Background
Oral squamous cell carcinoma (OSCC) is a major oral cavity health problem. Although many therapeutic strategies have been carried out [
1], the 5-year survival rate for these patients has remained at 50–60% for the last three decades [
2]. Tissue invasion and metastasis are exceedingly complex processes and are one of the hallmarks of cancer [
3]; thus, it is important to clarify the biological mechanism of tissue invasion and metastasis for grading the course of cancer and developing more effective therapies [
3,
4].
The epithelial-to-mesenchymal transition (EMT) is the cellular and molecular process through which cell-to-cell interactions and apico-basal polarity are lost and a mesenchymal phenotype is acquired, which are required for cell motility and basement membrane invasion during metastasis [
5,
6]. The EMT plays a critical role in embryogenesis and is associated with tissue remolding, wound healing, fibrosis, cancer progression and metastasis [
5,
7‐
9]. In the metastatic cascade of epithelial tumors, the EMT has been established as an important step [
10]. Furthermore, researchers have shown that the EMT is associated with the dedifferentiation program that leads to malignant carcinoma [
5], as the EMT confers invasive cancer cells an efficient migration ability and a selective advantage to reach distant locations [
9,
10]. Transcriptional repression of the E-cadherin gene can lead to the loss of the epithelial phenotype and the functional loss of E-cadherin is one of the hallmarks of EMT [
5]. In particular, transcriptional repressor has recently emerged as a fundamental mechanism for the silencing of CDH1 (the gene that encodes E-cadherin), such as the Snail (Snail1 and Slug), ZEB (ZEB1 and ZEB2) and basic helix-loop-helix (bHLH: Twist) families [
6,
11].
Matrix metalloproteinases (MMPs) are zinc-dependent endopeptidases. MMPs are involved in degrading extracellular matrix (ECM) in normal physiological processes, such as embryonic development, reproduction and tissue remodeling, as well as in disease processes, such as arthritis and metastasis [
12,
13]. There are over 23 MMPs identified in humans, which are subdivided into soluble MMPs and membrane-type MMPs (MT-MMPs) [
14,
15]. While MT1-MMP has a common MMP domain structure with a signal peptide, a pro-peptide, catalytic and hemopexin-like domains, it also has unique insertions. One of the insertions is at the C-terminus and contains a hydrophobic amino-acid sequence that acts as a transmembrane domain [
16,
17]. As a member of the MMPs, MT1-MMP is closely associated with cancer invasiveness and the promotion of cell migration [
16,
18‐
20]. Recent researches have emerged to indicate that cell surface MT1-MMP has been recognized as an inducer of EMT in cancer cells [
21,
22]. The researches on MT1-MMP further demonstrated that MT1-MMP via cleaving E-cadherin induced an EMT in transfected breast cancer [
21], which was shown to be dependent on up-regulation of Wnt5a in prostate cancer cells [
22]. However, the molecular transcriptional mechanism related to MT1-MMP as an inducer of EMT remains poorly understood, and the association of MT1-MMP and EMT has not been reported in oral cancer cells. Thus, we examined whether MT1-MMP-induced EMT through mediation of transcriptional repression of E-cadherin in OSCC.
Recently, studies of neoplastic tissues have provided evidence of self-renewing, stem-like cells within tumors, which have been called cancer stem cells (CSCs) [
23]. Increasing evidence suggests that EMT bestows carcinoma cells at the tumor front with cancer stem cell (CSC)-like properties and plays an important role in initiating CSCs [
24,
25]. Furthermore, CSCs have been identified in head and neck SCC [
4,
25]. However, an association specifying the EMT and CSCs induced by MT1-MMP in SCC9 cells has not been investigated.
Based on the above studies, we demonstrate the molecular mechanisms in OSCC that are involved in the overexpression of MT1-MMP by the cancer cells that induces an EMT and leads to the acquisition of CSC-like properties by the cancer cells. These studies may provide new avenues of research with potential clinical implications.
