The present study was approved by the Ethics Committee of Tianjin Nankai Hospital (Tianjin, China). Clinical data were collected between July 2016 and July 2017, and a total of 150 patients were recruited to the study from Tianjin Nankai Hospital. Written informed consent was obtained from all patients. Clinical information, including age, sex, differentiation grade and tumor size, was obtained (Table I). Tissue samples of gastric cancer and para-tumor tissue were excised from the patients with gastric cancer during resection. Tissue samples were fixed immediately in 4% neutral-buffered formalin at room temperature for 24 h and subsequently processed to prepare paraffin-embedded sections (1×1×0.5 cm).

FOXK2 expression data in gastric cancer were collated from the Human Protein Atlas (15).

Among the samples, 22 tumors were classified as well differentiated, 102 as moderately differentiated and 26 as poorly differentiated. Formalin-fixed tissue samples were prepared prior to immunohistochemical staining. Samples were stained using the avidin-biotin complex method. Primary antibodies specific for FOXK2 (1:500; cat. no. ab50946; Abcam, Cambridge, MA, USA) were incubated with sections overnight at 4°C. The samples were then incubated with a biotin-conjugated goat anti-rabbit immunoglobulin G secondary antibody (1:100; cat. no. TA130016; OriGene Technologies, Inc., Beijing, China) at 37°C for 1 h. The expression of FOXK2 was detected by coloration with DAB, and the procedure was performed as previously described (16). Staining intensity was scored as follows with an inverted microscope (Olympus Corporation, Tokyo, Japan): 0, negative; 1, weakly positive; 2, moderately positive and 3, strongly positive. The percentage of FOXK2-positive cells was therefore scored as 0 (0%), 1 (1–25%), 2 (26–50%) or 3 (>50%).

The human gastric cancer cell line BGC-823 was obtained from the China Academia Sinica Cell Repository (Shanghai, China). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Hyclone; GE Healthcare Life Sciences, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) under 5% CO₂ at 37°C. FOXK2 (cat. no. ab50946; Abcam), cleaved caspase-3 (cat. no. ab2302; Abcam), E-cadherin (cat. no. ab15148; Abcam) and N-cadherin (cat. no. ab18203; Abcam) antibodies were purchased from Abcam. GAPDH antibody (cat. no. TA802519) was purchased from OriGene Technologies, Inc.

For transient cell transfection, gastric cancer cells (1.5×10⁶ cells/well) were seeded in 6-well plates and cultured overnight, and subsequently transfected with FOXK2 plasmids encoding human FOXK2 or empty vector. The FOXK2 overexpression plasmid (plasmid no. S57120) and empty vector were purchased from GenScript (Nanjing, China). FOXK2 small interfering (si)RNA (si-FOXK2; 50 nM; sense, 5′-GAGTTCGAGTATCTGATGA-3 and antisense, 5′-GCGAACACGTACACTGTCT-3′) and negative control siRNA (50 nM; cat. no. siN05815122147-1-5) were purchased from Guangzhou RiboBio Co., Ltd. (Guangzhou, China) and transfected with X-tremeGENE™ 9 DNA transfection reagent (cat. no. 06365787001; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) and X-tremeGENE™ siRNA transfection reagent (cat. no. 04476093001; Sigma-Aldrich; Merck KGaA). Plasmid transfection (1 µg/well) was performed with Lipofectamine^® 3000 (cat. no. L300001; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Cells were collected and seeded for assays 48 h post-transfection.

Antibodies against FOXK2, cleaved caspase-3, N-cadherin, E-cadherin and GAPDH were used for western blot analysis, as described previously (14). The cells were washed with PBS three times and protein was extracted in radioimmunoprecipitation assay lysis buffer (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China) with 1% phenylmethanesulfonyl fluoride. Protein concentration was determined with a bicinchoninic acid protein assay. Proteins (40 µg/lane) were separated by 10% SDS-PAGE. The separated proteins were transferred to polyvinylidene fluoride membranes and blocked in 5% bovine serum albumin (Beijing Solarbio Science & Technology Co., Ltd.) at 37°C for 1 h. Following this, membranes were incubated with primary antibodies against FOXK2, cleaved caspase-3, N-cadherin, E-cadherin and GAPDH (1:1,000 dilution) at 4°C for 12 h, and subsequent incubation with horseradish peroxidase-conjugated secondary antibody (1:2,000 dilution; cat. no. ZDR-5306; OriGene Technologies, Inc., Beijing, China) at room temperature for 1 h. Protein expression was visualized using with an enhanced chemiluminescence kit (cat. no. WBKLS0500; Merck KGaA). Bands were analyzed with ImageJ software (version 1.51j8; National Institutes of Health, Bethesda, MD, USA).

