A total of 79 ESCC specimens (55 paraffin-embedded and 24 fresh) with adjacent free-cancerous esophageal tissues that were frozen and stored in liquid nitrogen were histopathologically diagnosed at the Department of Pathology of Chongqing Medical University (Chongqing, China), and the Department of Cardiothoracic Surgery of The First Affiliated Hospital of Chongqing Medical University (Chongqing, China), respectively, from 2013 to 2015. Clinical and tumor (T)-lymph node (N)-metastasis (M) staging was performed in accordance with the guidelines of the 6th edition of the International Union Against Cancer (UICC). None of the patients received radiation or chemotherapy before surgery. Clinical patient information is summarized in Table I. The study protocol was approved by the Institutional Research Ethics Committee of The First Affiliated Hospital of Chongqing Medical University and informed consent was obtained from all patients.

Immunohistochemical analysis of 55 human paraffin-embedded ESCC samples was performed to study the expression profiles of TNFAIP2, β-catenin and E-cadherin using anti-human TNFAIP2 antibody (dilution, 1:100; cat. no. sc-30138; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), β-catenin antibody (dilution, 1:100; cat. no. 8480) and E-cadherin antibody (dilution, 1:400; cat. no. 3195) (both from Cell Signaling Technology, Inc., Danvers, MA, USA). The procedure was similar to previously described methods (11,13–15). The histopathological features and patient data of the samples were evaluated via the H-scores method by two board-certified pathologists who were blinded to patient data. The intensity of specific staining was defined as negative (0), weakly positive (1+), somewhat positive (2+), strongly positive (3+) and very strongly positive (4+). H-score = Σ P_i (i + 1), where P_i is the percentage of stained cells in each intensity category; i = 1-4. Total scores ≥300 were characterized as a high-level expression (16,17).

Protein extraction and western blot analysis were performed as previously described (15,18). Briefly, the cells were lysed with lysis buffer containing protease inhibitors and the lysates (40 µg/lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Then, proteins were transferred to a polyvinylidene fluoride membrane, which was incubated with appropriate secondary antibodies [i.e., anti-TNFAIP2 (dilution, 1:1,000; cat. no. sc-30138; Santa Cruz Biotechnology, Inc.), β-catenin (1:1,000; cat. no. 8480), E-cadherin (dilution, 1:1,000; cat. no. 3195; both from Cell Signaling Technology, Inc.), c-myc (dilution, 1:1,000; cat. no. ab32072; Abcam, Cambridge, UK), cyclin D1 (dilution, 1:1,000; cat. no. 2978), MMP-7 (dilution, 1:1,000; cat. no. 3801), p-GSK-3β (dilution, 1:1,000; cat. no. 9322), GSK-3β (1:1,000; cat. no. 12456), Snail (dilution, 1:1,000; cat. no. 3879) (all from Cell Signaling Technology, Inc.), and β-actin, respectively] that were detected and visualized using an enhanced chemiluminescence detection system (BeyoECL Plus; Beyotime Institute of Biotechnology, Shanghai, China).

Eca109 and Kyse150 cells, both ESCC cell lines, were maintained in Roswell Park Memorial Institute (RPMI)-1640 medium (Corning, Inc., Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS), 100 U/ml of penicillin, and 100 µg/ml of streptomycin, and incubated in a humidified atmosphere of 5% CO₂ at 37̊C.

rh-TNFα (cat. no. 10602; Sino Biological, Inc., Beijing, China) was diluted with complete medium to a concentration of 25 ng/ml. Cells were treated with rh-TNFα for 48 h and TNFAIP2 mRNA and protein expression levels were detected by quantitative real-time PCR (qRT-PCR) and western blot analysis, respectively. All experiments were repeated three times. The results of qRT-PCR were normalized with β-actin using the 2^−∆∆Ct method (19).

Lentivirus RNA interference of TNFAIP2 (LV-RNAi TNFAIP2 shRNA, 5′-TCTTCACCAAAGGGAAGAA-3′); and a negative control lentiviral vector (LV-CON shRNA, 5′-TTCTCCGAACGTGTCACGT-3′) were designed and synthesized by GeneChem Co., Ltd. (Shanghai, China). The lentivirus titers were 5×10⁸ infectious U/ml.

