Breast cancer samples and the corresponding adjacent ovarian tissues were obtained from 32 patients with primary BC who underwent surgery at China-Japan Union Hospital of Jilin University (Changchun, China) from July 2012 to August 2014. Normal breast tissues adjacent to the tumor were taken from 3 cm away from the tumor cells. All of the samples and matched clinical information were collected after obtaining prior written informed consent from the patients. The samples were immediately snap-frozen in liquid nitrogen and stored at −80°C until use. No patients received chemotherapy or radiotherapy prior to surgery. This study is approved by Institutional Ethics Committees of Jilin University.

The non-cancerous human mammary epithelial cell line MCF-10A, breast cancer cell line and human breast cancer cell lines MCF-7, MDA-MB-231, BT-549 and MDA-MB-453 were obtained from the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). All cells were maintained in RPMI-1640 medium (Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (FBS, Gibco BRL, Gaithersburg, MD, USA), 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C in a humidified chamber supplemented with 5% CO₂.

Total RNA was extracted from tissue and cells using TRIzol (Invitrogen) according to the manufacturer's instructions. For miRNA reverse transcription, cDNA was synthesized using One Step Prime script miRNA cDNA Synthesis kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. For mRNA reverse transcription, cDNA was synthesized using the Primer Script RT reagent kit (Takara Bio, Japan). Quantitative PCR was performed using Fast SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) under ABI 7900 Sequence Detection System (Life Technologies, NY, USA). U6 snRNA and GAPDH was used as an endogenous control. The specific primers for miRNA-126 and U6 were purchased from Applied Biosystems. The specific primers for GAPDH and SOX4 are as follows: GAPDH (sense), 5′-TCAACGACCACTTTGTCAAGCTCA-3′; antisense: 5′-GCTGGTGGTCCAGGGGTCTTACT-3′; SOX4 (sense), 5′-AGCGACAAGATCCCTTTCATTC-3′; antisense: 5′-CGTTGCCGGACTTCACCTT-3′. The comparative 2^−ΔΔCt method was used for relative quantification and statistical analysis. All experiments were performed in triplicate.

miRNA-338-3p mimics and negative control mimics (miR-NC) were purchased from GenePharma (Shanghai, China). MDA-MB-231 cells were seeded into cell culture plates 20 h before transfection to ensure 70% cell confluence at the moment of transfection. Transfection of miRNA mimics into MDA-MB-231 cells was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's procedure at the final concentration of 100 nM. At 48 h post-transfection, transfection efficiencies were evaluated in every experiment by qRT-PCR and western blot analysis.

The transfected cells (5×10³ cells/well) were plated into 96-well plates. At 24, 48 and 72 h post-transfection, MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) were added into cells and cultured for 4 h at 37°C, followed by removal of the culture medium and the addition of 150 μl dimethyl sulfoxide (DMSO, Sigma-Aldrich). The absorbance at 490 nm (OD490 nm) was measured with a spec-trophometer.

For colony formation, the transfected cells were seeded into a 6-well plate at a density of 1,000 cells/well. The medium was changed every three days. Approximately 2 weeks later, the clones were washed with 1X PBS and stained with 1.0% crystal violet (Sigma) for 5 min. Finally the clones were photographed and counted.

Cell cycle and apoptosis assay was performed on MDA-MB-231 cells 48 h after transfection. For cell cycle assay, the transfected cells at 1×10⁶ cells per well were cultured in 12-well plates in triplicate and cultured for 24 h. Then the cells were collected by trypsinization and washed in PBS and fixed in ice-cold ethanol overnight at 4°C, followed by incubated in 1 ml of staining solution (20 μg/ml propidium iodide and 10 U/ml RNaseA) for 30 min at room temperature. Cell cycle distribution was assayed by fluorescence-activated cell sorting based on FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). All experiments were performed in triplicate.

Cell apoptosis analysis was performed with Annexin V-FITC Apoptosis Detection kit I (BD Biosciences) according to the manufacturer's instructions using FACSCalibur flow cytometer (BD Biosciences).

Transfected cells (5×10³) were seeded into 24-well tissue culture plates for 48 h. Thereafter, a linear wound of cellular monolayer was created in the confluent cells. After wounding, the debris was removed by washing the cells with PBS. Migration of cells into the wound was observed at 0 and 24 h using an IX51 inverted microscope (Olympus, Tokyo, Japan). Individual cells were quantified as an average of at least five fields for each experiment.

Cell invasion assays were performed using 24-well Transwell Permeable Supports with 8-μM pores (Corning, Lowell, MA, USA). Briefly, 2×10⁴ transfected cells were suspended in serum-free medium and seeded into the Transwell inserts coated with growth factor reduced Matrigel (BD Biosciences, Bedford, MA, USA). Bottom wells were filled with media containing complete media. The invasion assay was performed for 24 h at 37°C in a 5% CO₂ incubator. After incubation, migrating cells present on the lower surface of the membrane were fixed in 70% ethanol for 30 min and stained with 2% crystal violet for 10 min on a glass slide. Cells from 10 random fields were counted under an IX51 inverted microscope (Olympus).

