SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

One of the most fundamental biological processes in tumor metastasis is the process of epithelial–mesenchymal transition (EMT). During EMT, zinc-finger-family of transcription factors such as Snail, Slug and Twist, and matrix metalloproteinases (MMPs) are upregulated, and this correlates with increased tumor cell invasion and motility. We previously obtained a highly invasive A431-III tumor subline, which is a rich source of MMP-9 and observed a plausible link between MMP levels and the promotion of EMT. To gain further understanding of EMT, we investigated the contribution of distinct MMPs to the induction of EMT. Exposing A431, cervical carcinoma parental cells, to MMP-9 stimulated a phenotypic alteration and cells became spindle-like as shown for A431-III cells. In the present communication, we document changes in gene expression profiles of A431-P and A431-III cells, including those of genes involved in cell adhesion, cytoskeleton reorganization, polarity, migration and transcription. Treatment of both A431-P and A431-III cells with GM6001, a broad spectrum MMP inhibitor, resulted in the diminution of vimentin and fibronectin, indicating a role for MMP-9 in the induction of EMT. Abrogation of MMP-9-mediated cell–cell contact in both A431-P and A431-III cells using MMP-9 siRNA resulted in decreased cell invasion, motility and altered cytoskeleton arrangement together with a reduction in Snail expression. Knockdown of Snail resulted in similar changes along with diminished MMP-9 expression. These data suggest a higher capacity of MMP-9 than that of Snail in eliciting the development of EMT in A431 cells. Based on these findings, we speculate that the overexpression of MMP-9 in A431-III cells might directly induce (or stimulate) EMT and that the transcriptional factor, Snail, could cooperatively engage in this phenomenon. (Cancer Sci 2011; 102: 815–827)

The spread of cancer through metastasis is considered to be responsible for the majority of cancer mortalities.(1) Alteration in matrix remodeling-related proteolysis in cancers is linked to unregulated tumor growth and cancer cell invasion and metastasis. Matrix metalloproteinases (MMPs) are the most noteworthy proteolytic enzymes that are associated with tumorigenesis.(2) The MMPs belong to a rapidly growing family of zinc-dependent endopeptidases that currently includes more than 25 members classified as collagenases, gelatinases, stromelysins, matrilysins and membrane-associated MMP.(3,4) Enhanced levels of certain MMPs are associated with cancer growth and are regarded as prime candidates functioning during tumor invasion, metastasis and angiogenesis, and, in some instances, overexpression correlates with poor clinical outcomes.(4) It is believed that MMPs degrade the extracellular matrix (ECM) and thus enable tumor cells to migrate, invade and spread to various secondary sites, where they form metastases.(5,6) Tumor cells require the action of more than one MMP and more general degradative enzymes to cross the tissue barriers they encounter in the process of invasion. MMPs regulate the tumor microenvironment, and their expression and activation is increased in almost all human cancers compared with the respective normal tissues.(3,7)

Emerging evidence indicates that MMPs can stimulate processes associated with epithelial–mesenchymal transition (EMT) with enhanced tumor cell invasion and metastasis potential.(8) Most reports suggest that the predominance of MMP-2, -3 and -9 proteins correlate with worse prognosis. These specific proteinases can disrupt cell adhesion by processing the components of cell–cell and cell-ECM contacts, and interfere with the function of other full-length E-cadherin molecules.(9,10) E-cadherin is a transmembrane glycoprotein involved in calcium-dependent intercellular adhesion, and is specifically associated with epithelial cell-to-cell adhesion.(11) MMP procession of E-cadherin contributes to the initiation of EMT, detachment of carcinoma cells and their transfer into the stroma, which allows stationary epithelial cells to become motile.(8,12,13) The differentiated epithelial phenotype is characterized by structural and functional polarization with junctional complex formed mediating intercellular adhesion and polarity. E-cadherin, the major component of adherens junctions, forms homophilic contacts between neighboring cells.(14,15) Aberration of E-cadherin caused by MMPs would interfere with cell adhesion and polarity. Though much progress has been made in recent years toward the understanding of EMT,(16) further study of the controlling mechanisms is of crucial importance. It is not clear as to what extent diverse MMPs are acting as specific inducers of EMT or whether there are functional overlaps between different MMPs.(8) One major reason is our lack of precise knowledge regarding which specific MMPs are mediating pro-tumoural process. Those MMPs associated with EMT clearly constitute one of the most attractive therapeutic targets.

Previously, we have obtained a highly invasive A431-III tumor subline derived from parental A431 (A431-P) tumor cells by three successive passages through a Boyden chamber with EHS matrigel-coated membrane support.(17) The A431-III displayed increased secretion of MMP-3 and -9 (the prominence of which is associated with dismal cancer prognosis) and expressed less E-cadherin.(17) These results indicate that MMPs might associate with the induction of EMT. It is worth noting that downregulation of E-cadherin is a critical step in EMT process. The expression of a number of zinc-finger-family of transcription factors, including Snail, Slug and Twist has been shown to correlate with E-cadherin Downregulation.(18) In addition, Snail and Slug function to promote cell fate changes during development, leading to the production of mesenchymal cells with greater migratory potential.(19) Snail has been found to be expressed at the invasive front of carcinomas and linked to lymph node metastasis and tumor relapse, thus implying its important role in tumor progression.(20) Slug has also been detected at sites of EMT and is known to play a role in the maintenance of semi-differentiated structures.(21) As EMT is triggered by E-cadherin repressors, targeting Snail and Slug members represent a potential tool for intercepting EMT and invasion.(22) Also, the regulation of MMPs by transcription factors (Snail, Ets, β-catenin) have clearly been elucidated.(23)

