Highly invasive A431-III cells, which are derived from parental A431-P cells, were originally isolated by three successive passages through a Boyden chamber using a Matrigel-coated membrane support. The greater invasion potential shown by A431-III cells was due to their increased ability to spread/migrate, which was associated with enhanced MMP activity. The tumor progression events evoked by A431-P cells compared to A431-III cells may help identify useful strategies for evaluating the epithelial–mesenchymal transition (EMT) and these cell lines could be a reliable model for evaluating tumor metastasis events. Using this approach, we evaluated the effects of luteolin and quercetin using the A431-P/A431-III EMT model. These flavonoids reversed cadherin switching, downregulated EMT markers, and nullified the invasion ability of A431-III cells. Overexpression of MMP-9 resulted in induction of the EMT in A431-P cells and this could be reversed by treating with luteolin or quercetin. Cotreatment of A431-P and A431-III cells with epidermal growth factor (EGF) plus luteolin or quercetin resulted in a more epithelial-like morphology, led to reduced levels of EGF-induced markers of EMT, and caused the restoration of cell–cell junctions. E-cadherin was decreased by EGF, but increased by luteolin and quercetin. Our results suggest that luteolin and quercetin are potentially beneficial agents that target and prevent the occurrence of EMT in epidermal carcinoma cells. These chemicals also have the ability to attenuate tumor progression in A431-III cells. Luteolin and quercetin show inherent potential as chemopreventive/antineoplastic agents and do this by abating tumor progression through a reversal of EMT. (Cancer Sci 2011; 102: 1829–1839)
The flavonoids, generally classified as phenylchromones, are polyphenolic compounds that are recognized as integral components of the human diet. They are ubiquitous constituents of flowering plants, particularly food plants.(1–3) Several plants, spices, and herbs that contain flavonoid derivatives are applied as disease preventive and therapeutic agents in traditional medicine in Asia, and have been used as such for thousands of years.(4) The much lower risk of colon, prostate, and breast cancers in Asians (i.e. Japanese and Chinese), who consume more vegetables, fruits, and tea than those in the United States,(5) raises the question of whether flavonoid components mediate protective effects when diets are rich in these foodstuffs and may be natural chemopreventive/anticancer agents. Plant flavonoids have been recognized for some time as possessing antitumor and differentiating effects.(4,6–8) Two dietary flavonoid constituents, luteolin (Lu), a flavone, and quercetin (Qu), a flavonol, generally appear to be the most potent among plant flavonoids in terms of their in vitro biological activities.(9) They show a variety of anticancer effects, such as cell growth and kinase activity inhibition, apoptosis induction, differentiation, suppression of the secretion of MMPs, and a reduction in tumor cell adhesion, invasive behavior, metastasis, and angiogenesis.(4,6–8) In earlier studies, we documented the antiproliferative effects of Lu and Qu.(6–9) An appreciation of the diverse anticancer activities exerted by these two flavonoids has prompted us to evaluate their impact on events such as tumor progression and invasion.
Cancer cells undergo dramatic and dynamic changes in their adhesion molecule expression profile during the course of tumor progression and these changes result in their detachment from their original tissue and the acquisition of a highly motile and invasive phenotype. A hallmark of this change, also known as the epithelial–mesenchymal transition (EMT), is intimate linkage with changes in the expression of the adhesion molecules that regulate the interactions of cancer cells with both the ECM and their neighboring cells. Mounting evidence documents that the detachment of carcinoma cells and their transfer into the stroma recapitulates the EMT, a developmental process that allows stationary epithelial cells to become motile.(10,11) A common feature of cancers of epithelial origin (i.e. carcinomas) is increased expression, often de novo expression, of N-cadherin and thus the concomitant downregulation of E-cadherin, a major component of adherens junctions that brings about homophilic contact between neighboring cells. The loss of E-cadherin and the de novo expression of N-cadherin is termed cadherin switching.(12,13) This phenomenon is correlated with the acquisition of invasiveness and metastatic potential by the cancer cells. Cadherin switching has been observed in biopsies from a wide range of types of cancers, including melanoma, breast, and prostate cancers,(12) bladder carcinoma,(14) some types of ovarian(15) and gastric carcinomas,(16,17) and adrenal tumors.(18) N-cadherin promotes tumor cell survival, migration, and invasion, and high level expression is often associated with a poor prognosis. N-cadherin is also expressed in endothelial cells; it plays an essential role in the maturation and stabilization of normal vessels and tumor-associated angiogenic vessels. Increasing experimental evidence suggests that N-cadherin is a potential therapeutic target when treating cancer.(19)
Highly invasive A431 cells used in this study were selected using a Boyden chamber assay and the cell line thus obtained is termed A431-III. This cell line expresses high levels of MMP-9 and has a greater invasion ability than the original parental A431 cells (A431-P).(20) The high invasive ability of A431-III cells, when contrasted with A431-P cells, is likely to be due to cadherin switching and the EMT. In this study, we first compared the A431-P cell line with the A431-III cell line as an EMT model, then measured the effects of Lu and Qu on cadherin switching, downregulation of EMT markers, and the suppression of the invasion ability of A431-III cells. We then used transient overexpression of MMP-9 in A431-P to mimic A431-III cells in order to determine the effects of Lu and Qu on MMP-9 overexpression in terms of the EMT phenomena in A431-P cells. Additionally, the effects of Lu and Qu on the potential of A431-P/A431-III cells to undergo the EMT on exposure to epidermal growth factor (EGF) were also determined.
