Epithelial-to-mesenchymal transition (EMT) is the process of disaggregating structured epithelial units to enable cell motility and morphogenesis occurring in the restricted events such as embryonic development and wound healing. For the retention of epithelial integrity in the adult organism, E-cadherin plays an important role for the formation of stable cell-cell adhesion. Loss of E-cadherin expression is a primal molecular event on the progression of tumour.1 During the invasive process, tumour cells frequently undergo EMT, which comprises breakdown of E-cadherin-mediated interaction as an essential process.2 Snail, a zinc-finger transcription factor, triggers EMT through direct repression of E-cadherin.3, 4 δEF-1, Twist and TGF-β also induce EMT in tumour cells and it accelerates malignancy of tumour.5, 6 In addition, maintenance of cell polarity is also fundamental for the identity of epithelial tissues. ΔNp63α, the predominant isoform of p63 which belongs to the p53 family, in basal keratinocytes plays a crucial role in the formation of stratified squamous epithelial structures by regulating asymmetric division.7 We previously found that loss of p63 is accompanied by EMT in human squamous cell carcinoma (SCC) cells, and demonstrated that forced depletion of p63 leads to the acquisition of high-invasive capability independent of E-cadherin.8 Significantly, the fact that re-expression of ΔNp63α indeed prevents the invasiveness selectively in Snail-expressing SCC cells with EMT features indicates a certain mechanism of the ΔNp63α-dependent regulation for tumour invasion, consequently, providing a new therapeutic target.8
Basic helix-loop-helix (bHLH) transcriptional regulator controls the cellular differentiation, growth programs and tumourigenesis.9 E2A belongs to the bHLH family, which functions as dimers with the E-protein though an HLH domain and bind their DNA-binding domains to E-box, N-box or Ets sites present in the promoter regions of regulated genes.10 During the EMT process in epithelial cells, E2A represses E-cadherin expression by direct binding to the E-cadherin promoter region.11 The Id (inhibitor of differentiation) family, helix-loop-helix proteins, which lacks DNA-binding domain functions a dominant-negative regulator of bHLH through the formation of inactive heterodimers with intact bHLH and shows important differences in their expression pattern and binding specificity for bHLH.12 Id-1 has recently been proposed as a key regulator in tumour progression as well as cell proliferation, differentiation and tumourigenesis.13 Id-1 expression is found to be associated with activation of angiogenesis,14, 15 induction of tumour invasion,16 and expressions of matrix-metalloproteinases (MMPs).17–20 More recently, the Snail-induced Id-1 expression in MDCK epithelial cells and the decrease of Id-1 expression in breast cancer cells by Snail knockdown have been described.21, 22 Interaction between Id-1 and Caveolin-1, a binding partner of Id-1, plays a key role in EMT and cell migratory.23 It has also been documented that TGF- β down -regulates Id-1 expression via a transcriptional repressor, ATF-3,24, 25 and down-regulation of Id stimulates E2A-dependent repression of E-cadherin, leading to EMT in mammary epithelial cells.26 These observations imply that Id and EMT participate in similar processes of cellular and tumour invasion.
Here we report a novel molecular mechanism that mediates the ΔNp63α-induced Id-3 expression in suppression for the MMP-2-induced invasiveness of SCC cells with EMT features, supporting evidence for the ΔNp63α-dependent regulation of tumour invasion.
Material and methods
Reagents and antibodies
The MMP-2 inhibitor I was obtained from Calbiochem. Commercially available antibodies were as follows: anti-V5 tag (Invitrogen), anti-p63 which are specific for ΔNp63 isoforms (Ab-1, Oncogene Research Products), anti-Id-1 (BD biosciences), anti-Id-2 (Santa Cruz Biotechnology), anti-Id-3 (Santa Cruz Biotechnology), anti-E2A (Santa Cruz Biotechnology), anti-E-cadherin (Santa Cruz Biotechnology), anti-N-cadherin (Santa Cruz Biotechnology), anti-MMP-2 (Cell Signaling Technology), and anti-α-tubulin (Zymed Laboratories).
