We collected tissues specimens of 60 gastric cancer patients from consecutive surgical cases in Department of Surgical Oncology, The First Affiliated Hospital, Xi'an Jiaotong University between 2004 and 2009. The patients included 37 male and 23 female patients (ranging from 45 to 68 years of age). All of the patients were assessed according to the system for staging primary tumor/regional lymph nodes/distant metastasis (TNM) described in the AJCC Cancer Staging Manual. None of these 60 patients received neoadjuvant or adjuvant chemotherapy before the operation. Paraffin-embedded tumor specimens were carefully collected immediately after the surgery. The clinicopathological characteristics of the patients are summarized in Table I. This study was approved by the Protection of Human Subjects Committee of First Affiliated Hospital, Xi'an Jiaotong University and complies with the Helsinki declaration. The BGC-823, MKN-45, SGC-7901, AGS, MGC-803, and GES1 cell lines used were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA).

The tissues specimens were fixed in neutral buffered formalin and embedded in paraffin wax. The 4-mm sections were cut and mounted on charged glass slides. Antigen retrieval was performed using citrate buffer at pH 6.0. Immunohistochemical staining was performed with rabbit anti-human FOXQ1 (bs-16175R, Beijing Bioss Biotechnology, Beijing, China). The streptavidin-peroxidase technique (SP-9001 Golden Bridge Int., Beijing, China) was used. An irrelevant rabbit antiserum served as a negative control. The sections were stained with 0.02% diaminobenzidine (DAB) solution followed by counterstaining with hematoxylin.

The evaluation of FOXQ1 expression was performed independently by two experienced pathologists who were blinded to the clinical data with consensus. Sections were observed under a light microscope (Carl Zeiss Axio Scope A1 microscope) at high magnification (×400). The staining results of FOXQ1 were scored semi-quantitatively by calculating the immunostaining intensity and the percentage of positive malignant cells. The percentage of positive malignant cells was determined in at least 5 areas under ×400 magnifications and averaged. The mean percentage was scored as: 0 (0–5%); 1 (6–25%); 2 (26–50%); 3 (51–75%), and 4 (76–100%). The staining intensities were scored as follows: no coloring, 0 point; slightly yellow, 1 point; brownish-yellow, 2 points; and tan, 3 points. Finally, the staining score was obtained by calculating the product of the staining intensity and the positive cell percentage, where ≤5 was defined as low expression and ≥6 as high expression.

The FOXQ1 shRNA lentiviral particle containing FOXQ1 shRNA sequences was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA), MKN-45 cells were infected with shFOXQ1 lentiviral particles and a negative control for 48 h and followed by 2 mg/ml puromycin selection. Full-length FOXQ1 cDNA was cloned into pCMV6/AC/GFP vector (OriGene, Rockville, MD, USA). Cell lines were separately transfected with plasmids and selected by GFP sorting. Cells were then grown in complete medium containing 200 µg/ml G418 (Roche Diagnostics, Mannheim, Germany). qRT-PCR and western blotting was used to confirm the presence of the plasmids.

Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcribed into cDNA using the Reverse Transcription kit from Takara. After adjusting the cDNA concentration in all groups, qRT-PCR was performed using the CFX96 Real-time PCR Detection System (Bio-Rad, Hercules, CA, USA) with SYBR Green. The PCR conditions were as follows: pre-denaturation at 95°C for 30 sec; 35 cycles of denaturation (95°C for 5 sec), annealing (55–60°C for 30 sec) and extension (72°C for 1 min); final extension at 72°C for 10 min. The relative level of gene expression was calculated using the ∆∆Ct method with normalization to GAPDH. All experiments were performed in triplicate. The primers used are listed in Table II.

Protein expression levels were analyzed by western blotting standard protocols. Briefly, 20 µl of total protein extracts was resolved by denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% non-fat milk and then incubated with primary antibodies against FOXQ1 (ab51340; Abcam, Cambridge, MA, USA), E-cadherin (AF0131; Affinity, USA), vimentin (5741; Cell Signaling Technology, Inc., Beverly, MA, USA) and β-actin (T0022; Affinity). The blots were washed and probed with the respective secondary peroxidase-conjugated antibodies. Signals were detected using the chemiluminescence solvent (Thermo Scientific, Rockford, IL, USA).