To determine the multiplicity of infection (MOI) of the Eca109 and Kyse150 cell lines, a pre-experiment was performed. Briefly, Eca109 and Kyse150 cells (2×10³ cells/well) were cultured in 96-well plates for 12 h, and then infected with 0.2 µl (MOI=10), 0.4 µl (MOI=20) or 0.6 µl (MOI=30) of lentivirus diluted with serum-free medium to 1×10⁸ infectious U/ml, respectively. After 12 h, the medium was replaced with fresh complete medium. After three days, 95% transduction efficiencies of the Eca109 and Kyse150 cells were infected at an MOI of 30 and 20, respectively.

According to the results of the pre-experiment, the Kyse150 and Eca109 cells (2×10⁴ cells/well) were cultured in 6-well plates and transfected with 4 and 6 µl of the diluted lentivirus, respectively. To obtain stable transfection, puromycin (1 µg/ml) was added to the cultures at 72 h after transfection. The RNAi effectiveness was validated by qRT-PCR and western blot analysis.

Total RNA was isolated from esophagus carcinoma tissues and Ec109 and Kyse150 cells, using RNAiso Plus Total RNA extraction reagent (cat. no. 9108; Takara Bio, Inc., Otsu, Japan). The PrimeScript™ RT reagent kit with gDNA Eraser (Perfect Real-Time) (cat. no. RR047A; Takara Bio, Inc.) was used for reverse transcription of total RNA (1 µg). qRT-PCR was performed with an Applied Biosystems 7500 Real-Time PCR System (Life Technologies, Carlsbad, CA, USA) using SYBR^® Premix Ex Taq™ II kit (Tli RnaseH Plus) (cat. no. RR820A, Takara Bio, Inc.) according to the manufacturers instructions. β-actin was used as an endogenous control to normalize the relative expression level of TNFAIP2. The following primer pairs were used to amplify the mRNA of TNFAIP2 and β-actin: (F) 5′-CCTGCTCTCCCTACGC-3′ and (R) 5′-CGTCCAAGATGCTCCG-3′ (11); and (F) 5′-CTCCTCCTGAGCGCAAGTACTC-3′ and (R) 5′-TCCTGCTTGCTGATCCACATC-3 (13), respectively.

Migration and invasion assays were performed as described by Huang et al and Whitington et al (15,20). Briefly, lentivirus-transfected Eca109 and Kyse150 cells were plated at concentrations of 5×10⁴ and 1×10⁵ cells/8-µm Transwell insert with or without Matrigel coating (both from Corning, Inc.) and incubated for 48 h, respectively. Cells that had penetrated the membranes were quantified by counting the number of cells in five random microscopic fields. All experiments were performed in triplicate.

LV-RNAi TNFAIP2 and LV-CON group cells were trypsinized and resuspended at 2×10⁴ cells/ml. Then, 25 µl (for colony formation; total volume, 2 ml) and 100 µl (for proliferation) of the cell suspensions were added to each well of 6- and 96-well plates, respectively. Cells plated in the 6-well plates were incubated for 10–14 days and the formed colonies were fixed with 4% paraformaldehyde and stained with crystal violet for 20 min, respectively. After 24, 48 and 72 h, 10 µl of Cell Counting Kit-8 (CCK-8; reagent) (Dojindo Molecular Technologies, Inc., Tokyo, Japan) was added to each well of the 96-well plates. Cell proliferation was determined using a microplate reader (Infinite 200 PROs; Tecan Trading AG, Männedorf, Switzerland) at designed time points by reading the absorbance at 450 nm after CCK-8 treatment for 2 h. Values of proliferation were obtained from five replicate wells for each treatment and time point. Each experiment was performed in triplicate.

