This article is mentioned in:

Introduction

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common type of human cancer worldwide 1). Oral squamous cell carcinoma (OSCC), which is a subset of HNSCC, is characterized by a high risk of lymph node metastasis and local invasion. Although systemic therapeutic strategies, including chemo- and radiotherapy, have been developed for the treatment of patients with OSCC, the 5-year overall survival rate remains at ~50% (2), due to uncontrolled local adjacent tissue invasion and neck lymph node metastasis (3-5).

Signal transducer and activator of transcription 3 (STAT3) has been identified as a negative prognostic factor in human cancer, including OSCC (6). Upon cytokine stimulation, STAT3 translocates into the cell nucleus where it triggers the transcription of various downstream target genes, which have important roles in cancer invasion and metastasis (7,8). Although previous studies have provided information regarding the oncogenic role of STAT3 in OSCC (6-8), the detailed regulatory mechanism underlying the invasion-metastasis cascade remains poorly understood. Therefore, determining the underlying molecular network of advanced OSCC is required, in order to identify more efficient molecular targets with therapeutic potential.

Enhancer of zeste homolog 2 (EZH2) is a component of polycomb repressive complex 2 (PRC2), which is known to catalyze trimethylation of lysine 27 in histone 3 (H3K27me3) and consequently induce transcriptional repression of target genes (9,10). A previous study reported that EZH2 may mediate the oncogenic role of STAT3 in cancer invasion (11). Furthermore, phosphorylated (p)-EZH2 (Ser21) is able to enhance STAT3 activity through epigenetic modification (12,13). Members of the microRNA (miR)-200 family, which includes miR-200a, miR-200b, miR-200c, miR-141 and miR-429, are frequently silenced in advanced carcinoma (14-16) and serve tumor-suppressive roles in cell behaviors, such as invasion and metastasis (14,17-20). Emerging evidence has suggested that miR-200 family members may regulate epithelial-to-mesenchymal transition (EMT) by targeting the E-cadherin repressors zinc finger E-box binding homeobox (ZEB)1 and ZEB2 (21,22). For example, overexpression of miR-429 has been reported to reduce tumor invasion and metastasis in ovarian cancer via the initiation of mesenchymal-epithelial transition (MET) (20). Notably, a previous study demonstrated that miR-200a, miR-200b and miR-429 (miR-200a/b/429) were globally silenced by EZH2 activation through epigenetic modification in gastric cancer (23).

Based on these previous findings, the present study took integrated approaches to determine the crosstalk between STAT3 and the EZH2/miR-200a/b/429 axis in OSCC. The results indicated that targeting STAT3 could significantly inhibit EMT-mediated invasion in vitro, via the EZH2/miR-200a/b/429 axis. Furthermore, EZH2 suppression markedly reduced the oncogenic role of STAT3 in tumor invasion. In addition, animal models suggested that depletion of STAT3 and EZH2, or elevated miR-200a/b/429, could globally suppress OSCC invasion and metastasis in vivo.

Materials and methods

Cell culture

The OSCC cell lines SCC25 and SCC15 were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 supplemented with 10% fetal bovine serum (FBS) (both from Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and penicillin (100 U/ml)/streptomycin (100 μg/ml) (HyClone; GE Healthcare, Logan, UT, USA). Cells were maintained at 37°C in an atmosphere containing 5% CO2, and were regularly checked for mycoplasma contamination.

Cell groups

Cells were divided into the following groups: i) Empty vector (EV) group, cells were infected with lenti-viruses containing EV (multiplicity of infection, 10) at 37°C for 24 h; ii) EV + Stattic group, cells were infected with lentiviruses containing EV (multiplicity of infection, 10) at 37°C for 24 h and were then treated with Stattic [half maximal inhibitory concentration (IC50) for each cell line] at 37°C for 24 h; iii) STAT3 group, cells were infected with lentiviruses containing STAT3 vector (multiplicity of infection, 10) at 37°C for 24 h; iv) negative control (NC) small interfering (si)RNA (si-NC) group, cells were transfected with si-NC (20 nM) at 37°C for 24 h; v) si-EZH2 group, cells were transfected with si-EZH2 (20 nM) at 37°C for 24 h; vi) si-NC + EV group, cells were transfected with si-NC (20 nM) at 37°C for 24 h and were then infected with lentiviruses containing EV (multiplicity of infection, 10) at 37°C for 24 h; vii) si-NC + STAT3 group, cells were transfected with si-NC (20 nM) at 37°C for 24 h and were then infected with lentiviruses containing STAT3 vector (multiplicity of infection, 10) at 37°C for 24 h; viii) si-EZH2 + STAT3 group, cells were transfected with si-EZH2 (20nM) at 37°C for 24 h and were then infected with lentiviruses containing STAT3 vector (multiplicity of infection, 10) at 37°C for 24 h.

Antibodies and reagents

Dimethyl sulfoxide (DMSO) was purchased from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany). The STAT3 phosphorylation inhibitor Stattic and the EZH2 inhibitor 3-deazaneplanocin A (DZNeP) were purchased from Selleck Chemicals (Houston, TX, USA). pLVX-IRES-puro vectors were purchased from Shanghai GeneChem Co., Ltd. (Shanghai, China) to establish stable SCC25 or SCC15 clones overexpressing STAT3. The primary antibodies used in the present study are listed in Table I.

Table I Primary antibodies used in the present study.

Table I

Primary antibodies used in the present study.

