The present study was approved by the Ethics Committee of Subei Hospital (Jiangsu, China), and written informed consent was obtained from all patients.

Ten pairs of formalin-fixed, paraffin-embedded (FFPE) NSCLC tissues (4 patients with squamous cell carcinoma and 6 adenocarcinoma) and their corresponding non-tumorous lung tissues were collected from May to October, 2012 at Subei Hospital. Another 48 FFPE NSCLC tissues were obtained from February, 2008 to December, 2009. On enrollment, tissue samples used in this study were collected prior to treatment with chemoradiotherapy. Each sample was confirmed by histopathologic evaluation through hematoxylin and eosin (H&E) staining. Clinical data were extracted at the time of resection, and patients who entered into the registry were prospectively followed up for the ascertainment of vital status through April 30, 2012.

Human lung cancer A549 and H1299 cells and normal bronchial epithelial BEAS-2B cells were purchased from the Cell Resource Center, Shanghai Institute of Biochemistry. BEAS-2B cells were derived by transforming human bronchial epithelial cells with an adenovirus 12-simian virus 40 construct (23). The cell lines were maintained at 37°C in a humidified air atmosphere containing 5% carbon dioxide in RPMI-1640 (Wisent, Quebec, Canada) supplemented with 10% fetal bovine serum (Wisent).

For cultured cells, total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA). Total tissue RNA was extracted from FFPE tissue sections using the miRNeasy FFPE kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. Briefly, all paraffin was removed from freshly cut FFPE tissue sections by treating with deparaffinization solution. Samples were under protease digestion to release RNA from the sections and short incubation to reverse formalin cross-linking of the released nucleic acids. DNA was removed from the RNA sample by a final DNase digestion step. Total RNA including miRNA was eventually dissolved in 20 μl of RNase-free water. The concentration of RNA was measured using a NanoDrop-1000 (Thermo Fisher Scientific, Waltham, MA, USA), and RNA integrity was determined by 1.5% agarose gel electrophoresis.

cDNA synthesis for miRNA-148a validation was performed according to the manufacturer’s recommendation. Briefly, RNA was reverse-transcribed (RT) using a ReverAid First Strand cDNA kit (Thermo Fisher Scientific) in combination with a stem-loop primer for miRNA-148a. Under our experimental conditions, we used U6 snRNA as an internal control for normalizing the expression levels of miRNA-148a. The RT primers were designed as follows: miRNA-148a, 5′-GTCGTATCCAG TGCAGGGTCCGAGGTATTCGCACTGGATACGACACA AAG-3′ and U6, 5′-TGGTGTCGTGGAGTCG-3′. In addition, 1.0 μg total RNA was combined with 4.0 μl reaction buffer (5X), 1.0 μl RiboLock RNase inhibitor, 2.0 μl dNTP mix (10 mM each), 1.0 μl ReverAid M-MuLV reverse transcriptase in a total reaction volume of 20 μl. Reactions were carried out using the ABI Prism 7900HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) using the following conditions: 25°C for 5 min, 42°C for 60 min, 70°C for 5 min, and then hold at 4°C. After the RT reaction, miRNA expression was detected using LightCycler^® 480 SYBR-Green I Master on the LightCycler^® 480 Real-Time PCR system (both from Roche Applied Science, Indianapolis, IN, USA). cDNA products were diluted at 20X, and the 10-μl PCR mixture contained 1 μl diluted RT product, 5 μl SYBR-Green Master Mix, 2 μl RNase-free water, 1 μl forward, and 1 μl reverse primers. The PCR primers for miRNA-148a and U6 were designed as follows: miRNA-148a forward, 5′-TCAGTGCACTACAGAACTT TGT-3′ and reverse, 5′-GCTGTCAACGATACGCTACGT-3′; U6 forward, 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′. The reaction was incubated at 95°C for 10 min followed by 40 cycles of 95°C for 15 sec and 65°C for 60 sec. The relative miRNA-148a expression was calculated using the 2^−ΔCt method, where ΔCt is the difference in threshold cycles (Ct) for the target and reference = Ct(miRNA-148a) − Ct(U6). The RT and PCR primers were synthesized by Shenggong Biotech Co., Ltd. (Shanghai, China).