Methods
Cell cultrue, plasmid construction and transfection
Human oral squamous cell carcinoma SCC9 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were maintained in a mixture of Dulbecco’s Modified Eagle’s medium and Ham’s F12 medium (1:1) (Invitrogen, Burlington, Ontario, Canada) supplemented with 10% fetal bovine serum (FBS, Invitrogen), 400 ng/ml hydrocortisone (Sigma-Aldrich, St Louis, MO, USA) and penicillin (100 U/ml)/streptomycin (100 μg/ml) (Invitrogen). A full-length cDNA for human MT1-MMP (NM_004995.2) was amplified using RT-PCR and then ligated into the PCR2.1-TOPO vector. The constructed PCR gene product was cloned into the pEGFP-N1 vector. The final gene synthesis was verified via sequencing and amplified using DH5α competent cells. The Endo-free Plasmid Mini Kit II (OMEGA) was used for all plasmid preparations. For transfection experiments, cells were maintained in six-well plates (Corning, Lowell, MA, USA) and cultured to 80% confluence, after which the medium was changed to serum-free medium for overnight incubation. The cells were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. G418 (400 μg/ml; Invitrogen) was added to the media 48 h after transfection. The cells were allowed to grow in the presence of G418 for two weeks, and clones were picked for growth on plates to confluence. Thus, stably expressing empty vector--SCC9-pEGFP-N cells (SCC9-N) and a vector encoding human MT1-MMP--SCC9-pEGFP-M cells (SCC9-M) were obtained for our study.
For the experiment of addition of inhibitors, 2×105/ml SCC9-M cells were added to six-well plates (Corning). The cells were then treated with 5 nM tissue inhibitor of metalloproteinase (TIMP)-1 (Calbiochem, Darmstadt, Germany), 5 nM of TIMP2 (Calbiochem) and incubated for three days at 37°C.
Real-time RT-PCR
Total RNA was extracted from cells using the TRIzol reagent (Invitrogen). For cDNA synthesis, mRNA was reverse-transcribed into cDNA using the 5×PrimeScript RT Master Mix (TaKaRa) at 37°C for 15 min and 85°C for 5 s according to the manufacturer’s protocol. Gene expression was quantified by real-time quantitative PCR using 2×SYBR Premix Ex Taq (TaKaRa) with a 7300 ABI Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) under the conditions of 95°C for 30 s, 95°C for 5 s, and 60°C for 31 s for 40 cycles. The relative gene expression was calculated using the 2(−ΔΔCT) method. Briefly, the resultant mRNA was normalized to its own GAPDH [
26]. The following primers were utilized for the real-time RT-PCR. GAPDH (5
′-GAAGGTGAAGGTCGGAGTC-3
′, 5
′-GAGATGGTGATGGGATTTC -3
′), MT1-MMP (5
′-GGAACCCTGTAGCTTTGTGTCTGTC-3
′, 5
′-TGAGGGTCCTGCCTTCAAGTG-3
′), E-cadherin (5
′-TACACTGCCCAGGAGCCAGA-3
′, 5
′-TGGCACCAGTGTCCGGATTA-3
′), β-catenin (5
′-GCTGAAGGTGCTATCTGTCTGCTC-3
′, 5
′-TGAACAAGACGTTGACTTGGATCTG-3
′), cytokeratin18 (5
′-AGGAGTATGAGGCCCTGCTGAA-3
′, 5
′-TTGCATGGAGTTGCTGCTGTC-3
′), vimentin (5
′-TGAGTACCGGAGACAGGTGCAG-3
′, 5
′-TAGCAGCTTCAACGGCAAAGTTC-3
′), fibronectin (5
′ –TGCCTTGCACGATGATATGGA-3
′, 5
′-CTTGTGGGTGTGACCTGAGTGAA-3
′), snail (5
′-GACCACTATGCCGCGCTCTT-3
′, 5
′-TCGCTGTAGTTAGGCTTCCGATT-3
′), slug (5
′-ATGCATATTCGGACCCACACATTAC-3
′, 5
′-AGATTTGACCTGTCTGCAAATGCTC-3
′), Twist (5
′-GGAGTCCGCAGTCTTACGAG-3
′, 5
′-TCTGGAGGACCTGGTAGAGG-3
′), ZEB1 (5
′-GAAAGTGATCCAGCCAAATGGAA-3
′, 5
′-TTTGGGCGGTGTAGAATCAGAG-3
′), ZEB2 (5
′-AAATGCACAGAGTGTGGCAAGG-3
′, 5
′-CTGCTGATGTGCGAACTGTAGGA-3
′) and CDH1 (5
′-AGATGGTGTGATTACAGTCAAAAGG-3
′, 5
′-CAGGCGTAGACCAAGAAAT-3
′).