Primary antibody	Catalog no.	Vendor	Application
STAT3	ab119352	Abcam	WB/IHC
p-STAT3 (Tyr705)	ab76315	Abcam	WB/IHC
Twist	ab50581	Abcam	WB
EZH2	5246	Cell Signaling Technology, Inc.	WB/IHC
H3K27me3	9733	Cell Signaling Technology, Inc.	WB/IHC
Histone 3	4499	Cell Signaling Technology, Inc.	WB
E-cadherin	610181	BD Biosciences	WB/IF/IHC
N-cadherin	610920	BD Biosciences	WB/IF/IHC
Vimentin	33541	AbSci	WB/IF/IHC
MMP2	sc-13595	Santa Cruz Biotechnology, Inc.	WB
MMP9	sc-21733	Santa Cruz Biotechnology, Inc.	WB
GAPDH	TA-08	OriGene Technologies, Inc.	WB

[i] Abcam, Cambridge, UK; Cell Signaling Technology, Inc., Danvers, MA, USA, BD Biosciences, Franklin Lakes, NJ, USA; AbSci, Vancouver, WA, USA; Santa Cruz Biotechnology, Inc., Dallas, TX, USA; OriGene Technologies, Inc., Beijing, China. EZH2, enhancer of zeste homolog 2; H3K27me3, trimethylation of lysine 27 in histone H3; IF, immunofluorescence; IHC, immunohistochemistry; MMP, matrix metalloproteinase; STAT3, signal transducer and activator of transcription 3; p-STAT3, phosphorylated-STAT3; WB, western blotting.

MTT assay

SCC25 and SCC15 cell lines (5,000 cells/well) were seeded into 96-well plates and incubated at 37°C for 24 h to allow for stabilization, after which, the cells were exposed to Stattic (0.1, 0.5, 1, 2, 4, 6, 8 or 10 μmol/l) at 37°C for 24 h. Cell viability was measured using an MTT assay (5 mg/ml; Sigma-Aldrich; Merck KGaA). The MTT crystals were dissolved in DMSO, and absorbance was assessed at 490 nm using a microplate reader (Model 680; Bio-Rad Laboratories, Inc., Hercules, CA, USA). IC50 was calculated using SPSS software (version 16.0; SPSS, Inc., Chicago, IL, USA).

Transwell assay

After transfection or drug administration, the Transwell assays were performed in 24-well culture plates; Matrigel-coated inserts were used to analyze invasion and uncoated inserts were used to analyze migration. Briefly, SCC25 or SCC15 cells (50,000 cells/well) were plated into Transwell inserts (Corning Incorporated, Corning, NY, USA) for invasion or migration assay, which had been coated with or without Matrigel (BD Biosciences, Franklin Lakes, NJ, USA), in FBS-free medium. The lower chamber was filled with complete growth medium containing 20% FBS. The cells were incubated at 37°C for 16 h, after which the cells that had passed through the membrane were fixed with 4% paraformaldehyde and were stained with 0.1% crystal violet (both from Beijing Solarbio Science & Technology Co., Ltd., Beijing, China). Cells were counted in four separate fields in three independent experiments, using an inverted microscope (DMI6000B; Leica Microsystems, Inc., Buffalo Grove, IL, USA).

Wound-healing assay

A wound-healing assay was conducted to confirm the results of the Transwell assay. Briefly, 2×105 SCC25 and SCC15 cells/well were plated in 6-well plates. Once the cells had reached 80% confluence, scratches were generated in the cell layer using a 10 μl pipette tip, thus creating a wound field ~400 mm wide, based on the scale plate in the microscope. Photomicrographs were captured of live cells at 0 and 48 h, using an inverted microscope (DMI6000B; Leica Microsystems, Inc.), and the distance migrated was determined within an appropriate time.

RNA isolation and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

Total RNA was isolated using TRIzol® (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Equal amounts of RNA were reverse transcribed into cDNA using the PrimeScript 1st Strand cDNA Synthesis kit (Nanjing KeyGen BioTech Co., Ltd., Nanjing, China) according to the manufacturer's protocol. qPCR analysis was conducted using the SYBR Green PCR Master Mix (Nanjing KeyGen BioTech Co., Ltd.) according to the manufacturer's protocol (thermocycling: 95°C initial denaturation for 2 min; followed by 40 cycles of denaturation at 95°C for 15 sec, and annealing and extension at 60°C for 60 sec), was employed to detect the relative expression levels of miR-200b, miR-200a and miR-429. U6 was used as a loading control, and the 2−ΔΔCq method (24) was used to calculate relative gene abundance. Primers used in the present study are shown in Table II.

Table II List of primers and oligonucleotides used in the present study.

Table II

List of primers and oligonucleotides used in the present study.

Primer	Sequence
miR-200a
RT	5′-CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGACATCGTT-3′
PCR	F: 5′-ACACTCCAGCTGGGTAACACTGTCTGGTAA-3′
	R: 5′-TGGTGTCGTGGAGTCG-3′
miR-200b
RT	5′-CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGTCATCATT-3′
PCR	F: 5′-ACACTCCAGCTGGGTAATACTGCCTGGTAA-3′
	R: 5′-TGGTGTCGTGGAGTCG-3′
miR-429
RT	5′-CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGACGGTTTT-3′
PCR	F: 5′-ACACTCCAGCTGGGTAATACTGTCTGGTAA-3′
	R: 5′-TGGTGTCGTGGAGTCG-3′
U6
PCR	F: 5′-CTCGCTTCGGCAGCACA-3′
	R: 5′-AACGCTTCACGAATTTGCGT-3′

[i] F, forward; miR, microRNA; PCR, polymerase chain reaction; R, reverse; RT, reverse transcription.