Total tissue DNA was extracted from FFPE tissue scrolls using a Qiagen EpiTect Plus FFPE Bisulfite kit (Qiagen) according to the manufacturer’s instructions. Briefly, FFPE tissue scrolls were deparaffinized, followed by proteinase digestion and decrosslinking. The DNA bisulfite reaction was then set up and performed using the ABI Prism 7900HT Fast Real-Time PCR system. Upon completion of the bisulfite conversion, modified DNA was purified by several steps of clean-up and eventually eluted. The modified DNA concentration was measured using a NanoDrop-1000 and stored at −20°C for further analysis.

As previously described, the methylation rate of miRNA-148a was determined by qMS-PCR (24). We subjected 10 ng of the modified DNA to PCR amplification with ABI Prism 7900HT Fast Real-Time PCR system and primers: miRNA-148a forward (5-TGGGTA TTTGTTTTTGTTGATTG-3) and reverse (5-ACTACACTTA AACCCCCTCTAACC-3). Reactions were carried out using the following conditions: 90°C for 5 min, 40 cycles of 95°C for 15 sec, 60°C for 30 sec and 72°C for 15 sec; final extension of 10 min at 72°C. The PCR products were diluted into water at 500X, and 1 μl diluted products was then subjected to qMS-PCR with LightCycler^® 480 SYBR-Green I Master. The RT primers were designed as follows: unmethylated-specific miRNA-148a forward, 5-TATGATTTGTTTTATTATTGG TT-3 and reverse, 5-AACACTAACAACATCAACAACC-3; methylated-specific miRNA-148a forward, 5-TGATTCGTTT TATTATCGGTC-3 and reverse, 5-AACACTAACGACATC GACG-3. The reaction was conducted at 95°C for 10 min followed by 40 cycles of 95°C for 15 sec and 60°C for 60 sec in the LightCycler^® 480 Real-Time PCR system. The methylation level was calculated with the formula 2^−ΔCt, where ΔCt is the difference in the threshold cycle values for methylated and unmethylated = Ct(methylated) − Ct(unmethylated). The methylation level was expressed as a percentage.

For treatment with DNA methylation inhibitor 5-aza-2′-deoxycytidine (5-aza-dC; Sigma, St. Louis, MO, USA), A549 and H1299 cells were seeded into 24-well plates on day 0 and exposed to 5-aza-dC at a final concentration of 5 μmol/l from day 1 to 3 and then harvested for qRT-PCR assay of miRNA-148a expression on day 4.

miRNA-148a mimics and negative control oligonucleotides were synthesized by GenePharma Co., Ltd. (Shanghai, China). Cell transfection was performed using Lipofectamine 2000 reagents (Invitrogen Life Technologies) according to the manufacturer’s protocol. Twenty-four hours after transfection, cells were collected for qRT-PCR analysis and functional assay.

miRNA-148a mimic-transfected cells were grown to confluence in a 24-well plate and wounded by dragging a 200-μl pipette tip through the monolayer. Cells were then washed to remove cellular debris and allowed to migrate for 24 h. Images were captured at 0 and 24 h after wounding under an Eclipse Ti-U (Nikon, Kanagawa, Japan) inverted microscope. The relative surface traveled by the leading edge was assessed using Image-Pro Plus version 6.0 software. Three independent experiments were performed.

For the migration assays, 2.5×10⁴ cells were added into the upper chamber of the insert (8-μm pore size; Corning Incorporated, Corning, NY, USA). For the invasion assays, cells were added into the upper chamber of the insert pre-coated with Matrigel (BD Biosciences, Bedford, MA, USA). In both assays, cells were plated in medium without serum, and medium containing 10% fetal bovine serum in the lower chamber served as a chemoattractant. After 40 h of incubation, the cells that did not migrate or invade through the pores were carefully wiped off with cotton wool. The inserts were then stained with 20% methanol and 0.2% crystal violet, imaged with an Eclipse Ti-U inverted microscope, and counted by Image-Pro Plus version 6.0.

