Fibroblastic galectin-1-fostered invasion and metastasis are mediated by TGF-β1-induced epithelial-mesenchymal transition in gastric cancer

CAFs overexpresses GAL-1/ LGALS1 and promotes lymph node metastasis in GC

To confirm that LGALS1 is associated with the malignant behavior of GC, quantitative real-time reverse transcription PCR (qRT-PCR) was performed initially to assess LGALS1 mRNA expression in 15 matched gastric cancer tissues (GCT) and non- gastric cancer tissues (NGCT). LGALS1 mRNA expression was significantly different between NGCT and GCT: GCT exhibited significantly higher levels of LGALS1 mRNA than that in NGCT (P < 0.01; Figure 1A). We then examined GAL-1 protein levels in 15 pairs of GCT and NGCT using western blotting (WB), the results of which were consistent with the qRT-PCR results, demonstrating higher levels of GAL-1 in GCT than in NGCT (P < 0.01; Figure 1B). Besides, the expression of GAL-1 was detected in GCT and in adjacent NGCT of 127 patients with GC using immunohistochemistry (IHC). Image Pro Plus (Media cybernetics, San Diego, CA, USA) was used to evaluated the digital images of IHC. GAL-1 median IHC scores in GCT and NGCT were 78.29 (9.51–186.24) and 31.09 (5.89–123.45), respectively. GAL-1 expression was significantly different between GCT and NGCT (Figure 1C, 1D; P < 0.01), and GAL-1 was mainly expressed by stroma tissue of GC (Figure 1C). To further confirm which cells produced GAL-1/ LGALS1, we used IHC to detect alpha smooth muscle actin (SMA-α) in GCT from 127 patients with GC. We found that GAL-1 was expressed in CAFs with a SMA-α protein positive phenotype (Figure 1C). Moreover, we found that GC cells expressed low or no GAL-1 (Figure 1C). Surprisingly, we found that the GAL-1 was highly expressed in GC cells from metastatic lymph node tissues when detected using IHC (Figure 1E). In GCT, SMA-α expression correlated positively with GAL-1 expression (r = 0.963, P < 0.01; Figure 1F), and residual analysis showed that GAL-1/SMA-α proteins fitted the regression model hypothesis (Supplementary Figure 1A). We classified the GAL-1 IHC scores as positive and negative using Receiver Operating Curve (ROC) statistics, and observed a significantly higher lymph node metastasis rate in the GAL-1 positive group (64/86) compared with that in the GAL-1 negative group (7/41) (P < 0.01; Figure 1G).

Figure 1. GAL-1/LGALS1 is overexpressed in CAFs and promotes lymph node metastasis in GC tissues. (A) GCT exhibited significantly higher levels of the LGALS1 mRNA than that in NGCT. (B) GAL-1 is overexpressed in GC tissues. (C–E) Representative images of IHC for GAL-1 and SMA-α protein levels in GCT, NGCT, and metastatic LN. (F) The IHC score of GAL-1 correlated positively with the IHC score of SMA-α in GC tissues (r = 0.963; P < 0.01). (G) The lymph node metastasis rate of the GAL-1-positive group was significantly higher than that in the GAL-1-negative group (P < 0.01).

CAFs express GAL-1/LGALS1 and promote GC cell line metastasis and invasion abilities in vitro

To further clarify the origin of GAL-1/LGALS1, we cultured CAFs from human GC tissue, and the GC cell lines (SGC-7901, AGS, BGC-823, and MGC-803) in vitro and investigated the GAL-1 levels in all cells using WB, which confirmed strong GAL-1 expression in CAFs, but lower levels in GC cells (Figure 2A). Meanwhile, the results of qRT-PCR analysis of LGALS1 mRNA expression in all the cell lines were consistent with the WB results (Figure 2B). MGC-803 cells and SGC-7901 were then treated with conditioned medium (CM) from CAFs (CM-CAFs) for 72 h, and then GAL-1 and LGALS1 mRNA levels were monitored using WB and qRT-PCR. Compared with untreated MGC-803 and SGC-7901 cells, CM-CAFs treatment of SGC-7901 and MGC-803 cells resulted in significant increases in the GAL-1 protein and LGALS1 mRNA levels (Figure 2C, 2D). SGC-7901 and MGC-803 cell proliferation increased after CM-CAFs treatment for 48 h, and the proliferation effect was abolished when the medium contained 10 μg/mL mitomycin C (Figure 2E). CM-CAFs treatment of SGC-7901 and MGC-803 cells resulted in significantly increased migration abilities compared with those of the wild-type (untreated) control (Figure 2F); the fold changes in migration are shown in Figure 2G (P < 0.01). Transwell assays showed that MGC-803 and SGC-7901 cells treated with CM-CAFs had increased cell invasion and migration abilities (P < 0.01, Figure 2H, 2I).