Materials and Methods
Materials. A431 cells (human epidermal carcinoma cells; named A431-P in this study) were obtained from ATCC (Manassas, VA, USA). The A431-III cells were isolated in our laboratory from the parental A431 tumor cells (A431-P) using a Boyden chamber.(20) RPMI-1640 and FCS were obtained from Gibco (Grand Island, NY, USA). Anti-Snail, anti-Twist, and anti-E-cadherin (HECD1) antibodies were obtained from Abcam (Cambridge, MA, USA). Anti-N-cadherin antibody was purchased from Abgent (San Diego, CA, USA). Anti-lamin A, anti-p-Akt(Ser 473) and anti-p-GSK3β(Ser9) antibodies were obtained from GeneTex (Irvine, CA, USA). Growth factor-reduced EHS Matrigel was acquired from BD (Franklin Lake, NJ, USA). Anti-fibronectin and anti-β-actin antibodies were purchased from Sigma (St. Louis, MO, USA). Anti-vimentin (V9) antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The PCR forward and reverse primers were purchased from Purigo Biotech (Taipei, Taiwan). Luteolin was purchased from Toronto Research Chemicals (North York, ON, Canada). Agarose and DMSO were purchased from E. Merck (Darmstadt, Germany). Epidermal growth factor was obtained from Upstate Biotechnology (Lake Placid, NY, USA) and dissolved in RPMI-1640 medium. Unless otherwise indicated, all other reagents were obtained from Sigma. Luteolin and Qu were dissolved in 100% DMSO and their concentrations were adjusted to 100 mM as stock solutions.
Preparation of cell lysate. 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 removed by centrifugation at 14 000g for 20 min at 4°C. Nuclei were isolated by hypotonic treatment (swelling the cells in hypotonic buffer; 10 mM HEPES at pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM PMSF, 1 mM Na3VO4, and 10 μg/mL leupeptin) for 15 min on ice, followed by centrifugation for 30 s. After centrifugation, the sample of nuclei were lysed in lysis buffer (20 mM HEPES at pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, and 1 mM DTT) with added inhibitors in order to obtain nuclear extracts. Protein concentrations were quantified using a Bio-Rad protein assay kit (Bio-Rad, Hercules, CA, USA). All samples were divided into 50 μL aliquots and stored at −80°C for further study.
Western blot analysis. The cell lysate samples were mixed with 5× sample buffer and boiled for 5 min. They were then separated by 10% SDS-PAGE, and transferred to a PVDF membrane (Millipore, Bedford, MA, USA). Next the membrane was 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 blot was incubated with secondary antibody conjugated with HRP (Millipore, Billarica, MA, USA). Finally, the membrane was washed with TBST, and the presence of immunoreactive bands detected using an ECL reagent kit (Millipore) and exposure to Fujifilm (Tokyo, Japan). Relative quantification of the ECL signals on the X-ray film was analyzed using ImageJ (http://rsb.info.nih.gov/ij/index.html, NIH, USA).
Immunofluorescence microscopy. Cancer cells were plated onto 6-well plates containing glass coverslips without fibronectin coating and allowed to adhere for 12 h. Following agent treatment, the cells were fixed with 4% formaldehyde for 15 min. The fixed cells were next permeabilized in 0.2% Triton X-100, washed with PBS, and blocked with 5% BSA for 1 h at room temperature. After thorough rinsing with PBS, the samples were incubated with anti-E-cadherin antibody for 1.5 h, followed by secondary antibody conjugated with AlexaFluor 555 (Invitrogen, Carlsbad, CA, USA) for 30 min. The same cells were stained with phalloidin conjugated with AlexaFluor 488 and DAPI to detect F-actin and cell nuclei, respectively. The coverslips were then mounted with mounting media on microscope slides and the images visualized by confocal image microscopy (Carl Zeiss, Thornwood, NY, USA). All images were obtained using a 67× oil-immersion objective and a digital camera with LSM 5 Image Browser software (Carl Zeiss).