Western blotting

Western blot analysis was employed to assess protein expression. Cell lysates were prepared in radioimmunoprecipitation assay buffer supplemented with protease and phosphatase inhibitors (both from Beijing Solarbio Science & Technology Co., Ltd.). Protein concentrations were determined using a bicinchoninic acid (BCA) protein assay kit (Micro BCA Protein Assay kit; Thermo Fisher Scientific, Inc.). The protein samples (20-30 μg) were separated by 10% SDS-PAGE and were transferred onto polyvinylidene difluoride membranes (Merck KGaA). The membranes were blocked with 5% non-fat milk in Tris-buffered saline containing 0.1% Tween-20 at room temperature for 2 h. Subsequently, the membranes were incubated with the following antibodies (1:1,000 dilutions) overnight at 4°C: STAT3, p-STAT3 (Tyr705), Twist (Abcam, Cambridge, UK), EZH2, H3K27me3 (Cell Signaling Technology, Inc., Danvers, MA, USA), E-cadherin, N-cadherin (BD Biosciences), Vimentin (AbSci, Vancouver, WA, USA), matrix metallo-proteinase (MMP)2 and MMP9 (Santa Cruz Biotechnology, Inc., Dallas, TX, USA). GAPDH (OriGene Technologies, Inc., Beijing, China) and Histone 3 (Cell Signaling Technology, Inc.) were used as internal controls (Table I). Horseradish peroxidase-conjugated anti-rabbit (sc-2004) or anti-mouse (sc-2005) immunoglobulin G antibodies (Santa Cruz Biotechnology, Inc.) were used as secondary antibodies (1:5,000 dilutions), and blots were incubated with them at room temperature for 1 h. ImageJ software (Version 5.2.5; National Institutes of Health, Bethesda, MD, USA) was used to semi-quantify the relative expression levels of the target proteins, which were normalized to the respective internal controls.

Immunofluorescence staining

For immunofluorescence staining, OSCC cells were grown on 18-mm coverslips for 24 h following treatment. Immunofluorescence staining was conducted using primary antibodies against E-cadherin, N-cadherin (1:100 dilutions; BD Biosciences) and Vimentin (1:100 dilution; AbSci), overnight at 4°C. The cells were then washed with PBS and incubated with AlexaFluor 488 (anti-rabbit: #4412; anti-mouse: #4408) or AlexaFluor 594 (anti-rabbit: #8889; anti-mouse: #8890) secondary antibodies (1:500 dilutions; Cell Signaling Technology, Inc.) at room temperature for 1 h. For stress fiber formation assessment, cells were stained with DyLight™ 594-phalloidin (1:500 dilution; #12877; Cell Signaling Technology, Inc.) at room temperature for 1 h. The nuclei were stained using DAPI (Thermo Fisher Scientific, Inc.), and each slide was visualized using an FV-1000 laser scanning confocal microscope (Olympus Corporation, Tokyo, Japan).

Lentivirus packaging and transduction

Total RNA from both HNSCC cell lines (SCC15 and SCC25) was isolated using TRIzol® (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. RNA was reverse transcribed into cDNA using the PrimeScript 1st Strand cDNA Synthesis kit (Nanjing KeyGen BioTech Co., Ltd.) according to the manufacturer's protocol. Then, STAT3 coding sequence was amplified by PCR, using the SYBR-Green PCR Master Mix (Nanjing KeyGen BioTech Co., Ltd.) according to the manufacturer's instructions, and cloned into the pLVX-IRES-puro vector (Shanghai GeneChem Co., Ltd.). All constructs were verified by sequencing (Sanger chain termination method), which was conducted by AuGCT (Beijing, China). Lentivirus was prepared in 293T cells (ATCC) using Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. For transduction, cells (30-40% confluence) were infected with lentiviruses (multiplicity of infection, 10) at 37°C for 24 h. SCC25 or SCC15 clones stably overexpressing STAT3 were labeled as STAT3, and an EV was used as a negative control.

Transient transfection

SCC25 and SCC15 cell lines were transfected with NC siRNA or siRNA against EZH2 (20 nM; Guangzhou Ribobio Co., Ltd., Guangzhou, China), which were labeled as si-NC (5′-UUCUCCGAACGUGUCACGU-3′) or si-EZH2 (#1, 5′-GCUGGAAUCAAAGGAUACA-3′; #2, 5′-GTGCCCTTGTGTGATAGCACAA-3′; #3, 5′-GGCACT TTCATTGAAGAACTAA-3′), respectively, using Lipo-fectamine 2000 reagent, according to the manufacturer's protocol. Briefly, cells (~40% confluence) were transfected with 20 nM siRNAs at 37°C for 24 h. During transfection, FBS-free Opti-MEM medium was employed, and after 6 h, the medium was replaced with DMEM-F12 or DMEM.

In vivo orthotopic tumor model

To establish orthotopic models, 5×106 SCC15 cells, which were transduced with luciferase lentivirus (Promega Corporation, Madison, WI, USA), were injected into the base of the oral cavity of 4-week-old, male BALB/c-nu mice (n=40; weight, ~15 g) (4,25). The animals were obtained from the Animal Center at the Cancer Institute, Chinese Academy of Medical Science (Beijing, China). Mice were maintained under the following conditions: Temperature, 18-22°C; humidity, 50-60%; 12-h light/dark cycle; food and water access, ad libitum. To verify tumor establishment, fluorescence images were captured 7 days after injection using an IVIS Lumina Imaging system (PerkinElmer, Inc., Waltham, MA, USA). Mice were randomly assigned to five groups (n=8/group): DMSO (20 μl dissolved in 100 μl PBS), Stattic (3.75 mg/kg) (26), DZNeP (1 mg/kg) (27), NC mimics (5′-UUUGUACUACACAAAAGUACUG-3′, 10 nM/mice) or miR-429 mimics (5′-UAAUACUGUCUGGUAAAACCGU-3′, 10 nM/mice; Guangzhou Ribobio Co., Ltd.); treatments were administered by intraperitoneal injection every 3 days for 21 days. Bioluminescence imaging was conducted to measure tumor volume each week, and body weight was measured daily. Finally, the animals were sacrificed by cervical dislocation and the orthotopic tumors were collected for further examination. Tumor volume was finally measured using a caliper (Volume = long diameter x short diameter2/2). Tumor specimens were used for IHC and RT-qPCR. The present study was approved by the Institutional Animal Care and Use Committee of Tianjin Medical University Cancer Institute and Hospital (Tianjin, China).