Figure 2. CAFs expression of GAL-1/ LGALS1 promotes the invasion and metastasis ability of GC cell lines in vitro. (A, B) GAL-1 protein and LGALS1 mRNA levels in CAFs and GC cell lines, GAL-1/LGALS1 levels were high in CAFs. (C, D) Treatment of MGC-803 and SGC-7901 cells with CM-CAFs significantly increased GAL-1 protein levels and LGALS1 mRNA expression. (E) CM-CAFs increased the proliferation of MGC-803 and SGC-7901 cells, and the proliferation effect was abolished when the medium contained 10 μg/mL mitomycin C. (F, G) MGC-803 and SGC-7901 cells treated with CM-CAFs exhibited a significantly enhanced migration capacity compared with the wild-type control (P < 0.01). Magnification: ×40. (H, I) Transwell assay showing that MGC-803 and SGC-7901 cells treated with CM-CAFs had increased cell invasion and migration abilities (P < 0.01). Magnification: ×200.

GAL-1/LGALS1 promotes EMT in GC, and EMT promotes lymph node metastasis of GC

To determine whether GAL-1/LGALS1 promotes the migration and invasion of GC cells in an EMT-dependent manner, IHC of GC tissue was used to examine the EMT-related biomarkers Vimentin and E-Cadherin. In 85 cases, GC cells stained positive for E-cadherin and stromal cells stained positive for Vimentin, indicating that EMT had not occurred (Figure 3A). In 42 cases, GC and stromal cells stained positive for Vimentin, and E-cadherin levels were reduced, indicating that EMT had occurred (Figure 3B). The EMT rate in the GAL-1 positive group was 41.86% (36/86), while that in the GAL-1 negative group was 14.63% (6/41), suggesting that GAL-1 promotes EMT significantly in GC (P < 0.05, Figure 3C). In addition, we found that GC cases with EMT were more prone to lymph node metastasis: Among 71 patients with GC had lymph node metastasis, 30 with EMT (42.25%, 30/71), while there were 12 cases with EMT among 56 cases of GC hadn’t lymph node metastasis (21.43%, 12/56) (P < 0.05, Figure 3D). GC cells with or without EMT in metastatic lymph nodes are shown in Figure 3E, 3F.

Figure 3. GAL-1/LGALS1 promotes EMT in GC, and EMT promotes lymph node metastasis of GC. (A) Representative images of GCT without EMT. (B) Representative images of GCT with EMT. (C) GAL-1/LGALS1 significantly promoted EMT in GC (P < 0.05). (D) GC with EMT was more prone to lymph node metastasis (P < 0.05). (E, F) Representative images of GC cells with or without EMT in metastatic lymph nodes.

GAL-1/LGALS1 activates TGF-β/Smad signaling pathways in GC tissues

To investigate how GAL-1/LGALS1 promotes EMT in GC mechanistically, we examined the protein levels of GAL-1, TGF-β1, and phosphorylated (p)-Smad2/3 in GCT and NGCT using IHC. Representative IHC images of IHC for p-Smad2/3, TGF-β1, and GAL-1, protein levels in GCT and NGCT are shown in Figure 4A. GCT and NGCT showed significant differences in p-Smad2/3 and TGF-β1 levels (P < 0.01; Figure 4B, 4C). For TGF-β1, the median IHC score for was 108.26 (8.741–234.56) in GC tissues, 127.22 (56.67–234.56) in GAL-1-positive GC tissues, and 43.47 (8.741–97.12) in GAL-1-negative GCT. In GAL-1-positive GCT, the IHC scores for TGF-β1 were significantly higher than those in GAL-1-negative GCT (P <0.01; Figure 4D). In GC tissues, the TGF-β1 IHC scores correlated positively with those of GAL-1 (r = 0.97; P < 0.01; Figure 4E). Residual analysis demonstrated that GAL-1/TGF-β1 proteins levels fitted the regression model hypothesis (Supplementary Figure 1B).

