We propose that Snail-mediated promotion of EMT in epithelial cancer cells by alcohol may be one mechanism for alcohol promotion of cancer progression and poor outcome of colon and breast cancers in alcoholics.

Relevant to our MMP data, Snail has been shown to directly stimulate MMP-2 and MMP-9 expression (Jorda et al., 2005; Yokoyama et al., 2003). Knockdown of Snail in a mouse tumor model resulted in fewer tumors at least in part due to reduced MMP-9 (Olmeda et al., 2007). MMP-3 or MMP-9 can also further promote EMT via Snail (Radisky et al., 2005). In addition, MMP-7 is a target of β-catenin/TCF signaling which is another pathway important in EMT also tied to Snail (Brabletz et al., 1999; Peinado et al., 2007). Increased Snail activity can stimulate β-catenin indirectly by repressing E-cadherin or directly by interacting with β-catenin (Stemmer et al., 2008). Our data supports a mechanism in which alcohol stimulation of activated Snail nuclear localization as well as increased Snail mRNA expression results in promoting EMT and cancer progression by increasing expression of vimentin as well as MMP-2, MMP-7, and MMP-9 in cancer cells.

Experiments in the APCMin mouse model of colon cancer also support our hypothesis for alcohol promoting cancer progression through stimulation of Snail and EMT. Snail is overexpressed in human colon cancer (Roy et al., 2005) and adenoma formation in APCMin mice is inhibited by Snail knockdown (Roy et al., 2004). Significantly, adenoma formation in these mice is enhanced by drinking alcohol (Roy et al., 2002) and recently this same adenoma formation in APCMin mice was shown to occur by an EMT-dependent mechanism with early expression of vimentin (Chen et al., 2008).

Our data suggest that alcohol stimulation of EGFR signaling may promote activation of Snail and EMT. Abnormalities in EGFR signaling or expression are associated with many human cancers (Hynes, 2007). Several different cell signaling pathways have been shown to stimulate induction of EMT including signaling by receptor tyrosine kinases such as the EGFR (Peinado et al., 2007; Thiery and Sleeman, 2006). Chronic stimulation with EGF can result in activation of Snail and EMT (Ackland et al., 2003; Peinado et al., 2007). Blocking EGFR signaling inhibits alcohol-stimulated Snail mRNA expression in our study and Snail mediated colon cancer metastases in mice (Mann et al., 2006). EMT in cervical cancer also correlates with EGFR and Snail overexpression (Lee et al., 2008). Our data suggest alcohol activates EGFR signaling through TACE-mediated transactivation. TACE-mediated EGFR transactivation promotes EMT-mediated cancer progression in breast epithelial cells (Kenny and Bissell, 2007). Our data also show alcohol-stimulated ERK1/2 activation through the EGFR. Overexpression of ERK1/2 also promotes EMT and promoter analysis of Snail reveals a key stimulatory role for ERK1/2 (Barbera et al., 2004; Schramek et al., 2003). In addition, our data show that alcohol stimulates phosphorylation/activation of Snail at Ser246 probably by PAK1 (Yang et al., 2005). EGFR signaling has also been shown to activate PAK1 (Kumar et al., 2006). In further support of a role for EGFR-EMT signaling in our model, an EGFR inhibitor also inhibited alcohol stimulated cell migration in our study. Therefore data from others as well as our own data support a role for alcohol-stimulation of EGFR signaling in activation of Snail nuclear localization and mRNA expression and subsequent promotion of EMT.

It should also be noted that several epidemiological studies suggest that alcohol abuse not only promotes cancer progression and metastasis, it may also increase the risk of certain cancers like squamous epithelial cancer of esophagus, head and neck cancer and also breast and colon cancer (Seitz and Stickel, 2007). Furthermore, several experimental studies suggested that Snail and EMT are not only involved in cancer progression but also in cancer initiation (Hugo et al., 2007; Thiery, 2002; Thiery and Sleeman, 2006; Turley et al., 2008; Weinberg, 2008). For example, overexpression of Snail in transgenic mice results in multiple tumor types and EMT itself is associated with increased genomic instability which in turn could contribute to cancer initiation as well as cancer progression (Radisky et al., 2005). Our findings in normal, nontransformed intestinal cells that alcohol promotes EMT in noncancer cells support the notion that alcohol-induced stimulation of EMT pathway in normal epithelial cells may be a mechanism for increased risk of cancers in alcoholics. Further studies are needed to confirm our findings in different noncancer cell lines and primary epithelial cells.

In summary, collectively our data present evidence for a novel pathway for alcohol promotion of cancer progression through induction of the EMT gene program in cancer cells. To our knowledge these data are the first to show that alcohol activates and increases mRNA expression of the key EMT transcription factor Snail or to show EGFR transactivation and induction of MMPs by alcohol in epithelial cells. Together with the vimentin, E-cadherin reporter, and cell migration (a surrogate marker for metastasis) data our study strongly supports the hypothesis that alcohol induces the EMT program in cancer cells and promotes metastasis. The mechanisms for exactly how EMT promotes cancer progression are currently being unraveled. Future studies will be needed to identify the individual roles of EGFR, MMP, and Snail signaling in alcohol induction of EMT as well as the possible roles for other EMT transcription factors. Identification of this novel mechanism for alcohol promotion of cancer progression could open new avenues for prevention and treatment of alcohol-related cancers.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References

We thank Lionel Larue and Antonio Garcia de Herreros for providing reagents. This study was supported in part by NIH funding through grant AA013745 (to A.K.).

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Conflict of interest
  8. References