Trypsinized cells were resuspended, washed, centrifuged with phosphate-buffered saline (PBS) at 4̊C three times, and then fixed with ice-cold 70% ethanol overnight at 4̊C for the cell cycle assay or resuspended in 1,000 µl of PBS for the cell apoptosis assay. The cell cycle and apoptosis assays were conducted at Chongqing Medical University Life Sciences Institute. Each assay was independently repeated three times.

SPSS 22 statistical software package (SPSS, Inc., Chicago, IL, USA) was used to perform all statistical analyses. The statistical significance of the qRT-PCR results for TNFAIP2 gene expression and for the observations on the basis of cell proliferation, colony formation, migration, invasion, cycle and apoptosis were assessed using Student's t-tests (two-tailed). The Pearson χ² test was applied to evaluate the relationships between TNFAIP2 expression and expression of β-catenin and E-cadherin and various clinicopathological characteristics. The Kaplan-Meier method and the log-rank test were used to plot and compare the survival curves, respectively. Data shown are presented as the mean ± SD or SEM from triplicate experiments as indicated. Significant differences were P-value <0.05.

TNFAIP2 was identified as an angiogenic factor and can be induced in umbilical vein endothelial cells by TNF-α (9). Furthermore, TNFAIP2 regulates invasion and metastasis in nasopharyngeal carcinoma (11), and its genetic polymorphism was found to influence ESCC, squamous cell carcinoma of the head and neck and gastric cancer (21–23). However, the mechanisms underlying TNFAIP2 regulation in ESCC remain unknown. Hence, immunohistochemical staining with anti-TNFAIP2 antibody was used to examine TNFAIP2 protein expression in 55 paraffin-embedded ESCC specimens which showed high levels in 34 (62%) samples and low levels in 21 (38%) samples, respectively. In contrast, TNFAIP2 was undetectable in the adjacent normal epithelium (Fig. 1A). Moreover, TNFAIP2 mRNA expression was quantified by qRT-PCR, which showed that mRNA levels in 16 (67%) of 24 samples were higher than levels noted in the adjacent normal tissues (Fig. 1B).

The patient characteristics and clinical features shown in Table I were evaluated to identify correlations between TNFAIP2 and clinicopathological characteristics of ESCC patients, which showed that there were no significant correlations between TNFAIP2 expression and age, gender and cellular differentiation. However, there were significant associations between TNFAIP2 expression and T, N and UICC stage. As shown in Fig. 1A, TNFAIP2 expression gradually increased along with the T stage.

Kaplan-Meier survival analysis and the log-rank test were used to investigate potential impacts of TNFAIP2 expression on the DFS of patients. The median follow-up period was 20 months and at the time of the present study 28 patients survived and 27 had died. Kaplan-Meier survival curves showed that DFS was obviously higher with low TNFAIP2 expression than high expression (Fig. 1C). Moreover, univariate Cox regression analysis revealed a correlation between DFS and TNFAIP2 expression (p=0.023; regression coefficient=1.047; hazard ratio=2.849; 95% confidence interval (CI)=1.152–7.044). These results indicate that elevated TNFAIP2 expression may be a risk factor for ESCC-associated mortality.

The results of the present study showed that aberrant TNFAIP2 expression in ESCC clinical specimens was correlated with intratumoral T, N and UICC stage. qRT-PCR and western blotting were used to detect THFAIP2 expression levels in Eca109, Eca9706, TE-10, Kyse150 and Kyse510 cells. As shown in Fig. 2A, Eca109 and Kyse150 were used to study their biological functions influenced by TNFAIP2 in vitro. In this experiment, Eca109, Kyse150, Kyse510 and TE-10 cells were treated with 25 ng/ml of rh-TNFα for 48 h. As shown by the results in Fig. 2B, there were significant differences in TNFAIP2 mRNA and protein expression levels between treated and untreated cells, indicating that TNFAIP2 can be induced and regulated by TNFα.

Cell colony formation, proliferation, cell cycle, apoptosis, migration and invasive ability play significant roles in tumor tumorigenic and metastatic properties. Therefore, lentiviral RNA interference (LV-RNAi) was used to knock down TNFAIP2 in the ESCC cell lines Eca109 and Kyse150 using techniques to study biological functions of cell lines in vitro with an empty lentiviral vector as a negative control (LV-CON). The efficiencies of knockdown and transfection were >75 and 90% at the mRNA and protein levels (Fig. 2C), respectively.