Since A431-III cells overexpress MMP-9, the goal of this study was to gain insights into the potential role of MMP-9 in the regulation of EMT process. We first characterized the EMT-related markers in both A431-P and A431-III cells, including fibronectin, vimentin, N-cadherin, Snail, Twist and E-cadherin using immunoblotting, RT-PCR and microarray analysis. We then explored the potential role of MMP-9 in EMT by blockade of MMP-9 in A431-III cells using small interference RNA and MMP inhibitor, and overexpression of MMP-9 in A431-P cells. This study reveals that enhanced expression of MMP-9 may facilitate EMT process, and further presents a novel finding the existence of feedback-loop regulation between MMP-9 and Snail.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Materials.  The A431 tumor cell line was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). MMP-9 siRNA, Snail siRNA, and non-specific siRNA were purchased from Invitrogen (Carlsbad, CA, USA). Anti-Snail, anti-Twist and anti-E-cadherin (HECD1) were obtained from Abcam (Cambridge, MA, USA); Anti-N-cadherin was purchased from Abgent (San Diego, CA, USA). Growth factor-reduced EHS matrigel was acquired from BD (Franklin Lake, NJ, USA). Anti-fibronectin and anti-β-actin were purchased from Sigma (St Louis, MO, USA). Anti-vimentin (V9) and anti-β-catenin was obtained from Santa Cruz (Santa Cruz, CA, USA). E-cadherin neutralizing antibody (nAb) was obtained from Invitrogen. Human recombinant MMP-9 was obtained from R&D System (Minneapolis, MN, USA). MMP-9 inhibitor GM6001 was obtained from Millipore (Billarica, MA, USA). PCR forward and reverse primers were purchased from Purigo Biotech (Taipei, Taiwan). Unless otherwise indicated, all other reagents were obtained from Sigma.

Preparation of cell lysates.  Tumor cells were harvested and washed three times with PBS. The cells were then lysed in gold lysis buffer, containing 20 mM Tris–HCl, (pH 7.9), 1 mM EGTA, 0.8% NaCl, 0.1 mM β-glycerylphosphate, 1 mM sodium pyrophosphate, 10 mM NaF, 1 mM Na4P2O7, 1 mM Na3VO4, 10% glycerol, 1% Triton X-100, 1 mM PMSF, 10 μg/mL aprotinin, and 10 μg/mL leupeptin. Insoluble material was collected by centrifugation at 14 000g for 20 min at 4°C. The protein concentration was determined according to the method of Bradford(24) and adjusted to 5 μg/μL. The samples were then divided into 50 μL aliquots and stored at −80°C for further study.

Gene expression microarray analysis.  The procedure of gene expression microarray analysis was provided by manufacturer Welgene Biotech (Taipei, Taiwan). Briefly, both A431-P and A431-III cells (1 × 106) were plated onto 100-mm dishes and allowed to grow in the presence of 10% FBS. After 24 h, the total RNA of A431-P and A431-III cells were extracted by 3 mL Trizol reagent. The gene expression microarray analysis of A431-P and A431-III cells were performed by Welgene Biotech. For data analysis, we selected the genes whose median expression was up or down at least by twofold in A431P cells with respect to A431-III.

Western blot.  The cell lysate samples were mixed with 5× sample buffer and boiled for 5 min, separated on 10% SDS-polyacrylamide gels (PAGE), and then transferred to PVDF membrane (Millipore). The membrane blots were blocked in PBS containing 5% BSA for 1 h at room temperature, and incubated with primary antibody overnight at 4°C. After washing with TBST containing 20 mM Tris–HCl (pH 7.6), 0.8% (w/v) NaCl and 0.25% Tween-20, the blots were incubated with secondary antibody conjugated with horseradish peroxidase (Millipore). Then the membranes were washed with TBST, and immunoreacted bands were detected with ECL reagents (Millipore) and exposed on Fujifilm (Tokyo, Japan). Relative quantification of ECL signals on X-ray film was analyzed by using Image J (NIH, Bethesda, MD, USA).

Reverse transcriptase-polymerase chain reaction.  Total RNA was isolated by using PureLink RNA Mini Kit (Invitrogen), and reverse transcribed by using the MMLV High Performance Reverse Transcriptase kit (Epicentre, Madison, WI, USA). The PCR program was performed for 20–40 cycles by denaturing at 94°C for 30 s, annealing at 55–60°C for 30 s, and extension at 72°C for 30–60 s. The annealing temperature, extension time and reaction cycles were adjusted by different genes. Forward and reverse primers for genes cDNA amplification were listed on the Table 1. The PCR products were run on 1.2% agarose gels, stained with SYBR safe DNA stain (Invitrogen), and visualized by UV.

Table 1.   The forward and reverse primers of genes
Gene nameForward and reverse primerAmplified size (bps)
  1. MMP, matrix metalloproteinases.

MMP-95′-TCTTCCCTGGAGACCTGAGAAC-3′428
5′-GACACCAAACTGGATGACGATG-3′
E-cadherin5′-CCTTAGAGGTGGGTGACTACAA-3′567
5′-TCAGACTAGCAGCTTCGGAAC-3′
N-cadherin5′-GCTTCTGGTGAAATCGCATTA-3′409
5′-AGTCTCTCTTCTGCCTTTGTAG-3′
Desmoplakin5′-GTGGCTCTATGATGCTAAACGC-3′833
5′-GTAGAACCCTCAACCTTTCAATCTCGTA-3′
Cytokeratin5′-GGACTGTGAGGCAGAACCTAGA-3′405
5′-CTTTGACCTCGGCGATGATACT-3′
Occludin5′-AGTGGTTCAGGAGCTTCCATTA-3′396
5′-TATTCATCAGCAGCAGCCATGT-3′
MMP-35′-GGTCTCTTTCACTCAGCCAACA-3′383
5′-ACGAGGTCCTTGCTAGTAAC-3′
Fibronectin5′-GGTGACACTTATGAGCGTCCTAAA-3′316
5′-TACAATCTACCATCATCCAGCCT-3′
Vimentin5′-TGGCACGTCTTGACCTTGAA-3′750
5′-GGTCATCGTGATGCTGAGAA-3′
Snail15′-GCTCCTTCGTCCTTCTCCTCTA-3′390
5′-GGCACTGGTACTTCTTGACA-3′
Snail25′-CTGGTCAAGAAGCATTTCAACG-3′536
5′-GGTAATGTGTGGGTCCGAATA-3′
Twist5′-AGATGTCATTGTTTCCAGAGAAGG-3′232
5′-CTATCAGAATGCAGAGGTGTGAG-3′
ZEB15′-CACACTCTGGGTCTTATTCTCAAC-3′329
5′-CTCTTCTACCTCTTCTTCTTCCACTT-3′
ZEB25′-CCTAGATGATATGACAGACTCCGAC-3′513
5′-TCCTCGCCTTCTTTCTCGTG-3′
E475′-ATCTGCATCCTCCTTCTCCTCA-3′330
5′-TCCAGGTGGTCTTCTATCTTACTCT-3′
GAPDH5′-CCATCACTGCCACCCAGAAGA-3′439
5′-TCCACCACCCTGTTGCTGTA-3′
β-actin5′-GCTCGTCGTCGACAACGGCTC-3′353
5′-CAAACATGATCTGGGTCATCTTCTC-3′