Reverse transcription–PCR. Total RNA was isolated using a PureLink RNA Mini kit (Invitrogen), and reverse transcribed using an MMLV High Performance Reverse Transcriptase kit (Epicentre, Madison, WI, USA). The PCR program was carried out for 25–40 cycles by denaturing at 94°C for 30 s, annealing at 55–60°C for 30–40 s, and extension at 72°C for 60 s. The annealing temperature, extension time, and reaction cycles were adjusted to fit the various different genes. The forward and reverse primers, respectively, for cDNA amplification were: E-cadherin, 5′-CCTTAGAGGTGGGTGACTACAA-3′ and 5′-TCAGACTAGCAGCTTCGGAAC-3′; N-cadherin, 5′-GCTTCTGGTGAAATCGCATTA-3′ and 5′-AGTCTCTCTTCTGCCTTTGTAG-3′; Twist, 5′-AGATGTCATTGTTTCCAGAGAAGG-3′ and 5′-CTATCAGAATGCAGAGGTGTGAG-3′; Snail, 5′-GCTCCTTCGTCCTTCTCCTCTA-3′ and 5′-GGCACTGGTACTTCTTGACA-3′; MMP-9, 5′-TCTTCCCTGGAGACCTGAGAAC-3′ and 5′-GACACCAAACTGGATGACGATG-3′; and GADPH, 5′-CCATCACTGCCACCCAGAAGA-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′. The PCR products were separated on 1.2% agarose gels, stained with SYBR safe DNA stain (Invitrogen), and visualized by UV.
Gelatin zymography. Gelatinases secreted from cultured cells were measured using gelatin zymograph. In brief, samples of conditioned media and total cell lysates were subjected to electrophoresis on 3–18% linear gradient SDS-polyacrylamide gels copolymerized with 0.1% gelatin (Sigma). The volume of each medium sample analyzed was normalized according to cell number and total cell lysates that contained 50 μg protein were analyzed. 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 solution (10% acetic acid, 20% ethanol, 70% dd H2O) for 1 h, and destained in the same solution without dye. A clear zone on the gel indicated the presence of gelatinase activity, which was quantified using densitometry.
In vitro wound-healing migration assay. A431-III cells were plated onto 6-well culture plates in RPMI-1640 containing 10% FCS (2 × 106 cells/well). After 24 h, the cell monolayer was wounded by manual scratching with a pipette tip, washed with PBS, photographed by phase contrast microscopy (0 h) using an Olympus IX70 camera (Tokyo, Japan), then treated with or without 20 μM flavonoids (Lu and Qu). The damaged monolayers were incubated at 37°C for 24 h. Cells were photographed at 24 h after wounding using the same camera system. All experiments were carried out in triplicate for each treatment group.
In vitro invasion assay. In vitro invasiveness was investigated according to a previously described procedure(21) with a few modifications. In brief, the filter of a 24-well Transwell unit was coated with 0.1 mL of 0.6 mg/mL EHS Matrigel. The lower compartment contained RPMI-1640 with 10% FCS as a chemoattractant. A431-III cells were placed in the upper compartment (1 × 105 cells/0.5 mL RPMI-1640 containing 0.1% BSA) and these cells were then incubated with or without flavonoids (Lu and Qu) at 37°C for 48 h. After incubation, the filters were fixed with 3% glutaraldehyde in PBS and stained with crystal violet. Cells on the upper surface of the filter were gently scraped off, and those that penetrated through the Matrigel to the lower surface of the filter were counted under a microscope (200×). Each treatment was assayed in triplicate, and two independent experiments were carried out.
