From January 2011 to December 2012, 193 patients (150 males and 43 females) were enrolled at The First Hospital of Jilin University with gastric cancer and positive lymph node metastasis. Patients underwent gastrectomy with D2 lymph node dissection. The diagnosis of was confirmed by histopathological analysis. Patients with distant metastasis, incomplete clinical or pathological data, impaired organ function, other malignant tumors in the previous year, or were undergoing preoperative neoadjuvant chemotherapy/emergency surgery for obstruction and perforation were excluded. A summary of the patient clinical data is presented in Table I. Primary gastric cancer tissues, adjacent non-cancerous gastric tissues, positive metastatic lymph node tissues and corresponding negative metastatic lymph node tissues were obtained, fixed in 10% formalin at room temperature for 48 h, and embedded in paraffin. The present study was approved by the Research Ethics Boards of The First Hospital of Jilin University, and all patients provided informed consent.

Tissue sections (4 µm) cut from the paraffin-embedded samples mentioned above were deparaffinized with xylene and rehydrated with graded xylene and serial ethanol concentrations. For antigen retrieval, the sections were incubated with citrate buffer (0.01 M, pH 6.0) and microwaved at 95°C for 10 min. After rinsing thrice with PBS (pH 7.2), 3% H₂O₂ in methanol was used to block endogenous peroxidase activity at room temperature for 15 min. To reduce non-specific reactions, 5% bovine serum albumin (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 0.3% Triton X-100 were added and incubated with the sections at room temperature for 1 h. Subsequently, the sections were probed with appropriate antibodies against ADAM8 (1:100; cat. no. 23778-1-AP; ProteinTech Group, Inc., Chicago, IL, USA), ADAM9 (1:75; cat. no. PA5-25959; Thermo Fisher Scientific, Inc.), ADAM10 (1:250; cat no. PA5-28161; Thermo Fisher Scientific, Inc.), ADAM12 (1:100; cat. no. 14139-1-AP), ADAM17 (1:300; cat. no. 20259-1-AP) and GAPDH (1:100; cat. no. 10494-1-AP; all ProteinTech Group, Inc.) overnight at 4°C. PBS was used as the negative control. Following three rinses with PBS (pH 7.2), the sections were probed with horseradish peroxidase (HRP)-labeled Peroxidase AffiniPure goat anti-rabbit IgG (1:300; cat. no. 111-035-045) or Peroxidase AffiniPure HRP-labeled goat anti-mouse IgG (1:400; cat. no. 115-035-003; both Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) secondary antibodies for 30 min at room temperature. Following three washes, the sections were stained with 3,3′-diaminobenzidine (cat. no. DA1010; Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) at room temperature for 1 min, counterstained with hematoxylin (cat. no. 517-28-2; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 50 sec, dehydrated with graded concentrations of ethanol and xylene, and coverslipped.

Under a light microscope (magnification, ×40), the immunostaining intensity for each protein was reviewed and independently scored by two pathologists who were blinded to the clinical data and scored independently according to the staining intensity and the proportion of stained tumor cells, as previously described (20) with minor modifications. According to the staining intensity, they were scored as follows: No staining =0; light yellow (weak staining) =1; yellow brown (moderate staining) =2; and brown (strong staining) =3. The scores were expressed in terms of the proportion of cell staining as follows: scores of 0, 1, 2, 3 and 4 indicated 0, ≤25, 26-50, 51-75, and ≥75% positive cells, respectively. Thus, the two combined scores (from the two independent pathologists) were taken as the final score: 0-1, 2-3, 4-5 and 6-7 indicated negat ive (−), weak positive (+), strong positive (++), and very strong positive (+++), respectively. In the statistical analysis, (+ +) and (+ + +) were classified as the positive group, while (−) and (+) were classified as the negative group.

Human gastric cancer cell lines KATO III and AGS were purchased from the China Center for Type Culture Collection (Wuhan University, Wuhan, China). SGC-7901 and BGC-823 cell lines were obtained from the Shanghai Institute of Cell Biology of the Chinese Academy of Science (Shanghai, China). All cells were grown in RPMI-1640 medium (cat. no. 11995500BT) containing 10% fetal bovine serum (cat. no. 10099-141; both Gibco; Thermo Fisher Scientific, Inc.) and a mixture of penicillin (100 U/ml) and streptomycin (100 µg/ml) at 37°C until 80-90% confluence.