Transfection of small interference RNA.  MMP-9 siRNA, Snail1 siRNA and non-specific siRNA were dissolved in 1000 μL RNase-free water provide by the manufacturer to a stock concentration of 20 μM. A431-P and A431-III cells (2.5 × 105 cells) were plated into 60-mm culture dishes and allowed to adhere for 12 h. Then 6 μL Lifofectamine 2000 reagent (Invitrogen) was added to 300 μL serum-free RPMI-1640 medium, thoroughly mixed, and incubated at room temperature for 5 min. Then 6 μL of siRNA was added to 300 μL serum-free RPMI-1640 medium, thoroughly mixed and then combined with the diluted Lipofectamine 2000. The siRNA/Lipofectamine complex was gently mixed and incubated at room temperature for 20 min, and then added into 60 mm culture dish containing 2.4 mL serum-free RPMI-1640 medium for 24 h. All assays were performed 48 h after transfection.

Snail gene construction and transfection.  The full length cDNA encoding hSnail1 was isolated from human cervical epithelial cancer cell A431-III sub-line cDNA by RT-PCR using specific primers (hSnail-F, 5′-ACT ATG CCG CGC TCT TTC CTC GTC AGG AAG-3′; hSnail-R, 5′-AAG TGG GGC ACT CAG GAG GGA ATT CCA TGG-3′). The full-length of hSnail1 was cloned into pGEMT-Easy vector (Promega, San Luis Obispo, CA, USA) and identified by DNA sequencing. The coding region of hSnail1 was digested with restricted enzyme SalI and following removed the 5′ overhangs to form the blunt end and then digested with restricted enzyme XhoI. The pcDNA3 vector was digested with restricted enzyme EcoRI and following removed 3′ overhangs to form the blunt end and then digested with restricted enzyme XhoI. Ligation of restricted enzyme digested hSnail1 and pcDNA3 vector to generate the pcDNA3-hSnail1 plasmid.

A431-P cells were seeded into 6 cm cultured dishes and then transfected with 4 μg of plasmid with Xfect transfection reagent (Clontech, Mountain View, CA, USA) following the manufacturer’s instructions. The expression of Snail was screening by western blotting and RT-PCR.

Gelatin zymography.  Gelatinases secreted from cultured cells were measured using gelatin zymograph. In brief, samples of conditioned media were subjected to electrophoresis on 8% mini SDS-polyacrylamide gels copolymerized with 0.1% gelatin (Sigma). The volume of each medium sample analyzed was normalized according to the cell number. After electrophoresis, the gels were washed for 60 min in 2.5% Triton X-100, and incubated in reaction buffer (50 mM Tris–HCl, pH 8.0, containing 5 mM CaCl2, 0.02% NaN3) at 37°C for 24 h. The gels were then stained with Coomassie Blue R-250 in 10% acetic acid/20% ethanol for 1 h, and de-stained in the same solution without dye. A clear zone on the gel indicated the presence of gelatinase activity, which was quantified using densitometry.

Treatments of A431-P and A431-III cells with activated MMP-9.  A431-P and A431-III cells (5 × 104) were plated onto 24-well plate and allowed to adhere for 12 h. Recombinant human MMP-9 protein (R&D Systems) were activated with p-aminophenylmercuric acetate (APMA; Sigma) at a final concentration of 1 mM at 37°C overnight. Following activation, 2 μL of activated MMP-9 (final concentration of 0.5 mM APMA) was applied to A431-P and A431-III cells and incubated at 37°C for 24 h. The image of treated A431-P and A431-III cells were photographed by phase contrast microscopy.

E-cadherin neutralization.  A431-P and A431-III cells (1.5 × 105 cells) were seeded onto 6-well plates and allowed to adhere for 12 h. Before treatment of E-cadherin neutralizing antibody, the cells were starved in serum-free medium for 6 h at 37°C. Following starvation, 2 g of E-cadherin neutralizing antibody (nAb) was applied to A431-P and A431-III cells for additional 18 h. The image of treated A431-P and A431-III cells were photographed by phase contrast microscopy.

Immunofluorescence staining.  A431-P and A431-III cells (1.5 × 105 cells) were plated onto 6-well plate containing glass coverslips without coating FN and allowed to adhere for 12 h. Following MMP-9 siRNA, Snail siRNA and non-specific siRNA treatment, the cells were fixed with 4% paraformaldehyde in PBS for 15 min. After rinsing with PBS, cells were quenched with 100 mM glycine in PBS for 10 min. The quenched cells were washed with PBS, permeablized with 0.1% Triton X-100 in PBS for 10 min. The permeablized cells were then incubated with 3% BSA in PBS to block non-specific binding for 1 h at room temperature. After thorough rinsing with PBS, cells were incubated with the anti-vimentin and anti-fibronectin antibodies at 4°C overnight according to the manufacturer instructions. Then, the cells were incubated with fluorescently labeled secondary antibodies for 1 h at room temperature in the dark. After rinsing with PBS, the cells were then stained with DAPI in PBS for 5 min at room temperature. The coverslips were then mounted with mounting media on microslides and visualized the images by confocal image microsopy.

