Tissues of tongue squamous cell carcinoma and the matched normal counterparts were obtained from surgical specimens collected immediately after resection from patients undergoing primary surgical treatment of oral tongue carcinoma at the Department of Stomatology, The First Affiliated Hospital of the PLA General Hospital, and Department of Stomatology, Wuhan General Hospital of Guangzhou Command. The samples were flash frozen in liquid nitrogen and stored at −80°C. Histology of the tissues was evaluated by the hospital pathologist. Written consent of tissue donation for research purposes was obtained from patients before tissue collection, and the protocol was approved by the Ethics Committees of The First Affiliated Hospital of the PLA General Hospital, and Wuhan General Hospital of Guangzhou Command.

Human tongue squamous cell carcinoma cell lines HSC-3, SCC-4 and Cal27, and human normal oral keratinocytes (hNOKs) were routinely cultured in RPMI-1640 medium (Hyclone Laboratories, Logan, UT, USA) supplemented with 10% fetal bovine serum (Hyclone) at 37°C in a 5% CO₂ incubator. For transient transfection, the cells were transfected with miR-26b mimics/inhibitor and their respective negative control duplexes (Ambion, Austin, TX, USA) using Lipofectamine 2000 (Invitrogen Life Technologies, Carlsbad, CA, USA). After a 24-h transfection, the cells were collected for quantitative real-time RT-PCR (qRT-PCR) analysis or further processing.

The TaqMan stem-loop RT-PCR method was used to assess the expression of mature miR-26b with kits from Applied Biosystems (Foster City, CA, USA). The real-time PCR results, recorded as threshold cycle numbers (Ct), were normalized against an internal control (U6). For relative expression levels, the 2^−ΔCt method was used as previously described (18). Experiments were carried out in triplicate for each data point, and analysis was carried out by using Bio-Rad IQ software.

Cells transfected with miR-26b mimics/inhibitor were harvested 24 h later by trypsinization, washed with ice-cold PBS, fixed in 70% ethanol and stored at 4°C. Following overnight incubation, cells were washed and resuspended in propidium iodide (PI) staining buffer. DNA content was evaluated by flow cytometry (XL-MCL; Coulter Epics, Miami, FL, USA).

Detection of apoptotic cells by flow cytometry was performed as described previously (19). Cells transfected with miR-26b mimics/inhibitor were harvested 24 h later and then Annexin V/PI analysis by flow cytometry (XL-MCL; Coulter Epics) was performed.

Transwell chambers (8-μm pore size, Corning, Inc., Corning, NY, USA) were used in the migration and invasion analysis. For migration assays, 1×10⁵ cells were plated in the top chamber lined with a non-coated membrane. For invasion assays, the chamber inserts were coated with 200 μg/ml of Matrigel and dried overnight under sterile conditions. Then, 1×10⁵ cells were plated in the top chamber. In both assays, cells were suspended in medium without serum or growth factors, and medium supplemented with serum was used as a chemoattractant in the lower chamber. After incubation at 37°C for 48 h, the top chambers were wiped with cotton wool to remove the non-migrating or non-invasive cells. The invasive cells on the underside of the membrane were fixed in 100% methanol for 10 min, air-dried, stained with 0.1% crystal violet and counted under a microscope. The mean of triplicate assays for each experimental condition was used.

Total cell lysate was prepared in 1× SDS buffer. Equal amounts of protein were analyzed by western blotting, using antibodies against COX-2, VEGF-C, cyclin D1 and β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

The 3′UTR segments of PTGS2 mRNA containing the miR-26b binding sites were amplified by PCR from human genomic DNA and inserted into the pMIR-Report luciferase reporter vector (Ambion) and named pMIR-PTGS2-3′UTR. A mutant version from the site of perfect complementarity was also generated and named pMIR-mut-PTGS2-3′UTR. The recombinant reporter vectors with wild-type or mutant PTGS2 3′UTR were then cotransfected with miR-26b mimics or control into SCC-4 cells, respectively, using Lipofectamine™ 2000. The luciferase assay was performed according to the manufacturer’s instructions.

