MicroRNA-542-3p inhibits oral squamous cell carcinoma progression by inhibiting ILK/TGF-β1/Smad2/3 signaling

In this study, we investigated the effects of microRNA-542-3p (miR-542-3p) on ILK/TGF-β1/Smad2/3 signaling and oral squamous cell carcinoma (OSCC) progression. Levels of miR-542-3p were lower in OSCC tissues (n=108) than adjacent normal tissues, whereas levels of ILK, TGF-β1 and Smad2/3 were higher. Patients with undifferentiated tumors, advanced TNM stage and lymph node metastasis showed low miR-542-3p levels. This was accompanied by high ILK expression and poor survival. Dual luciferase reporter assays of SCC-9 cells showed that miR-542-3p inhibited ILK gene expression by binding to its 3’UTR at 233-240 bp. SCC-9 cells transfected with miR-542-3p mimics exhibited elevated miR-542-3p and decreased ILK, TGF-β1 and Smad2/3 expression. They also showed reduced self-renewal (fewer CD44+ cells and tumor-spheres), invasiveness, migration, proliferation and survival. Conversely, miR-542-3p inhibitors promoted increased self-renewal (more CD44+ cells and tumor-spheres), invasiveness, migration, proliferation and survival. In xenograft experiments with nude mice, SCC-9 cells transfected with miR-542-3p mimics or siRNA-ILK yielded tumors with smaller volumes and weights than control tumors. These results demonstrate that miR-542-3p is a tumor suppressor that inhibits ILK/TGF-β1/Smad2/3 signaling, thereby inhibiting OSCC progression.


INTRODUCTION
Oral cancer ranks sixth among all cancers with 90% of oral cancer patients diagnosed as oral squamous cell carcinoma (OSCC) [1]. It includes malignancies in the oral cavity, lip and pharynx. OSCC has the highest mortality rates among all carcinomas and is the most common head and neck cancer [1,2]. The incidence of OSCC is between the ages of 45 and 75 years, and is gradually increasing [3,4]. The etiology of OSCC involves both intrinsic and extrinsic factors with smoking, alcohol consumption and human papilloma virus infection being the main risk factors [5]. Poor therapeutic response, recurrence and metastasis are commonly encountered despite considerable advances in surgery, radiotherapy and chemotherapy [6]. OSCC progression involves multistep transformational changes with multiple genetic mutations [7]. In recent years, many biomarkers have been investigated to improve the power and cost-effectiveness of clinical applications. These include miRNAs that have been recognized as novel targets in OSCC diagnosis and therapy [8]. MicroRNAs (miRNAs) are short (17~25 nucleotides) non-coding RNAs that regulate gene expression through the posttranscriptional repression of their target mRNAs and influence various physiological and pathological processes [9][10][11]. They regulate cell proliferation, differentiation, apoptosis, angiogenesis, stress resistance and self-renewal of stem cells [12]. The microRNA-542-3p (miR-542-3p) has been recognized as a tumor suppressor gene in numerous cancers [13,14]. Transforming growth factor-β1 (TGF-β1) plays an important part in human diseases by downstream activation of TGF-β1/Smad signaling pathway and is secreted by many cell types [11]. TGF-β1/Smad signaling pathway is a major signaling pathway that regulates cancer cell growth and proliferation [15]. Integrin-linked kinase (ILK) regulates many key biological processes including cell cycle, growth and angiogenesis; its high expression enhances tumorigenesis [16][17][18]. Over expression of ILK is associated with the recurrence and poor survival for OSCC patients [19]. ILK is also the target gene of miR-542-3p [20]. A previous study demonstrated that miR-204-5p regulated proliferation and metastasis of OSCC cells by inhibiting CXCR4 [21]. MiR-483-5p was considered as an independent prognostic factor, and its high expression in OSCC patients had diagnostic and prognostic significance for OSCC [22]. Although miR-542-3p serves as a tumor suppressor in esophageal squamous cell carcinoma (ESCC), its role in OSCC is unclear [23]. Therefore, in this study, we investigated the significance of miR-542-3p in OSCC progression and survival through ILK/TGF-β1/Smad2/3 signaling pathway.

