All of the patients included in the study underwent surgical resection procedures at the Thoracic Surgery Department of the Affiliated Hospital of Qingdao University, Qingdao, China. One hundred and sixty-four eligible patients diagnosed with a primary tumor in stages I, II or III were enrolled into the present study between February 2009 and May 2013. None of the patients had received any treatment for lung cancer prior to surgery, and all patients had a preoperative computed tomography scan. Patients with diffuse lobar or multifocal bronchioalveolar carcinoma and patients who had malignancy within the previous 5 years were ineligible. The NSCLC stage was based on the American Joint Committee on Cancer (AJCC) TNM System. Tumors were graded as well, moderately or poorly differentiated. After the surgical specimens were obtained, individual-matched normal lung tissues adjacent to the proximal excision margin were harvested. A total of 164 pairs of specimens were used to generate paraffin-embedded tissue microarrays. The remaining specimens were immediately snap-frozen in liquid nitrogen and stored at −80°C until further analysis.

The A549 human lung cancer cell line, NCL-H460 cells and BEAS-2E normal lung epithelial cells were obtained from the Cell Resource Center of Shanghai Institutes for Biological Sciences. The cell lines were routinely cultured in DMEM-1640 medium (Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Gibco), 100 U/ml penicillin-G, and 100 μg/ml streptomycin, and were cultivated in a humidified incubator with 5% CO₂ at 37°C.

The miR-124 mimics (miR-124) and negative control mimics (miR-NC) were all purchased from Beyotime Institute of Biotechnology (Co., Ltd., Shanghai, China). The cells were seeded in 12-well plates and grown to 40% confluence before transfection. The oligonucleotides were transfected at a final concentration of 50 nM using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer’s instructions. The cells were divided into three groups according to the treatment and were named as follows: i) miR-124 (transfected with miR-124-mimics); ii) miR-NC (transfected with negative control mimics); and iii) untransfected control. RNA or protein was isolated 48 h after transfection and used for qRT-PCR or western blotting, respectively. The transfections were repeated in triplicate a minimum of 3 times.

Cell proliferation was determined using methyl thiazolyl tetrazolium (MTT) assays. Briefly, A549 cells (5×10³ cells/well) were plated in 96-well plates with DMEM containing 10% FBS. After 24 h of culturing, the A549 cells were transfected with miR-124 or miR-NC mimics as described above. The MTT assay was completed 0, 12, 24, 48 and 72 h after gene transfection. The absorbance in each well was measured with a microplate reader set at 450 and 630 nM. All experiments were performed in triplicate and repeated three times. For the apoptosis analysis, cultured cells were harvested by trypsinization and washed twice with ice-cold PBS. Forty-eight hours after transfection, the cells (5×10⁵ cells/ml) from each sample were stained with Annexin V-FITC and PI using the protocol included in the Apoptosis Detection kit I (BD Biosciences, San Jose, CA, USA). A fluorescence-activated cell sorting FACSCalibur instrument (BD Biosciences) was used for this assay, and the data generated were subsequently analyzed with CellQuest software (BD Biosciences). The experiments were performed in triplicate and repeated three times.

The 3′ UTR of STAT3 containing wild-type or mutant miR-124 target sites was amplified by PCR and cloned into the XhoI/NotI site of the pmiR-RB-REPORT vector. The XhoI/NotI site was downstream of the Renilla reporter gene. The mutant constructs were created by mutating the seed regions of the miR-124 sites (5′-GUGCCU-3′). The primers for the wild-type STAT3 3′ UTR were 5′-CCCTCGAGGGAGCTGAGAACGGAAGCTGCA-3′ (forward) and 5′-ATTTGCGGCCGCAGCTGTTCTGCCTCACCTGTGGG-3′ (reverse), while the mutagenesis primers were 5′-CCCTGATATCACATCCACAGAAACAACCT-3′ (reverse) and 5′-AGGTTGTTTCTGTGGATGTGATATCAGGG-3′ (forward). All of the constructs were confirmed using restriction digests and DNA sequencing. For the luciferase reporter assay, the A549 cell line was co-transfected with luciferase constructs containing the 3′ UTR of STAT3 (with wild-type or mutant miR-124 binding sites) and miR-124 or miR-NC mimics. All of the transfections were conducted in triplicate with Lipofectamine 2000. Luciferase activity was measured 48 h after transfection using a dual luciferase reporter system (Promega, Madison, WI, USA) according to the manufacturer’s protocol and normalized to the corresponding Renilla luciferase activity.

Total RNA was extracted from fresh tissues or cells using TRIzol reagent (Gibco). The concentration and purity of the RNA was determined using a NanoDrop^® ND-1000 spectrophotometer (Thermo Fisher Scientific, Inc., Wilmington, DE, USA). miR-124 expression was determined using a TaqMan microRNA assay kit (Applied Biosystems, Foster City, CA, USA). To determine the expression of STAT3 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 1 μg of total RNA was subjected to first-strand cDNA synthesis (Takara, Shiga, Japan) and examined using a PrimeScript RT reagent kit (Takara). Real-time PCR was performed using SYBR-Green PCR Master Mix (Takara) on the ABI 7500HT System. All of the primers used in the present study were designed by Sangon Biotech Co., Ltd. (Shanghai, China). The expression of U6 and GAPDH was used for normalization. The relative expression of mRNA was determined using the 2^−ΔΔCt method (21). All of the real-time PCR reactions were carried out in triplicate. A list of primers used in the reactions is presented in Table I.