The expression vector plasmid
The Id-3 full-length cDNA (GenBank Accession Number: X69111) was amplified by RT-PCR with Pfx polymerase (Invitrogen) and cloned into the BamH I and Xho I sites of pcDNA3.1-V5/His-tagged expression vector (Invitrogen). Primers for amplification were 5′-GGATCCACCATGAAGGCGCTGAGCCC-3′ and 5′-CTCGAGTCTGTGGCAAAAGCTCCTT-3′. The sequence of the PCR product was verified by sequencing. The Snail expression vector, pcDNA3-mm SnailHA, was kindly provided by Dr. de Herreros (Universitat Pompeu Fabra, Barcelona, Spain). Both the ΔNp63α expression vector, pcDNA6-ΔNp63α V5/His-tagged, and the Ets-1 expression vector, pcDNA3.1-Ets-1 V5/His-tagged, have previously been generated.8, 27
Cell lines and cell culture
The human oral SCC cell lines, OM-1 and HOC313, have been described and the forced Snail-expressing OM-1 cell line and the forced ΔNp63α-expressing HOC313 cell line have been generated previously.8, 28 The p63 knockdowned OM-1 cell line and the control siRNA OM-1 cell line which expressed constitutive p63-silencing shRNA (p63 siRNA) and control shRNA (control siRNA), respectively, were also obtained and described previously.8 The Id-3 expression vector or the empty pcDNA3.1-V5/His vector as control was transfected into HOC313 cells, and stable cell clones were established by G418 selection. All cell lines were cultured at 37°C in a humidified atmosphere of 5 % CO2 in air and maintained with DMEM (Sigma) supplemented with 10% fetal bovine serum (FBS) (Sigma).
Total RNAs were isolated from the cells in 70–80% confluence with Trizol (Invitrogen). First-strand synthesis was performed with First-strand cDNA synthesis kit (Roche). Semi-quantitative RT-PCR reactions (20 μl) were amplified with 30 cycles of denaturing at 95°C for 30 sec, annealing for 30 sec and extension at 72°C for 1 min. Primers and annealing temperatures were follows:
MMP-2, 5′-GAGCTGCAACCTGTTTGTGCTGAA-3′ and 5′-ACGAGCAAAGGCATCATCCACTGT-3′, 58°C;
MMP-9, 5′-AACCAATCTCACCGACAG-3′ and 5′-AAA GGCGTCGTCAATCAC-3′, 54°C;
Ets-1, 5′-GGGTGACGACTTCTTGTTTG-3′ and 5′-GTTAA TGGAGTCAACCCAGC-3′, 56°C;
SIP1, 5′-ACCTTACGAGTGCCCAA-3′ and 5′-TTCAGTGT GGGAAACCCAG-3′, 56°C;
δEF-1, 5′-AAAGCGCTTCTCACACTCTG-3′ and 5′-TGTC CTAAGCTGCTTGCTTG-3′, 56°C;
G3PDH, 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCC ACCACCCTGTTGCTGTA-3′, 56°C.
PCR products were analyzed by 1.8% agarose gel electrophoresis and sequenced to verify their identity.
Real-time quantitative RT-PCR was carried out using LineGene software. The fluorescence was detected by Fluorescence Quantitative Detection System (BioFlux, Tokyo, Japan) and the detection was performed by measuring the binding of a fluorescence dye, SYBR Green I (TOYOBO, Osaka, Japan), to double-stranded DNA. The quantification of each gene amplification relative to an internal control, G3PDH, was calculated by delta threshold cycle (dCt) method as described previously.29 Primer sets were follows:
Id-1, 5′-AGGTGGTGCGCTGTCTGT-3′ and 5′-GATTCC GAGTTCAGCTCCAA-3′;
Id-2, 5′-TATTGTCAGCCTGCATCACC-3′ and 5′-AGAA CACCCTGGGAAGATGA-3′;
Id-3, 5′-GAACGCAGTCTGGCCATC-3′ and 5′-AAGCTCC TTTTGTCGTTGGA-3′.