In vitro wound-healing migration assay.  Both A431 and A431-III cells transfected with MMP-9 or Snail siRNA were plated onto 6-well culture plates in RPMI-1640 containing 10% FBS (2 × 106 cells/well). After 24 h, the cell monolayer was wounded by manual scratching with a pipette tip, washed with PBS, photographed in phase contrast using an Olympus IX70 camera (Tokyo, Japan), and then incubated at 37°C for 24 h. Cells were photographed at 0 and 24 h after wound scratch under phase contrast microscope using an Olympus IX70 camera. Experiments were carried out in triplicate per treatment group.

Statistical analysis.  Quantitative data from three to six independent experiments are expressed as means (±SEM). Unpaired Student’s t-tests were used to analyze between group differences. P < 0.05 was considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Characterization of the EMT-related markers in A431-P and A431-III cells.  We have demonstrated that A431-III subline displays increased spreading, migratory and invasive capacity together with enhanced secretion of MMP-9 as compared with A431-P cells.(17) Also, MMP-9 is capable of cleaving E-cadherin.(25) We hypothesized MMP-9 might be responsible for the induction and acceleration of EMT.

A431-III cells had reduced cell–cell contact together with protruding filopodia, whereas A431-P cells grew cohesively and formed colonies (Fig. 1A). We therefore characterized potential genes involved in EMT events first by RNA microarray analysis, and then further analyzed the selected parameters by RT-PCR, immunoblotting and gelatin zymography. A431-III cells exhibited increased mRNA and protein levels of mesenchymal cell markers (MMP-9, MMP-3, fibronection, vimentin and N-cadherin) and MMP-9 activity, and reduced levels of epithelial cell markers (E-cadherin, desmoplakin, occludin and cytokeratin) in A431-III cells (Fig. 1B,C). It is well documented that E-cadherin is negatively regulated by a number of zinc-finger-family of transcription factors, including Snail, Slug and Twist.(26,27) Snail in particular, is also found to be capable of regulating the expression of MMPs.(23) We demonstrate that A431-III cells exhibited a significant increase in the mRNA and protein levels of Snail and Twist, but not Slug (Fig. 1D).

image

Figure 1.  Expression of epithelial–mesenchymal transition (EMT) markers in A431-P and A431-III cells. (A) Pattern of cell spreading on fibronectin. Phase-contrast photographs show that A431-III cells displayed fibroblastic morphology with increased cell spreading, while A431-P cells form a compact sheet-like colony. (B) Expression of mesenchymal cell markers analyzed by microarray, RT-PCR and immunoblotting, and gelatin zymography analysis of matrix metalloproteinases (MMP)-9 secreted activity. β-actin served as internal control. (C) Expression of epithelial cell markers analyzed by microarray, RT-PCR and immunoblotting. (D) Expression of transcription factors involved in EMT process analyzed by microarray, RT-PCR and immunoblotting. Shown in each panel are representative data from three independent experiments.

Download figure to PowerPoint

Potential role of MMP-9 in facilitation of EMT process.  We undertook the task of understanding the potential role of MMP-9 in facilitation of EMT process in two ways. As known, A431-III cells expressed relative high MMP-9 and less E-cadherin compared to A431-P cells (Fig. 1). Therefore, first approach was targeted on A431-P cells through increasing the MMP-9 level by providing exogenous MMP-9, and reducing the extent of cell–cell contact by blockade E-cadherin function using neutralizing antibody. Exposing A431-P cells to exogenous MMP-9 caused alteration in cell morphology with fibroblastic shape, and cell scattering (Fig. 2A), suggesting that MMP-9 may promote EMT by reducing the extent of cell–cell contact and enhancing cell motility. To further elucidate possible events associated with the disruption of cell–cell contact, A431-P cells were given with the E-cadherin nAb. This resulted in the loss of cell–cell adhesion, and cells became fibroblastic and dispersed (Fig. 2B). Also, this effect of E-cadherin nAb was reversible when antibody was removed from the culture medium (data not shown). In order to ascertain the link between cell phenotypes and the extent of E-cadherin/β-catenin colocalization at the plasma membrane, we performed immunofluorescence analysis using confocal microscopy. A431-P cells displayed a higher extent and more compact colocalization between E-cadherin (red) and β-catenin (green) as compared with that in A431-III cells (Fig. 2C). These data further suggest that A431-III acquires motility and invasiveness.

image

Figure 2.  Regulatory effect of matrix metalloproteinases (MMP)-9 and E-cadherin on epithelial–mesenchymal transition (EMT)-related morphological changes in A431-P cells. A431-P cells were treated with (A) 0.5 μM MMP-9 for 48 h, or (B) 2 μg E-cadherin neutralizing antibody (nAb). In both experiments, phase-contrast photographs show that both treatments caused cell contact became loose and more fibroblastic, distinct from the control that formed compact colonies. (C) Immunofluroescence microscopy images of cells double immunostained of E-cadherin (red) and β-catenin (green).