Western blotting and shedding of the E-cadherin ectodomain
Cells were lysed using a RIPA lysis buffer (Beyotime). Total protein (30 μg) from each sample was subjected to the SDS-PAGE and then transferred to PVDF membranes (Millipore, Billerica, MA, USA), which were blocked for 2 h at room temperature with 5% nonfat milk in PBST. The following antibodies were used to detect bands on the protein blots: anti-β-actin (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-MT1-MMP (1:500, Abcam, Cambridge, MA, USA), anti-E-cadherin (1:1000, Cell Signaling Technology, Danvers, MA, USA), anti-β-catenin (1:500, Santa Cruz Biotechnology), anti-cytokeratin18 (1:500, Bioworld Technology, MN, USA), anti-vimentin (1:500, Santa Cruz Biotechnology), anti-fibronectin (1:500, Santa Cruz Biotechnology), anti-Snail (1:500, Abcam), anti-Slug (1;1000, Cell Signaling Technology), anti-Twist (1:500, Abcam), anti-ZEB1 (1:300, Abcam) and anti-ZEB2 (1:500, Novus Biologicals, Littleton, USA). Immunoreactive material was visualized using the Immun-Star WesternC Kit (Bio-Rad, Hercules, CA, USA) products and bands were detected via exposure to film (Kodak, Japan). For detecting the expression of extracellular E-cadherin, the cells were cultured with serum-free medium for 24 h. Next, the conditioned medium was collected via centrifugation and concentrated 10-fold with a VirTis freeze dryer (SP Scientific, NY, USA). An immunoblot was performed as described above using an anti-E-cadherin ectodomain antibody (1:500, Santa Cruz Biotechnology). All western bolt analyses were performed at least three independent experiments.
Immunofluorescence
Cells were cultured on glass coverslips, fixed in 4% paraformaldehyde (PFA) for 20 min at room temperature, permeabilized with 1% Triton X-100 for 15 min and blocked with goat serum albumin for 30 min 37°C, followed by an overnight incubation at 4°C with antibodies specific for E-cadherin (1:100, Cell Signaling Technology) and vimentin (1:100, Santa Cruz Biotechnology), or cytokeratin 18 (1:100, Bioworld technology) and fibronectin (1:100, Santa Cruz Biotechnology). The appropriate secondary antibodies (diluted 1:50) were then used, and then nuclei were stained by 4, 6-diamidino-2-phenylindole (DAPI; 1:1000, Invitrogen) for 2 min. Immunofluorescence was visualized using a Zeiss LSM-710 laser-scanning confocal microscopy.
Adhesion, invasion and wound healing assays
The cells were plated in six-well plates (Corning) at a density of 4×105 per well and then trypsinized after 1 and 2 h, respectively. The attached cells were counted under an inverted microscope (Olympus), and the adherent rate of the three different cell populations was calculated. The cell invasion was assessed using Transwell filters with 6.5-mm diameters and 8-μM pore sizes (Costar, Lowell, MA, USA). The filters were precoated for 30 min at 37°C with 50 μL per square centimeter of growth surface with Matrigel Basement Membrane Matrix (BD Biosciences, MA, USA) diluted with serum-free medium (1:3) according to the manufacturer’s procedures. The cells (3×105) were resuspended with 100 μl serum-free medium inoculated in the upper chamber while 500 μl medium containing 10% FBS was placed in the lower chamber. The plates were placed at 37°C in 5% CO2 for 24 h. The chambers were fixed with 4% PFA and stained with 0.1% crystal violet (Beyotime) for 30 min. The non-migratory cells were removed, and the migratory cells were counted as those presenting on the lower surface of the upper chamber. Images of at least ten random fields per chamber were captured (×100 magnification). For the wound healing assay, the cells were allowed to grow to 90% confluence and then wounded by scratching with a pipette tip in the central area. Floating cells and debris were removed, and the medium was changed to serum-free. The cells were incubated for 48 h to allow cells to grow and close the wound. Photographs were taken at the same position of the wound at the indicated time points.
Flow cytometry
For flow cytometric cell-cycle analysis, the cells were synchronized with serum-free medium for 24 h, released and then cultured for three days. The cells were detached from the culture plate with trypsin, washed with PBS, and then resuspended in 75% alcohol. The prepared cells were stained with 100 mg/ml of propidium iodide (BD Pharmingen, San Jose, CA, USA) prior to analysis using flow cytometry with a BD FACS Calibur (BD Biosciences) and CellQuest Pro software (BD Biosciences). For surface marker analysis, the cells were collected and then labeled with human-fluorochrome-conjugated anti-CD24-PE (10 μl per test, Beckman Coulter, Los Angeles, CA, USA), anti-CD44-APC (20 μl per test, BD Pharmingen), anti-CD133-PE (10 μl per test, Miltenyi Biotech, Auburn, CA, USA). The corresponding mouse immunoglobulins conjugated to PE or APC (BD Pharmingen) were used as isotype controls in each experiment. For apoptosis analysis, the cells were dealed with mitomycin at concentration gradients of 16 and 128 mg/ml for 24 h. Then the prepared cells were collected and stained with PE Annexin V Apoptosis Detection Kit I (BD Pharmingen) for 15 min according to the manufacturer’s protocol. The rate of apoptosis cells was relative to each untreated group.