Risk factors for prognosis of OSCC patients
Kaplan-Meier survival curves showed that patients with low miR-542-3p expression (average expression of miR-542-3P was 0.67, which used as cutoff value) had lower survival rates compared to high miR-542-3p expression group ( Figure 3A). Similarly, patients with high ILK expression (20% used as cutoff value) had lower survival compared to those with low ILK expression ( Figure 3B). Kaplan-Meier survival curves based on clinicopathological factors showed that (1) patients in III/ IV stages had lower survival compared to those in the I/ II stages; (2) patients with low differentiation tumors had lower survival rates compared to those with the moderate and high differentiation tumors; and (3) patients with lymph node metastasis had lower survival rates compared to those without lymph node metastasis ( Figure 3C-3E). COX proportional hazard model analysis showed that miR-542-3p and ILK expression and lymph node metastasis indicated poor prognosis for OSCC patients (both P < 0.05; Table 2).

MiR-542-3p decreases SCC-9 cell proliferation
MTT assay demonstrated that cell proliferation was similar in the mimic control, inhibitor control, si-ILK and the blank groups (P > 0.05; Figure 7). Moreover, cell proliferation decreased in the miR-542-3p mimics and si-

MiR-542-3p mimics and ILK siRNA inhibit SCC-9 xenograft tumor growth in nude mice
As shown in Figure 11A-11B, the tumor volume and weight in nude mice xenografted with SCC-9 cells transfected with miR-542-3p mimics and siRNA-ILK groups were similar (P > 0.05) and lower compared to the control group (all P < 0.05).
MiR-542-3p has a tumor suppressor role in osteoblast cells and ESCC tissues, and gastric and bladder cancers [13,[23][24][25]. Recent studies showed low miR-542-3p expression  in bladder cancer tissues; its inhibition promoted tumor cell growth [13,14]. Dual-luciferase reporter gene assay demonstrated that ILK was the target gene of miR-542-3p and overexpression of miR-542-3p negatively regulated ILK expression. Oneyama et al. also demonstrated that miR-542-3p performed tumor suppressor function in many human tumors and controlled ILK-mediated cell adhesion and invasion during tumor progression [20].
Most importantly, we demonstrated that miR-542-3p inhibition increased the proportion of CD44 + SCC-9 cells, which suggested that miR-542-3p inhibited the proliferation of SCC-9 cells. Besides, compared with the blank group, the numbers of tumor spheres decreased in the miR-542-3p mimics and siRNA-ILK groups and increased in the miR-542-3p inhibitors group. Moreover, compared with the miR-542-3p inhibitors group, the number of tumor spheres decreased in the miR-542-3p inhibitors + siRNA-ILK group. This suggested that miR-542-3p inhibited self-renewal and differentiation of SCC-9 cells by down-regulating ILK/TGF-β1/Smad2/3 signaling pathway. Tumor sphere expands and enriches tumor cells, and tumor sphere formation is analyzed to test selfrenewal and differentiation of stem cells [26].
Cancer stem cells (CSCs) are either derived from transformation of normal stem cells that possess self-renewal capacity or from differentiated cells that acquire self-renewal properties as a result of genetic or epigenetic alterations [27]. The CSCs share characteristics of normal stem cells including self-renewal, ability to differentiate and activate anti-apoptotic pathways. Hence, CSCs can initiate tumors that are resistant to chemotherapy and metastasize to other tissues where they further differentiate and generate tumors [28]. The overexpression of miR-542-3p significantly inhibited the activation of β-catenin and nuclear factor-κB signaling pathways that promote carcinogenesis [25]. TGF-β signaling pathway also plays a very critical role in regulating cell growth, differentiation, and development of biological systems [29,30]. TGF-β1 promotes tumor cell growth through the Smad-ILK signaling pathway by increasing Smad2/3 and ILK expression [31]. TGF-β1 expression is regulated by ILK activation, which responds to various pathological and physiological stimuli via TGF-β1 [32].
Additionally, our study found that miR-542-3p inhibited the SCC-9 cell proliferation by suppressing ILK expression. Inhibition of ILK resulted in treatment-related delays in tumor growth, reduced tumor angiogenesis and functionality of tumor vasculature [33]. Moreover, overexpression of miR-542-3p inhibited the migration and invasion of SCC-9 and arrested SCC-9 cells in G0/
A previous study showed that miR-542-3p expression decreased cell proliferation by inhibiting G1 and G2/M cell cycle phases [34]. MiR-542-3p was a tumor suppressor in melanomas by reducing cell migration and invasion [35]. Our study showed decreased invasion of SCC-9 cells in the miR-542-3p mimics and siRNA-ILK groups. This suggested that over expression of miR-542-3p and low expression of ILK decreased the invasion of SCC-9 cells. ILK regulates cell proliferation, differentiation, migration and apoptosis [36]. ILK plays critical roles in progression and metastasis of OSCC, and is a critical prognostic indicator for OSCC patients [36]. Bedal et al. demonstrated that high expression of ILK promoted tumor invasion, tumor grade, clinical stage and positive lymph node status, and increased the risk and complications of disease recurrence in OSCC [37]. The over expression of ILK improved the invasiveness of tumor cells by inducing the expression of invasion-related genes [38,39]. Since our study was based on only one cell line, namely, SCC-9, further clinical research is necessary to demonstrate the efficacy of treating OSCC by regulating miR-542-3p. Also, the role of TGF-β1/Smad signaling pathway on the proliferation and survival of OSCC cells needs further experimentation.