Seventy-two hours after transfection, cells were collected and homogenized in lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.02% sodium azide, 100 mg/l PMSF, 1 mg/l aprotinin) and centrifuged at 11,000 × g for 15 min. The supernatant was collected and the protein content was determined by BCA protein reagent (Pierce Chemical Co., Rockford, IL, USA). The proteins (50 μg/lane) were separated on 10% SDS-PAGE gels and transferred to PVDF membranes. Membranes were blocked with 5% nonfat milk and incubated with primary antibodies against DNMT1 (Abcam, Cambridge, MA, USA), E-cadherin and β-actin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) overnight at 4°C. After washing, the membranes were subjected to continued incubation with goat anti-rabbit or anti-mouse secondary antibody (Santa Cruz Biotechnology, Inc.) and visualized with enhanced chemiluminescence.

Genomic DNA was obtained from the cell lines and modified with sodium bisulfite as described above. The modified DNAs were specifically amplified by methylated and unmethylated primers of E-cadherin as previously described (25). The PCR products were separated by high-resolution 4% agarose E-gels (Invitrogen Life Technologies) and visualized under UV illumination.

The results were analyzed using the SPSS 13.0 statistical software. Data are shown as means ± standard deviation (SD), and the Student’s t-test was used for statistical analysis. Correlation between miRNA-148a expression and the methylation level of the DNA region encoding miRNA-148a in tissues was evaluated by Pearson’s correlation. Descriptive statistics for the study population were generated by miRNA-148a expression for demographic and clinicopathologic characteristics. Categorical variables were compared using analysis of variance test. Comparing factors for miRNA-148a expression included gender, age, pathology, smoking status, tumor size, histologic grade, T stage, lymph node metastasis and clinical stage. Univariate Kaplan-Meier method was performed to estimate disease-free survival and overall survival according to miRNA-148a expression. Survival differences according to expression were analyzed using the log-rank test. For the analysis of disease-free and overall survival, patients who died prior to recurrence were considered censored at death. Multivariable Cox proportional hazards model analysis of factors potentially related to survival was used to identify the significant influence of miRNA-148a expression on survival, adjusted for gender, age, histologic grade, T stage and smoking status. P<0.05 was considered to indicate a statistically significant result.

The expression levels of miRNA-148a in A549 and H1299 lung cancer cells and in BEAS-2B normal lung bronchial epithelial cells were determined by qRT-PCR. Our results showed that the levels of miRNA-148a were dramatically downregulated in lung cancer A549 and H1299 cells when compared with BEAS-2B cells (Fig. 1A). Next, we analyzed the expression levels of miRNA-148a in 10 pairs of FFPE lung cancer tissue specimens and matched adjacent normal lung tissues. Consistent with the results in the cell lines, the expression levels of miRNA-148a in lung cancer tissues were significantly decreased when compared to that in the matched normal lung tissues (Fig. 1B).

There are many factors that reduce the expression of miRNAs. CpG plot software (http://bioweb.pasteur.fr) reported two CpG islands located close to the miRNA-148a gene, and the hypermethylation of the DNA region encoding miRNA-148a was reported to be responsible for its repression (24). We retrospectively tested miRNA-148a expression in 48 FFPE lung cancer tissues using qRT-PCR. We also analyzed the methylation status of the DNA region encoding miRNA-148a in the same population by means of qMS-PCR assay and found that 38 tumor tissue samples were available for further analysis. Among the analyzed tissues, we found an inverse correlation between the expression level of miRNA-148a and the methylation rates of the miRNA-148a encoding region in lung cancer tissues as evaluated by Pearson’s regression (P=0.044, Fig. 2). The correlation coefficient was −0.329. To further identify whether miRNA-148a was aberrantly inhibited by DNA hypermethylation, we treated A549 and H1299 cells with 5 μmol/l 5-aza-dC for 72 h and then analyzed the expression of miRNA-148a. As shown in Fig. 3, the expression of miRNA-148a was increased in both A549 and H1299 cells, suggesting that DNA methylation was involved in the silencing of miRNA-148a in lung cancer.