Cell invasion was measured using Transwell chambers (Millipore, Billerica, USA) coated with Matrigel. Cells from different groups were suspended in serum free RPMI-1640 and were added into the upper compartment of the chamber; the bottom chamber was filled with RPMI-1640 containing 10% FBS as a chemoattractant. After incubation in a 5% CO₂ humidified chamber at 37°C for 24 h, the cells were removed from the upper surface of the filter with a cotton swab. The invaded cells were then fixed and stained using 0.1% crystal violet. The cells were quantified from five different fields under a light microscope. The experiment was repeated in triplicate.

Statistical analysis was done using the SPSS software package (version 16.0, SPSS Institute) or Prism (GraphPad Software, Inc., La Jolla, CA, USA). The association between staining index and other categorical factors potentially predictive of prognosis was analyzed using the Chi-square test and Fisher's exact test. Overall survival (OS) was defined as the time from the date of surgery to the date of last follow-up or death from any cause. Survival curve and median survival were estimated by the Kaplan-Meier method. Their differences were verified by log-rank test. Differences between groups were assessed using an unpaired, two-tailed Student's t-test; P<0.05 was considered significant.

FOXQ1 protein level was examined in 60 gastric cancer tissues and their corresponding normal gastric epithelium tissues using immunohistochemistry. The clinicopathological characteristics of the gastric cancer patients are described in Table I. Respective photomicrographs of immunohistochemical staining of FOXQ1 are shown in Fig. 1. Increased FOXQ1 expression was detected in 37 of the 60 tumor tissue samples (61.7%) and in 14 (23.9%) of the 60 adjacent matched tumor tissues P<0.05. High-expression of FOXQ1 was closely related to the histological differentiation, pTNM stage, and lymphatic metastasis, but not to the age, gender, smoking, or alcohol consumption of the patients.

To investigate the prognostic significance of FOXQ1 in GC, the correlation between FOXQ1 expression and overall survival time was assessed by Kaplan-Meier analysis along with log-rank test. Kaplan-Meier analysis showed that patients with higher FOXQ1 expression had significantly poorer prognosis than those with lower FOXQ1 expression (mean overall survival 32.0 vs. 43.0 months, P=0.002, log-rank test; Fig. 2).

To determine if expression of FOXQ1 is increased in gastric cancer cell lines compared with human immortalized gastric epithelial cell line, we investigated the protein and mRNA level of FOXQ1 in five human gastric cancer cell lines (BGC-823, MKN-45, SGC-7901, AGS, MGC-803) and one human immortalized gastric epithelial cell line (GES-1). Western blot and qRT-PCR results showed higher expression of FOXQ1 in MKN-45 and AGS cell lines than BGC-823, SGC-7901, and MGC-803 cell lines compared to GES-1 cell line (Fig. 3). MKN-45 and BGC-823, with the highest and the lowest expression levels of FOXQ1, were chose for further investigation.

Overexpressing FOXQ1 expressed higher levels of mesenchymal marker, vimentin and lower levels of epithelial marker E-cadherin compared with their vector control at both protein and mRNA levels (Fig. 4A and B). To investigate whether the FOXQ1-induced EMT-like phenotype could be translated into enhanced metastatic ability of the BGC-823 cells, the invasion of BGC-823-FOXQ1 cells was tested. FOXQ1-overexpressing BGC-823 cells displayed increased motility through the Matrigel in Transwell invasion assays, compared with the vector control cells (Fig. 4C and D).

FOXQ1 was knocked down by shRNA in the high metastatic potential cell line MKN-45, which expresses high levels of FOXQ1. Knockdown of FOXQ1 in the cell line resulted in a significant decrease at both protein and mRNA level. Protein and mRNA expression of the mesenchymal marker vimentin was also decreased. The level of epithelial marker E-cadherin was increased (Fig. 5A and B). Furthermore, knockdown of FOXQ1 in MKN-45 cells resulted in decreased ability of the cells to invade through the Matrigel in the invasion assay (Fig. 5C and D).

To elucidate the mechanism of FOXQ1 that leads to EMT and invasiveness, we hypothesized that FOXQ1 regulates the expression of Snail. Ectopic FOXQ1 expression greatly increased Snail at both the mRNA and the protein levels (Fig. 6A and B). Conversely, knockdown of FOXQ1 reduced Snail expression (Fig. 6C and D).