OSCC patient specimen
OSCC tissues were obtained from 108 patients who underwent surgical treatment in First Affiliated Hospital of Zhengzhou University and Henan Cancer Hospital/ Affiliated Cancer Hospital of Zhengzhou University between August 2012 and December 2015. Among the 108 OSCC patients, 64 had moderately or highly differentiated squamous carcinoma and 44 were poorly differentiated or undifferentiated. The patients included 29 females and 79 males with an average age of 58.52 ± 14.47 years. Sixtynine patients had a history of smoking and 49 patients had lymph node metastasis. Based on Tumor Node Metastasis (TNM) staging system of Union International Against Cancer (UIAC) in 2002, 60 cases were stage I/II and 48 were stage III/IV [40]. The normal tissues adjacent to the tumors were used as controls. The patients never had any radiotherapy, chemotherapy or related treatment before surgical treatment with complete clinical data. This study was approved by the Ethics Committee of First Affiliated Hospital of Zhengzhou University and Henan Cancer Hospital/Affiliated Cancer Hospital of Zhengzhou University. Informed consent was obtained from all study subjects and/or their legal guardians.

Immunohistochemistry
The PV-9000 two-step method for immunohistochemical staining was used to determine the expression of ILK, TGF-β1 and Smad2/3 proteins in OSCC tissues. Each tissue microarray paraffin block was cut into serial 4μm sections, dewaxed and hydrated with gradient alcohol followed by antigen retrieval in a microwave. Endogenous peroxidase (PO) was blocked by incubation with 3% hydrogen peroxide (H 2 O 2 ). Then, the sections were incubated with mouse anti-human ILK antibody (1:50; Wuhan Boster Biological Technology Ltd., Wuhan, China) at 4°C overnight. Then, they were incubated with polymerase auxiliary agent for 20 min at room temperature followed by horseradish peroxidase-labeled goat anti-mouse secondary antibody (Abcam, Cambridge, MA, USA) for 30 min at room temperature. Diaminobenzidine (DAB) (Sigma-Aldrich Biological Technology Co., Ltd., USA) was used for developing. The sections were counterstained with hematoxylin and sealed. Phosphate buffered saline (PBS) replaced the primary antibody in the negative control. The cells with visible yellow or brown cytoplasm or cell membrane were identified as positive. Four high power fields (× 400) were randomly selected in each slide and 200 OSCC cells were counted in each field. Proportion of positive cells was used to determine the scores. Cells with more than 25% brown and brownish yellow granules in the cytoplasm were designated as positive., Positive expression rate was defined as positive cell numbers relative to total cell numbers. The scoring was performed independently by two individuals who were blinded to the source of the specimens.

Patient follow-up and overall survival
Follow-up was conducted by out-patient consultation or over the phone. The deadline of follow up was January 2017. Overall survival (OS) was defined as the time from first diagnosis to either death or the final time of follow-up (January 2017). At the end of follow-up (12 months), 101 patients completed follow-up data, whereas 7 patients died. Hence, the effective follow-up rate was 93.5.0%. USA) based on CD44 immunofluorescent labeling. The CD44 + cells corresponded to OSCC cells whereas the rest were non-stem-like cells. These cells were preserved and frozen for further use or used for further experiments.
The dual-luciferase reporter assay was performed to test the luciferase activity of samples. The primary medium was removed 48h after transfection. The SCC-9 cells were washed twice by PBS and lysed with 100 μl passive lysis buffer (PLB) at room temperature for 15 min. Then, the cell lysate was obtained. LARII and Stop & Glo® Reagent (100 μl) was added into the luminous tube or plate with 20μl cell lysate and the firefly and renilla luciferase activities were measured in a luminescence detector (Turner BioSystems Co, Ltd, USA, Model: Modulus™). The relative activity of the two luciferases was calculated (ΔCT). The experiment was repeated thrice.
SCC-9 cells were grown in a 25 cm 2 culture flask until cell density reached 30-50% in complete medium. Then, the cells were mixed with 100μl serum-free culture medium in sterile EP tubes and incubated with 5μl lipofectamine 2000 (Invitrogen Biotech Co., Ltd., New York, USA) for 5 min at room temperature. Then, 1μg DNA was diluted in culture medium without serum. The diluted lipofectamine 2000 and DNA mixture was mixed well and incubated for 20 min at room temperature to form DNA-liposome complexes. The cells in the culture flask were washed with culture medium without serum. Serumfree DMEM (with no antibiotics) was added to the DNAliposome complexes and mixed with the cells such that plasmid concentration was 50nM. The transfected cells were grown in an incubator with 5% CO 2 at 37°C. After 6~8 h, the liquid was replaced and complete medium was added for continuous culture. The SCC-9 stem-like cells labeled with CD44 + were sorted out by flow cytometry.

Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was extracted from the OSCC and adjacent normal tissues and SCC-9 cells by Trizol method. The optical density (OD) values at 260 nm and 280 nm were recorded by the ND-1000 ultravioletvisible spectrophotometer (Nanodrop Technologies Inc., Wilmington, USA). The quality of total RNA was measured and the RNA concentration was adjusted in samples by diluting with ultra pure water. Then, 1μg RNA was reverse transcribed by two-step method according to instructions in the reverse transcription kit (Fermentas Inc., Hanover, MD, USA). The first strand cDNA synthesis conditions were as follows: 70°C for 10 min, ice bath for 2 min, 42°C for 60 min, 70°C for 10 min. The first strand cDNA samples were stored at -80 ºC for further use.
The qRT-PCR reaction was performed according to the instructions of reverse transcription kit (Fermentas Inc., Hanover, MD, USA) using TaqMan probes. The primer sequences for qRT-PCR are presented in Table 4. The reaction conditions were as follows: pre-denaturation at 95°C for 30 s, denaturation at 95°C for 10 s, annealing at 60°C for 20 s, and extension at 70°C for 10 s (40 cycles). The qRT-PCR was performed in Bio-Rad iQ5 (Bio-Rad, Inc., Hercules, CA, USA). U6 was used as internal reference of miR-542-3p [41], and GAPDH was used as the internal reference of other target genes. 2 -ΔΔCt was used to calculate the relative expression of the targeted genes. Each experiment was performed in triplicates.

Western blotting
Total protein lysates from the OSCC and adjacent normal tissues and SCC-9 cells from the 7 groups were prepared using the RIPA lysis buffer (Beyotime Biotechnology Co., Shanghai, China). Protein concentration was quantified by BCA protein assay kit (Thermo Fisher Scientific, San Jose, California, USA). Equal amounts of protein lysates (20 μl) were boiled with SDS-PAGE buffer for 5 min at 100°C and separated on a 10% SDS-PAGE gel. and transferred into the PVDF membrane at 48 V for 3.5 h. The PVDF membranes were blocked with 5% bovine serum albumin (BSA) for 2 h at room temperature and then washed by 1× Tris-buffered saline with Tween20 (TBST). Then, the membranes were probed with primary antibodies namely, mouse anti-human ILK (1: 2000), TGF-β1 (1: 1000), Smad2/3 (1: 1000) and β-actin (1: 1000) (Cell Signaling Technologies, Beverly, MA, USA) that were prepared with 5% BSA for 1 h at room temperature and rinsed in 1× TBST. Then, the blots were incubated with secondary goat anti-rabbit antibody (1: 4000; Cell Signaling Technologies, Beverly, MA, USA) for an hour at room temperature, rinsed with 1 × TBST and then developed with ECL reagent. The protein bands were quantified relative to β-actin to determine the relative expression of the target proteins.

MTT assay
Transfected SCC9 cells were harvested by digesting with 0.25% trypsin when they reached 80% confluence. Then, the cells were seeded into a 96-well plate at 3 × 10 3~6 × 10 3 cells per well in 200μl medium (6 wells per condition). Then, 20μl of 5mg/ml MTT solution (Sigma, USA) was added to each well and the cells were further cultured for 4 h. Then, after aspirating out the medium, 150μl Dimethyl sulfoxide (DMSO) (Sigma Company, USA) was added and the plate was gently shaken for 10 min. The OD value of the lysates was measured at 490 nm in the enzyme linked immunosorbent assay instrument at 24, 48 and 72 h, respectively. Cell viability was determined from the standard curve that was concurrently generated.