Immunohistochemistry (IHC) and hematoxylin and eosin (H&E) staining

For IHC, tumor samples were fixed with 40% formalin in methanol for 2 days at room temperature and were embedded in paraffin. Subsequently, formalin-fixed, paraffin-embedded SCC15-derived OSCC orthotopic tumor samples were deparaffinized, rehydrated and incubated with primary antibodies (1:100 dilutions) overnight at 4°C. Subsequently, tissue sections were incubated with a biotin-labeled secondary antibody (PV-9000; OriGene Technologies, Inc.) for 40 min at 37°C. Cells were visualized using ABC-peroxidase and diaminobenzidine (Vector Laboratories, Inc., Burlingame, CA, USA) at room temperature for 30 sec, counterstained with hematoxylin and visualized using light microscopy. The primary antibodies used in this investigation are listed in Table I. In addition, H&E staining was performed as previously described (28).

Database analysis

The online software cBioPortal (www.cbio-portal.org) and Gene Expression Profiling Interactive Analysis (GEPIA; gepia.cancer-pku.cn) were used to analyze TCGA (The Cancer Genome Atlas) database.

Statistical analysis

SPSS 16.0 (SPSS, Inc.) was used for all statistical analyses. Statistical comparisons between two groups were made using Student's t-test. Differences between more than two groups were determined by two-way analysis of variance followed by Dunnett's test. All data are presented as the means ± standard deviation, and represent the average of at least three experiments performed in triplicate. P<0.05 was considered to indicate a statistically significant difference.

Results

Inhibition of STAT3 suppresses OSCC tumor invasion and migration in vitro

According to previous studies, Stattic, a STAT3-selective phosphorylation inhibitor, may inhibit the expression levels of p-STAT3 (Tyr705), without any effect on total STAT3 (26,29). As shown in Fig. 1A, western blot analysis demonstrated that treatment with Stattic (24 h IC50 for SCC15: 1.979 μM; IC50 for SCC25: 1.388 μM; Fig. 1B) markedly suppressed MMP2 and MMP9 expression in SCC25 and SCC15 cell lines, which may be induced by inhibition of p-STAT3 (Tyr705). Furthermore, STAT3 silencing significantly impaired the invasive and migratory capabilities of OSCC cells, according to the results of a Transwell assay and scratch test analysis in vitro (Fig. 1C and D). Meanwhile, stable OSCC clones overexpressing STAT3 were generated, in order to verify the function of STAT3 in OSCC cells. Stable SCC25 or SCC15 clones overexpressing STAT3 were labeled STAT3, and cells infected with EV were used as a negative control. Compared with in the EV group, cells overexpressing STAT3 possessed markedly increased MMP2 and MMP9 protein expression, leading to a significant elevation in tumor cell migration and invasion (Fig. 1A, C and D).

Figure 1 STAT3 contributes to invasion and migration of OSCC cells. (A) Western blot detection of STAT3, p-STAT3 (Tyr705), MMP2 and MMP9 expression. GAPDH was used as a loading control. (B) MTT analysis was used to determine the IC50 values for both OSCC cell lines. (C) Targeting STAT3 significantly inhibited SCC25 and SCC15 cell invasion and migration ability, as determined by Transwell assay (magnification, ×40). (D) Cell migration ability was measured by a wound-healing assay (magnification, ×40). ***P<0.001 vs. EV group; ###P<0.001 EV + Stattic group vs. STAT3 group. EV, empty vector; IC50, half maximal inhibitory concentration; MMP, matrix metalloproteinase; OSCC, oral squamous cell carcinoma; p-STAT3, phosphorylated-STAT3; STAT3, signal transducer and activator of transcription 3.

Regulatory role of the EZH2/miR-200/a/b/429 axis in EMT of OSCC

Previous studies have indicated that downregulation of EZH2 may increase miR-200b/a/429 expression in human cancers (23,30). In order to explore the regulatory role in OSCC, two EZH2 siRNAs (si#1 and si#2) were used to inhibit the expression of EZH2. Subsequently si#2 was selected for further analysis (Fig. 2A). In siRNA-transfected OSCC cells, EZH2 expression was markedly inhibited, as was H3K27me3 (Fig. 2B). p-EZH2 (Ser21) has previously been reported to significantly enhance STAT3 activity through epigenetic modification (12,13). In the present study, p-STAT3 expression was suppressed by EZH2 attenuation. Furthermore, qPCR was used to detect miR-200-a/b/429 expression in both cell lines. Compared with in the untreated SCC15 and SCC25 cells, EZH2-depleted cells exhibited significantly increased miR-200-b/a/429 expression (Fig. 2C).

Figure 2 EZH2/miR-200/b/a/429 axis regulates the invasiveness of OSCC cells in vitro. (A) Following transfection with EZH2 siRNAs (si#1 and si#2) for 3 days, the expression levels of EZH2 were detected using western blotting. (B) OSCC cells underwent western blot analysis using antibodies specific to STAT3, p-STAT3 (Tyr705), EZH2, H3K27me3, H3 and GAPDH. EZH2 siRNA markedly inhibited EZH2 and H3K27me3 expression. In addition, p-STAT3 (Tyr705), but not STAT3, was suppressed by EZH2 depletion. (C) EZH2 siRNA induced miR-200b/a/429 expression. (D) Results of a Transwell assay demonstrated that EZH2 knockdown markedly attenuated migration (without Matrigel) and invasion (with Matrigel) of SCC25 and SCC15 cells (scale bar, 100 μm; magnification, ×40). (E) Scratch assay demonstrated that EZH2 silencing markedly delayed wound healing in SCC25 and SCC15 cells (magnification, ×40). *P<0.05 and ***P<0.001 vs. si-NC group. EZH2, enhancer of zeste homolog 3; H3, histone 3; H3K27me3, trimethylation of lysine 27 in H3; miR-200b/a/429, microRNA-200b, -200a and -429; NC, negative control; OSCC, oral squamous cell carcinoma; p-STAT3, phosphorylated-STAT3; siRNA/si, small interfering RNA; STAT3, signal transducer and activator of transcription 3.