Download figure to PowerPoint

Second approach was to blockade endogenous MMP-9 function in A431-P and A431-III cells through siRNA knockdown of MMP-9 and MMP inhibitor to determine MMP-9 effect on EMT markers. Knockdown of MMP-9 in A431-III cells reduced the mRNA and protein levels of vimentin and fibronection; whereas knockdown of MMP-9 in A431-P cells reduced the mRNA level of vimentin and fibronection, and no obvious effect was observed at the protein level (Fig. 3A,B). Confocal microscopy analysis reveals that A431-III cells had dispersed distribution pattern of fibronectin and vimentin and, interestingly, MMP-9 knockdown downregulated vimentin and fibronectin with fibronectin mainly present at the perinuclear area resembling that of A431-P cells (Fig. 3C). Since MMP-9 plays an important role in tumor invasion and migration, we were interested to learn if MMP-9 knockdown cells would reduce cell motility. We further employed in vitro wound healing assay to examine the effect of MMP-9 knockdown on migratory behavior of A431 cells. MMP-9 siRNA knockdown A431-III cells repopulated the wound area slower than the respective control cells, while this was not observed in A431-P cells (Fig. 3D–F). This is positively correlated with the effect of MMP-9 knockdown on the protein level of fibronectin and vimentin (Fig. 3B). In addition, in vitro invasion assay demonstrated that knockdown of endogenous MMP-9 significantly suppressed the invasive activity of A431-III cells (data not shown). To further confirm whether MMP-9 contributes to EMT process, a broad spectrum MMP inhibitor, GM6001 was utilized. In general, A431 cells exposed to GM6001 had similar phenotypes as those of MMP-9 siRNA knockdown cells. A431-III cells treated with GM6001 had reduced levels of vimentin and fibronectin with fibronectin mainly present at the perinuclear area resembling that of A431-P cells (Fig. 4A). GM6001 inhibitory effect on MMP-9 activity in A431 cells was affirmed (Fig. 4B). In vitro wound healing assay indicated that A431-III cells treated with GM6001 repopulated the wound area slower than the respective control cells, while this was not observed in A431-P cells (Fig. 4C,D). These observations were confirmed by quantitative cell count analysis where the number of spindle-shaped cells was found to be diminished after GM6001 treatment. These results together implicate depletion of MMP-9 caused impaired expression of mesenchymal cell marker proteins and diminished cell motility.

image

Figure 3.  Effect of matrix metalloproteinases (MMP)-9 knockdown on mesenchymal cell markers in A431-III cells. Both A431-P and A431-III cells were plated onto culture wares, and allowed to grow in the presence of 10% FBS culture medium. After 24 h, cells were treated with 40 nM MMP-9 siRNA or control siRNA. Forty-eight hours later, (A) RNAs were isolated and analyzed by RT-PCR for fibronectin, vimentin and MMP-9. (B) Cell lysates were prepared and analyzed by immunoblotting for fibronectin and vimentin with β-actin used as an internal control. (C) Cells were prepared for confocal microscopy analysis, immunostained with fibronectin (green) and vimentin (red) antibody, and DAPI staining for nucleus. (D) Conditioned media were collected and analyzed for MMP-9 activity using gelatin zymography. (E) After MMP-9 siRNA transfection, A431 cells were plated onto 6-well plate and allowed to grow in the presence of 10% FBS for 24 h. Wound healing assay was conducted, cells were observed for migration into the wound area. Phase-contrast images were taken 18 h later to assess cell migration. (F) Quantitative analysis of cell migration was determined by measuring the gap distance. Data are presented as the mean (±SD) percentage of migration distance (n = 20). * and #, respectively indicates a significant difference compared with the respective control in the presence of or absence of MMP-9 siRNA knockdown (P < 0.05).

Download figure to PowerPoint

image

Figure 4.  Inhibition of matrix metalloproteinases (MMP)-9 activity on mesenchymal cell markers in A431-III cells. Both A431-P and A431-III cells were plated onto culture wares, and allowed to grow in the presence of 10% FBS culture medium. After 24 h, cells were treated with 25 μM GM6001. Twenty-four hours later, (A) Cells were prepared for confocal microscopy analysis, immunostained with fibronectin (green) and vimentin (red) antibody, and DAPI staining for nucleus. (B) Conditioned media were collected and analyzed for MMP-9 activity using gelatin zymography. (C) A431 cells were plated onto six-well plate and allowed to grow in the presence of 10% FBS for 24 h. After GM6001 treatment, wound healing assay was conducted, cells were observed for migration into the wound area. Phase-contrast images were taken 24 h later to assess cell migration. (D) Quantitative analysis of cell migration was determined by measuring the gap distance. Data are presented as the mean (±SD) percentage of migration distance (n = 20). * and #, respectively indicates a significant difference compared with the respective control in the presence of or absence of MMP-9 siRNA knockdown (P < 0.05).

Download figure to PowerPoint

Relationship between MMP-9 and Snail on the EMT process.  As levels of vimentin and fibronectin were clearly affected by MMP-9 knockdown in A431-III cells (Fig. 3), here we further analyzed the effect of Snail siRNA knockdown on the expression of these two mesenchymal cell markers and MMP-9. Knockdown of Snail resulted in significant downregulation of vimentin and fibronectin protein levels in A431-III cells, and modest effect was observed in A431-P cells (Fig. 5A,B). The reduction of vimentin and fibronectin levels at the surface and in the cytoplasma of Snail depleted cells was also clearly observed by immunofluorescence (Fig. 5C). Additionally, gelatin zymography assay demonstrated an appreciable reduction of secreted MMP-9 activity after Snail knockdown in both A431-P and A431-III cells (Fig. 5D). Furthermore, in vitro wound healing assay revealed that Snail siRNA knockdown A431-III cells repopulated the wound area slower than the respective control cells, while this was not observed in A431-P cells (Fig. 5E,F). Also, in vitro invasion assay showed that knockdown of endogenous Snail suppressed the invasive activity of A431-III cells (data not shown). These results indicate that depletion of Snail displayed similar phenotypes as depletion of MMP-9, reduced expression of mesenchymal cell marker proteins and impairment of cell motility.

image

Figure 5.  Effect of Snail knockdown on mesenchymal cell markers in A431-III cells. Both A431-P and A431-III cells were plated onto culture wares, and allowed to grow in the presence of 10% FBS culture medium. After 24 h, cells were treated with 40 nM Snail siRNA or control siRNA. Forty-eight hours later, (A) RNAs were isolated and analyzed by RT-PCR for fibronectin, vimentin and Snail. (B) Cell lysates were prepared and analyzed by immunoblotting for fibronectin and vimentin with β-actin used as an internal control. (C) Cells were prepared for confocal microscopy analysis, immunostained with fibronectin (green) and vimentin (red) antibody, and DAPI staining for nucleus. (D) Conditioned media were collected and analyzed for matrix metalloproteinases (MMP)-9 activity using gelatin zymography. (E) After Snail siRNA transfection, A431 cells were plated onto 6-well plate and allowed to grow in the presence of 10% FBS for 24 h. Wound healing assay was conducted, cells were observed for migration into the wound area. Phase-contrast images were taken 24 h later to assess cell migration. (F) Quantitative analysis of cell migration was determined by measuring the gap distance. Data are presented as the mean (±SD) percentage of migration distance (n = 20). * and #, respectively indicates a significant difference compared with the respective control in the presence of or absence of MMP-9 siRNA knockdown (P < 0.05).