Migration assay of SCC-9 cells
The bottom of the 6-well plate was marked with lines every 0.5~1 cm followed by seeding 3 × 10 4 transfected SCC-9 cells for culture overnight. When the cells reached 80~90% confluence, the cells were scratched perpendicular to the line. The cells were further cultured for 48 h. Eight fields with scratches were randomly selected and analyzed for cell movement around the scratch in each well and photographed. Then the Motic Images Advanced 3.2 software was used to measure the relative width of scratch to determine the migration ability of cells. Each experiment was repeated thrice.

Transwell invasion assay
The matrigel (Corning Glass Works, Corning, N.Y., USA) was dissolved overnight at 4°C and then diluted with serum-free DMEM medium (1: 3). Then, 30μl diluted matrigel was added to the upper chamber of each Transwell chamber (Corning Incorporated, US) and the bottom chamber. The cell suspension (3 × 10 4 ) was added into Transwell upper chamber, and 0.5 ml DMEM with 10% FBS was added to the lower chamber. After 48 h of culturing, the numbers of cells in the bottom chamber were photographed under the inverted microscope. The mean number of cells was determined by selecting 5 fields per well for each group. The total number of cells in each group passing through the matrigel represented the ability of invasion for the particular group. All experiments were repeated thrice.

Flow cytometry
After transfection for 48 h, the SCC9 cells were harvested and fixed in cold 75% ethanol (chilled at -20°C) at 4°C overnight. After centrifugation, the cells were washed by cold PBS twice to remove excess ethanol. RNAase was added to the cell suspension and incubated in a water bath without light for 30 min followed by 5μg/ ml propidiom iodide (PI) for staining and mixed well. The cells were then subjected to flow cytometry and the proportion of cells in G0/G1, S and G2/M phase of the cell cycle were analyzed from the FACS plots. Each experiment was performed in triplicate.
Annexin V/PI double staining was used to detect cellular apoptosis. The SCC-9 cells after 48 h of transfection were harvested, adjusted to 1 × 10 6 /ml in 0.5 ml binding buffer and incubated for 15 mins at room temperature in the dark with 1.25μl AnnexinV-FITC (Nanjing Keygen Biotech. Co., Ltd., Jiangsu, China). Then the cells were pelleted by centrifugation at 1000 rpm for 5 min. The cells were then re-suspended with 0.5 ml chilled binding buffer, followed by addition of 10μl PI and then analyzed by flow cytometry (BD Biosciences Company, NJ, USA).

Tumorsphere formation assay
After trypsinization, the SCC-9 cells were washed twice with PBS and once with serum-free DMEM medium. Then, 1000 cells were seeded into a culture flask and grown for a week. The cells were photographed every 3 days under a microscope. Finally, the number of tumorspheres and the number of cells in each tumor sphere were counted and compared.

Tumor xenografts in nude mice
Fifteen specific pathogen free (SPF) nude mice (Hunan Silaike Jingda Experimental Animal Co., Ltd., Changsha, China) that weighed between 18~25g and 5~6 weeks of age were maintained in a SPF sterile laminar flow chamber with constant temperature (22°C-25°C) and humidity (55 ± 5%), sterile food, water, and bedding. The mice were divided into the control, miR-542-3p mimic and siRNA-ILK groups (5 mice in each group). Then, they were injected with 0.2 ml of corresponding cell suspensions (10 6 / ml) at the back with a hypodermic needle. The maximum diameter (a) and the maximum transverse diameter (b) of tumors were measured every three days. The volume (V = ab 2 /2) of tumors was calculated, and other parameters like the diet, shape and defecation of the nude mice was documented. After 28 days, all mice were sacrificed by cervical dislocation and the tumor was surgically removed, weighed and the final tumor volume was calculated.

Statistical analysis
Data were analyzed using the SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). The experimental data was represented as mean ± SD. Comparison between multiple groups was analyzed using variance analysis. The two pair comparison was tested with least significant difference (LSD)-t test. Enumeration data were displayed as frequency. Categorical data were compared by chi square test and ranked data were displayed by rank sum test. Correlation was analyzed by Spearman correlation analysis; survival time was measured by Kaplan-Meier curve; comparison among groups was analyzed by Log-rank test; multivariate analysis was performed using Cox regression model. P < 0.05 indicated statistical significance.

Author contributions
B.Q. designed the study. J.H.C. and B.X.H. collated the data, designed and developed the database, carried out data analyses and produced the initial draft of the manuscript. A.K.Y.L. contributed to drafting the manuscript. All authors have read and approved the final submitted manuscript.