MMP-9 gene construction and transfection. The full-length cDNA encoding MMP-9 was isolated from the human cervical epithelial cancer cell A431-III subline cDNA by RT-PCR using specific primers (MMP-9-F, 5′-ACAACAGCAGCTGCAGTCAGACACCTCTGC-3′; MMP-9-R, 5′-AGTCCTCAGGGCACTGCAGGATGTCATAGG-3′), then cloned into pGEMT-Easy vector (Promega, San Luis Obispo, CA, USA); the product was characterized by DNA sequencing. Next the pcDNA3-MMP-9 plasmid was constructed such that it contained only the coding region, as described below. To do this, the coding region of MMP-9 cDNA was obtained by sequential steps, first by restriction enzyme NcoI digestion, which was followed by removal of the 5′ overhangs to form blunt ends; the resulting product was then digested with SalI. Parallel to this, the pcDNA3 vector was first digested with EcoRI, which was followed by removal of the 3′ overhangs to form blunt ends; the resulting product was then digested with XhoI. The restricted enzyme-digested MMP-9 cDNA and pcDNA3 vector were then ligated to generate the pcDNA3-MMP-9 plasmid.
A431-P cells were transfected with 4 μg pCDNA3-MMP-9 plasmid using Xfect transfection reagent (Clontech, Mountain View, CA, USA) according to the manufacturer’s instructions. After transfection for 5 h, the medium was changed with fresh medium and the incubation continued for an additional 24 h. The cells were then either treated or not treated with flavonoid (20 μM Lu or Qu) for 24 or 48 h to determine the influence of the compounds on EMT-related morphology and key proteins.
Statistical analysis. The results from three to six independent experiments are expressed as the mean ± SD. Statistical significance between groups was determined by means of an unpaired Student’s t-test. A probability of P < 0.05 was considered significant for all tests.
The highly invasive A431-III cells previously isolated from the parental A431 epidermoid cancerous cells (A431-P) in our laboratory displayed a mesenchymal-like morphology, whereas the A431-P cells showed an epithelial-like morphology.(20) Based on the fact that the two dietary flavonoids Lu and Qu seem to exert diverse anticancer activity,(4,6–8) we investigated their impact on EMT during tumor progression using the A431-P/A431-III cell model system.
Effect of flavonoids on characteristic markers of EMT in A431-III cells. We first determined the effect that flavonoids had on the morphology of A431-III cells by the conditions used in our previous study to show that a specific treatment of A431 cells was able to reduced cell proliferative activity.(8) When A431-III cells were treated with 20 μM Lu or Qu for 48 h, this caused their cell morphology to change from mesenchymal-like to epithelial-like (Fig. 1A). This prompted us to examine whether the two dietary flavonoids are able to reverse the EMT process. Consistent with our recent study,(22) in the absence of either flavonoid, the highly invasive A431-III cells expressed higher levels of a range of mesenchymal-like markers, including fibronectin, vimentin, Twist, Snail, and N-cadherin compared to A431-P cells, and this was complemented by lower levels of the epithelial-like marker E-cadherin in the same cells (Fig. 1B). Interestingly, treatment of A431-III cells with either flavonoid (Lu or Qu) downregulated the various mesenchymal-like markers (fibronectin, vimentin, Twist, Snail, and N-cadherin), and upregulated the epithelial-like marker E-cadherin (Fig. 1B). E-cadherin was found to be prominently localized at the cell–cell junctions in A431-P cells, but this distribution was largely lost in A431-III cells (Fig. 1C). Exposure of A431-III cells to Lu or Qu for 48 h led to the cell–cell junctions being reformed with E-cadherin predominantly localized at these sites; there was also more F-actin present in the cell cortical area and fewer actin stress fibers present (Fig. 1C). When taken together, these results support the hypothesis that treatment with either flavonoid (Lu or Qu) was able to reverse the EMT process in A431-III cells.
Effect of flavonoids on invasion/metastasis-associated activities in A431-III cells. Cadherin switching (from E-cadherin to N-cadherin) and increased MMP expression are associated with the acquisition of invasiveness and metastatic potential by cancer cells.(12,23–25) As shown in Figure 1(B), the highly invasive A431-III cells expressed higher levels of N-cadherin and lower levels of E-cadherin compared to A431-P cells, and this cadherin switching was reversed by flavonoid treatment (20 μM Lu or Qu for 48 h). We further examined the dose–response curve of these flavonoids on cadherin switching and various key transcription factors using the shorter treatment period of 24 h to avoid possible cytotoxic effects. Luteolin or Qu markedly reduced N-cadherin levels and moderately increased E-cadherin levels in terms of both protein and mRNA in a dose-dependent manner (Fig. 2A,B). Regardless of the E-cadherin upregulation status of the cells, treatment with Lu or Qu downregulated N-cadherin expression and reversed cadherin switching in A431-III cells. Our recent study(22) revealed overexpression in the highly invasive A431-III cells of two transcription factors, namely Twist and Snail, which are repressors of E-cadherin and served as EMT markers. We therefore investigated the effects of Lu or Qu on the expression of these two transcription factors. Treatment with Lu or Qu dramatically reduced the protein and mRNA levels of Twist and Snail in A431-III cells in a dose-dependent manner (Fig. 2C,D).