Figure 4. GAL-1/LGALS1 activates TGF-β/Smad signaling pathways in GC tissues. (A) Representative images of IHC for GAL-1, TGF-β1, and p-Smad2/3 protein levels in GCT and NGCT. Magnification: ×400. (B, C) Significant differences in TGF-β1 and p-Smad2/3 levels were observed between GCT and NGCT (all P < 0.01). (D) The TGF-β1 IHC scores in GAL-1-positive GCT were significantly higher than those in GAL-1-negative GCT (P < 0.01). (E) The TGF-β1 IHC scores correlate positively with the GAL-1 IHC scores in GC tissues (r = 0.97; P < 0.01).

GAL-1/LGALS1 induces EMT via TGF-β/Smad pathways in vitro

To identify if GAL-1 activates Smad/TGF-β signaling to induce EMT in GC, SGC-7901 and MGC-803 cells were transfected using LV- LGALS1-OE or LV-LGALS1-RNAi to overexpress or silence LGALS1, respectively, as described previously [16]. Western blotting showed increased levels of Vimentin and decreased levels of E-cadherin in MGC-803 cells overexpressing LGALS1 (OE-LGALS1) compared with those in non-transfected MGC-803 cells (wild-type), and the levels of TGF-β1 and p-Smad2/3 also increased in MGC-803 OE-LGALS1 cells (all P < 0.01, Figure 5A). The results in SGC-7901 cells were consistent with those in MGC-803 cells (all P < 0.01, Figure 5B). In MGC-803 cells silenced for LGALS1, E-cadherin levels increased, and Vimentin, TGF-β1, and p-Smad2/3 levels decreased (all P < 0.01, Figure 5C). SGC-7901 cells showed similar results (all P < 0.01, Figure 5D).

Figure 5. GAL-1/ LGALS1 induces EMT through TGF-β/Smad signaling pathways in vitro. (A, B) WB showing that OE-LGALS1 efficiently increased the expression of TGF-β1 and p-Smad2/3 and induced EMT in MGC-803 and SGC-7901 cells compared with the wild-type cells (all P < 0.01). (C, D) Silencing LGALS1 in MGC-803 and SGA-7901 cells efficiently decreased the levels of TGF-β1 and p-Smad2/3 and inhibited EMT (all P < 0.01).

In GC cells, GAL-1/LGALS1 promotes invasion and migration in vitro via the TGF-β/Smad pathways

To investigate whether GAL-1/LGALS1 promotes GC invasion and migration by activating TGF-β/Smad signaling, ITD1, a specific antagonist of TGF-β/Smad signaling was employed to investigate the relationship between the induction of GC cell migration and invasion by GAL-1 and TGF-β/Smad pathway activation. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays suggested that neither ITD1 at 10 μM nor the empty cassette of the viral vector system affected the cell number at 48 h. OE-LGALS1 increased the number of cells; however, the proliferation effect was abolished using 10 μg/mL mitomycin C. Compared with the negative control-transfected MGC-803 cells (OE-con) and wild-type control (WC) cells, OE-LGALS1 MGC-803 cells showed significantly increased migration (P < 0.01, Figure 6A). However, this migration capacity was abolished when the medium contained 10 μM ITD1(P < 0.01, Figure 6A). The fold changes in migration are shown in Figure 6B.

Figure 6. GAL-1/ LGALS1 promotes the migration and invasion of GC cells in vitro through TGF-β/Smad signaling pathways. (A–D) OE-LGALS1 significantly enhanced the migration capacity of MGC-803 and SGC-7901 cells compared with WC and OE-con. The migration capacity was abolished when the medium contained 10 μM ITD1 (P < 0.01). Magnification: ×40. (E) Transwell assay showing that MGC-803 and SGC-7901 cells increased their invasive ability after transfection with LV-LGALS1-OE, and 10μM ITD1 abolished this increase in invasive ability (n = 3). Magnification: ×200.