Prediction of the miR-338-3p targets was performed using two publicly available algorithms: TargetScan6.2 (http://www.targetscan.org/), miRanda (http://www.microrna.org/).

The 3′-UTR of SOX4 was amplified and cloned downstream of pGL3/Luciferase vector (Wt SOX4 3′-UTR). Then the mutant 3′-UTR of SOX4 (several nucleotides within the binding sites were mutanted) was amplified using pGL3/Luc-SOX4 3′-UTR as the template, and was cloned downstream of the pGL3/Luciferase vector (Mut SOX4 3′-UTR). For the luciferase reporter assay, the cells were co-transfected with miR-338-3p mimic or miR-NC and Wt SOX4 3′-UTR or Mut SOX4 3′-UTR, together with the controls. Forty-eight hours post-transfection, cells were lysed using 1X passive lysis buffer (Promega, Madison, WI, USA) and lysates were analyzed using the Dual-Glo Luciferase Reporter Assay System (Promega) on the Synergy4 multi-mode microplate reader (BioTeK).

Total protein was extracted by using RIPA buffer (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) from cells harvested 48 h after transfection, separated in 10% SDS polyacrylamide gels, and electrophoretically transferred to nitrocellulose membranes (NC membranes, Invitrogen, Carlsbad, CA, USA), blocked in 4% dry milk at room temperature for 1 h, and immunostained with primary antibodies at 4°C overnight using anti-SOX4 (1:1,000, Santa Cruz); and anti-GAPDH antibody (1;5,000, Santa Cruz). After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse secondary antibodies (1:5,000; Santa Cruz Biotechnology). The protein bands were visualized on X-ray film with a chemiluminescent detection system (Beyotime, Shanghai, China). Blots were stripped and reprobed with anti-GAPDH to control for loading variations.

Twenty female BALB/c mice (18–20-g, 4–5-week-old) were obtained from Experimental Animal Center of Changchun Biological Institute (Changchun, China), and kept under specific pathogen-free (SPF) conditions. This study and all experimental protocols were approved and the methods were performed in accordance with the guidelines of the Animal Care and Use Committee of Jilin University.

MDA-MB-231 cells (2×10⁶), stably expressing miR-338-3p or miR-NC, were suspended in 100 μl PBS and then injected subcutaneously into the posterior flank of female BALB/c athymic nude mice. Tumor volumes in mice were measured with a slide caliper every week until scarifice according to the formula: Volume (mm³) = 1/2 × width² × length. Five weeks after injection, mice were sacrificed, and tumor tissues were resected and weighed. In addition, SOX4 protein expression level of tumor tissue was determined by western blot analysis.

All the data are shown as mean ± SD (standard deviation) and the experiments were repeated three times. The difference was determined by two-tailed Student's t-test. Statistical analysis was performed with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA, USA). P<0.05 was considered statistically significant.

To determine the role of miR-338-3p in breast cancer progression, we examined miR-338-3p expression in breast cancer tissue samples and corresponding normal tissue sample by qRT-PCR. miR-338-3p was downregulated in the breast cancer samples compared with the paired normal breast tissues (Fig. 1A). To investigate the clinical relevance of miR-338-3p in breast cancer, we divided the 32 patients to high miR-338-3p expression group (n=17) and low miR-338-3p expression group (n=15) using the mean value (0.45±0.04) of relative expression levels as a cutoff. The correlation between the miR-338-3p expression level and clinical and pathologic characteristics of breast cancer is listed in Table 1. In 12 cases presenting as advanced stage III, 9 (75.0%) of the cases have low-level miR-338-3p expression in cancer tissues; while in 20 early stages (stages I and II), only 6 (30.0%) presented with low levels of miR-338-3p expression. In the 13 cases of breast cancer patients with lymph node metastasis, 10 (76.9%) exhibited low miR-338-3p expression, however only 5 (26.3%) of 19 cases of cancers without lymph node metastasis presented low-level miR-338-3p expression. No correlation was observed between the miR-338-3p level and the age, tumor size or pathologic grade status of breast cancer.

In addition, the expression of miR-338-3p was detected in a panel of breast cancer cell lines and non-cancerous breast epithelial cell line (MCF-10A). It was found that miR-338-3p expression was downregulated in breast cancer cell lines compared to normal non-cancerous breast epithelial cell line MCF-10A (Fig. 1B). The data above showed that miR-338-3p decreases in both breast cancer tissues and cell lines, and that its expression is inversely correlated with the metastatic abilities of breast cancer cells. We selected MDA-MB-231 cells, which showed the lowest expression of miR-338-3p, to conduct further experiments.