Download figure to PowerPoint

We also overexpressed the Snail gene in A431-P cells to verify the role of Snail in EMT processes in A431-P and A431-III sub-line cell model. As shown in Figure 6(A), pcDNA3-Snail-transfected A431-P cells had reduced cell–cell contact together with protruding filopodia, whereas pcDNA3-transfected A431-P cells grew cohesively and formed colonies. Overexpression of Snail gene resulted in significant down-regulation of E-cadherin and upregulation of vimentin, N-cadherin and fibronectin levels in A431-P cells (Fig. 6B,C). Additionally, gelatin zymography assay demonstrated an appreciable induction of secreted MMP-9 activity after Snail overexpression in A431-P cells (Fig. 6D). In addition, in vitro wound healing assay revealed that Snail-overexpressed A431-P cells repopulated the wound area faster than the respective control cells (Fig. 6E,F). Furthermore, in vitro invasion assay showed that overexpression of Snail gene promoted the invasive activity of A431-III cells (data not shown). These results indicated that Snail plays an important role in induction of EMT process in A431 cells.

image

Figure 6.  Effect of Snail overexpression on EMT markers in A431-P cells. A431-P cells were plated onto culture wares, and allowed to grow in the presence of 10% FBS culture medium. After 24 h, cells were transfected with pcDNA3 or pcDNA3-Snail. Forty-eight hours later, (A) Phase-contrast photographs show that pcDNA3-Snail transfected A431-P cells displayed fibroblastic morphology with increased cell spreading, while pcDNA3 transfected A431-P cells form a compact sheet-like colony. (B) RNAs were isolated and analyzed by RT-PCR for EMT markers. (C) Cell lysates were prepared and analyzed by immunoblotting for EMT markers with β-actin used as an internal control. (D) Conditioned media were collected and analyzed for matrix metalloproteinases (MMP)-9 activity using gelatin zymography. (E) After pcDNA3-Snail transfection, A431 cells were plated onto six-well plate and allowed to grow in the presence of 10% FBS for 24 h. Wound healing assay was conducted, cells were observed for migration into the wound area. Phase-contrast images were taken 24 h later to assess cell migration. (F) Quantitative analysis of cell migration was determined by measuring the gap distance. Data are presented as the mean (±SD) percentage of migration distance (n = 20). *Significant difference compared with the respective control in the presence Snail overexpression (P < 0.05).

Download figure to PowerPoint

To explore the link between MMP-9 expression and Snail induction, we used siRNA knockdown of MMP-9 and Snail in A431 cells to study their mutual relationship. Interestingly, we observed that knockdown of MMP-9 markedly reduced the expression of Snail, and knockdown of Snail also suppressed the expression of MMP-9 (Fig. 7). This provides a novel concept that a positive regulatory loop exists between MMP-9 and Snail.

image

Figure 7.  The interplay of matrix metalloproteinases (MMP)-9 and Snail in A431 cells. Cells were plated onto 60-mm culture dish, and allowed to grow in the presence of 10% FBS culture medium. After 24 h, cells were treated with (A) 40 nM MMP-9 siRNA or control siRNA, or (B) 40 nM Snail1 siRNA or control siRNA. After 48 h, mRNA levels of Snail1 and MMP-9 were detected by RT-PCR with GAPDH served as an internal control. Three independent experiments were conducted.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Previously we successfully separated a sub-line A431-III derived from the A431 parental cell line using Bowden chamber assay apparatus.(17) One important rationale behind the selection is that the tumor cells secrete higher amounts of MMPs in order to penetrate through the chamber in a shorter period of time. Our data suggest such highly invasive tumor cells do exist within the parental population. We observed A41-III is a very stable sub-line since we obtained 5 years ago and have passed 10 passages already. We suggest A431-III cells could serve as a powerful approach in establishing strategies to block the spread of cancers from in situ tumors.(17)

Crucial and dynamic changes in proteolysis in tumors are associated with unregulated cancer growth, tissue invasion and metastasis, among other events. Membrane-associated MMPs, which are zinc-dependent endopeptidases, are known to digest the ECM, thus equipping cancer cells to migrate, invade and spread to distant and distal sites at which they form metastasis.(3,5) MMPs are known for their versatility in function and have been implicated in processes leading to cancer growth, tumor invasion and metastasis, and are also considered to play a predominant role in tumor angiogenesis.(1) These contentions are supported by the fact the levels of some MMPs are critically elevated in certain tumor types.(28,29) Among this family, MMP-9 plays a key role in cell invasion and migration. Recent studies have shown that MMPs stimulate transcription factors expression, such as Snail, Twist and consequent EMT in mammary epithelial cells.(23,30)

Previous findings from our laboratory reported that a highly invasive A431-III subline displayed a preponderant secretion of MMP-9 and exhibited increased spreading, migratory and invasive activities compared to that in A431-P cells.(17) The factors contributing to these enhanced metastatic features have not been fully elucidated. Nonetheless, MMP-9 elicited altered morphology might be one of the causes behind enhanced metastasis. As shown in Figure 1, the A431-III cells showed a robust induction of MMP-9 and this induction exhibited EMT as characterized by increased cell scattering and decreased E-cadherin and elevated N-cadherin expression (Fig. 1B,C). It is well known that various cell types constitutively express a latent form of MMP-2, whereas MMP-9 is inducible.(31) Both A431-P and A431-III cells secret almost the same quantities of MMP-2 and, therefore, MMP-2 does not seem to replicate what MMP-9 does.

