A total of 64 patients who underwent surgical resection at the Fourth Hospital of Hebei Medical University between January 2006 and January 2008 were enrolled in this study. All patients were diagnosed with ESCC based on pathology, and none of the patients received radiotherapy, chemotherapy or biological therapy before surgery. Complete clinical data and follow-up examinations were available for all patients. The study included 48 men and 16 women (median age, 57.5 years; range, 37–78 years). The median follow-up was 31 months, with a range from 7 to 105 months. Tumor-node-metastasis (TNM) stage was determined using criteria of the World Health Organization (WHO) and the International Union Against Cancer (UICC). Tumor masses as well as adjacent normal tissues, which were at least 5 cm distal to tumor margins, were snap-frozen in liquid nitrogen for quantitative real-time reverse transcriptase-PCR (qRT-PCR) assay or fixed in 10% neutral formalin solution for immunohistochemistry. This study protocol was approved by the Ethics Department of the Fourth Hospital of Hebei Medical University (Shijiazhuang, China), and informed consent was obtained from patients or their families.

Human esophageal squamous cell lines Eca109, TE-1, TE-13, KYSE30, and KYSE170 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cell lines were cultured in RPMI-1640 (Gibco, Carlsbad, CA, USA) supplemented with 25 mM HEPES, 2 mM L-glutamine, 10% fetal bovine serum (FBS; Sijiqing, Hangzhou, China), 100 U/ml penicillin and 100 g/ml streptomycin. All cells were incubated at 37°C in a humidified atmosphere with 5% CO₂.

Total RNA from tissue specimens and cultured cells was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. RNA was reverse-transcribed into cDNA using the RevertAid First Strand cDNA Synthesis kit (Thermo Fisher Scientific, Shanghai, China) according to the manufacturer's instructions. Amplification of the generated cDNA was performed on Mx3005P Real-Time PCR Cycler (Stratagene Corp.; Agilent Technologies, Inc., USA) using SYBR-Green PCR Master Mix (Promega, Madison, WI, USA). Levels of mRNA were normalized to those of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The primer sequences used in this study are shown in Table I. The PCR conditions were 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec, 60°C for 30 sec, and 72°C for 30 sec. Verification of specific product amplification was determined by melting curve analysis. Relative expression of TIM-3 mRNA in carcinoma tissues and matched adjacent normal tissues was calculated using the 2^-ΔCt method (17). The fold change of each mRNA was evaluated using the 2^−ΔΔCt method (18).

Formalin-fixed tissues were embedded in paraffin, and sections (4 µm thick) were cut from each paraffin block. Paraffin sections were dewaxed in xylene twice for 15 min each and rehydrated through a graduated series of alcohol solutions (5 min per step). Antigen was retrieved by boiling sections in citrate buffer (pH 6.0) for 5 min in a pressure cooker, and endogenous peroxidase activity was blocked by incubation in 3% H₂O₂ for 20 min. Sections were incubated at 37°C for 60 min in 10% normal goat serum to block non-specific background staining, and incubated overnight at 4°C with rabbit anti-human TIM-3 polyclonal antibody (1:100 dilution; Abcam, Southampton, UK). Subsequently, the sections were incubated with a biotinylated secondary antibody, followed by the streptavidin-peroxidase complex. Chromogen [3,3′-diaminobenzidine (DAB)] was added, after which sections were counterstained with Mayer's hematoxylin. Negative control sections were treated as described, except that primary antibody was omitted.

Two pathologists blinded to clinical information independently assessed the immunohistochemistry results. Five visual fields were randomly selected in each section, and 200 tumor cells were counted in each visual field at ×200 magnification. Cells were regarded as positive if the cell membrane and/or cytoplasm stained brown. The percentage of positive cells and the intensity of staining were evaluated in all sections, and the final score for each section was derived from the multiplication of the two. The scoring system for the percentage of positive cells was as follows: 1 point, ≤33%; 2 points, >33 to ≤66%; 3 points, >66%. The scoring system for staining intensity was as follows: 1 point, absent/weak staining; 2 points, moderate staining; 3 points, strong staining. Sections with a final overall score of ≤3 were classified as the TIM-3 low expression group; other sections were classified as the TIM-3 high expression group (19).

To extract total protein, the cells were collected and lysed in RIPA buffer supplemented with PMSF (Beyotime, Shanghai, China) at 4°C for 30 min. Total protein concentrations were measured using the BCA Protein Assay kit (Beyotime). Equal amounts of protein were separated by 10% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with 5% non-fat dry milk at room temperature for 1 h and incubated at 4°C overnight with the respective antibodies: rabbit anti-GAPDH, TIM-3, matrix metalloproteinase (MMP)-9, tissue inhibitor of metalloproteinase (TIMP)-1, E-cadherin, N-cadherin, vimentin, Snail, Akt, GSK-3β, p-Akt, and p-GSK-3β (Abcam). The membranes were washed with TBST three times for 5 min each, and incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. The bands were quantitated in grayscale using Image-Pro Plus software 6.0 (Media Cybernetics, Silver Spring, MD, USA).