As miR-338-3p was downregulated in breast cancer tissues and cell lines, we transfected miR-338-3p mimic into breast cancer cells. The results from qRT-PCR assay showed that miR-338-3p mimic could significantly upregulate the level of miR-338-3p expression in breast cancer cells compared to miR-NC group (Fig. 2A). To confirm miR-338-3p effect on proliferation and colony formation in breast cancer cells, MTT assay and colony formation assay were performed. It was found that overexpression of miR-338-3p significantly inhibited cell proliferation (Fig. 2B) and colony formation (Fig. 2C) in breast cancer cells.

The above results showed that miR-338-3p expression is inversely correlated with the meta-static ability of breast cancer cells, therefore, to investigate the miR-338-3p effect on metastasis in vitro, migration and invasion assays were performed in MDA-MB-231 cells transfected with miR-338-3p mimic or miR-NC by wound healing and transwell chamber assay, respectively. As expected, overexpression of miR-338-3p in breast cancer cells significantly inhibited cell migration (Fig. 3A) and invasion (Fig. 3B) of breast cancer cells. Collectively, these results suggested that miR-338-3p can efficiently inhibited migration and invasion of breast cancer cells.

To further verify the role of miR-338-3p in breast cancer cells, we tested the cell cycle and apoptosis effects in breast cancer cells by flow cytometry. Cell cycle assay showed that the percentage of G1 phase cells increased, and the percentage of S phase cells decreased in breast cancer cells transfected with miR-338-3p mimic compared to cells transfected with miR-NC (Fig. 4A). Cell apoptosis assay showed that overexpression of miR-338-3p significantly induced cell apoptosis in breast cancer cells (Fig. 4B).

To explore the mechanism underlying the growth inhibition by miR-338-3p in breast cancer cells, we used publicly available algorithms (Targetscan6.2 and miRanda) to help identify miR-338-3p targets in human breast cancer. We found that sex-determining region Y-box 4 (SOX4) was a putative target of miR-338-3p. To confirm this possibility, the miR-338-3p binding sequences present at the 3′-UTR of SOX4 mRNA (WT-3′-UTR SOX4) or its mutant site (Mut-3′-UTR-SOX4) were subcloned downstream of the luciferase reporter gene in pGL3 vector (Fig. 5A) and then co-transfected into MDA-MB-231 cells, along with miR-338-3p mimic or miR-NC for luciferase assay evaluation. Luciferase assay further revealed that breast cancer cells transfected with miR-338-3p mimic repressed wild-type 3′-UTR-SOX4 reporter activity (P<0.01), while miR-338-3p mimic had no inhibition effect on the mutant 3′-UTR-SOX4 reporter activity (Fig. 5B), indicting the direct regulation of miR-338-3p in the 3′-UTR of SOX4 mRNA. To further confirm that SOX4 acts as a target of miR-338-3p, we examine the effect of miR-338-3p on SOX4 expression by qRT-PCR and western blot analysis. As predicted, qRT-PCR and western blotting showed that ectopic miR-338-3p in BC cells downregulated SOX4 expression on mRNA level (Fig. 5C) and protein level (Fig. 5D).

As the above results show that miR-338-3p is downregulated in BC and targets SOX4 by binding to its 3′-UTR, we next determined whether SOX4 expression is negatively associated with miR-338-3p levels in the BC tissue samples. qRT-PCR and western blot assay showed that the expression of SOX4 on mRNA and protein level was significantly higher in BC tissues than in matched normal tissues (P<0.05) (Fig. 6A and B). In addition, a statistically significant inverse correlation was found between expression levels of miR-338-3p and SOX4 mRNA in BC tissue by Spearman's correlation analysis (r=−0.6431, P<0.001, Fig. 6C).

Finally, we tested whether ectopic expression of miR-338-3p could influence the growth of breast tumor in vivo. MDA-MB-231 cells with stable expression either of miR-338-3p or miR-NC were selected and injected subcutaneously into nude mice, and the tumor formation was monitored. MDA-MB-231 cells transfected with miR-NC showed progressive growth, while MDA-MB-231 cells transfected with miR-338-3p mimic inhibited tumor growth relative to miR-NC group. After 35 days, the nude mice were sacrificed, and the tumor weights and volume were measured (Fig. 7A–C). It was found that overexpression of miR-338-3p can significantly suppress the tumor growth of xenografts in nude mice (Fig. 7A), and decrease the tumor volume (Fig. 7B) and tumor weight (Fig. 7C) compared to miR-NC group, indicating the suppressive function of miR-338-3p on breast cancer tumorigenicity in vivo. In addition, we also determined SOX4 protein expression in tumor tissue by western blot analysis. We found that SOX4 protein expression was decreased in the xenograft tumors of miR-338-3p mimic group compared to the xenograft tumors of miR-NC group (Fig. 7D), suggesting that miR-338-3p suppressed breast cancer tumorigenicity in vivo by targeting SOX4.