Total RNA was isolated with TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Equal amounts of RNA were converted into cDNA with a PrimeScript RT reagent kit (Promega Corporation, Madison, WI, USA) according to manufacturer's protocol. Relative expression levels of FOXK2 were determined by PCR with a GoTaq^® Real-Time PCR system (Promega Corporation). The primer sequences for FOXK2 were as follows: Forward, 5′-AAGAACGGGGTATTCGTGGAC-3′ and reverse, 5′-CTCGGGAACCTGAATGTGC-3′. The reference gene was GAPDH, and the primer sequences for GAPDH were as follows: Forward, 5′-ACAACTTTGGTATCGTGGAAGG-3′ and reverse, 5′-GCCATCACGCCACAGTTTC-3′. The PCR conditions were as follows: 30 cycles of 94°C for 30 sec, 56°C for 30 sec, 72°C for 90 sec, and a final extension at 72°C for 5 min. Quantification was performed using the 2^−∆∆Cq method (17).

Colony formation and Cell Counting Kit (CCK)-8 assays. Tumor cells transfected with si-FOXK2 or FOXK2 plasmid were cultured at 2,000 cells/well in 6-well plates. The cells were allowed to grow for 14 days, and the medium was changed every 3 days. Colonies were subsequently fixed with 4% paraformaldehyde for 1 h at room temperature, stained with 0.5% crystal violet for 10 min at room temperature and counted under an inverted microscope (Olympus Corporation).

For assessment of cell proliferation, a CCK-8 assay (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) was performed according to the manufacturer's protocol. To determine the effect of FOXK2 on cell proliferation, tumor cells were transfected with si-FOXK2 or FOXK2 plasmid, seeded into a 96-well plate at a density of 2×10³ cells in 100 µl culture medium containing 10% FBS and cultured overnight. All experiments were performed in triplicate. The medium was subsequently replaced with 100 µl fresh medium containing 10% CCK-8 reagent, and cells were incubated for 3.5 h at 37°C. At 0, 24, 48, 72, 96 and 120 h, absorbance was measured at 450 nm using an ELx800 microplate reader (BioTek Instruments, Inc., Winooski, VT, USA).

To determine the effect of FOXK2 on cell proliferation, cells were also assessed with a colony formation assay. Tumor cells transfected with si-FOXK2 or FOXK2 plasmid were seeded into each well of a 12-well plate at a density of 5×10⁴ cells in 2 ml culture medium containing 10% FBS, and were cultured overnight. All experiments were performed in triplicate. The cells were incubated at 37°C. At 0, 24, 48, 72, 96 and 120 h, the cells were counted following trypsin enzyme digestion.

Annexin V/propidium iodide (PI) staining was performed to quantify cell apoptosis. Transfected cells (8×10⁶) in the logarithmic growth phase were collected and subjected to Annexin V/PI staining using an Annexin V-fluorescein isothiocyanate (FITC) Apoptosis Detection kit (BioVision, Inc., Milpitas, CA, USA) according to the manufacturer's protocol. The resulting fluorescence was measured by flow cytometry using a FACS flow cytometer (BD Biosciences, San Jose, CA, USA). Obtained data were analyzed via BD Cell Quest Pro™ software (version 5.1; BD Biosciences).

The ability of cells to invade was assessed using Matrigel-coated Transwell membranes (BD Biosciences). After 30 min incubation at 37°C, the Matrigel solidified and served as an extracellular matrix for tumor cell invasion analysis. The upper chamber contained ~5×10⁴ cells in 200 µl DMEM without serum. The lower chamber contained DMEM with 10% FBS. The cells were then incubated for 48 h at 37°C in an atmosphere containing 5% CO₂. Cell invasion to the underside of the Matrigel-coated membrane was subsequently fixed with 4% paraformaldehyde for 10 min at room temperature, stained with 0.5% crystal violet for 10 min at room temperature followed by imaging and counting under an inverted microscope at ×200 magnification (Olympus Corporation). The results are expressed as the average number of invasive cells per field.

BGC-823 cells were transfected with si-FOXK2 or FOXK2 plasmid at 37°C for 48 h, seeded into 6-well plates (5×10⁶ cells/well), cultured in serum-free medium and a straight wound was created using a pipette tip. Cells were further incubated at 37°C in an atmosphere containing 5% CO₂ for 24 h. Images of wound healing were captured under an inverted microscope at ×100 magnification (Olympus Corporation, Tokyo, Japan). The percentage of wound closure was the rate of migration distance compare to the control group. Results were analyzed with ImageJ software (version 1.51j8; National Institutes of Health).

Data are presented as the means ± standard deviation of three independent experiments. Statistical analyses were performed in SPSS version 16.0 (SPSS, Inc., Chicago, IL, USA). Two groups were compared using Student's t-test, whereas multiple groups were compared using one-way analysis of variance followed by the Student-Newman-Keuls method. Survival curves were plotted using the Kaplan-Meier method, and the differences between the survival curves were examined by the log-rank test. Cox proportional hazards models were used to identify factors with an independent influence on survival (18–20). P<0.05 was considered to indicate a statistically significant difference.