Luciferase reporter assay
The Id-3 promoter region of nucleotide (nt) −2368 to +126 was amplified with Pfx polymerase (Invitrogen) from genomic DNA of normal human fibroblasts with primers, 5′-GGTACCT GAGGCATCAGCTGCAGTAG-3′ and 5′-CTCGAGGCTGGG GAGTGAGTCCAGAG-3′. PCR product was cloned into the Kpn I and Xho I sites of pGL3-basic vector (Promega). The E-cadherin reporter construct has kindly been provided from Dr. Frans van Roy and described previously.30 The MMP-2 reporter construct containing an Ets-1 binding site (nt −1255 to +22) has been generated as reported previously.27 Cells were co-transfected with the reporter construct and 1 ng of phRL-CMV (Promega) as an inner control for transfection efficiency using FuGENE 6 (Roche). Either empty vector, pcDNA6-ΔNp63α V5/His-tagged, pcDNA3.1-Ets-1 V5/His-tagged, pcDNA3.1-Id-3 V5/His-tagged or pcDNA3-mm SnailHA was further transfected. At 24 hr after transfection, cells were lysed with passive lysis buffer, and the promoter activity was measured with a Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer's protocol. The results correspond to the mean of at least 3 independent experiments.
Gelatinase substrate gel electrophoresis was performed with gels containing 12% polyacrylamide and 0.1% gelatin. Cells were incubated in a serum free medium for 4 hr and the conditioned medium were collected. After electrophoresis, the gels were incubated in reaction solution (20 mM Tris-HCl (pH 7.6), 10 mM CaCl2 and 0.04% NaN3) for 24 hr at 37°C and then stained with Coomassie Blue.
MatriGel™ cell invasion assay
Cell invasion activity was measured with BioCoat MatriGel™ Invasion Chamber (Becton Dickinson) according to the manufacturer's protocol. Briefly, cells were suspended in DMEM medium containing 5 × 105 cells/ml. Cell suspension (500 μl) was added to each chambers containing an 8-μm pore size PET membrane with MatriGel™ basement membrane and incubated for 30 hr at 37°C and 5% CO2 atmosphere. Cells on the bottom surface of the membrane were fixed with 4% paraformaldehyde, and stained with hematoxylin and counted as the invading cells.
In vitro three-dimensional culture
Three-dimensional (3D) cultures of epithelial cells with contracted collagen I gel containing fibroblasts were performed as described previously.8 Briefly, immortalised fibroblasts were suspended in a mixture of type I collagen (Koken, Tokyo, Japan) and DMEM medium containing 10% Fetal bovine serum and seeded in 12-well culture dishes. The collagen was allowed to solidify by incubating at 37°C for 1 hr. The final concentrations of collagen and immortalised fibroblasts were 1 mg/ml and 1 × 106 cells/ml, respectively. Epithelial cells (1 × 106) suspended in 1 ml of culture medium were seeded on the collagen gel. After incubation at 37°C for 1 hr, the gels were removed from the sides and bottoms of dishes and floated in the medium. After 1 week of incubation, the contracted gel was placed on a nylon mesh, and culture medium was added until the fluid level reached the upper edge of the gel. The gels were incubated under air-liquid interface culture for 1 more week. Culture medium was changed every second day. The gel was fixed with Mildform® (Wako, Osaka, Japan), embedded in paraffin and stained with hematoxylin-eosin (HE).
Real-time quantitative RT-PCR and MatriGel™ cell invasion assay were carried out with no less than 3 replicates per experiment and statistical analysis of the experiments was performed using Microsoft excel. A Students t-test was used for comparison between 2 groups. Significance was defined as a p value <0.01.