The present study also aimed to determine whether cell motility was governed by the EZH2/miR-200-b/a/429 regulatory network. The results of Transwell and wound healing assays indicated that EZH2 silencing markedly inhibited the invasion and migration of OSCC cell lines (Fig. 2D and E). To confirm these results, immunofluorescence staining was conducted to directly visualize the effects of EZH2 abrogation on E-cadherin expression, localization and cell morphology. As shown in Fig. 3, si-EZH2-transfected SCC15 and SCC25 cells possessed an epithelial phenotype, as indicated by an increase in E-cadherin expression on the cell membrane and rearrangement of F-actin to a relatively cortical pattern, which is a hallmark of the epithelial phenotype.

Figure 3 EZH2 silencing affects the EMT process in OSCC cells. (A) Protein expression levels of EMT-associated markers were analyzed by western blotting and were normalized to GAPDH. (B) F-actin staining exhibited a stress-fiber pattern in NC siRNA-transfected cells, whereas a cortical pattern was observed in EZH2 siRNA-transfected cells, as assessed by immunofluorescence (scale bar, 20 μm). (C) E-cadherin, N-cadherin and Vimentin staining of OSCC cells transfected with si-EZH2 or si-NC. Cell nuclei were stained with DAPI (scale bar, 20 μm). EMT, epithelial-mesenchymal transition; EZH2, enhancer of zeste homolog 3; MMP, matrix metalloproteinase; NC, negative control; OSCC, oral squamous cell carcinoma; siRNA/si, small interfering RNA.

Targeting STAT3 signaling regulates EZH2 and miR-200b/a/429, thus affecting EMT in OSCC in vitro

A previous study verified that the EZH2/miR-200b/a/429 axis is a critical regulatory network associated with tumor motility. Notably, STAT3 has been reported to transcriptionally regulate EZH2 in human colorectal cancer (11). Based on these findings, the present study aimed to determine whether STAT3 could trigger OSCC invasion via the EZH2/miR-200b/a/429 axis.

Analysis of TCGA demonstrated that EZH2 expression was positively correlated with STAT3 expression in human HNSCC specimens (Fig. 4A), which was in line with the outcome of the OSCC dataset in TCGA analyzed by GEPIA. Moreover, western blot analysis demonstrated that treatment with Stattic markedly suppressed p-STAT3 (Tyr705, Fig. 1A) and EZH2 (Fig. 4B), and inhibited EZH2-induced H3K27me3 in both OSCC cell lines (Fig. 4B). Notably, STAT3 stable clones exhibited increased EZH2 expression (Fig. 4B). Subsequently, the present study examined whether targeting STAT3 could regulate the expression levels of miR-200b/a/429. The expression levels of miR-200b/a/429 were detected prior to and following treatment with Stattic using qPCR. The results of qPCR revealed that Stattic-induced STAT3 inhibition markedly enhanced mir-200b/a/429 expression (Fig. 4C). Conversely, overexpression of STAT3 globally silenced the tumor suppressor miRNAs, miR-200b/a/429, compared with in the EV group (Fig. 4C). These data indicated that STAT3 may inhibit the expression levels of mir-200b/a/429 in OSCC cells via EZH2-induced epigenetic silencing.

Figure 4 EZH2/miR-200/b/a/429 axis mediates the prometastatic role of STAT3 in an EMT-dependent manner. (A) The Cancer Genome Atlas data were used to determine the correlation between STAT3 and EZH2 in head and neck squamous cell carcinoma (P<0.05). (B) Western blot analysis of EZH2, H3K27me3 and EMT markers following Stattic treatment or STAT3 transduction. (C) Stattic upregulated miR-200b/a/429 expression, whereas ectopic STAT3 overexpression significantly attenuated mir-200b/a/429 expression. (D) Expression of the epithelial marker E-cadherin, and the mesenchymal markers N-cadherin and Vimentin, were detected by immunofluorescence staining (scale bar, 20 μm). ***P<0.001 vs. EV group; ###P<0.001 EV + Stattic group vs. STAT3 group. EMT, epithelial-mesenchymal transition; EV, empty vector; EZH2, enhancer of zeste homolog 3; H3, histone 3; H3K27me3, trimethylation of lysine 27 in H3; miR-200b/a/429, microRNA-200b, -200a and -429; STAT3, signal transducer and activator of transcription 3.

SCC25 and SCC15 cells treated with EV, Stattic or a STAT3 plasmid were examined by immunofluorescence; Stattic treatment resulted in less regularly shaped cells compared with in the control group (Fig. 4D). Furthermore, STAT3-depleted cells exhibited reduced formation of robust actin stress fibers, suggesting that features of mesenchymal cells and mobility features were decreased (Fig. 4D). In addition, in STAT3-depleted cells E-cadherin expression was increased, whereas N-cadherin and Vimentin expression were decreased (Fig. 4D). Conversely, STAT3 enhancement markedly induced the mesenchymal phenotype, which was characterized by an absence of E-cadherin on the cell membrane and rearrangement of F-actin from a cortical to stress-fiber pattern, thus indicating the presence of robust mesenchymal features. Subsequent western blot analysis verified the immunofluorescence results of EMT markers (Fig. 4B).

EZH2 silencing counteracts STAT3 activation in OSCC cells in vitro

To verify the underlying mechanism by which STAT3 governs OSCC invasion and migration, the role of EZH2 was determined in STAT3-mediated cell biology. When OSCC cells overexpressing STAT3 were transfected with si-EZH2 for 72 h, the effects of STAT3 on H3K27me3 and miR-200b/a/429 were significantly attenuated in vitro (Fig. 5A and B). Furthermore, invasion assays demonstrated that EZH2 knockdown markedly reversed the oncogenic effects of STAT3 on tumor invasion and migration (Fig. 5C and D).