To confirm GAL-1-mediated promotion of migration in vitro, these experiments were repeated in SGC-7901 cells. The migration of OE-LGALS1 SGC-7901 cells was enhanced compared with that of the WC and OE-con SGC-7901 groups, and 10 μM ITD1 abolished this migration capacity (Figure 6C, 6D). As shown in Figure 6E, Transwell assays demonstrated that LV-LGALS1-OE-transfected SGC-7901 and MGC-803 cells had increased invasiveness, and the addition of 10 μM ITD1 abolished this capacity for invasion.

GAL-1/LGALS1 promotes GC cell metastasis via TGF-β/Smad signaling in vivo

In vitro experiments and clinical analysis showed that GAL-1/LGALS1 promotes EMT-mediated enhancement of GC metastasis, invasion, and migration via TGF-β/Smad signaling. We established lung metastasis and subcutaneous GC implantation models in athymic mice (n = 6 per group) to further determine whether GAL-1/LGALS1 promotes GC growth and metastasis through EMT induced by activating TGF-β/Smad signaling. At 21 days after subcutaneous implantation of GC cells, in the LGALS1 overexpression group (OE-LGALS1), the tumors were larger and heavier compared with those in the WC group. Significantly smaller and lighter tumors were formed by OE-LGALS1 MGC-803 cells treated with ITD1 compared with those formed in the WC group (Figure 7A–7C; P < 0.01). After day 12, the tumors in the OE-LGALS1 group had a significantly larger volume compared with those in the WC group, and the tumor volumes in the OE-LGALS1 MGC-803 cells treated with ITD1 were significantly lower than those in the WC group (P < 0.01 and P < 0.05; Figure 7B). According to the western blotting results, Vimentin levels in the OE-LGALS1 group were increased, and E-cadherin levels were decreased, which suggested that LGALS1 overexpression promoted EMT. Compared with those in the WC group, TGF-β1 and p-Smad2/3 levels were elevated in the OE-LGALS1 group (P < 0.01, Figure 7D). However, ITD1 treatment reduced the levels of p-Smad2/3 significantly (P < 0.01, Figure 7D), and inhibited EMT in the implanted GC cells (Figure 7D).

Figure 7. GAL-1/ LGALS1 promotes GC cell metastasis in vivo through the TGF-β/Smad signaling pathway. (A) OE-LGALS1 induced MGC-803 to form subcutaneous xenograft tumors with larger volumes (B) and weighs (C) (expressed as the mean ± SE). * P <0.05, ** P< 0.01, n = 6. (D) OE-LGALS1 increased the levels of TGF-β1 and p-Smad2/3, and induced EMT in the subcutaneous xenograft tumor, ITD1 could inhibit this effect. Metastases were frequent in the (E). (F) Representative images of metastasis (yellow arrows) in the lungs at 50 days after inoculation, and representative images of H&E staining. Original magnification: ×400. (G) Immunostaining showing GAL-1, E-cadherin, vimentin, TGF-β1 and p-Smad2/3 levels in pulmonary metastases. Magnification: ×400.

In all the mice from the OE-LGALS1 group, larger pulmonary metastases were observed after fifty days in the lung metastasis models. Four cases of pulmonary metastasis were found in the WC group, while no pulmonary metastases were observed in the ITD1-treated OE-LGALS1 group (P = 0.002; Figure 7E). Hematoxylin and eosin (H&E) staining was used to identify pulmonary metastases in all groups (Figure 7F). IHC staining of TGF-β/Smad markers in pulmonary metastases, showed significantly increased GAL-1, Vimentin, TGF-β1, and p-Smad 2/3 levels and decreased E-cadherin levels in the pulmonary metastatic tissues from the OE-LGALS1 group (Figure 7G). Our results suggested an important role of GAL-1 in GC metastasis and EMT, and the TGF-β/Smad signaling pathway potentially contributes to this process.

Patients

We enrolled 127 patients with primary gastric adenocarcinoma into the present study, of whom 89 were male and 38 were female, and their average age was 64.26 ± 9.98(range: 38–87) years. None of the patients had received neoadjuvant radiotherapy or chemotherapy, and all patients underwent radical gastrectomy at the Department of Gastrointestinal Surgery, Taizhou Clinical Medical School of Nanjing Medical University (Taizhou People’s Hospital) of Jiangsu province.

