A total of 30 pairs of esophageal squamous cell carcinoma (ESCC) and non-tumor adjacent tissues (>5 cm away from tumor tissues) were obtained from Baoji Central Hospital (Shaanxi, China) between June 2015 and November 2017. All patients were diagnosed with ESCC by a pathological evaluation, and they had not received biotherapy, chemotherapy or any other treatment prior to the initiation of the study. Ethical approval was granted by the Ethics Committee of the Baoji Central Hospital and written informed consent was received from each patient in accordance with the institutional guidelines.

The KYSE30 human esophageal cancer cell line (The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences) and the human squamous epithelial cell line Het-1A (BeNa Culture Collection; Beijing Beina Chunglian Biotechnology Research Institute) were cultured in Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc.) at 37°C in the presence of 5% CO₂ and penicillin-streptomycin (Sigma-Aldrich; Merck KGaA). miR-106b-3p mimics, inhibitor and scramble (NC) were purchased from Shanghai GenePharma Co., Ltd. KYSE30 cells were seeded into 6-well plates at an initial density of 2×10⁵ cells/well. Cells were transfected with 40 nM NC or miR-106b-3p inhibitor or miR-106b-3p mimics sequences using Lipofectamine^® 2000 reagent (Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. After 24 h transfection, subsequent experimentation was conducted and the transfection efficiency of miR-106b-3p was assessed. The sequences were as follows: miR-106b-3p mimics, 5′-TAAAGTGCTGACAGTGCAGAT−3′; miR-106b-3p inhibitor, 5′-AUCUGCACUGUCAGCACUUUA−3′; and NC, 5′-CAGUACUUUUGUGUAGUACAA−3′.

Small interfering (si)RNA, siRNA-1 TGM3 and siRNA-2 TMG3, were designed and purchased from Shanghai GenePharma Co., Ltd. KYSE30 cells were seeded into a 6-well plate at a density of 2×10⁵ cells/well. Subsequently, 40 nM siRNA-1 TGM3 and siRNA-2 TMG3 sequences were transfected into cells with Lipofectamine 2000 reagent, according to the manufacturer's instructions. siRNA-1 TGM3 sequence, 5′-TATGAATTCTGTACGGGAGGCCACCAGCGC−3′; siRNA-2 TGM3 sequence, 5′-TATGAATTCTGTACGGGAGGCCACCAGCGC−3′.

Total RNA was extracted from tissues and cells using TRIzol^® reagent (Thermo Fisher Scientific, Inc.). Subsequently, cDNA was synthesized using the miScript Reverse Transcription kit at room temperature for 10 min, and qPCR was performed using the miScript SYBR-Green PCR kit (both purchased from Qiagen, Inc.). The U6 small nuclear RNA was used for normalization. Determination of the relative levels of TGM3 was performed using the TaqMan™ PCR Master Mix (Applied Biosystems; Thermo Fisher Scientific, Inc.), while glyceraldehyde phosphate dehydrogenase (GAPDH) functioned as an internal control. The relative expression levels were calculated according to the 2^−ΔΔCq method (20).

The transfected cells were seeded into a 96-well plate at a density of 1×10⁴ cells/well and cultivated for 24 h at 37°C in the presence of 5% CO₂. As performed previously, cells were treated with 4 µmol/l cisplatin for an additional 24 h (21). Cell viability was detected using Cell Counting Kit-8 reagent (CCK-8; Dojindo Molecular Technologies, Inc.), according to the manufacturer's instructions, and the absorbance was measured at 490 nm with a microplate reader.

The transfected cells were seeded into a 96-well plate at a density of 1×10⁵ cells/well. Medium containing 4 µmol/l cisplatin was added for 48 h at 37°C, in the presence of 5% CO₂. Following trypsinization and washing with PBS (Thermo Fisher Scientific, Inc.), the cells were double stained with FITC Annexin V and propidium iodide (PI) for 10 min in the dark at room temperature. Finally, the apoptotic rate at the early and late period was determined via a FACScan flow cytometer and analyzed with CellQuest software version 5.2.1 (both from BD Biosciences).