To knock down TIM-3 expression in the Eca109 and TE-1 cells, shRNA targeting TIM-3 or negative control shRNA (Applied BioProbes; GeneCopoeia Co.) was transfected into Eca109 and TE-1 cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. The sequences of the specific shRNA against TIM-3 were as follows: forward, 5′-GGGACTCTAGATTGGCCAATG-3′ and reverse, 5′-TAAGACTACGGCGCGAAGC-3′. Stable trans fected clones were selected for 2 weeks using 3 µg/ml puromycin (Sigma, St. Louis, MO, USA), and monoclonal cells were picked up with tips. Subsequently, the resistant clones were amplified in conventional medium with 3 µg/ml puromycin.

Cell viability was evaluated using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulphophenyl)-2H-tetrazolium (MTS) assay. Cells were seeded into 96-well plates at a density of 4×10³ cells/well for 24, 48 or 72 h. At each time point, 20 µl of MTS solution (Promega) was added to each well and incubated at 37°C for 3 h in the dark. The absorbance of each well was measured at 492 nm using a spectrophotometer (Thermo Fisher Scientific). Each experiment included six replications and was repeated three times.

Cells were seeded at a density of 3×10³ cells/well in 6-well culture plates and cultured at 37°C for 1 week. Colonies were washed with phosphate-buffered saline (PBS), fixed with 4% neutral formalin and stained with 0.1% crystal violet (Sigma). Colonies containing >50 cells were counted. Each experiment was performed in triplicate.

Cell migration and invasion assays were performed in 24-well Transwell cell culture chambers (Corning, Cambridge, MA, USA) containing a polycarbonate membrane filter with 8.0-µm pores. For the cell invasion assay, 30 µl of Matrigel (1:7 dilution; Becton-Dickinson, San Jose, CA, USA) was added to the upper chamber and dried overnight at 37°C in a 5% CO₂ incubator. A total of 5×10⁴ cells were resuspended in 200 µl serum-free RPMI-1640 medium and seeded into the upper chamber. The lower chamber was filled with 600 µl of RPMI-1640 medium containing 20% FBS. After incubation at 37°C with 5% CO₂ for 24 h, non-invading cells still on the upper side of the chamber were wiped with a cotton swab. Invaded cells on the bottom of the polycarbonate membrane were fixed for 20 min with 4% neutral formalin, and stained for 30 min with 0.1% crystal violet. For the cell migration assay, we performed a similar procedure as described above, except that the upper chamber was not coated with Matrigel. Each experiment was performed in triplicate. The number of cells that migrated or invaded through the polycarbonate membrane was counted in five randomly selected visual fields at ×200 magnification using a high-resolution inverted microscope.

Cells were collected and resuspended in 100 µl 1X binding buffer at a density of 1×10⁶ cells/ml. Cell apoptosis was detected using the Annexin V-PE/7-AAD Apoptosis Detection kit (Becton-Dickinson) according to the manufacturer's instructions. The Annexin V-PE and 7-AAD stained cells were examined by flow cytometry on a BD FACScalibur (BD Biosciences).

All statistical analyses were analyzed using SPSS 17.0 (IBM, Chicago, IL, USA), and figures were generated using GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA). Differences in mean values between groups were assessed for significance using the Student's t-test or one-way analysis of variance (ANOVA). Differences in TIM-3 expression between ESCC and matched adjacent normal tissues were assessed using the χ² test. The relationships between TIM-3 expression and clinicopathological features were analyzed using the χ² test or Fisher's exact test. Overall survival was calculated using the Kaplan-Meier method, and the differences between groups were assessed for significance using the log-rank test. Data are presented as means ± SD from at least three independent experiments. P<0.05 was considered statistically significant.

TIM-3 mRNA levels were analyzed by qRT-PCR in 40 pairs of ESCC and matched adjacent normal tissues. The relative mRNA expression of TIM-3 was significantly higher in the tumor tissues than that in the adjacent normal tissues (0.002±0.0007 vs. 0.001±0.0004; P<0.001, Fig. 1A).

TIM-3 protein levels were analyzed by immunohistochemistry in 64 pairs of ESCC and adjacent normal tissue specimens. TIM-3 protein was localized predominantly in the cellular membrane and the cytoplasm of tumor cells (Fig. 1B–D). TIM-3-positive staining (overall score >1) was present in 55 of the 64 ESCC tissue sections (85.9%), compared to only 7 of the 64 normal tissue sections (10.9%). Of the 55 TIM-3-positive ESCC samples, 19 (34.5%) showed low TIM-3 expression (overall score ≤3), and 36 (65.5%) showed high TIM-3 expression (overall score >3), whereas all 7 of the TIM-3-positive normal tissue samples showed low TIM-3 expression. Statistical analysis of these results indicated that TIM-3 protein expression was significantly higher in the ESCC tissues than that in the adjacent normal tissues (P<0.001, Table II).