We have recently shown that the high migratory ability of A431-III cells is attributable in part to high expression levels for MMP-9 and Snail, as evidenced by siRNA knockdown.(22) In addition to the above effects, treatment with either Lu or Qu reduced the level of Snail in A431-III cells (Fig. 2C,D). Thus, we were interested to explore the effect of these flavonoids on MMP-9 activity, cell migration, and invasion ability in A431-III cells. Cells were treated with different concentrations (10–40 μM) of Lu or Qu for 24 h, and MMP levels were assessed by gelatin zymography. The results showed that treatment with either Lu or Qu dose-dependently decreased MMP-9 activity in the conditioned media. Although the activity of MMP-9 in the cell lysate was too weak to be detected, treatment of A431-III cells with either Lu or Qu caused a notable increase in MMP-9 gelatinase levels within the cells (Fig. 2E). This result indicates that treatment with either flavonoid was able to suppress the secretion of MMP-9 and thereby increase the level of the protein within the cell. We also tested the mRNA level of MMP-9 by RT-PCR and the results showed that treatment with either Lu or Qu was able to dose-dependently decrease MMP-9 mRNA expression (Fig. 2F). Next, a wound healing assay was used to test the effect of either flavonoid on cell migratory activity. A431-III cells were grown to confluent density in 6-well culture plates, then the cell monolayer scratched with a pipette tip to create a wound. The cells were then incubated with 20 μM Lu or Qu for 24 h and a significant delay in closure of the wound gap was detected (Fig. 2G). Finally, we used a Boyden chamber invasion assay to test the effects of either flavonoid on cancer cell invasion ability. Treatment with either Lu or Qu (10 and 20 μM for 48 h) produced significant inhibition of the invasive ability of A431-III cells (Fig. 2H). In summary, when the above results are taken together, they indicated that treatment with either Lu or Qu is able to inhibit the invasion/metastasis-associated activities of A431-III cells.
Effects of flavonoids on MMP-9 overexpression induced during EMT in A431-P cells. Our recent study show that A431-III cells highly express MMP-9 and Snail compared to A431-P cells, and these proteins are mutual positive regulators that promote EMT behavior.(22) We therefore used transient overexpression of MMP-9 in A431-P to mimic the situation in A431-III cells by the presence of higher expression levels of MMP-9, and examined the effect of either flavonoid on various EMT-related activities. A431-P cells were transfected with pcDNA3 (as a control) or pcDNA3-MMP-9. After 24 h, the cells were treated or not treated with either flavonoid (20 μM Lu or Qu) for 24 or 48 h to determine the effect on MMP-9 activity (by zymography analysis and RT-PCR), in terms of EMT-related cell morphology and on the levels of other key molecules. The increased expression of MMP-9 in the pcDNA3-MMP-9-transfected A431-P cells compared to control A431-P cells was confirmed by determining the secreted MMP-9 activity level and mRNA level of the two types of cell. It was found that treatment with either Lu or Qu greatly reduced the increased expression of MMP-9 at both the protein and mRNA level (Fig. 3A,B). In addition, the pcDNA3-transfected control A431-P cells retained an epithelial-like morphology similar to that of the non-transfected A431-P cells. However, interestingly, the pcDNA3-MMP-9-transfected cells displayed a mesenchymal-like morphology similar to that of A431-III cells, and this morphological change was reversed by treatment with either Lu or Qu (Fig. 3C). Using immunofluorescence microscopy, we next evaluated the effect of either flavonoid on cell morphology by examining the adherens junctions, which were immunostained for E-cadherin, and for the presence of F-actin, which was stained with phalloidin. Overexpression of MMP-9 in A431-P cells led to the loss of cell–cell junctions, which are clearly present in the control cells (Fig. 3D). In the control A431-P cells, E-cadherin was localized prominently at the cell–cell junctions and F-actin was present in the cortical region. This contrasted with the situation in MMP-9 overexpressed cells where E-cadherin was largely lost and the F-actin formed stress fibers (Fig. 3D). Treatment of pcDNA3-MMP-9-transfected A431-P cells with either Lu or Qu led to the restoration of the cell–cell junctions, which now regained E-cadherin localized at junction sites (Fig. 3D). Consistent with the morphological changes, MMP-9 overexpression in A431-P cells also altered a range of EMT markers, namely upregulating fibronectin, vimentin, Twist, Snail, and N-cadherin protein levels and downregulating the level of E-cadherin; these changes were also reversed by treatment with either Lu or Qu (Fig. 3E). In summary, these results together imply that either of the dietary flavonoids Lu and Qu, independently, are able to reverse the MMP-9 overexpression-induced EMT-related activities of A431-P cells.
