Lung cell lines HBE, A549, H1299, SPC-α-1, PC-9, LIEP-α-2 and HTB182 were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA) and cultured in RPMI-1640 medium supplemented with 10% fetal calf serum (FCS; Invitrogen, Carlsbad, CA, USA), 100 IU/ml penicillin and 100 IU/ml streptomycin at 37 ºC in a humid atmosphere with 5% CO₂.

Tissue samples from 45 cases of histology-confirmed lung adenocarcinoma were obtained from the Cardiothoracic Surgery Department of Xingya Hospital, Central South University (Hunan, China). The patients had never received radiation therapy or chemotherapy before surgery and pathologically confirmed as lung adenocarcinoma. The present study was approved by the Ethics Committee of Xiangya Hospital and written informed consent was obtained from each patient before the participation.

Small interfering RNAs (siRNAs) targeting expression of ZEB1 and β-catenin as well as miR-33b mimic and miRNA mimic NC were synthesized by Shanghai GenePhama Co., Ltd. (Shanghai, China). Unrelated sequences were used as a negative control and were provided by Shanghai GenePhama (Table I). The miR-33b was generated by lentiviral transduction (Shanghai Genechem Co., Ltd., Shanghai, China). The sequences are listed in Table II. Each recombinant lentivirus was purified by ultracentrifugation and titer of these viral particles was assessed for multiplicity of infection (MOI). Similarly, miR-33b mimic and miRNA mimic NC were named as pGC-FU-miR-33b and pGC-FU-NC-LV, respectively. The anti-miR-33b and negative control were named as pFU-GW-miR-33b and pFU-GW-RANI-NC, respectively.

Lung cell lines were grown overnight and then infected with different lentiviral MOIs (20 for A549 and 2 for H1299) for 24, 48 and 72 h. The cells were then subjected to different assays.

Lung cancer A549 and H1299 cells were seeded into 96-well plates and then infected with different lentiviruses overnight. The cells were grown for 0, 24, 48, 72 and 96 h and then subjected to the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, the cell culture was supplemented with 20 μl of 0.5 mg/ml of MTT and cells were incubated for 4 h at 37ºC. Next, the culture medium was replaced with 200 μl of dimethyl sulfoxide to dissolve the precipitates for 20 min at the room temperature. The optical density was measured at 490 nm using a multi-well plate reader. The experiments were performed in triplicate and repeated at least three times. Data are summarized as mean ± SE and normalized as % of control experiments.

Lung cells (1×10⁴) in serum-free medium were plated in the top chamber of each Transwell insert (Corning Costar, Corning, NY, USA). The complete growth medium was added to the bottom chambers and cells were grown for 24 h at 37ºC. At the end of experiments, cells that remained in the upper chambers were removed using a cotton swab. The cells that migrated or invaded to the bottom chambers and attached to the lower surface of the membrane were fixed and stained with a dye solution containing 20% methanol violet and 0.1% crystal violet and counted under a microscope.

Transwell filter without Matrigel (BD Biosciences, Bedford MA, USA) was used for migration assay, while Transwell filter coated with Matrigel was used for tumor cell invasion assay.

Total RNA from A549 and H1299 cells was isolated using the TRIzol reagent (Invitrogen) and reversely transcribed into cDNA using the All-in-One™ First Strand cDNA Synthesis kit (AORT-0050; GeneCopoeia, Shanghai, China) and miRNA First Strand cDNA Synthesis kit (AMRT-0050; GeneCopoeia) according to the manufacturer's instructions. qPCR amplification was performed using the All-in-One™ qPCR mix (AOPR-1200; GeneCopoeia) and the All-in-One™ miRNA qRT-PCR detection kit (AOMD-Q050; GeneCopoeia) on an ABI 7300HT real-time PCR system (Applied Biosystems, Foster City, CA, USA). The qPCR cycle was set to an initial 95ºC for 5 min, 40 cycles of 95ºC for 15 sec, 60ºC for 30 sec, and a final melt-curve analysis (60–95ºC). Primers used for qPCR are listed in Table III. U6 was used as the endogenous control to normalize miRNA level, while GAPDH mRNA was used to normalize mRNA levels using the 2^−ΔΔCT method. Primers for miR-33b (HmiRQP0432) and U6 (HmiRQP900) were purchased from GeneCopoeia.