ΔNp63α positively regulates Id-3 expression
To understand the molecular mechanism underlying the attenuation confined to the invasiveness by re-introduction of ΔNp63α in human SCC cell line with EMT phenotype (i.e., HOC313),8 we first explored downstream targets of ΔNp63α. Real-time quantitative RT-PCR detected moderate but significant increase of Id-3 mRNA expression (p = 0.0076) in the ΔNp63α-expressing HOC313 cells among the Id family (Fig. 1a). The comparable increase in Id-3 protein (Fig. 1b) ensured significant ΔNp63-dependent Id-3 expression. To confirm the ΔNp63α-dependent Id-3 expression, we next examined the Id-3 promoter activity in response to ΔNp63α by luciferase reporter assay (Fig. 1c). The Id-3 promoter activity was remarkably elevated in HOC313 cells in which the ΔNp63α expression vector was co-transfected (Fig. 1c).
Id-3-dependent E-cadherin promoter activity was completely suppressed by Snail
To test possible roles of Id-3 in the context of the ΔNp63α-dependent suppression of the invasiveness of EMT cell, we generated the HOC313 cell line enforced to express exogenous Id-3 stably. Previous reports have demonstrated that Id regulates E-cadherin expression through the binding to E2A, a bHLH transcriptional repressor of E-cadherin.12 However, immunoblot analyses revealed that neither the ectopic ΔNp63α nor Id-3 elicited E-cadherin expression in HOC313 cells (Fig. 2a). Accordingly, HOC313 cells displayed similar cell morphology (Fig. 2b) despite of the enforced presence of Id-3.
Since our established SCC cell lines with EMT features (e.g., HOC313) commonly express Snail, of which expression was scarce in SCC cell lines with epithelial features (e.g., A431 and OM-1),28 we postulated that Snail suppresses E-cadherin expression beyond the potential promotion through Id-3-E2A axis in HOC313 cells. To elucidate the mechanism for the constant suppression of E-cadherin regardless of the enforced presence of Id-3, E-cadherin reporter gene assay was performed in HEK293 cells. Despite the basal transcriptional activity of E-cadherin promoter was remarkably enhanced by Id-3, Snail completely suppressed the Id-3-dependent promoter activity (Fig. 2c). Furthermore, real-time quantitative RT-PCR and immnoblot analysis showed significant down-regulation of Id-3 expression (p = 0.0025) and comparable decrease of Id-3 protein expression in the forced Snail-expressing OM-1 cells which underwent EMT (Fig. 2d).28
These data suggest that Snail is sufficient to override the Id-3-E2A axis-mediated E-cadherin expression. In fact, the Id-3-overexpressing HOC313 cells still represent EMT features such as mesenchymal cell-like morphology and “cadherin-switch” as observed in ΔNp63α-overexpressing HOC313 cells.8
Id-3 inhibits the invasiveness of SCC cells with EMT phenotype
Next we investigated whether the enhanced presence of Id-3 could impose regression of the invasiveness on HOC313 cells. The enhanced expression of Id-3 weakened the high-invasive behaviour of HOC313 cells significantly detected (p = 0.00053) in MatriGel™ invasion assay (Fig. 3a) and decreased the invaded cells into the dermis-mimicking layer consisting of the type I collagen in in vitro 3D culture (Fig. 3b). These results predict a bona fide regulator for matrix invasion as well as a target gene of ΔNp63α because a possible elevation of the cell motility in consequence of the enhanced expression of Id-3 was excluded by the result of wound healing assay on the culture dish (Fig. 3c).