Effects of flavonoids on EGF-induced EMT in A431-P and A431-III cells. To further explore the impact of flavonoid (Lu or Qu) treatment on the EMT process, we next used an EGF induction system. Our unpublished preliminary results revealed an EGF induction of morphology change in A431 cells requires at least 3 days in culture under low serum conditions (1% FCS), and that treatment with 20 μM Lu or Qu might cause cell death. Based on these results, we treated both A431-P and A431-III cells with EGF for 3 days in the absence or presence of 10 μM Lu or Qu. The A431-P cells cultured under low serum conditions retained an epithelial-like morphology, whereas the cells treated with EGF displayed a mesenchymal-like morphology after 3 days of treatment (Fig. 4A). This morphological alteration in the EGF-treated A431-P cells was reversed by concomitant treatment with either Lu or Qu (Fig. 4A). This is supported by observations that EGF induced the loss of cell–cell junctions, where E-cadherin was prominently localized, and caused an upregulation of various EMT markers including fibronectin, Twist, Snail, and N-cadherin, as well as a moderate downregulation of E-cadherin (Fig. 4B,C). Furthermore, these EGF-induced changes were also repressed by concomitant treatment with either Lu or Qu (Fig. 4B,C). We next examined the influence of either flavanoid on EGF-induced EMT-related changes in A431-III cells, and observed similar results to those seen in A431-P cells (Fig. 5). In summary, these results suggested that treatment with dietary flavonoids Lu or Qu is able to reverse EGF-induced EMT-related activities in A431-P and A431-III cells.
Inhibitory effect of flavonoids on Akt Ser473 phosphorylation and various downstream targets. It has been suggested that Akt, once activated, is able to transduce signals from growth factors and oncogenes to downstream targets that modulate cell growth, survival, migration, invasion, and angiogenesis.(26,27) Phosphorylation at Ser473 activates Akt to phosphorylate these downstream targets through its kinase activity. These findings prompted us to determine whether treatment with either Lu or Qu was able to directly inhibit the phosphorylation of Akt and thus reduce phosphorylation of GSK3β. The cell lysates obtained from A431-III cells treated with 20 μM Lu or Qu for 24 h were subjected to immunoblotting analysis. Treatment with either Lu or Qu greatly reduced Akt phosphorylation by 42% and 29%, respectively (Fig. 6). As GSK3β is a well-known direct target of Akt, we determined whether this reduced Akt phosphorylation was able to decrease the phosphorylation level of GSK3β. This was done by immunoblotting with anti-p-GSK3β antibody and we observed that phosphorylation of GSK3β was decreased 60% on treatment with Lu and by 43% on treatment with Qu (Fig. 6). The results indicate that a reduction in phosphorylation of GSK3β accompanied the lower level of Akt phosphorylation caused by treatment with either Lu or Qu.
The term EMT describes a process by which stationary epithelial cells lose their characteristic polarity, disassemble their cell–cell junctions, and become increasingly motile.(10,11) A comparison of the properties of A431-P cells with those of A431-III cells shows that A431-III cells display a mesenchymal-like morphology, express a far greater level of MMP-9, and have significantly elevated migration/invasion capacity, which corroborates our previous results.(20) Furthermore, A431-III cells express much higher levels of various EMT markers, including fibronectin, vimentin, Twist, and Snail, which are likely causes of E-cadherin switching to N-cadherin and the loss of cell–cell junctions.(22) Based on these findings, it seemed likely that a comparison of the tumor progression events evoked in A431-P cells with those in A431-III cells may be a useful strategy when investigating EMT. In these circumstances, these cells might be useful as a reliable model for the evaluation of tumor progression events and for investigating their mitigation. Using this approach, we found that treatment with either Lu or Qu caused A431-III cells to become more epithelial-like in morphology, resulted in a downregulation of EMT markers, and initiated a restoration of the cell–cell junctions (Fig. 1). Lu and Qu, therefore, appear to be potentially good agents able to target and prevent the occurrence of the EMT in epidermal carcinoma cells.