We analyzed the correlations between TIM-3 protein expression and numerous clinicopathological parameters, including age, gender, tumor size, TNM stage, lymphatic metastasis, and depth of tumor invasion (Table III). Our results revealed that the expression of TIM-3 protein was significantly associated with TNM stage (P=0.008), lymph node metastasis (P=0.042) and depth of tumor invasion (P=0.042). However, no significant differences were observed between the TIM-3 protein expression and the other clinico-pathological parameters.

Patients were divided into two groups (TIM-3 low expression group or TIM-3 high expression group) as described in the 'Materials and methods' section. Kaplan-Meier survival analysis indicated that patients in the TIM-3 high expression group had a significantly shorter overall survival than patients in the TIM-3 low expression group (P=0.001, Fig. 2). The overall 5-year survival rate for patients in the TIM-3 high expression group was 9% with a median survival time of 22 months, while the overall 5-year survival rate for patients in the TIM-3 low expression group was 19.2% with a median survival time of 41 months.

The expression levels of TIM-3 mRNA and protein in various ESCC cell lines, including KYSE30, TE-13, Eca109, KYSE170 and TE-1, were evaluated by qRT-PCR and western blotting, respectively. These results showed that TIM-3 expression was relatively highest in the Eca109 and TE-1 cell lines (Fig. 3A and B). Therefore, we selected these two cell lines to knock down TIM-3 expression and explore the effects on cancer cell function. We established shRNA-mediated stable Eca109-shTIM-3 and TE-1-shTIM-3 cells, and their respective negative controls (Eca109-NC and TE-1-NC cells). Transfection efficiency was evaluated by qRT-PCR and western blotting. The mRNA and protein levels of TIM-3 were markedly decreased in the Eca109-shTIM-3 and TE-1-shTIM-3 cells than levels in their corresponding parental or negative control cells (P<0.001, Fig. 3C–F).

Effects of TIM-3 knockdown on ESCC proliferation were examined by MTS and colony formation assays. In the MTS assay, the absorbance at 492 nm was significantly lower in the Eca109-shTIM-3 and TE-1-shTIM-3 cells than those in their corresponding parental or negative control cells at 24, 48 and 72 h (P<0.01, Fig. 4A and C). Consistent with these results, colony formation potential was significantly decreased in the Eca109-shTIM-3 and TE-1-shTIM-3 cells (P<0.001, Fig. 4B and D). In contrast to its effects on proliferation, TIM-3 knockdown did not significantly affect cell apoptosis based on an Annexin V-PE/7-AAD staining assay (Fig. 4E).

To characterize the effects of TIM-3 knockdown on the migration and invasion of ESCC cells, we performed Transwell migration and Matrigel invasion assays. In the Transwell migration assay, the number of Eca109-shTIM-3 cells which migrated through the Transwell was dramatically decreased compared with the number of Eca109-NC and untransfected Eca109 cells which migrated (P<0.001, Fig. 5A). Similarly, in a Transwell invasion assay, the number of Eca109-shTIM-3 cells which invaded through the Transwell was significantly decreased compared with the number of Eca109-NC and untransfected Eca109 cells which invaded (P<0.001, Fig. 5A). Similar results were obtained in the TE-1 cells (P<0.001, Fig. 5B).

MMPs and their inhibitors, termed TIMPs, have been reported to be closely associated with tumor growth, invasion and metastasis (20,21). Among the MMP members, in particular MMP-9 plays a prominent role in ESCC cell invasion and metastasis (22,23). Therefore, we investigated whether TIM-3 knockdown affects the expression of MMP-9 and TIMP-1 in Eca109 and TE-1 cells. qRT-PCR and western blotting revealed that TIM-3 depletion significantly downregulated MMP-9 and upregulated TIMP-1 in both cell lines (Fig. 6A–C).

Epithelial-mesenchymal transition (EMT) is a key step in the metastatic process of many solid tumors and represents a hallmark of this event (24). We aimed to ascertain whether TIM-3 depletion would inhibit the metastatic potential of ESCC cells in part by inhibiting EMT. Indeed, we found that TIM-3 depletion upregulated the epithelial marker E-cadherin and downregulated the mesenchymal markers N-cadherin and vimentin in both the Eca109 and TE-1 cells (Fig. 6A–C).

Accumulating evidence has demonstrated that the Akt/GSK-3β/Snail signaling pathway can drive EMT to facilitate tumor progression (25,26). To elucidate the potential signaling pathways involved in TIM-3-mediated EMT, we investigated the effects of TIM-3 knockdown on the Akt/GSK-3β/Snail signaling pathway. Western blotting revealed that the protein level of the active form of Akt (phosphorylated on Ser473), but not total Akt, the active form of GSK-3β (phosphorylated on Ser9), but not total GSK-3β were markedly decreased by knockdown of TIM-3 (Fig. 7). In addition, EMT-related transcription factor Snail was downregulated after TIM-3 knockdown (Fig. 7). These findings suggest that TIM-3 depletion acts via the Akt/GSK-3β/Snail signaling pathway to inhibit the EMT in ESCC.