The data are expressed as means ± SEM. Differences were compared by one-way ANOVA analysis followed by the LSD t-test. Survival curves were plotted by the Kaplan-Meier’s method, and the log-rank test was carried out to compare differences in survival. All statistical analyses were performed using SPSS 19.0 software (SPSS Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

To investigate the role of miR-26b in human tongue squamous cell carcinoma, we first compared the expression levels of miR-26b between carcinoma tissue samples and paired adjacent non-neoplastic mucosal tissues from 30 cases of tongue squamous cell carcinoma patients. Consistent with the microarray results presented by Wong et al (9), the qRT-PCR results verified that the miR-26b expression level in tongue squamous cell carcinoma tissues (−13.21±0.46) was significantly lower than that in the non-neoplastic mucosal tissues (−11.06±0.50) (P<0.05, t=8.355, paired t-test) (Fig. 1A). Correlations between the miR-26b expression level and clinicopathologic characteristics of the tongue squamous cell carcinoma cases are summarized in Table I. Statistically significant associations between miR-26b expression and clinical stage and between miR-26b expression and metastasis were observed in the present study. The median expression of miR-26b was −14.18±0.58 in the 15 cases with advanced stage (stage III and IV) disease, whereas the median expression was −12.24±0.63 (P<0.05, Mann-Whitney U test) in the other 15 cases with early-stage (stages I and II) disease. In the 14 cases with lymph node metastasis, the median expression of miR-26b was −14.28±0.62, which was significantly lower than the median expression (−12.24±0.63) in the 16 non-metastatic cases (P<0.05). The expression of miR-26b in the tongue squamous cell carcinoma patients did not correlate with age, gender, tumor size, or cell differentiation. Moreover, we examined whether the level of miR-26b expression was associated with survival in the patients with tongue squamous cell carcinoma. Patients were subsequently divided into low expression (n=15) and high expression groups (n=15) based on miR-26b levels greater or less than the mean (−13.21) (Fig. 1B). Kaplan-Meier’s survival analysis revealed that patients whose primary tumors displayed low expression of miR-26b had a shorter median survival time. The 5-year survival rate of patients with low miR-26 expression was 26.7%, which was significantly lower than the survival rate in patients with high miR-26b expression (53.3%, P<0.05, log-rank test, Fig. 1C).

We further detected the expression of miR-26b in hNOKs and 3 human tongue squamous cell carcinoma cell lines, HSC-3, SCC-4 and Cal27. As shown in Fig. 2A, all of the 3 carcinoma cell lines had a relatively low miR-26b expression compared to that of the hNOKs. Among the 3 carcinoma cell lines, SCC-4 cells had a moderate expression of miR-26b. Therefore, the SCC-4 cells were chosen to evaluate the effects of miR-26b expression on the oncogenicity of human tongue squamous cell carcinoma cells. Fig. 2B shows that miR-26b expression was strongly reduced or increased in the SCC-4 cells after transfection of the miR-26b-specific inhibitor or mimics. By flow cytometry, a significant accumulation of cells in the G1 phase was observed in the miR-26b-upregulated SCC-4 cells, while inhibition of miR-26b expression led to increased cell proliferation (Fig. 2C, P<0.05). No significant influence on apoptosis was observed in the SCC-4 cells after treatment with the miR-26b specific inhibitor or mimics (Fig. 2D). We further performed Transwell assays to determine whether altered miR-26b expression could influence the metastatic phenotype of the SCC-4 cells. As shown in Fig. 2E and F, miR-26b inhibition significantly increased the ability of the SCC-4 cells to migrate and invade, while miR-26b upregulation suppressed the migration and invasion (P<0.05).

Using bioinformatics analysis, such as miRanda and TargetScan, we found that miR-26b contained specific binding sequences for the 3′UTR region of the COX-2 gene (PTGS2) (Fig. 3A). In agreement with the computer prediction, COX-2 protein levels were significantly suppressed in the miR-26b mimic group whereas these levels were increased in the miR-26b inhibitor group (Fig. 3B). In order to further validate the relationship between miR-26b and COX-2, a dual luciferase reporter assay was performed in the SCC-4 cells. The sequences of the 3′UTR of wild-type and mutated COX-2 are shown in Fig. 3C. No reduction in luciferase activity was observed in the SCC-4 cells transfected with miR-26b mimics and mutated COX-2, but an ~45% reduction in luciferase activity was observed in wild-type COX-2 (P<0.05, Fig. 3D).

We further investigated the influence of nimesulide, a COX-2-specific inhibitor, on the proliferation, migration and invasion of SCC-4 cells (20). As shown in Fig. 4A–C, nimesulide abolished the promoting effects of miR-26b inhibition on the proliferation, migration and invasion in SCC-4 cells, suggested the critical role of COX-2 in miR-26b-regulated oncogenesis (P<0.05). Previous studies revealed that COX-2 could promote the proliferation and metastasis of human tongue squamous cell carcinoma cells through the VEGF-C pathway (21–25). In the present study, we investigated the expression changes of VEGF-C in SCC-4 cells by western blot analysis. As shown in Fig. 4D, the protein level of VEGF-C was significantly increased in the miR-26b-silenced SCC-4 cells, however, this was reversed by treatment of nimesulide. We also investigated the expression changes in cyclin D1, a key cell cycle-regulatory protein. Western blot analysis results indicated that cyclin D1 was also significantly upregulated in the miR-26b-silenced SCC-4 cells and this upregulation was reversed by nimesulide.