Tissues were lysed using RIPA buffer (Sigma, St. Louis, MO, USA). The soluble protein concentration was determined by the Bradford method. Total protein (10 μg) was run on a 10–15% SDS-PAGE gel and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were blocked with 1% Tween 20-PBS at 4°C overnight before being washed and incubated with primary antibodies against STAT3 and GAPDH (Bioworld Co., Dublin, OH, USA; diluted 1:500 in TBS-A) for 1 h at room temperature. After washing, the membranes were incubated with the secondary antibodies (Abcam Co., Cambridge, UK; diluted 1:1,000 in TBS-A) for 1 h at room temperature. The immunoblots were developed using an electrochemiluminescence kit and imaged with the Vilber Fusion FX5 automatic gel imaging analysis system (Vilber, Marne La Vallée, France).

Immunohistochemistry was performed using standard techniques. Briefly, 4-μm thick paraffin-embedded specimens were dewaxed in xylene and rehydrated in graded alcohols. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide. Antigen retrieval was accomplished in citrate buffer (pH 6.0) using a microwave. A polyclonal rabbit anti-human STAT3 antibody (1:200 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA) was added and the samples were incubated overnight at 4°C. After washing, the sections were incubated with a secondary antibody for 40 min at 37°C and subsequently incubated with streptavidin-conjugated horseradish peroxidase (HRP) and diaminobenzidine (DAB) and counterstained with hematoxylin. The primary antibody was replaced by an isotope IgG in the negative control. Representative viable tissue sections were scored semi-quantitatively using light microscopy. The proportion of positively stained cells was graded as follows: score 0, completely negative samples; score 1, samples with up to 10% positive cells; score 2, samples with 11–50% positive cells; and score 3, samples with 50% positive cells. The staining intensity in tumor cells was scored as 0, no staining, 1, weak staining, 2, moderate staining or 3, strong staining. The staining index was calculated as the product of the staining intensity score and the proportion score, which had a value of 0, 1, 2, 3, 4, 6 or 9. The final score was achieved by comparing the scores obtained by different observers, and any discrepancies were resolved by consensus. In the case of a disagreement, the slides were reexamined and the observers reached a consensus. A staining index ≥4 was defined as high expression, and an index <4 was defined as low expression.

Follow-up was carried out by telephone or written correspondence, while clinical examinations and chest radiography were performed every 3 months for 3 years and every 6 months thereafter. Disease-free survival (DFS) was evaluated for the period from the date of operation to the appearance of tumor progression. Overall survival (OS) was evaluated for the period from the date of surgery to the date of death (or last follow-up). The study concluded on December 31, 2013. The overall follow-up rate was 100%, and the median follow-up time was 31 months with a range from 2 months to 58 months.

All of the specimens were obtained from the Central Laboratory of the Affiliated Hospital of Qingdao University, and all of the patients provided written informed consent for the use of their tissues. The present study was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University.

All of the analyses were performed using SPSS 19.0 software (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism 5.0 (GraphPad Software, San Diego, CA, USA). All of the numerical data were expressed as the mean ± SD. P-values <0.05 were defined as statistically significant. The results were analyzed using the Student’s t-test or one-way analysis of variance (ANOVA). The Chi-square test was used to analyze the relationship between miR-124 or STAT3 expression and clinicopathological symptoms. Bivariate correlations between the study variables were calculated using Spearman’s rank-order correlation coefficients. The survival curves were analyzed using Kaplan-Meier methodology, and the difference in distribution was evaluated with the log-rank test. The influence of each variable on survival was examined by univariate and multivariate Cox regression analysis.

A total of 164 patients were enrolled in the present study (Table II). The patients consisted of 115 men and 49 women with a median age of 59 years (range, 22–76 years). The NSCLC tumors consisted of 106 (64.6%) adenocarcinomas (AC), 52 (31.7%) squamous cell carcinomas (SCC) and 6 (3.7%) carcinomas of other cell types. The post-surgical pathologic stage was determined by the TNM classification. The number of patients with disease stages I, II and III was 47 (28.6%), 29 (17.7%) and 88 (53.7%), respectively.

We performed real-time PCR in the A549 and NCL-H460 NSCLC cell lines and the BEAS-2E normal lung epithelial line to confirm the expression level of miR-124. The results showed that miR-124 expression in both A549 and NCL-H460 cells was significantly lower than in normal lung epithelial BEAS-2E cells (P=0.0015 and P<0.001, respectively) (Fig. 1A), confirming the downregulation of miR-124 in NSCLC cells.