Figure 5 EZH2 silencing counteracts STAT3-induced invasion by targeting miR-200b/a/429. (A) STAT3 and EZH2 expression levels were evaluated by western blotting. (B) EZH2 depletion markedly reduced the inhibitory effects of STAT3 on miR-200b/a/429 expression. (C and D) EZH2 knockdown reduced the invasion and migration of oral squamous cell carcinoma cells overexpressing STAT3 (magnification, ×40). *P<0.05 and ***P<0.001 vs. si-NC + EV group; ###P<0.001 si-NC + STAT3 group vs. si-EZH2 + STAT3 group. EV, empty vector; EZH2, enhancer of zeste homolog 3; H3, histone 3; H3K27me3, tri-methylation of lysine 27 in H3; miR-200b/a/429, microRNA-200b, -200a and -429; p-STAT3, phosphorylated-STAT3; siRNA/si, small interfering RNA; STAT3, signal transducer and activator of transcription 3.

Western blot analysis was used to determine whether EZH2/miR-200/b/a/429 mediated the prometastatic effects of STAT3 on the expression of EMT markers. In both cell lines, EZH2 knockdown reduced the oncogenic effects of STAT3, leading to increased E-cadherin and decreased N-cadherin expression (Fig. 6A). To further confirm these results, immunofluorescence staining was performed to directly visualize EMT markers and cell morphology. As shown in Fig. 6B, si-EZH2-transfected OSCCs possessed epithelial cell features, as characterized by a typical cobblestone structure and membrane-localized E-cadherin. Conversely, cells transfected with si-NC possessed a mesenchymal phenotype following ectopic overexpression of STAT3. These results suggested that the EZH2/miR-200b/a/429 axis may contribute to the STAT3-directed OSCC invasion and migration.

Figure 6 EZH2 silencing impairs STAT3-induced EMT-mediated metastasis. (A) Protein expression levels of EMT-associated markers were analyzed using western blotting and were normalized to GAPDH. (B) Subcellular location and expression of the epithelial marker E-cadherin, and the mesenchymal markers N-cadherin and Vimentin, in oral squamous cell carcinoma cells. F-actin distribution was rearranged to a cortical pattern following EZH2 knockdown (scale bar, 20 μm). EZH2, enhancer of zeste homolog 3; siRNA/si, small interfering RNA; STAT3, signal transducer and activator of transcription 3.

Targeting STAT3 and EZH2 reduces OSCC tumor invasion in orthotopic mouse models

The results of the present study suggested that the EZH2/miR-200/b/a/429 axis may trigger the oncogenic role of STAT3 in OSCC. To further validate the therapeutic potential of targeting STAT3 and EZH2, a proof-of-concept experiment was employed using an orthotopic OSCC tumor model, derived from the SCC15 cell line (Fig. 7). Bioluminescence assay results (Fig. 7A) and tumor volumes (Fig. 7D) indicated that Stattic-treated (3.75 mg/kg) and DZNeP-treated (1 mg/kg) mice exhibited significantly reduced growth stasis compared with the DMSO group (Table III; P<0.05). None of the mice developed numerous tumors. No significant alteration in body weight (Fig. 7B) was observed in the mice treated with the inhibitors (initial weight: DMSO, 17.6 g; Stattic, 17.3 g; DZNeP, 18.3 g; weight at sacrifice: DMSO, 19.5g; Stattic, 19.4 g; DZNeP, 18.6 g). To evaluate the therapeutic effects on tumor invasion, the tumors were resected from each group. As shown in Fig. 7C, Stattic-treated and DZNeP-treated mice exhibited reduced local invasion compared with in the control group.

Figure 7 Targeting STAT3 and EZH2 inhibits tumorigenesis and invasiveness in an orthotopic OSCC model. (A) Representative bioluminescence images of mice implanted with orthotopic tumors, which were intraperitoneally treated with 3.75 mg/kg Stattic, 1 mg/kg DZNeP or DMSO every 3 days. (B) Based on body weight, no detectable toxicity was observed at the tested dose. (C) Representative photographs of the tumor resection procedure. Stattic and DZNeP treatment inhibited local invasion of orthotopic OSCC tumors. (D) Tumor diameter and volume were measured. (E) Quantitative polymerase chain reaction was used to detect miR-200b/a/429 expression in OSCC tumor sections. (F) Tumor samples from three distinct groups underwent immunohistochemistry for STAT3, p-STAT3 (Tyr705), EZH2, H3K27me3, MMP2, MMP6, E-cadherin, N-cadherin and Vimentin expression (scale bar, 100 μm; magnification, ×200). ***P<0.001 vs. DMSO group. DMSO, dimethyl sulfoxide; DZNeP, 3-deazaneplanocin A; EZH2, enhancer of zeste homolog 3; H3K27me3, trimethylation of lysine 27 in histone 3; H&E, hematoxylin and eosin; miR, microRNA; MMP, matrix metalloproteinase; p-STAT3, phosphorylated-STAT3; STAT3, signal transducer and activator of transcription 3.

Table III Bioluminescence data obtained from the animal study (107 p/sec/cm2/sr).

Table III

Bioluminescence data obtained from the animal study (107 p/sec/cm2/sr).