Total cellular protein was extracted using a lysis buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, and 1% NP-40. BCA protein assay kit (Pierce, Rockford, IL, USA) was used to determine concentration of the protein samples. Western blotting was conducted according to a previous study (30) with the following primary antibodies: ZEB1 (NBP2-23484; Novus Biologicals, Littleton, CO, USA, at a dilution of 1:800), vimentin (D21H3, 1:1,000; Cell Signaling Technology, Danvers, MA, USA), E-cadherin (ab76055, 1:1,000; Abcam, Cambridge, MA, USA), β-catenin (D10A8, 1:1,000; Cell Signaling Technology), GAPDH (14C10, 1:1,000; Cell Signaling Technology), and secondary goat anti-rabbit and goat anti-mouse (1:5,000) immunoglobulin G (Invitrogen).

Paraffin sections were obtained from the Xiangya Hospital of The Central South University and used for immunostaining expression of different proteins according to a protocol provided by Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd. (Beijing, China). Negative control sections were incubated with normal goat serum to replace the primary antibody (ZEB1, vimentin, E-cadherin and β-catenin). All primary antibodies were used at a dilution of 1:200. The immunostained tissue sections were then reviewed and scored under a microscope according to a previous study. Specifically, each sample was scored separately by two pathologists and reviewed until a consensus was reached. Tissue sections without staining or <5% staining were scored as 0; those with 5–30% staining as 1; with 31–70% staining as 2; and ≥71% staining as 3. The staining intensity was scored as 0 for negative staining, 1 for weak staining, 2 for moderate staining, and 3 for strong staining. The two scores were then combined to obtain the immunostaining score for each case, i.e., the final score of 0–1 was considered as negative gene expression (−); 2–3, weak expression (+); 4–5, moderate expression (++); and 6, strong expression (+++) (31).

Lung cells were grown on coverslips and infected with different lentiviruses for 24 h and then washed with phosphate-buffered saline (PBS) twice and fixed in freshly prepared 4% paraformaldehyde for 10 min at room temperature. For immunofluorescence staining, the cells were treated with 0.1% Triton X-100 for 10 min to permeabilize cell membrane and further incubated with 1% BSA for 1 h at room temperature. Next, the cells were incubated overnight with a primary antibody against ZEB1, vimentin, E-cadherin, or β-catenin at a dilution of 1:100 at 4ºC. The next day, cells were washed with PBS three times and incubated with a secondary antibody diluted 1:100 in PBS for 1 h at the room temperature. DAPI was used to visualize the nuclei by incubating cells at room temperature for 10 min. Subsequently, the cells were examined and scored under a fluorescence microscope (Leica Microsystems, Heidelberg, Germany).

The wild-type and mutant ZEB1 3′-untranslational region (UTR) luciferase reporters were obtained from Guangzhou RiboBio Co., Ltd. (Guangzhou, China). Human ZEB1 cDNA 3′-UTR region and mutated 3′-UTR region were generated with genomic DNA from 239T cells using PCR with the following primers ZEB1-3′-UTR wild-type, 5′-GGCGGCTCGAGATACTTGCCTTGGACTGTAG-3′ and 5′-AATGCGGCCGCAGTGAGGAATGTGAGAGTGA-3′ or ZEB1-3′-UTR mutant, 5′-CACAATAATTACGTTCAACAAGCCTGAACTGC-3′ and 5′-CTTGTTGAACGTAATTATTGTGAGATGGGAGT-3′, and then cloned into the XhoI and NotI sites of pmiR-RB-REPORT™ vector (Promega, Madison, WI, USA). After amplification and DNA-sequence confirmation, these vectors were named as pmiR-ZEB1-3′-UTR-wt and pmiR-ZEB1-3′-UTR-mut, respectively. For luciferase assay, HEK-293 cells were grown and co-transfected with these luciferase reporters and miR-33b mimic or a negative control vector for 36 h using Lipofectamine 2000 (Invitrogen). Subsequently, cells were subjected to protein extraction and the dual luciferase assay system (Promega) to measure the luciferase activity according to the kit instructions. The firefly luciferase activity of each sample was then normalized to Renilla luciferase activity.