To determine whether MMP-9 is essential for the induction of EMT, we treated A431-P cells with MMP-9. We were intrigued by the observation that MMP-9 treated A431-P cells were scattered (Fig. 2A). This was a very interesting finding because, in our study, the parental cells expressed less quantities of MMP-9 and more amounts of E-cadherin. How does treatment of A431-P cells lead to these striking and rigorous changes? It is evident that E-cadherin is a substrate of MMP-9. Exposure of A431-P cells to MMP-9 led to a loss of intact E-cadherin while inducing EMT-like morphological changes. Importantly, whether MMP-9 treated A431P cells exhibit changes in EMT markers needs to be further validated. A hallmark of EMT is the loss of the cell surface E-cadherin, which forms adherens junction between cells.(26) To characterize MMP-9 induced process of EMT, both A431-P and A431-III cells were stained with antibodies against E-cadherin and β-catenin. As shown in Figure 2(C), less and loose E-cadherin was detected on the A431-III cell surface that had undergone EMT but, as anticipated, E-cadherin and β-catenin staining on the cell surface of A431-P was abundant and intense. A431-III cells expressing MMP-9 developed spindle-like, fibroblastic morphology, and reduced cell–cell contacts, which correlated with reduced E-cadherin and β-catenin expression. This altered morphology is one of the hallmarks of cells that undergo the process of EMT. Therefore, we suggest that MMP-9 could cleave E-cadherin, and consequently impair cell–cell adhesion. In addition, E-cadherin expression is a dynamic process, a small decrease of expression would result in cell morphology. Consistent with this specific function of MMP-9, is a phenomenon that was reported in mammary epithelial cells, wherein MMP-3 induced EMT.(12,32)

In order to gain insights into the cellular events of EMT in A431 cells, in the present study, we integrated morphological data with the microarray results and biochemical analyses. We observed an acquisition of a subset of EMT manifestations in A431-III cells compared to that in A431-P cells. Consistent with previous findings, we noticed a marked and decisive reduction of E-cadherin, desmoplakin, cytokeratin and occludin in the microarray experiments, which was further confirmed by RT-PCR analysis of A431-III cells (Fig. 1C). Diminution in adherins junctions is a well-known characteristic of tumorigenesis, and we show here that the decreased expression of E-cadherin by MMP-9 occurs in the A431-III cells as opposed to A431-P (Figs 1,2C). However, questions still remain as to whether or not E-cadherin is directly cleaved by MMP-9 in other cells in vitro.

It should be emphasized that, in the present study, we were fortunate enough to obtain much important expression data for most of the EMT genes that are of immense interest. Vimentin, one of the major mesenchymal intermediate filaments,(33) showed a tremendous increase in microarray (33.83-fold, RT-PCR 3.7-fold) and Western blot (2.5-fold) analyses in A431-III as compared to that in A431-P (Fig. 1B). It is well known that the microarray data do not represent endogenously produced mRNA levels within cells, but instead are based on a standard algorithm computed from many factors such as signal intensity, number of genes present and fold changes. Thus, it seems not unusual that vimentin was found to increase 33.83-fold in microarray analysis, whereas RT-PCR data only show 3.7-fold increase. The higher expression of ECM proteins such as vimentin, fibronectin and vinculin, together with the upregulation of MMPs suggest that integrin signaling plays an important role in the invasion and migration of A431-III cells. Also, in the microarray data, we observed significantly enhanced N-cadherin expression in both microarray and RT-PCR analyses (Fig. 1B). The loss of E-cadherin and gain of N-cadherin expression is reminiscent of the cadherin switch(32) and this underpins the phenotypic changes that occurred in A431-III cells. However, it is still less understood how E-cadherin switch to other cadherins could promote a motile phenotype and tumor invasion, and this needs to be further elucidated.

We provide evidence that MMP-9 exhibited a specific action in hampering cell adhesion proteins such as E-cadherin and aforementioned elements in A431-III cells. However, we hasten to add that E-cadherin expression can be downregulated by a number of mechanisms, including DNA hypermethylation, mutation and transcriptional control.(22) Transcription factors have been implicated in the transcriptional repression of E-cadherin, including zinc finger proteins of the Snail/Slug family, Twist,(27,34) dEF1/ZEB1, SIP1, and the basic helix-loop-helix factor E12/E47.(35–37) These repressors can also act as molecular triggers of the EMT program by repressing a subset of common genes that encode cadherins, claudins, cytokines, integrins, mucins, plakophilin, occluding and zonula occludens proteins to promote EMT.(13) Among these transcription factors, we observed that the expression of Snail and Twist was sharply increased in A431-III sub-line (Fig. 1D). Snail has been found to be specifically overexpressed in tumor cells located at the invasive front, where it is responsible for the disruption of E-cadherin mediated cell–cell contacts and invasion,(21) and for the reduction of tumor proliferative activity as well as for the gain of resistance to apoptosis.(19) Peinado et al.(21) suggested a model of participation of the different E-cadherin repressors during EMT/invasion in which Snail and SIP1 would play a role in inducing the initial step of EMT leading to the initiation of the invasive process, whereas Slug, E47 and possibly ZEB1 would favor the maintenance of the migratory, invasive phenotype. The basic helix-loop-helix transcription factor Twist is a repressor of E-cadherin transcription and an organizer of EMT and, in addition, seems to have a critical role in the development of distant metastasis by promoting cancer cells to enter the bloodstream.(27)

As Snail initiates EMT and is upregulated in many human cancers, as is the case with MMPs, it is worth exploring the possibility that MMP-9-induced EMT involves the induction of Snail. We first implemented a siRNA strategy to reduce MMP-9 in A431-P and A431-III cells in order to confirm our hypothesis that MMP-9 might be an initiator of EMT. Cells were first transfected with MMP-9 siRNA, or with a nonspecific sequence. There was no significant difference in the growth rates of transfected and non-transfected cells. Relative to the control, the MMP-9 siRNA significantly decreased the expression of MMP-9 and simultaneously reduced the expression of the mesenchymal protein, vimentin and fibronectin (Fig. 3A,C), without influencing cell viability. Furthermore, knockdown MMP-9 resulted in reduced cell motility (Fig. 3E). Since fibronectin was found to modulate cell-to-cell adhesion and cell attachment and migration, our data indicated that A431-III cells expressed higher quantities of fibronectin and enhanced cell migration ability. Therefore, it is concluded that diminished expression of fibronectin resulted in the decrease of cell migration potential. Consideration of these data indicates that A431 cells in which MMP-9 activity has been depleted manifests changes characteristic of EMT.