Group	Mouse
	#1	#2	#3	#4	#5	#6	#7	#8
DMSO
D0	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
D7	0.980	1.040	0.450	1.010	1.050	0.660	1.122	0.990
D14	2.110	2.770	3.310	2.090	1.980	3.150	2.220	3.010
D21	5.310	5.010	6.500	5.420	5.120	5.990	4.500	4.620
D28	9.870	8.340	7.990	8.550	7.040	9.010	9.550	8.010
D35	14.110	15.550	14.770	14.010	15.310	14.980	14.410	15.010
D42	18.110	19.100	22.100	19.020	18.960	20.810	18.410	18.760
Stattic
D0	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
D7	0.410	0.350	0.220	0.390	0.370	0.200	0.430	0.340
D14	0.660	0.870	0.410	0.600	0.940	0.360	0.700	0.880
D21	1.710	2.240	0.880	1.820	2.020	1.980	1.410	2.010
D28	3.110	2.330	3.010	3.010	2.330	3.320	2.880	2.130
D35	5.120	6.110	5.110	5.010	6.410	5.020	5.310	6.010
D42	9.010	11.210	9.010	10.100	10.310	8.510	8.670	10.010
DZNeP
D0	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
D7	0.340	0.120	0.410	0.300	0.160	0.330	0.440	0.090
D14	0.440	0.560	0.320	0.410	0.590	0.300	0.450	0.580
D21	1.200	1.110	0.990	1.310	0.990	1.010	1.220	1.040
D28	1.400	1.890	1.550	3.010	1.120	2.780	1.250	1.110
D35	4.110	3.030	7.010	4.010	3.710	7.310	3.910	5.120
D42	9.010	5.020	6.990	7.040	5.010	9.210	10.910	5.210
Negative control mimic
D0	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
D7	0.980	1.040	0.980	0.870	1.120	1.010	0.990	1.120
D14	2.110	2.770	3.450	2.230	2.540	2.780	2.450	3.010
D21	5.310	5.010	4.000	6.010	4.980	3.780	5.870	4.120
D28	10.120	8.000	7.000	10.120	8.120	7.000	10.120	8.240
D35	14.110	15.120	16.000	13.990	15.610	16.870	14.210	14.980
D42	22.000	19.100	21.000	23.120	18.120	22.340	21.980	19.870
microRNA-429
D0	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000
D7	0.410	0.350	0.410	0.430	0.400	0.360	0.340	0.380
D14	0.660	0.870	0.660	0.570	0.890	0.760	0.550	0.910
D21	2.110	1.880	1.980	2.010	1.780	1.990	1.750	2.070
D28	4.000	2.010	3.110	4.010	2.450	3.410	3.780	2.310
D35	5.000	7.010	7.000	4.560	7.010	6.310	4.780	6.010
D42	8.000	8.000	11.000	7.670	8.120	12.010	7.890	10.110

[i] D, day; DMSO, dimethyl sulfoxide; DZNeP, 3-deazaneplanocin A.

Notably, IHC demonstrated that p-STAT3 (Tyr705), EZH2 and H3K27me3 expression in Stattic- and DZNeP-treated groups were markedly suppressed (Fig. 7F). In addition, the expression levels of mir-200b/a/429 were elevated in both groups (Fig. 7E). Notably, compared with in the control tumors, Stattic and DZNeP treatment was sufficient to attenuate the cell invasion proteins MMP2 and 9, and the EMT-associated biomarkers N-cadherin and Vimentin in vivo, alongside an increase in E-cadherin expression (Fig. 7F).

Systematic delivery of miR-200b/a/429 markedly suppresses N-cadherin and induces E-cadherin expression in vivo

The present study indicated that the EZH2/miR-200b/a/429 axis may mediate the oncogenic role of STAT3 in tumor invasion via EMT interaction. To further confirm these results, another in vivo experiment was conducted to verify the inhibitory role of miR-200b/a/429 in tumor invasion. Since miR-429 is the most sensitive miRNA, according to upstream stimulation in vitro in the present study, as assessed by RT-qPCR (Figs. 2C, 4C and 5B), miR-429 was selected for further analysis.

A total of 7 days after tumor implantation, miR-control (miR-Ctrl) and miR-429 were intraperitoneally injected every 3 days (Fig. 8). Body weight was assessed daily, and tumor volume was measured each week using bioluminescence imaging. As shown in Fig. 8A and D, delivery of miR-429 markedly reduced tumor volume compared with in the miR-Ctrl group (Table III). None of the mice developed numerous tumors. No significant alteration in body weight was observed during treatment (Fig. 8B) (initial weight: miR-Ctrl, 18.1 g; miR-429, 17.9 g; weight at sacrifice: miR-Ctrl, 19.32 g; miR-429, 18.7 g). Furthermore, miR-429 overexpression markedly reduced tumor invasion (Fig. 8C).

Figure 8 miR-429 inhibits tumor progression in the orthotopic mouse model of OSCC. (A) Representative bioluminescence images of mice implanted with orthotopic tumors and treated intraperitoneally with miR-Ctrl (10 nM/mice) or miR-429 (10 nM/mice) every 3 days. (B) Quantification of body weight in control and miR-429-treated mice. (C) miR-429 inhibits OSCC local invasion in an orthotopic mouse model. (D) Tumor diameter and volume were measured. (E) Quantitative polymerase chain reaction was used to detect the expression levels of mir-429 in OSCC tumor sections. (F) Tumor samples from control and miR-429-treated mice underwent immunohistochemistry for MMP2, MMP9, E-cadherin, N-cadherin and Vimentin (scale bar, 100 μm magnification, ×200). ***P<0.001 vs. miR-Ctrl group. Ctrl, control; H&E, hematoxylin and eosin; miR, microRNA; MMP, matrix metalloproteinase; NC, negative control.

RT-qPCR was used to compare the expression levels of mir-429 between the two groups (Fig. 8E). In addition, IHC was conducted to determine whether systemic delivery of miR-429 affected the expression levels of EMT markers. As shown in Fig. 8F, miR-429 treatment markedly inhibited N-cadherin and Vimentin expression, and increased E-cadherin expression, compared with in the miR-Ctrl group. Furthermore, MMP2 and MMP9 were attenuated by miR-429 administration. These results indicated that elevating miR-200b/a/429 expression may be considered a potential therapeutic strategy for the treatment of patients with OSCC.

Discussion

Increasing attention has recently been devoted to the aberrant activation of STAT3, due to its critical role in the progression of numerous human malignancies, including lung cancer (31), pancreatic cancer (32,33), colorectal cancer (34) and OSCC (35-37). Extensive functional studies have characterized STAT3 as a potential therapeutic target in OSCC. The present study demonstrated that targeting STAT3 could inhibit OSCC invasion and metastasis in vitro and in vivo via the EZH2/miR-200b/a/429 axis.