All mouse experiments were carried out according to the regulations of the Animal Center of Central South University (Hunan, China), which supplied the 4-week-old female BALB/c nude mice. For animal experiments, lung adenocarcinoma H1299 cell line was infected with lentivirus carrying miR-33b or negative control for 36 h and then injected subcutaneously into the flank region of these nude mice for 1×10⁷ in 200 μl volume per mouse (n=6). The length (a) and the width (b) of the tumor xenografts in mice were measured. The xenograft volumes were calculated using the formula, V (mm³) = a×b²/2. The mice were sacrificed on day 21 and tumor xenografts were removed, photographed, weighed and sectioned for hematoxylin and eosin (H&E) staining, immunohistochemistry, qRT-PCR and western blot analysis.

All data were summarized as mean ± SD derived from three independent experiments, unless specifically stated. Comparative data between the groups were analyzed for statistical significance using the two-tailed paired Student's t-test. Comparison of categorical variables between the groups was statistically analyzed using the χ² test. Pearson's test was performed to calculate the association between miR-33b and ZEB1 expression. P<0.05 was considered statistically significant.

We first assessed miR-33b expression in NSCLC cell lines and found that the levels of miR-33b were significantly reduced in NSCLC cell lines compared with those of normal bronchial epithelium HBE cells (Fig. 1A). We then analyzed miR-33b expression in 45-paired primary lung adenocarcinoma and distant normal tissues (at least 5 cm distal to the tumor lesion). Compared with normal tissues, lung adenocarcinoma tissues expressed much lower levels of miR-33b (0.033±0.027 vs. 0.073±0.061, P<0.01; Fig. 1B). We then associated miR-33b expression with the patient clinicopathological features and found that reduced miR-33b expression was associated with tumor lymph node metastasis, but not with age, gender, or tumor stage (Fig. 1C and Table IV).

We then assessed the effects of miR-33b restoration on suppression of lung adenocarcinoma cell proliferation, migration, and invasion and tumor cell EMT in vitro by infected H1299 cells with miR-33b mimics and miR-NC. The data showed high transfection efficiency and restored miR33b expression in the infected cells (Fig. 2A). miR-33b upregulated E-cadherin expression. However, the expression of vimentin, another marker of mesenchymal phenotype, was decreased (Fig. 2B–D). The miR-33b restoration in lung cancer cells significantly reduced tumor cell viability (Fig. 2E) and migration and invasion capacity (Fig. 2F). In contrast, knockdown of miR-33b expression promoted lung adenocarcinoma cell proliferation, invasion and migration and EMT in vitro (Fig. 3).

MicroRNA regulates the expression of protein-coding genes in cells. Therefore, we explored the molecular mechanism by which miR-33b suppressed lung cancer metastasis. We first performed gene target analysis using online tools (miRanda and miRbase) to identify potential miR-33b-binding genes and found that miR-33b binds to ZEB1 3′-UTR region (Fig. 4A). We then validated it using several different experiments. First, we associated miR-33b with ZEB1 expression in lung adenocarcinoma cell lines transfected with miR-33b, miR-NC, anti-miR-33b and anti-miR-NC. We found that miR-33b overexpression significantly reduced ZEB1 expression in the cells, whereas knockdown of miR-33b expression significantly upregulated ZEB1 expression (Fig. 4B). We then constructed a luciferase reporter plasmid carrying ZEB1 3′-UTR (Fig. 4C). The luciferase activity was significantly decreased in the Luc-ZEB1-3′-UTR-transfected cells compared with that of the potential target site of mutant miR-33b in ZEB1 3′-UTR and negative control cells. In addition, immunofluorescence showed exclusive expression and cellular localization of ZEB1 protein in H1299 cells (Fig. 4D).