Colony formation was significantly inhibited in the LV-RNAi group, as compared with the LV-CON group (Fig. 3A1 and A2). LV-RNAi TNFAIP2 resulted in obvious decreased proliferation of the Eca109 and Kyse150 cells compared with the LV-CON group (Fig. 3A3 and A4).

A Transwell assay was employed to examine the role of TNFAIP2 in cellular migration and invasion. As shown by the Transwell assay results presented in Fig. 3B1-B4, cell migration and invasive ability were significantly declined in the LV-RNAi group as compared with the LV-CON group. To investigate potential mechanisms underlying LV-RNAi-induced inhibition of colony formation and proliferation, cell cycle distribution and apoptosis of cells in the LV-RNAi and LV-CON groups were detected by flow cytometry, which showed that in the LV-RNAi group, the Eca109 and Kyse150 cells were mainly arrested in the G0/G1 phase and the percentages of cells in the G2 and S phases were decreased (Fig. 3C1-C4), but, there were no significant differences in apoptosis between the LV-RNAi and LV-CON groups (p=0.292). Notably, these findings demonstrated that an abundance of TNFAIP2 plays an oncogenic role in ESCC.

Several pathways were investigated to elucidate the molecular mechanisms underlying LV-RNAi-mediated inhibition of tumorigenic and metastatic properties. The results showed discrepancies between the interrelated members of the Wnt/β-catenin signaling pathway expressed in the LV-RNAi TNFAIP2 group and LV-CON group cells. On the assumption that β-catenin expression could be upregulated by TNFAIP2, ESCC cell lines stably transfected with lentiviral particles carrying TNFAIP2 RNAi were used to identify the underlying mechanisms by western blot and qRT-PCR analyses of the cell lysates. Obviously, when TNFAIP2 expression was downregulated, the protein expression levels of β-catenin were decreased synchronously (Fig. 3D). Moreover, expression levels of p-GSK-3β (upstream of β-catenin), c-Myc, cyclin D1, matrix metalloproteinase-7 (MMP-7), Snail (downstream) and E-cadherin were evaluated by western blot analysis, which indicated that expression of c-Myc, cyclin D1, MMP-7 and Snail decreased, while expression levels of E-cadherin and p-GSK-3β were increased in the LV-RNAi group, as compared with the LV-CON group (Fig. 3D).

To elucidate potential associations between TNFAIP2 and β-catenin, immunohistochemical staining analysis with β-catenin antibody was conducted using 55 paraffin-embedded specimens. As shown in Fig. 4A, β-catenin expression (in cytoplasm or nucleus) was relatively high in 36 (65%) samples and relatively low in 19 (35%). β-catenin expression was found to be upregulated in the cytoplasm of cells from tumors with an advanced T stage. The results of the Pearson Chi-squared test confirmed the correlations between TNFAIP2 and β-catenin expression profiles (Table I).

Based on the western blot results of Snail and E-cadherin, we hypothesized that TNFAIP2 regulated E-cadherin expression via β-catenin. Hence, immunohistochemical staining analysis with E-cadherin antibody was conducted with 55 paraffin-embedded specimens to identify correlations between the expression profiles of TNFAIP2 and E-cadherin. The results showed that low and high E-cadherin expression levels were detected in 37 (67%) and 18 (33%) samples, respectively, as compared with adjacent normal epithelium (Fig. 4B). Pearson Chi-squared test results revealed a correlation between TNFAIP2 and E-cadherin expression levels, as shown in Table I. Moreover, the Pearson Chi-squared test results showed that β-catenin and E-cadherin expression levels were correlated with T stage (p=0.001 and p<0.001, respectively) and immunohistochemical staining analysis (Fig. 4).

Based on these results, we extrapolated that TNFAIP2 operates through the Wnt/β-catenin signaling pathway.