Directed pericellular proteolysis is crucial in cell migration and invasion through the ECM.(38) We found that A431-III cells that overexpress MMP-9 enhanced invasion/migration through EHS coated transwell membrane and wound healing assay.(17) This enhanced migration was inhibited by GM6001 (Fig. 4C) indicating that MMP-9 is required by the cells in order to acquire migratory phenotype. In vitro invasion ability of A431-P and A431-III cells was inhibited by GM6001, further emphasizing the role of MMP-9 in the process of EMT (data not shown). It is yet to be determined whether MMP-9 itself completely possesses this activity. Our current data also reveal that GM6001 inhibited MMP-9 activity (Fig. 4B), which in turn, inhibited the expression of vimentin and fibronectin in A431-P and A431-III cells (Fig. 4A). However, MMP-9 siRNA and GM6001 did not produce measure change of E-cadherin in both protein and mRNA levels (Figs 3,4). These data further lend support to the role of MMP-9 in the modulation of EMT process.

Our microarray, RT-PCR and western blot data indicated that A431-III cells expressed higher amounts of Snail compared to that in A431-P. Snail is an established suppressor of E-cadherin and could modulate the expression of MMPs, and this capacity is an important determinant of the EMT process and of tumor progression.(39–41) We next sought to evaluate the relationship between the induction of Snail and the enunciation of downstream of EMT by specific transcript knockdown of Snail with siRNA. Our results indicate that downregulation of Snail by siRNA in both A431-P and A431-III cells inhibited the expression of Vimentin, and fibronectin (Fig. 5A,B). This is consistent the findings of another study, in which knockdown of Snail by siRNA in MDCK-Snail cells caused reversion to the epithelial phenotype marked by re-expression of E-cadherin and, downregulation of vimentin and fibronectin, and loss of invasiveness.(39,40) In the present study, we did not observe detectable changes in the expression of E-cadherin in Snail knockdown in A431-P and A431-III cells. It should be emphasized that E-cadherin repression is not sufficient in itself to induce EMT or invasive properties, as its re-expression in mesenchymal cells does not induce the reversion to the epithelial phenotype.(42) Knockdown of Snail also resulted in decreased cell invasion potential and motility (Fig. 5E). Additionally, we also demonstrated that overexpression of Snail in A431-P cells resulted in the decreased E-cadherin and increased EMT markers such as fibronectin, vimentin and N-cadherin. Overexpression of Snail leads to up-regulation of MMP-9 expression and enhances cell migration (Fig. 6). We suggest that A431 cells could undergo EMT by Snail expression. These data are in agreement with other reports, highlighting the role of Snail as an invasion inducer.(19,43)

Concomitant with the elucidation of vimentin and fibronectin protein expression and invasive ability by knockdown MMP-9 and Snail in this study, we examined how both MMP-9 and Snail would function in modulating the metastatic phenotype in A431-III cells. As shown in Figure 7, MMP-9 silencing by siRNA resulted in a dramatic suppression of the expression of Snail in both A431-P (−68%) and A431-III (−82%). However, knockdown of Snail by siRNA decreased expression of MMP-9 in A431-P (−55%) and A431-III (−63%) to a lesser extent than that observed for the diminution of the expression of Snail in these cells under conditions of MMP-9 silencing by siRNA. These data suggest that the effect of Snail on MMP-9 expression was less pronounced than the effect of MMP-9 on Snail levels. Our data suggest a functional interplay between MMP-9 and Snail. Overexpression of MMP-9 cleaves E-cadherin and disrupts cell–cell adhesion, then affecting E-cadherin-β-catenin complex and enhancing Snail expression. On the other hand, withdrawal of MMP-9 from A431 cells decreased the expression of Snail, and altered expression of a number of other EMT-related genes (Figs 3,4). It is worth to note that the loop between the expression of MMP-9 and Snail remains largely unknown. Our recent data suggest that the enhancement of Akt activity inhibits GSK-3β activity which in turn increases Snail and MMP-9 expression (unpublished data). Taking together, our data is in agreement with the findings documented for MMP-3, which was the first MMP that was reported to induce EMT.(12)

Matrix metalloproteinases have been viewed as one of the most promising targets of cancer treatment.(7) Clinical trials have focused on using MMP inhibitors to block the function of MMP activity. Unfortunately, many high-profile MMP-specific inhibitors in clinical trials yielded disappointing results. The failure of MMP inhibitor drug clinical trials in cancer was partly owing to the inadvertent inhibition of MMP anti-targets that counterbalanced the benefits of MMP target inhibition.(44) In addition, one drug might target two phases or dual aspects of a physiological function such as wound healing and angiogenesis. Therefore, there is a need to switch current focus and develop more therapeutic strategies targeting tumor progress pathways that are activated, regulated or induced by MMPs. Recent studies suggest that Snail, Slug, Twist and ZEB family of transcription factors are turned on during EMT process.(16,18) In the present study in A431-III cells, Snail expression might have been induced by the upregulation of MMP-9, resulting in portal invasion and the process of EMT. Our A431-III cell model system provides a rapid method to obtain EMT-like cells. Comparison of A431-P and A431-III cells provides a useful tool to identify signaling pathways and genes that are relevant to the regulation EMT. These data may provide impetus to fashion novel diagnostic markers and lend fillip in the development of therapeutic agents to target tumors.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

We thank Dr Chin-Chun Hung for technical help. This work was supported in part by grants from the National Science Council of Taiwan (NSC 94-2320-B-001-034 to M.T.L; NSC 95-2320-B-010-006 to J.J.H) and the Taiwan Academia Sinica Thematic Project (AS-96-TP-B06 to M.T.L.).

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References