Id-3 down-regulates MMP-2 expression for suppression of the invasiveness of the Snail-expressing SCC cells with EMT features
On the hypothesis that a pivotal role of Id-3 for the deprivation of the invasiveness corresponds to diminished extra-cellular matrix-associated proteolysis, we next inspected expressions of MMPs (Fig. 4a). Semi-quantitative RT-PCR proved the exogenous Id-3 to down-regulate MMP-2 expression in HOC313 cells (Fig. 4a). The immunoblot using a specific antibody to MMP-2 revealed that the exogenous Id-3 decreased the amount of secreted MMP-2 as well as intracellular MMP-2 (Fig. 4b). We also confirmed that the major gelatinase activity in HOC313 cells was by MMP-2 and that the secreted activity from HOC313 cells with the exogenous Id-3 was compromised in zymograph (Fig. 4c).
The exogenous Id-3 affected neither expression of Ets-1, SIP1 nor δEF-1 (Fig. 4a), all of which were essential for TGF-β-mediated extinguishment of E-cadherin and down-regulation of Id expression.31 We have previously demonstrated that Snail up-regulates expression of transcription factor Ets-1, promoting the Ets-1-induced MMP-2 expression.27 To verify whether Id-3 interferes with the Ets-1-driven MMP-2 transcription, we performed the MMP-2 reporter gene assay (Fig. 4d). In HEK293 cells, Ets-1 expression clearly enhanced the activity of the MMP-2 promoter, whereas the additional Id-3 expression unequivocally suppressed the Ets-1-dependent gain (Fig. 4d). Taken together, these results indicate that Id-3 compromises the invasiveness of HOC313 cells through suppression for the EMT-associated MMP-2-up-regulating machinery (i.e., Snail-dependent Ets-1 expression) despite of constant expressions of the direct down-regulators for E-cadherin.
Depletion of p63 down-regulates Id expression and involves MMP-2 up-regulation for the acquisition of invasiveness in SCC cells with epithelial features
We verified the role of ΔNp63-dependent Id-3 expression in SCC cell line with epithelial features (e.g., strong expression of E-cadherin) that acquired the invasive ability by siRNA-mediated p63 knockdown.8 The forced depletion of p63 significantlydown-regulated Id-1 to -3 expressions (p = 0.0099, 0.0053 and 0.0039, respectively) in OM-1 cells (Fig. 5a), although the forced presence of ΔNp63α up-regulated only Id-3 expression in HOC313 cells. Semi-quantitative RT-PCR showed pronounced increase of the MMP-2 mRNA expression in the p63 knockdowned OM-1 cells (Fig. 5b). Unexpectedly, Ets-1 expression was also elevated concomitantly with the p63 knockdown (Fig. 5b). We also confirmed comparable up-regulation of protein expressions for Id-3 as well as Id-1, Id-2, and MMP-2 and the secreted gelatinolytic activity by the p63 knockdown in OM-1 cells (Fig. 5c). To know whether the gained MMP-2 production actually connects with the acquired invasiveness by the elimination of p63, we carried out the MatriGel™ invasion assay applied to the p63 knockdowned OM-1 cells with or without MMP-2 inhibitor (Fig. 5d). The MMP-2 inhibitor conferred specific decrease upon the number of the invaded cells (p = 0.0062) for the p63 knockdowned OM-1 cells (Fig. 5d), confirming that MMP-2 was capable to regulate the invasiveness of SCC cells in the downstream of ΔNp63α-Id3 axis even without Snail-mediated machinery for EMT.