To determine whether miRNA-148a expression is associated with clinicopathologic features of lung cancer, we further analyzed the relationship between miRNA-148a expression and clinicopathologic characteristics of lung cancer. We found that patients with lymph node metastasis and advanced clinical stage had a significantly lower expression of miRNA-148a (P=0.043 and 0.018 respectively, Table I). There were no significant differences between expression of miRNA-148a and other clinicopathologic characteristics of the lung cancer cases including gender, age, pathology, smoking status, tumor size, histologic grade and T stage (Table I).

To further evaluate the correlation between miRNA-148a expression and the prognosis of NSCLC patients, Kaplan-Meier survival analysis and multivariable Cox proportional hazards analysis were performed. In this study, one patient died from pneumonia shortly after surgery. Thus, according to the median relative miRNA-148a expression (0.2638) in the remaining 47 tumor tissues, patients were classified into 2 groups: the low miRNA-148a expression group (n=24) with miRNA-148a expression less than or equal to the median; and the high miRNA-148a group (n=23), with miRNA-148a expression greater than the median. In the univariate Kaplan-Meier analysis, patients with low expression of miRNA-148a had a significantly shorter disease-free survival (P=0.030, Fig. 4A) and overall survival (P=0.006, Fig. 4B) when compared with the patients with high expression of miRNA-148a. In the multivariable Cox proportional hazards analysis, adjusting for gender, age, histologic grade, smoking status and T stage, our results showed that patients with low expression of miRNA-148a had a higher risk of tumor-related death (hazard ratio, 5.018; 95% CI, 1.257–20.028, P=0.022, Table II) compared with the patients with high expression of miRNA-148a; expression of miRNA-148a was not significantly associated with disease-free survival (P=0.178).

Since we observed that the downregulation of miRNA-148a in lung cancer was closely associated with lymph node metastasis and poor survival, we postulated that overexpression of miRNA-148a in lung cancer cells may exert inhibitory effects on cell invasion and metastasis. We measured the difference in the migratory capability of lung cancer A549 and H1299 cells using transient transfection. The wound-healing assay showed that miRNA-148a mimic-transfected A549 and H1299 cells had a significantly lower capability of migration than the negative control cells (Fig. 5A). Moreover, overexpression of miRNA-148a significantly suppressed the migratory and invasive abilities of cells as determined by Transwell migration and Matrigel invasion assays (Fig. 5B). These results strongly suggest that miRNA-148a downregulation was associated with high migration capability of lung cancer cells.

It was predicted that both the 3′UTR and the gene coding DNA sequence (CDS) of DNMT1 had putative miRNA-148a binding elements, and miRNA-148a was found to inhibit DNMT1 protein expression in CD4⁺ T cells and gastric cancer cells (19,26). To explore the molecular mechanism of miRNA-148a in lung cancer metastasis, we examined whether the expression of DNMT1 was regulated by miRNA-148a in lung cancer cells. As shown in Fig. 6A, our results showed that the protein levels of DNMT1 were significantly inhibited in both A549 and H1299 cells after transfection with the miRNA-148a mimics.

In lung cancer, a number of methylation-sensitive genes including E-cadherin that are linked to lung cancer metastasis are downregulated, and E-cadherin was found to be demethylated and overexpressed following treatment with the DNMT inhibitor in lung cancer cells (25). As described above, miRNA-148a inhibited DNMT1 protein expression in lung cancer cells. We next examined whether miRNA-148a induces the overexpression of E-cadherin. Results showed that the protein levels of E-cadherin were significantly upregulated in A549 and H1299 cells after transfection with miRNA-148a mimics (Fig. 6A). We further used MSP to determine the methylation status of the E-cadherin promoter post-transfection of miRNA-148a mimics in lung cancer cells. Consistent with a previous study (25), the results showed that the A549 and H1299 cells had a nearly complete methylation of the E-cadherin promoter (Fig. 6B). We also showed that restoration of E-cadherin was significantly associated with its promoter demethylation in the A549 and H1299 cell lines (Fig. 6B). Taken together, our results suggest that the inhibition of migration by miRNA-148a may be associated with DNMT1-mediated epigenetic regulation of E-cadherin.