The cells were plated in 100-mm dishes (Corning) at a density of 1000 cells per dish and cultured at 37°C for two weeks. The dishes were fixed in 4% PFA, stained with crystal violet, and photographed. The colonies were visualized under an inverted microscope (Olympus). Aggregations of more than 50 cells were defined as a colony.
MTT assay
The survival rate of cells was analyzed using an MTT (Sigma) assay, which is a colorimetric assay for measuring the activity of enzymes that reduce MTT to formazan dyes, producing a purple color. The MTT assay is the preferred method used to assess the viability and proliferation of cells [
27]. The SCC9-N and SCC9-M cells were plated in 96-well plates (Corning) at an initial density of 2×10
3 cells per well, and then synchronized with serum-free medium for 24 h. For consecutive culturing at 0, 1, 3, 5, 7, 9 d, the cells were treated with 5 mg/ml MTT and incubated at 37°C for 4 h, and then treated by dimethylsulfoxide (Sigma). The absorbance of samples in triplicate wells was measured with an automatic enzyme-linked immunosorbent assay reader (ELx800, BioTek Instruments, Inc., USA) at a wavelength of 490 nm. Population doubling time (PDT) was calculated according to Patterson formulation. For drug resistant experiment, the SCC9-N and SCC9-M cells were plated in 96-well plates (Corning) at the same density of 5×10
4 cells. After serum-starvation, mitomycin at concentration gradients of 16 and 128 mg/ml was added separately to the culture medium and maintained for 24 h. The absorbance of samples in triplicate wells was measured as introduced above. The survival rate of the cells relative to each untreated group was calculated. The data were analyzed using three independent experiments.
Statistical analysis
The data were representative of three or more independent experiments as the mean ± s.d. Statistical significance was assessed using one-way analysis of variance and Student’s unpaired t test. P-value <0.05 was considered significant.
Discussion
Most patients with OSCC die because of metastasis or recurrence of the tumor [
2]. However, key events mediating invasion and metastasis of this carcinoma are still undefined, although the linkage between an EMT and cancer invasion and metastasis has been understood for years [
5,
8‐
10]. Studies have suggested that EMT endows cells with stem cell-like traits [
26,
32,
33] and allows to become more invasive and migratory. Thus, our research was focused on the association of MT1-MMP, EMT and invasion and metastasis of oral carcinoma SCC9 cells; and, we made four novel observations. First, overexpression of MT1-MMP can induce oral cancer SCC9 cells to undergo EMT. Second, MT1-MMP-induced phenotypic changes in the SCC9 cells increased the level of Twist and ZEB and were dependent on repressing the transcription of E-cadherin. Third, this phenotype transformation resulted in a change in the biological properties of the cells, with the cells having decreased adhesion, high invasion but low proliferation ability. Fourth, these mesenchymal-like cells gained CSCs features.
MT1-MMP was recognized as a key mediator in both ECM remolding and cell migration during tumor progression [
17,
19]. Previous studies on MT1-MMP were focused on the relationship of its domain structures and cancer invasion and metastasis. Our study related to the connection of MT1-MMP and the EMT revealed that up-regulation of MT1-MMP can induce oral carcinoma SCC9 cells to undergo EMT via transcriptional repression of E-cadherin. Upon the overexpression of MT1-MMP, SCC9-M cells presented a fibroblast-like phenotype compared with the cubic epithelial phenotype of SCC9-N cells. In addition, analysis of the mRNA and protein levels verified that the SCC9-M cells underwent an EMT, in which decreased expression of epithelial markers (E-cadherin, β-catenin, cytokeratin 18) and increased expression of mesenchymal markers (vimentin, fibronectin) were observed. Furthermore, overexpression of MT1-MMP in SCC9 cells resulted in a change in the biological properties of the cells. The SCC9-M cells lost the need for cell-to-cell adhesion which contributed to cells becoming motile. As shown in invasion assay, the SCC9-M cells acquired a highly invasive ability. This in turn may allow cancer cells to cross the basement membrane and invade surrounding tissues. Importantly, recent studies demonstrated that MT1-MMP was essential for the invasive ability of cells, due to its broad-spectrum activity of degrading ECM components [
16,
17,
34,
35]. Our results verified that MT1-MMP promoted cancer cell invasion in OSCC through inducing the EMT.