The transfected cells were isolated and protein was extracted using RIPA buffer (Beijing Solarbio Science & Technology Co., Ltd.) containing Protease Inhibitor Cocktail (Bimake). The concentrations of proteins were measured using BCA reagent (Solarbio Life Sciences) and 20 µg/lane proteins were separated by 10% SDS-PAGE (Thermo Fisher Scientific, Inc.) and transferred to PVDF membranes (Beijing Solarbio Science & Technology Co., Ltd.). The membranes were blocked with 5% non-fat milk for 50 min at room temperature and the primary antibodies were added to the membranes at 37°C overnight. The primary antibodies were as follows: Rabbit anti-TGM3 (cat. no. NBP1- 86950; Bio-Techne), rabbit anti-Bcl-2 (cat. no. IMG-5685; Bio-Techne), rabbit anti-Bax (cat. no. AF820; Bio-Techne), rabbit anti-caspase 3 (cat. no. AF835; Bio-Techne) and rabbit anti-GAPDH (cat. no. IMG-5143A; Bio-Techne). Subsequently, the HRP-conjugated secondary antibody IgG H&L (1:1,000; cat. no. ab7090; Abcam) was added and cultured for another 2 h at room temperature. Finally, the signals were detected using the electrochemiluminescence assay (BD Pharmingen; BD Biosciences). GAPDH was used as the endogenous reference.

TargetScan 7.2 software (targetscan.org) was used to predict the potential binding sequences of miR-106b-3p and TGM3 according to a previous study (22). Het-1A cells were cultured in a 24-well plate at a density of 1×10⁴ cells/well. Afterwards, the full length of TGM3 was amplified from the cDNA of Het-1A cells, and inserted into the pGL3-Basic vector (Promega Corporation) in order to construct wild-type (WT) TGM3. Subsequently, Het-1A cells were co-transfected with WT and mutant TGM3 along with NC and miR-106b-3p mimics sequences using Lipofectamine 2000 reagent (Thermo Fisher Scientific, Inc.) at 37°C, in the presence of 5% CO₂. Following 48 h of cell culture after transfection, luciferase activity was determined using the dual-luciferase reporter gene assay kit (Promega Corporation) and normalized to Renilla luciferase enzyme activity, according to the manufacturer's protocol.

SPSS version 22.0 (IBM Corp.) was used to analyze the data. All data are presented as the mean ± standard deviation. The differences between groups were analyzed using the Student's t-test or one-way ANOVA followed by Newman-Keuls post hoc test. Pearson's correlation analysis was performed to measure the correlation between miR-106b-3p and TGM3 expression levels. P<0.05 was considered to indicate a statistically significant difference.

As shown in Fig. 1A, the expression levels of miR-106b-p were significantly increased in the ESCC tissues compared with the non-tumor adjacent tissues (P<0.01); whereas the mRNA and protein expression levels of TGM3 were significantly downregulated in ESCC tissues compared with those in non-tumor adjacent tissues (P<0.01; Fig. 1B-D).

As shown in Fig. 2A, TGM3 expression was negatively correlated with miR-106-3p (r=−0.2726, P=−0.0323). Furthermore, the sequences of TGM3 3′-untranslated region (UTR) contained the putative miR-106b-3p binding sites as predicted by TargetScan (Fig. 2B). Following transfection with WT and mutant TGM3 3′-UTR plasmids, luciferase activity was significantly decreased in the cells transfected with miR-106b-3p mimics (P<0.01; Fig. 2C).

The expression levels of miR-106b-3p were significantly decreased following transfection with the miR-106b-3p inhibitor, compared with those transfected with the NC (P<0.01; Fig. 3A). Furthermore, the mRNA and protein expression levels of TMG3 were significantly reduced following transfection with siRNA-1 TMG3 or siRNA-2 TMG3 sequences, compared with those noted in the siRNA group (P<0.01; Fig. 3B and C).

Following treatment with cisplatin, cell viability was significantly suppressed by transfection with the miR-106b-3p inhibitor (Fig. 4). However, this suppressive effect could be reversed by co-transfection with TMG3 siRNA (P<0.01).

The induction rate of apoptosis was significantly increased by transfection with the miR-106b-3p inhibitor. However, this effect was reversed by co-transfection with TMG3 siRNA (P<0.01; Fig. 5A-F).

Following treatment of the cells with cisplatin, the expression levels of Bcl-2 were decreased, whereas the expression levels of TGM3, Bax and caspase-3 were increased following transfection with the miR-106b-3p inhibitor. This variation could be reversed by co-transfection with TMG3 siRNA (Fig. 6).