Furthermore, we analyzed the expression of ZEB1 mRNA in additional lung adenocarcinoma cell lines (Fig. 5A) and in paired lung cancer and normal tissue specimens. Expression of ZEB1 mRNA in normal and cancerous tissues was 0.084±0.11 and 0.319±0.12, respectively; P<0.001; Fig. 5B). We then associated the expression of ZEB1 mRNA and miR-33b in tissue samples and found an inverse correlation in lung adenocarcinoma tissues (r=−0.9086, P<0.001; Fig. 5C). These data implied that miR-33b targeted ZEB1 expression in lung cancer cells. Furthermore, our immunostaining data involving tissue samples also showed that expression of ZEB1 protein was upregulated in lung adenocarcinoma tissues and lymph node metastasized tumor tissues compared with that of distant normal tissues (P<0.01; Fig. 5D).

In addition, we co-infected siRNA-NC, ZEB1-siRNA, and anti-miR-33b in the lung cancer A549 cell line. The data confirmed the infection efficiency using western blot analyses of their protein expression after lentivirus infection (Fig. 5E). Furthermore, knockdown of ZEB1 expression in A549 cells using ZEB1 siRNA resulted in decrease in tumor cell proliferation, migration, and invasion (Fig. 5F and G), whereas induction of ZEB1 expression after miR-33b knockdown induced tumor cell proliferation, migration and invasion (Fig. 5F and G). Taken altogether, our data indicate that miR-33b targets ZEB1 expression by binding to ZEB1 3′-UTR region to modulate proliferation, migration and invasion of lung adenocarcinoma cells.

Previous studies demonstrated that the WNT/β-catenin signaling is activated in NSCLC cell lines (32). Activated WNT/β-catenin signaling promoted the growth of lung cancer cells regulated by miR-21 (33). Thus, in the present study, we assessed the role of miR-33b in lung cancer cells or tissues via modulation of WNT/β-catenin pathway activity. We infected lung cancer A549 and H1299 cells with lentivirus carrying miR-33b, miR-NC, anti-miR-33b or anti-miR-NC. Our data showed that miR-33b overexpression significantly decreased β-catenin expression, whereas miR-33b knockdown increased β-catenin expression (Fig. 6A and B).

Furthermore, we co-transfected β-catenin siRNA, siRNA-NC, and anti-miR-33b in A549 cells (Fig. 6C). We found that knockdown of β-catenin expression decreased proliferation and invasion of A549 cells, whereas anti-miR-33b-upregulated β-catenin promoted A549 cell proliferation and invasion (Fig. 6D and E). Taken together, these results indicated that miR-33b expression may regulate the activity of WNT/β-catenin signaling.

After demonstrating the role of miR-33b in lung cancer cells by targeting ZEB1 expression and suppressing activity of the Wnt/β-catenin/ZEB1 pathway in vitro, we assessed the effects of miR-33b on tumor growth using the in vivo nude mouse xenograft assay. Our data showed that the mean volume of miR-33b-H1299 cell xenograft tumors was significantly smaller than that of control cell xenografts (Fig. 7A and B). Immunohistochemistry revealed that the expression of ZEB1, vimentin, E-cadherin and β-catenin genes was consistent with our in vitro data (Fig. 7C). In brief, expression of ZEB1, β-catenin and Vimentin proteins was downregulated, whereas E-cadherin levels were upregulated in miR-33b-overexpressed lung cancer cell xenografts. The western blot data further confirmed immunostaining data (Fig. 7D and E) demonstrating that miR-33b overexpression inhibited tumor growth, EMT and the Wnt/β-catenin pathway activity in vivo.