The EMT is an important step in the metastatic process of epithelial tumors [
10], for which recent studies have provided a more in-depth understanding of the molecular mechanisms involved. Loss of E-cadherin is central to EMT in cancer cells [
5]. Thus, in the present study, we focused our attention on the transcriptional repression of E-cadherin to explain how MT1-MMP caused EMT in SCC9 cells. We also done the research to identify whether MT1-MMP overexpression resulted in the shedding of E-cadherin to induce an EMT, similar to that reported in prior studies [
21,
22]. However, an examination of extracellular E-cadherin in the conditioned medium on SCC9-M cells was nearly undetectable, not similarly as previously reported in prostate cancer. Our results demonstrated that MT1-MMP played a role in dynamic silencing of CDH1 so that transcriptional repression of E-cadherin, leading to the loss of the epithelial phenotype of SCC9 cells to undergo EMT. Indeed, several transcription factors that strongly repress CDH1 (such as members of Snail, ZEB and bHLH families) have recently emerged, which are now thought to be involved in tumor progression [
36]. The Snail family (Snail and Slug) was first identified as inducers of EMT [
37,
38], and our previous work demonstrated that Snail had played an important role in inducing an EMT in SCC9 cells [
39]. However, in this study, no significant difference in the expression of Snail and Slug were observed in the three experimental cells. Furthermore, the level of mRNA and protein expression observed on SCC9-M cells and SCC9-M cells treated with TIMP1 and TIMP2 may demonstrate that the MT1-MMP-induced EMT change was associated with an increase of Twist and ZEB. Twist and ZEB genes are key inducers of EMT and are closely associated with tumor progression [
40‐
42]. However, further investigation is required, exploring the linkage between increased expression of both Twist and ZEB via MT1-MMP.
Furthermore, the more invasive SCC9-M cells did not have the ability to close the wounds in the wound healing assay. This result was consistent with our previous work [
39] and revealed that the SCC9-M cells exhibited a low growth ability, which was further validated by cell-cycle analysis and cell proliferation assay. The cell mitosis of SCC9-M was blocked at the G0/G1 phase leading to a low percentage of the cells residing in S-phase, suggesting a decreased ability of cell growth. The cell growth curve showed that the PDT in SCC9-M cells was significantly longer than in SCC9-N cells, which further demonstrated this point that the SCC9-M cells displayed lower proliferation ability. Although the SCC9-M cells had lower growth ability, they possessed the ability of self-renewal, as demonstrated in the colony-forming assay. The less visible and smaller colonies formed by SCC9-M cells further elucidated that the SCC9-M cells had lower proliferation ability. It has been proposed that most of the CSCs exist in the quiescent G0 cell phase, which allows an escape from anti-cancer drug targeting and resistance to apoptosis [
31]. In our study, the mesenchymal-like SCC9-M cells shared the ability of chemotherapeutic resistance to mitomycin. The flow cytometric apoptosis analysis confirmed that the SCC9-M cells are more resistant to apoptosis. These results demonstrated that there existed more SCC9-M cells in a relative quiescent state, and these non-dividing SCC9-M cells shared the ability to resistance to cell death which presented CSC-like properties. Recently, many CSC signatures have been reported, such as CD24, CD44, CD133 and so forth [
4,
25,
43,
44]. In the current study, the SCC9-M cells possessed CD24
low expression in contrast to the CD24
high expression of the SCC9-N cells, while both cell populations were CD44
high. The expression of CD133 was either 0.89% or 0.29% in SCC9-M and SCC9-N cells, respectively. This result was not the same as that of prior reports; however, the CSC surface markers are not consistent across various tumors. The marker may or may not be useful for identifying stem cells from the other organ or tumor type [
45]. Our results demonstrated that the cell surface marker CD44, while certified as a CSC marker in HNSCC, was not specific to oral SCC9 cells. Thus, it is not sufficient to define a stem cell solely based on surface markers, and multiple assays are required to isolate putative CSCs efficiently. Overall, our study demonstrated that the SCC9-M cells possessed CSC-like properties, including the ability to self-renew, resistance to chemotherapeutic agents and apoptosis, and expression of CSC markers.
Conclusions
In conclusion, our study demonstrated that MT1-MMP, through repressing the transcription of E-cadherin, induced less aggressive oral SCC9 cells to undergo an EMT, which converted the SCC9-M cells into exhibiting a mesenchymal-like phenotype, and to possess more invasive ability. Furthermore, this transformation revealed a connection with CSCs. Collectively, further detailed information related to the molecular requirements for EMT will contribute to a better understand of tumor progression and may suggest more efficient targets for future therapeutic development.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
CCY and LFZ carried out the experiment and performed the data analysis. XHX, TYN and JHY participated in the experiment. CCY, LKL, and LFZ designed the study, wrote and edited the manuscript. All authors read and approved the manuscript.