The coding sequence of ADAM17 was synthesized by Genewiz, Inc. (Suzhou, China). The overexpressed vector pcDNA-3-ADAM17 was then constructed by inserting the coding sequence of ADAM17 into the pcDNA-3 vector (Invitrogen; Thermo Fisher Scientific, Inc.) with the restriction sites KpnI (GGTACC) and EcoRI (GAATTC). The empty pcDNA-3 vector served as the control. Small interfering RNAs (siRNAs) targeting ADAM17 (siRNA-ADAM17) and negative control siRNAs (siRNA-NC) were synthesized by Shanghai GenePharma Co., Ltd. (Shanghai, China). The sequences were as follows: siRNA-ADAM17 forward, 5′-GGG CCG AAU AUA ACA UAG AdTdT-3′ and reverse, 5′-UCU AUG UUA UAU UCG GCC CdTdT-3′; siRNA-NC forward, 5′-UUC UCC GAA CGU GUC ACG UdTdT-3′ and reverse, 5′-ACG UGA CAC GUU CGG AGA AdTdT-3′.

These siRNAs were designed using BLOCK-iT™ RNAi Designer (Thermo Fisher Scientific, Inc.). For cell transfection, SGC-7901 and BGC-823 cells (2.4×10⁵/well) were seeded into a six-well plate and incubated for 24 h at 37°C. Next, 2 µg pcDNA-3-ADAM17 and 2 µg pcDNA-3 were transfected into BGC-823 cells using Lipofectamine^® 2000 reagent (cat. no. 11668-027; Invitrogen; Thermo Fisher Scientific, Inc.). si-RNA-ADAM17 (50 nM) BS siRNA-NC (50 nM) were transfected into SGC-7901 cells using the same method. Cells were incubated for another 48 h at 37°C before performing further experiments.

Total mRNA was extracted from transfected cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Reverse transcription of mRNA to cDNA was performed using the PrimeScript RT Master Mix kit (cat. no. RR036A; Takara Bio, Inc., Otsu, Japan) with the following conditions: 37°C for 15 sec, 85°C for 5 sec; then held at 4°C. To detect the expression levels of ADAM17 mRNA, qPCR was performed using a SYBR Green kit (cat. no. 4367659; Invitrogen) on an ABI ViiA7 PCR system (both Thermo Fisher Scientific, Inc.). Primers used for the amplification of targets were as follows: ADAM17 forward, 5′-ATC AAA CCC TTT CCT GCG-3′ and reverse, 5′-CAA ACC CAT CCT CGT CCA-3′; GAPDH forward, 5′-TGA CAA CTT TGG TAT CGT GGA AGG-3′ and reverse, 5′-AGG CAG GGA TGA TGT TCT GGA GAG-3′. The thermocycling conditions were as follows: 95°C for 10 min, followed by 40 cycles at 95°C for 15 sec and 60°C for 1 min, and finally 95°C for 15 sec, 60°C for 1 min, and 95°C for 15 sec. The relative expression of ADAM17 mRNA was normalized to GAPDH and was quantified using the 2^−ΔΔCq method (25).

Cells (5×10⁶/well) were seeded in a 96-well plate and incubated for 24 h and were subsequently transfected with pcDNA-3-ADAM17, pcDNA-3, siRNA-ADAM17 or siRNA-NC. Each treatment was performed in triplicate wells. After transfection for 48 h, cell viability of each group was assessed using a CCK-8 kit (cat. no. CK04, Dojindo Molecular Technologies, Inc., Kumamoto, Japan) following the manufacturers’ recommended protocols. Absorbance at 450 nm was measured on a microplate reader (BioTek Instruments, Inc., Winooski, VT, USA).

The cell migratory capacity of each group was evaluated using the scratch wound healing assay, as described previously (26). Briefly, transfected cells were seeded in culture dishes in triplicate and were incubated at 37°C until a confluent monolayer was formed (>90%). With a 10 µl sterile pipette tip, a ‘scratch’ of the cell monolayer was created in a straight line. The cells were washed thrice with PBS (pH 7.2) to remove cell fragments, and the cells were incubated with serum-free RPMI-1640 medium for 24 h at 37°C. The migrated cells were observed under an inverted light microscope (Nikon Corporation, Tokyo, Japan).