Cadherin switching has a profound effect on cell phenotype and behavior and is one aspect of the EMT process. N-cadherin actively promotes motility when expressed by epithelial cells.(28–30) Interestingly, cells that express significant amounts of E-cadherin, but only minor amounts of N-cadherin, still display accentuated motility. This suggests that N-cadherin possesses a dominant, marked, and decisive function in terms of cell motility that E-cadherin is unable to suppress.(29,30) The term “cadherin switching” usually refers to a switch from the expression of E-cadherin to that of N-cadherin, but also includes situations in which E-cadherin expression levels do not change significantly but the cells turn on (or increase) expression of N-cadherin.(31) Further evidence for such a deployment comes from studies involving the transforming growth factor (TGF)β1-induced motility of mammary epithelial cells. MCF10A cells undergo a classical TGFβ1-induced morphological change involving the downregulation of E-cadherin and the upregulation of other mesenchymal proteins that are typical of TGFβ1-stimulated mammary epithelial cells. However, if shRNA is used to block the ability to upregulate N-cadherin expression, the cells do not show increased motility in response to TGFβ1.(12) These observations indicate that N-cadherin specifically promotes cell motility and as a result it is implicated in cadherin switching when cell behavior is regulated. A431-III cells are highly invasive, which is possibly due to their high levels of N-cadherin expression. In this study, both flavonoids (Lu and Qu) independently were able to upregulate E-cadherin and downregulate N-cadherin in A431-III cells (Figs 1B,2A,B). Overexpression of MMP-9 in A431-P cells induced E-cadherin to N-cadherin switching and this could be reversed by treatment with either Lu or Qu (Fig. 3E). In addition, cotreatment of A431-P and A431-III cells with EGF plus either Lu or Qu caused a fall in the level of EGF-induced E-cadherin to N-cadherin switching (Figs 4C,5C). Together, these results support the hypothesis that both Lu and Qu independently are able to cause downregulation of N-cadherin and thus highlighting the potential of these flavonoids as tumor suppressive agents.
With a view to elucidating the possible impact of Lu and Qu on N- and E-cadherin levels in A431-III cells, we tested two transcription factors (Twist and Snail) that are also EMT markers. These proteins showed higher levels of expression in A431-III cells compared to A431-P cells. Treatment with either Lu or Qu induced E-cadherin expression, perhaps due to an immediate suppression of the repressors Twist and Snail (Fig. 2C,D). This might explain why both flavonoids have a greater effect on E-cadherin at the mRNA level that at the protein level (Fig. 2A,B). Overexpression of MMP-9 in A431-P cells also induced Twist and Snail that were reduced by flavonoids (Lu and Qu), and it may influence the expression of E-cadherin (Fig. 3E). In both A431-P and A431-III cells, the flavonoids (Lu and Qu)-induced E-cadherin expression that was inhibited by EGF might be due to the changes in Twist and Snail (Figs 4C,5C).
Although several transcription factors that regulate the expression of E-cadherin have been documented, the transcriptional regulators influencing N-cadherin expression are not yet fully characterized. An analysis of gastrulation in the Drosophila embryo has shown that Twist directly activates N-cadherin expression.(32) Moreover, overexpression of Twist has been reported in gastric cancer cells that have abnormally high N-cadherin levels.(16) In addition, Twist binds to an E-box in the first intron of the human N-cadherin gene and this upregulates expression of N-cadherin in prostate cancer cells.(33) Thus, Twist would seem to be an inducer of N-cadherin. Our results show that treatment with either Lu or Qu impairs the level of Twist, indicating that the suppression of N-cadherin could be due to a suppression of Twist (Figs 1B,2A–D,3E,4C,5C).
No evidence has been published indicating that N-cadherin expression is induced by Snail. However, Snail-expressing cells do acquire migratory and invasive properties through changes in their cytoskeleton and the induction of MMPs.(34) The Snail-induced EMT contributes to the increased invasion of these cells not only through an inhibition of cell–cell adhesion, but also by the upregulation of MMP expression in A431 cells.(35) Snail is also an enhancer of other EMT markers, which will result in increased cancer cell invasiveness. Therefore, Snail could be a tumor progression marker and an anti-invasive drug target.(34,36–38) Our recent study showed the overexpression of MMP-9 in A431-III cells might directly induce (or stimulate) EMT and that Snail may cooperatively engage in this phenomenon.(22) In this study, our results show that treatment with either Lu or Qu results in a suppression of Snail, the consequence of which could reduce MMP-9 secretion by A431-III cells (Fig. 2C–F).