STAT3 signaling is well known to contribute to cancer progression, invasion and metastasis. It has previously been suggested that targeting STAT3 may be considered a promising therapeutic strategy for the treatment of solid tumors. In the present study, inhibition of STAT3 decreased MMP2 and MMP9 expression, which was accompanied by reduced invasion and migration of OSCC cells. In addition, targeting STAT3 could inhibit tumor local invasion in vivo. Conversely, cells overexpressing STAT3 exhibited a robust malignant phenotype. These findings suggested that STAT3 may have a contributing role in OSCC invasion and migration.

It has been reported that EZH2 induces transcriptional repression of numerous target genes through the formation of a PRC2 complex with EED and SUZ12 proteins, and is highly relevant to cancer invasion and metastasis (38). In particular, the present integrated analysis of EZH2 highlighted its important role in OSCC. In the present study, EZH2 depletion markedly suppressed OSCC cell motility in vitro and in animal models. Subsequent western blotting and immunofluorescence staining revealed that EZH2 silencing significantly hindered EMT by retaining E-cadherin expression in the cell membrane. IHC staining also detected MET progression following DZNeP exposure in vivo.

Recent efforts have focused on the regulatory network between STAT3 and EZH2. In experimental models, EZH2 may significantly enhance STAT3 accumulation in tumor cells via methylation, thus promoting tumorigenicity in humans (12,13,39). In addition, STAT3 has been globally validated to transcriptionally regulate EZH2 in gastric cancer (40) and colorectal cancer (11). Therefore, a feed-forward loop between STAT3 and EZH2 has been successfully established and verified in human cancer. Therefore, the present study focused on the regulatory relationship between STAT3 and EZH2, in order to explore whether this crosstalk contributes to OSCC invasion and metastasis. TCGA network analysis confirmed the positive correlation between STAT3 and EZH2 among ~500 OSCC cases on the basis of transcriptome data. In the present study, treatment with Stattic significantly attenuated EZH2 expression via targeting STAT3. Furthermore, the present mechanistic investigation revealed that knockdown of EZH2 may partially reduce the oncogenic role of STAT3 in OSCC migration and invasion. In addition, STAT3-induced EMT was markedly suppressed by EZH2 knockdown. Knockdown of EZH2 also inhibited STAT3 phosphorylation, whereas inhibition or activation of STAT3 phosphorylation affected EZH2 levels accordingly, establishing a feed-forward loop between these two proteins. Consequently, the present study established a regulatory network between STAT3 and EZH2 by in vitro experiments and animal models. This finding may provide a rationale for the treatment of OSCC.

miR-200b/a/429 genes have been reported to serve as tumor suppressors in numerous types of human cancer (41,42) and have been demonstrated to be major regulators of EMT (14,19,43). Silencing miR-200 genes by methylation has been reported to promote tumor growth and invasion via EMT induction (44,45). The present results demonstrated that systemic delivery of miR-429 induced a marked decrease in OSCC cell growth and local invasion. Notably, EZH2 was validated to abrogate miR-200b/a/429 expression levels through epigenetic regulation. The EZH2/miR-200b/a/429 axis has previously been identified as a critical therapeutic target in cancers of epithelial origin (23). In the present study, EZH2 depletion significantly increased miR-200b/a/429 expression in vitro and in vivo. Together with the confirmed role of EZH2 in EMT, the EZH2/miR-200b/a/429 axis may serve a critical role in tumor invasion and EMT-directed metastasis. Furthermore, STAT3 was revealed to regulate miR-200b/a/429 in an EZH2-dependent manner. These data provide evidence to suggest that the EZH2/miR-200b/a/429 axis may mediate the oncogenic role of STAT3 in vitro and in vivo.

On the basis of increasing integrated research and experimental evidence, STAT3 has been revealed to serve a central role in cancer progression, and its overexpression is associated with tumor cell proliferation, invasiveness and metastasis in human cancers, including OSCC. Accordingly, the present study revealed the therapeutic effect of targeting STAT3, and the mechanism by which STAT3 exerts its oncogenic role in OSCC. The results indicated that STAT3 or EZH2 inhibition, as well as miR-429 systemic delivery, may markedly attenuate local tumor invasion and outgrowth, which may have potential significance for the identification of targeted therapies for patients with OSCC.

In conclusion, the present study is the first, to the best of our knowledge, to identify the mechanism by which STAT3 governs OSCC invasion and metastasis, and revealed a master regulatory network between STAT3 and the EZH2/miR-200b/a/429 axis in OSCC. Although STAT3-based therapeutics are in their infancy, these findings are still encouraging and suggest that STAT3 is a potential target; therefore, these findings may be translated clinically for the treatment of OSCC.

Glossary

Abbreviations

Abbreviations:

Acknowledgments

The authors would like to thank Dr Jie Zhao at Tianjin Hospital (Tianjin, China) for his kind assistance.

Notes

[1] Funding

This study was supported by the National Natural Science Foundation of China (grant no. 81572492) and the CSCO-Merck Serono Oncology Research Fund, SCORE (grant no. Y-MT2015-017).

[2] Availability of data and materials

All data generated or analyzed during this study are included in this published article.

[3] Authors' contributions

YWa, YWu, XZ, MZ, XQ and LK made substantial contributions to the conception and design of the study. ZL, CZ, LZ, JD, RJ, YQ and WG made substantial contributions to data acquisition. ZL, CJ, YS, JC, XW and YR made substantial contributions to data analysis and interpretation. MZ, LZ, SS, CW and JD were also involved in drafting the manuscript. LK, SS, CW, YQ, RJ and XQ were involved in revising the manuscript critically for important intellectual content. XZ, SS and CW also gave final approval of the version to be published. All authors read and approved the final manuscript.

[4] Ethics approval and consent to participate

The present study was approved by the Institutional Animal Care and Use Committee of Tianjin Medical University Cancer Institute and Hospital (Tianjin, China).

[5] Consent for publication

Not applicable.

[6] Competing interests

The authors declare that they have no competing interests.

References