Our present study is the first report to define a functional role of Id-3 specific to the modification of behaviours of SCC cells in vitro. Id family displays diverse effects with their transcriptional regulation by cell specificities in tumours.12 Expression level of Id-1 correlates with progressive behaviours in a variety of tumours13; nonetheless, no evidence has been found that involves tumour progression in the altered pattern of expression of Id-3 in different tumours.32
Snail represses E-cadherin to lead SCC cells to EMT, resulting in acquisition of invasive behaviour. Simultaneously it interferes with expression of ΔNp63α to increase their invasive activity. The fact that the depletion of p63 in SCC cells promotes their invasiveness without the Snail-directed repression of E-cadherin supports a unique ΔNp63α-loss-evoked signaling pathway in regulation for the invasiveness during the Snail-mediated EMT process. In this study, we identified Id-3 as a core molecule on the ΔNp63α-dependent regulation for the invasiveness of SCC cells. We found that expression level of Id-3 was concomitantly regulated by the existence of ΔNp63α (Figs. 1a and 1b) and the enforced expression of Id-3 impaired the invasive capacity of HOC313 cells through the down-regulation of MMP-2. Since the exogenous Id-3 could inhibit the Ets-1-mediated MMP-2 expression in HOC313 cells without changing the amount of endogenous Ets-1 (Fig. 4a), Id-3 might restrict the Ets-1 transcriptional activity through forming suppressive complex between Id and Ets family.31, 33, 34 Unexpectedly, the ablation of ΔNp63α in OM-1 cells increased the Ets-1 expression (Fig. 5b) accompanied by the Id-3 down-regulation, resulting in the adequate MMP-2 production in the absence of upstream regulator of EMT (i.e., Snail). In contrast, inhibition of the Ets-1 expression by the ΔNp63α would be counteracted by the Snail-mediated Ets-1 promotion27 which gave a good account of the inadequate suppressive effect to the Ets-1 expression by the exogenous ΔNp63α in HOC313 cells (Fig. 4a). Because the exogenous ΔNp63α in HOC313 cells was capable of inducing endogenous Id3 significantly but slightly (Fig. 2a), the exogenous ΔNp63α was less effective against the down-regulation of MMP-2 expression than the exogenous Id-3 (Fig. 4a); however, the comparable suppression for the invasiveness in Matrigel™ invasion assay was acquired (Fig. 3a). These findings suggest that ΔNp63α drives several axes including the Id-3-mediated MMP-2 suppression for regulating the invasiveness. This prediction was also supported by the fact that MMP inhibitor caused significant but inadequate suppression for the p63 knockdown-mediated invasiveness in OM-1 cells (Fig. 5d). Additionally, our previous observation that re-introduced ΔNp63α stratified HOC313 cells on the collagen gel in in vitro 3D culture proposes that ΔNp63α might provide SCC cells with epithelial integrity as well as prevention of invasion.8 Conversely, HOC313 cells with the enforced expression of Id-3 retained single layer on the collagen gel in the consequence of the compromised invasiveness, indicating Id-3 act as one of the downstream targets for ΔNp63α-mediated behaviours to restrain tumour invasion.
Figure 6 schematically represents that the Id-3-dependent pathway contributes a particular effect to the suppression of invasiveness in the ΔNp63α avenue to the regulation of invasiveness of SCC cells. Snail enables SCC cells to present EMT characteristic features and high-invasiveness (e.g., down-regulation of E-cadherin and Ets-1-dependent up-regulation of MMP-2). Simultaneously, Snail also represses the C/EBPα-driven ΔNp63α expression,8 resulting in the evading from Id-3-dependent surveillance to Ets-1-MMP-2 axis which accelerates the invasive behaviour. Although the exogenous ΔNp63α drove only Id-3 expression, the mechanisms how ΔNp63α positively regulates transcription of Id-3 is still unclear. The endogenous ΔNp63α also preserves the constitutive expressions of Id-1 and -2 in non-EMT phenotype OM-1 cells (Fig. 5a), offering the possible contributions of Ids to suppress the MMP-dependent invasiveness in non-EMT phenotype SCC cells in vitro. We are currently investigating other axes regulating by ΔNp63α for restraining the invasiveness of SCC cells. Finally, our study would provide the additional evidence to application of Id-3 as a potential therapeutic target selective to aggressive SCC.
We thank Dr. Kohei Miyazono (Department of Molecular Pathology, University of Tokyo) for valuable discussion and our colleagues at Department of Oral and Maxillofacial Surgery for encouragement to accomplish the study.