GSEA (www.broadin-stitute.org/gsea/index.jsp ) (27) is a computational method for interpreting gene expression data; it creates a molecular signature database based on known information on the positions, characteristics and biological functions of genes. To investigate the role of ADAM17 in gastric cancer, mRNA sequence data of stomach adenocarcinoma from The Cancer Genome Atlas (TCGA; tcga-data.nci.nih.gov/) was downloaded, which included 415 stomach adenocarcinoma samples. Key pathways that were enriched by gene sets significantly associated with ADAM17 were then analyzed by GSEA, and P<0.05 was set as the cut-off value.

Total protein was extracted from gastric cancer tissues and cells using radioimmunoprecipitation assay buffer (cat. no. PL007; Sangon Biotech Co., Ltd., Shanghai, China) for 20 min on ice. Protein concentration was determined with bicinchoninic acid protein assay kit (cat no. 23223; Pierce; Thermo Fisher Scientific, Inc.). For western blotting, protein extracts (30 µg/lane) were subjected to 10% SDS-PAGE and transferred onto a polyvinylidene fluoride membrane. Following blocking in 1X Tris-buffered saline/0.05% Tween-20 (TBST) supplemented with 5% non-fat milk at room temperature for 1 h, primary antibodies against ADAM17 (1:1,000; cat. no. 20259-1-AP; ProteinTech Group, Inc.), ADAM9 (1:1,000; cat. no. PA5-25959; Thermo Fisher Scientific, Inc.), neurogenic locus notch homolog protein (Notch)2 (1:1,000; cat. no. WL02409; Wanleibio Co., Ltd., Shanghai, China); glycogen synthase kinase (GSK)-3β (1:1,000; cat. no. 22104-1-AP; ProteinTech Group, Inc.), β-catenin (1:1,000; cat. no. 8480S; Cell Signaling Technology, Inc., Danvers, MA, USA) or β-actin (1:2,000; cat. no. 60008-1-Ig; ProteinTech Group, Inc.) were added to the membranes and incubated overnight at 4°C. β-actin was used as the internal control. After washing with 1X PBST thrice, HRP-labeled Peroxidase AffiniPure goat anti-rabbit IgG (1:5,000; cat. no. 111-035-045; Jackson ImmunoResearch Laboratories, Inc.) or Peroxidase AffiniPure goat anti-mouse IgG HRP (1:5,000; cat. no. 115-035-003; Jackson ImmunoResearch Laboratories, Inc.) was added to the membranes and incubated at room temperature for another 2 h. After rinsing five times with 1X PBST, protein bands were visualized using chemiluminescent HRP substrate (cat. no. WBKLS0500; Merck KGaA, Darmstadt, Germany). The intensity of each protein band was quantified and analyzed using Tanon Image Software version 1.10 (Tanon 1600R; Tanon Science and Technology Co., Ltd., Shanghai, China).

Univariate survival analysis for the comparison of survival times was conducted using the Kaplan-Meier method, and the log-rank test was applied to analyze the significance of difference between survival times of patients and their clinicopathological features, as well as between survival times and ADAM9 or ADAM17 expression. Multivariate survival analysis was performed for identifying significant variables associated with survival times using the Cox regression proportional hazards model. In addition, the risk scores of death were calculated with multivariate logistic regression analysis (28) According to the T staging, N staging and ADAM17 values, the risk score of each patient (ranging from 0 to 100) was calculated. The receiver operating characteristic (ROC) curve was applied, and the area under the curve (AUC) was determined by the MedCalc statistical software version 11.0 (MedCalc, Mariakerke, Belgium). Data were presented as the mean ± standard deviation, and the differences between two groups were compared using the Student’s t-test and among three or more groups were analyzed by one-way analysis of variance, followed by Fisher’s least significant difference test. Statistical analyses were carried out using SPSS 17.0 (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

Fig. 1 shows the immunohistochemical staining intensity for ADAM8, ADAM9, ADAM10, ADAM12 and ADAM17 in adjacent non-cancerous gastric tissues, gastric cancer tissues, positive metastatic lymph node tissues and negative metastatic lymph node tissues. From the combined score that was obtained from the staining intensity and the percentage of cell staining, the expression of ADAM9, ADAM10, and ADAM17 in gastric tumor tissues was significantly upregu-lated compared to those in adjacent normal tissues (P<0.05; Fig. 2A). Furthermore, the expression of ADAM8, ADAM9 and ADAM17 in positive metastatic lymph node tissues was also upregulated relative to those in the corresponding negative tissues (P<0.05; Fig. 2B). These data indicated that ADAM9 and ADAM17 were significantly upregulated in both gastric cancer and positive metastatic lymph node tissues. To further verify the immunohistochemistry results, the expression of ADAM9 and ADAM17 was detected by western blotting. As expected, ADAM9 and ADAM17 expression was significantly upregulated in primary gastric tumor tissues and positive metastatic lymph node tissues (P<0.05; Fig. 3).