We have also shown in this study that the phosphorylation of GSK3β and Akt is suppressed by treatment with either Lu or Qu (Fig. 6). Flavonoids act as cellular protein phosphorylation inhibitors,(8) therefore the fact that treatment with either Lu or Qu was able to suppress Snail and MMP-9 could be due to an inhibition of phosphorylation of Akt and GSK3β. In addition, the overexpression of the Twist gene implies a reduction in expression of tissue inhibitor of metalloproteinase-1 (TIMP-1), a naturally occurring specific inhibitor of MMPs.(39) Our results also show that treatment with either Lu or Qu suppresses MMP secretion perhaps through the Twist/TIMP-1 pathway (Fig. 2C–F). Furthermore, the interaction of N-cadherin with the fibroblast growth factor receptor is able to cause sustained MAPK–ERK activation, which would result in increased transcription of extracellular MMP-9.(40) In the present study, our findings indicated that suppression of MMP-9 secretion by either Lu or Qu may be due to inhibited N-cadherin expression (Fig. 2A,B,E,F). Luteolin- or Qu-mediated suppressed secretion of MMP-9 in A431-III cells appears to be a result of downregulation of Snail, Twist, and N-cadherin at the same time. In our previous study, we observed that treatment of A431 cells with a flavonoid causes a notable increase in cellular MMP-9 and decreases the protein’s presence in the conditioned media, whereas EGF treatment has the opposite effect.(8) In this study, we reproduced these findings and showed that treatment with either Lu or Qu was able to suppress the secretion of MMP-9 and cause an intracellular accumulation (Fig. 2E). Together, these results indicate that treatment with either Lu or Qu reversed the EMT and downregulated MMPs, which might be the reason for the reduction in wound healing and the impaired invasive potential of A431-III (Fig. 2).
We examined the effects of MMP-9 overexpression in A431-P cells in an attempt to mimic A431-III cells, and found that overexpression of MMP-9 did indeed induce EMT and that this event could be reversed by treatment with either Lu or Qu, just like in A431-P cells (Fig. 3). Consideration of these findings leads to the conclusion that Lu and Qu are multifunctional molecules (agents) that have the potential to modify multiple targets in A431-III cells and these modifications lead to an attenuation of cancer invasiveness.
A431 is a human epidermal carcinoma cell line that overexpresses EGFR.(41) Treatment with EGF is able to induce EMT in some cancer cells, including A431 cells.(42,43) Therefore, one would anticipate that a blockade of this activity or the expression of EGFR should be a potential target that allows cancer progression to be intercepted. In the current investigation, A431-P cells were treated with EGF (100 ng/mL) for 3 days and this led to a mesenchymal-like cell morphology with downregulation of E-cadherin, upregulation of various EMT markers, and an obliteration of cell–cell junctions (Fig. 4). Exposure of A431-P cells for 3 days to EGF (100 ng/mL) plus 10 μM Lu or Qu resulted in the cells retaining their epithelial-like morphology and remaining similar to the controls. Treatment with either Lu or Qu upregulated E-cadherin that was suppressed by EGF, inhibited EMT markers induced by EGF, and restored cell–cell junctions (Fig. 4). These results indicate that treatment with either Lu or Qu is able to undo the EGF-induced EMT that occurs in A431-P cells. In A431-III cells, EGF evoked an EMT process similar to that in A431-P cells. Contrary to our expectations, EGF had little discernible influence on the level of Snail in A431-III cells (Fig. 5C). This may be attributed to A431-III cells already possessing a high level of Snail that enables them to turn on EMT. Furthermore, cotreatment of A431-III cells with EGF (100 ng/mL) and either flavanoid (10 μM Lu or Qu) for 3 days, led to epithelial-like morphological changes, an inhibition in the manifestation of EMT markers induced by EGF, and a restoration of cell–cell junctions (Fig. 5).
In conclusion, our present results lucidly illustrate the biological potency of dietary flavonoids Lu and Qu and emphasize several activities of these compounds that are relevant to attenuation of tumor progression in the A431-III cell line. In general, these two polyphenolic flavonoids reversed cadherin switching, downregulated EMT markers, moderated the expunging of cell–cell interactions, and suppressed the invasiveness of the highly invasive A431-III cells, while also abating the EMT process. Thus Lu and Qu may have inherent potential as chemopreventive, antineoplastic, or chemotherapeutic agents that are able to attenuate tumor progression, particularly through a reversal of the EMT in cancer cells.
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,) and the Taiwan Academia Sinica Thematic Project (AS-96-TP-B06 to M.T.L.).
The authors have no conflicts of interest to declare.