To explore the expression and significance of miR-124 in lung cancer, we transfected miR-124 mimics or negative control mimics into both A549 and NCL-H460 cells. The expression level of miR-124 was detected using qRT-PCR 48 h after transfection. The cells transfected with miR-124 mimics showed a significant, 300-fold increase in miR-124 expression compared with miR-NC cells or untransfected cells (P<0.001) (Fig. 1B). We further investigated the effect of miR-124 on cell survival. The cell lines transfected with miR-124 mimics were significantly less viable 24, 48 and 72 h post-transfection compared with miR-NC cells or untransfected cells (Fig. 1C). The cells overexpressing miR-124 began to shrink in size and display vague membranes and an irregular shape 48 h after transfection (Fig. 1D and E). Few early or late apoptotic cells were detected in untransfected and miR-NC cells (0.2608±0.019 vs. 1.358±0.301%; P=0.059), whereas the cells transfected with miR-124 mimics exhibited a significantly higher percentage of early and late apoptotic cells (20.88±2.84; P<0.001) (Fig. 1F and G).

STAT3 has been reported as a potential target of miR-124 during T cell-mediated immune clearance of glioma (16). However, the relationship between miR-124 and STAT3 has not been validated in NSCLC. We used TargetScan target prediction software (http://www.targetscan.org/) to analyze the potential interaction between miR-124 and STAT3, which is a potential target of miR-124 and contains putative miR-124 target sites in the 3′ UTR (Fig. 2A). To confirm that STAT3 is a direct target of miR-124, we cloned the 3′ UTR of STAT3 into the pGL3-control vector and performed reporter assays. The result showed that the relative luciferase activity of the wild-type STAT3 3′ UTR reporter construct was notably decreased in A549-miR-124 cells compared to A549-miR-NC (P=0.0024) or untransfected cells (P=0.0032). However, the relative luciferase activity of the mutant STAT3 3′ UTR reporter construct was not significantly different from the luciferase activity of the A549-miR-NC cells and failed to respond to miR-124 (Fig. 2B). To further prove that STAT3 is a target of miR-124, we evaluated the expression levels of STAT3 in the three cell lines 48 h after transfection using real-time PCR and western blotting. The results from these assays show that STAT3 was significantly downregulated in A549-miR-124 cells compared to A549-miR-NC or untransfected cells (P<0.001) (Fig. 2C and D).

miR-124 and STAT3 levels were detected in all 164 pairs of NSCLC tissues and adjacent normal lung tissue by real-time PCR. U6 snRNA, which are not deregulated in lung cancer, was used for normalization (22). These results indicated that the expression of miR-124 was significantly lower in NSCLC tissues (0.351±0.129) than in adjacent normal lung tissues (0.661±0.304; P=0.0002) (Fig. 3A). The IHC analysis showed that STAT3 localized to the cytoplasm of cells (Fig. 3B). The relative expression of STAT3 was significantly higher in NSCLC tissues than adjacent normal lung tissues (2.858±1.264 vs. 1.476±0.672; P<0.001) (Fig. 3C). In addition, Spearman’s rank-order correlation analysis revealed that STAT3 expression in NSCLC tissues was inversely correlated with the expression of miR-124 (r_s=−0.2642; P=0.0006) (Fig. 3D).

The relationship between miR-124 or STAT3 expression and clinicopathological NSCLC characteristics is shown in Table II. The miR-124 expression levels in NSCLC tissues were closely associated with the differentiation grade (P=0.001), lymph node metastasis (P=0.017) and TNM stage (P=0.006). However, a statistical analysis revealed no significant correlations between miR-124 expression and gender, age, SI, location, pleura invasion or pathological type. As shown in Table II, STAT3 expression correlated significantly with tumor differentiation, lymph node metastasis and TNM stage (P<0.001, P=0.007 and 0.008, respectively). In contrast, STAT3 expression did not correlate with gender, age, SI, location, pleural invasion or pathological type.

Thirty of the patients in the study died of NSCLC cancer during the follow-up period. The 1-year overall survival and DFS rates in the miR-124 upregulated group were 87.5 and 81.2%, respectively, but these rates were only 68.1 and 57.8% in the miR-124 downregulated group. A Kaplan-Meier survival analysis was performed to further analyze the association between miR-124 and STAT3 expression and patient prognosis. The results of this analysis suggested that patients with low miR-124 or high STAT3 expression had poor OS and DFS (Fig. 4). Previously it was suggested that patients with positive lymph nodes are more prone to recurrence and metastasis compared with negative lymph node patients (23). Therefore, an additional analysis of the relationship between miR-124 expression and patient survival was performed in positive lymph node and negative lymph node patients. The results suggested that miR-124 expression was significantly associated with DFS in both positive and negative lymph node groups (P=0.035 vs. 0.025, respectively) (Fig. 5A and B). Furthermore, a Cox multivariate analysis showed that miR-124 expression (positive lymph node group, P=0.004; negative lymph node group, P=0.009), lymph node metastasis (positive lymph node group, P=0.037; negative lymph node group, P=0.048) and STAT3 expression (positive lymph node group, P=0.007; negative lymph node group, P=0.004) were significantly correlated with OS and DFS (Tables III and IV).