Univariate survival analysis for the comparison of survival times between patients with clinicopathological features was performed. As presented in Table I, the survival times of patients were significantly associated with vascular invasion, neural invasion, T staging, N staging and ADAM17 expression (P<0.05). Therefore, multivariate analysis was used to identify significant clini-copathological features associated with survival times. The results demonstrated that the survival times of patients strongly correlated with T staging, N staging and ADAM17 expression (P<0.05; Table II). Furthermore, multivariate logistic regression analysis showed similar results that the above three variables were independent predictors (P<0.05; Table III). The above three variables were then included in the ‘risk score’ calculation to measure the ‘mortality risk score’ of any given patients, as follows: Probability =1/[1+exp [5.454-0.993 (T staging)-0.720 (N staging)-0.771 (ADAM17)]].

The distribution of the risk score of 193 patients is presented in Fig. 4. The majority of scores were distributed between 60 and 100 (60-80, n=89; 80-100, n=43; Fig. 4A), suggesting that this model can separate the low- and high-risk groups, and the risk of mortality increased with the increase in risk scores (Fig. 4B). To further determine the optimal model to predict the risk of death, ROC curve analysis was performed. As shown in Table IV and Fig. 4C, the AUCs for the risk score model, T staging, N staging, and ADAM17 expression were 0.757, 0.625, 0.720 and 0.618, respectively. These data indicated that T staging, N staging and ADAM17 expression had independent prognostic value for predicting the risk of patients with gastric cancer.

Next, the expression of ADAM7 in gastric cancer cells was determined. As shown in Fig. 5A, ADAM17 expression in KATO III and SGC-7901 cells that had high metastatic potential was significantly higher than that in BGC-823 and AGS cells with low metastatic potential, suggesting that ADAM17 may be associated with tumor cell metastasis. Among the above four cell lines, ADAM17 expression was lowest in BGC-823 cells and highest in SGC-7901 cells. Therefore, BGC-823 and SGC-7901 cells were used for the overexpression and repression of ADAM17 expression, respectively. As expected, ADAM17 expression was significantly increased in pcDNA-3-ADAM17-transfected BGC-823 cells, compared with mock cells or pcDNA-3-transfected cells (P<0.05; Fig. 5B), while its expression was markedly reduced in siRNA-ADAM17-trans-fected SGC-7901 cells compared with either mock cells or siRNA-NC-transfected cells (P<0.05; Fig. 5C).

The effects of ADAM17 on cell viability in each group was assessed using the CCK-8 assay. As presented in Fig. 6A, the overexpression of ADAM17 in pcDNA-3-ADAM17-transfected BGC-823 cells resulted in a significant increase in cell viability, compared with mock cells or pcDNA-3-transfected cells (P<0.05). However, the cell viability of siRNA-ADAM17-transfected SGC-7901 cells was significantly reduced compared with the mock or siRNA-NC transfected cells (P<0.05; Fig. 6B). Next, the scratch wound healing assay was performed to study the effects of ADAM17 on cell migration. It was revealed that the number of migrated SGC-7901 cells was significantly decreased following ADAM17 silencing (P<0.05; Fig. 6D). However, the migratory capacity between pcDNA-3-ADAM17-transfected BGC-823 cells, mock cells and pcDNA-3-transfected cells did not show significant difference (P>0.05; Fig. 6C). Taken together, these results suggested that ADAM17 promoted gastric cancer cell viability and migration.

The results of GSEA revealed 12 pathways, including the Notch and Wnt signaling pathways (Fig. 7), that positively correlated with ADAM17 expression (Table V). Furthermore, 19 pathways associated with metabolism, including oxidative phosphorylation and phenylalanine metabolism, were negatively correlated with ADAM17 expression (Table V).

To further verify the results of GSEA, western blot analysis was performed to investigate the expression of key proteins involving in Notch and Wnt signaling pathways in SGC-7901 cells following the suppression of ADAM17 expression. The results demonstrated that suppression of ADAM17 in SGC-7901 cells resulted in a significant downregulation of Notch2, GSK-3β and β-catenin expression (Fig. 8), suggesting that suppression of ADAM17 in SGC-7901 cells inhibited the Notch and/or Wnt signaling pathways.