A total of 150 patients were examined (male:female, 1:1), and the associations between FOXK2 expression and clinicopathological characteristics were investigated. Among the samples, 22 tumors were well differentiated, 102 cases were moderately differentiated and 26 were poorly differentiated. It was demonstrated that the expression of FOXK2 was positively correlated with tumor differentiation (P<0.05; Table I). There was no significant association between FOXK2 expression with sex and age (P>0.05; Table I). Cox regression analyses revealed a significant association between overall survival and tumor size, FOXK2 expression (P<0.001; Table II). The prognostic value of FOXK2 expression with regards to the OS and PFS of patients was determined by survival analysis. The results revealed that high FOXK2 expression was associated with an improved prognosis (Fig. 1A).

Kaplan-Meier survival analysis was performed to analyze the association between FOXK2 mRNA expression and patient survival. Data on gastric cancer were obtained from the Pathology Atlas in The Human Protein Atlas (http://www.proteinatlas.org/). The results revealed that a high level of FOXK2 expression was associated with improved prognosis (P=0.0224; Fig. 1B).

FOXK2 expression was reduced in patients with high-grade gastric cancer. The proportion of FOXK2-positive cells in the intense, moderately and poorly differentiated gastric cancer tissues was 88.3±4.7, 53.7±6.5 and 29.7±9.5%, respectively (P<0.05; Fig. 1C). Compared with the para-tumor tissue, the relative mRNA expression levels of FOXK2 were significantly lower in the tumor tissue (Fig. 1D).

To investigate the functions of FOXK2, FOXK2 was downregulated in BGC-823 cells using siRNA, and FOXK2 was upregulated using a FOXK2 plasmid. Alterations in the expression of FOXK2 were verified by RT-qPCR and western blotting. Western blot analysis revealed that increased FOXK2 expression reduced the upregulation of cleaved caspase-3 (Fig. 2A-D).

The effects of FOXK2 on apoptosis were subsequently determined by western blot analysis and flow cytometry. The percentage of apoptotic cells in the si-FOXK2 group (5.03±0.44%) was significantly reduced compared with in the negative control (NC) group (7.56±0.34%; P<0.01; Fig. 2E). Furthermore, the percentage of apoptotic cells in the FOXK2 group (13.4±0.89%) was significantly increased compared with in the vector group (7.9±0.64%; P<0.01; Fig. 2F). Cells in the lower right quadrant were early apoptotic cells, and the upper right quadrant was designated as late apoptotic cells. These results indicated that upregulation of FOXK2 expression induced early apoptosis, whereas FOXK2 expression in gastric cancer cells had no effect on late apoptosis.

To investigate the effects of FOXK2 on gastric cancer cell migration and invasion, western blotting, Transwell and wound-healing assays were conducted. Western blot analysis revealed that FOXK2 knockdown increased the expression of N-cadherin and decreased the expression of E-cadherin (P<0.05; Fig. 3A). Conversely, FOXK2 upregulation decreased the expression of N-cadherin and increased the expression of E-cadherin (P<0.01; Fig. 3B). Transwell assays were conducted in order to investigate the effects of FOXK2 on gastric cancer invasion. The results revealed that the invasion rate was increased by 54.7% in BGC-823 cells transfected with si-FOXK2 (P<0.01; Fig. 3C). The invasion rate was decreased by 44.8% following FOXK2 plasmid transfection (P<0.05; Fig. 3D).

FOXK2 regulates gastric cancer cell proliferation and migration. To clarify the role of FOXK2 in gastric cancer, colony formation, CCK-8 and cell proliferation assays were performed to examine BGC-823 cell proliferation. The expression levels of FOXK2 in BGC-823 gastric cancer cells were downregulated by si-FOXK2 and upregulated by FOXK2 plasmid transfection. Colony formation, CCK-8 and cell proliferation assays demonstrated that si-FOXK2 significantly increased the growth of gastric cancer cells compared with the NC group (P<0.05; Fig. 4A), whereas FOXK2 plasmid transfection inhibited BGC-23 cell growth compared with the vector group (P<0.05; Fig. 4B). Average colony numbers in the NC and si-FOXK2 groups were 24.7±5.7 and 46±3.2, respectively (P<0.05). Average colony numbers in the vector and FOXK2 groups were 34±7.55 and 16.7±3.1, respectively (P<0.05; Fig. 4A and B).

Wound-healing assays were conducted to investigate the function of FOXK2 in gastric cancer cell migration. In BGC-823 cells, the migration rate was increased by 32% following si-FOXK2 transfection compared with the NC group (Fig. 4C). Migration rate was decreased by 47.3% following FOXK2 plasmid transfection (P<0.05). Taken together, these results indicated that FOXK2 upregulation inhibited the migration of gastric cancer cells (P<0.05; Fig. 4C and D).