Tumor and samples

For IHC, GC tissues and their matching adjacent gastric mucosa tissues were formalin-fixed and paraffin-embedded. For molecular analysis, GC tissues and their matching adjacent mucosal tissues were collected from 15 patients, and stored in liquid nitrogen. The Clinical Research Ethics Committee of Taizhou People’s Hospital (TZRY-EC-12-068) approved this study. All patients were provided with details regarding the assessment procedure, and all patients provided written, informed consent.

Cell lines and their culture

SGC-7901 and MGC-803 (human gastric adenocarcinoma cell lines) were obtained from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in Roswell Park Memorial Institute (RPMI) medium (Thermo Fisher Scientific, Waltham, MA, USA) containing 1% (V/V) penicillin and streptomycin (Gibco, Grand Island, NY, USA), and 10% (V/V) fetal bovine serum (FBS; Thermo Fisher Scientific). The cells were cultured in a humidified atmosphere with 5% (V/V) CO₂ at 37° C [25]. The cells were passaged via trypsinization when they reached 80% confluence.

Lentiviral transduction

Lentiviral transduction was caried out according to a previously published protocol [25]. Genechem Co. Ltd (Shanghai, China) constructed the lentiviral vectors to overexpress and silence LGALS1. The short hairpin RNA (shRNA) sequences and the lentiviral vector were designed in a previous report [16]. SGC-7901 or MGC-803 cells were seeded at 5 × 10⁴ cells per well in 6-well plates before lentiviral transduction. Cells were transduced with the lentiviral vectors and 10 μg/mL polybrene (Sigma-Aldrich, St. Louis, MO, USA) at a multiplicity of infection of 10. At 12 h after transfection, the medium was replaced. To select stably transduced cell lines Puromycin (2 μg/mL; Sigma-Aldrich) was added. After 48 h, medium containing puromycin at 0.5 μg/mL was used to culture the stably transduced cells. After 72 h, a fluorescence microscope (OLYMPUS-U-HGLGPS-IX73, Olympus, Tokyo, Japan) was used to assess the transduction efficiency, which was further confirmed using qRT-PCR and WB.

Reagents and antibodies

ITD-1, a TGF-β signaling pathway inhibitor, was purchased from Macklin (1884570, Shanghai, China). Abcam provided the anti-GAL-1 antibody (ab138513, Cambridge, UK); Cell Signaling Technology provided the anti-SMA-α antibody (56856, Danvers, MA, USA); anti-E-cadherin (bs-10009R), anti-Vimentin (bs-0756R), anti-TGF-β1 (bs-0086R), and anti-p-Smad2/3(bs-8853R) antibodies were purchased from Bioss (Beijing, China); Santa Cruz Biotechnology (Santa Cruz, CA, USA) provided the anti-glyceraldehyde phosphate dehydrogenase (GAPDH) antibody (sc-47724), the goat anti-rabbit IgG (sc-2357), the horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (sc-516102); Sigma Biotechnology (St. Louis, MO, USA) provided the MTT assay kit and dimethyl sulfoxide (DMSO).

RNA extraction and qRT-PCR

Total RNA was extracted using an RNeasy Mini Kit (Invitrogen, Waltham, MA, USA), and cDNA was synthesized from RNA using a Reverse transcription kit (Takara, Shiga, Japan). The qPCR reactions were carried out using a SYBR Green dye kit (Roche Diagnostics, Mannheim, Germany) and the reaction products were analyzed using an iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). The following thermocycling conditions were used: 95° C for 30 s; 40 cycles of 95° C for 5 s, 60° C for 30 s, and 72° C for 30 s. The reference control gene was GAPDH. The following primers were used: LGALS1 (forward): GCTGAACCTGGGCAAAGACAG an GTGGCGGTTGGGGAACTT (producing a 247 bp amplicon); and GAPDH (forward) TGACTTCAACAGCGACACCCA) and (reverse) CACCCTGTTGCTGTAGCCAAA (producing a 121 bp amplicon).

Western blotting

An extraction kit (Beyotime, Shanghai, China) was used to prepare total cell and nuclear extracts. 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate 20 μg of the extracts, followed by transfer of the separated proteins to a nitrocellulose membrane (GE Healthcare Life Sciences, Pittsburgh, PA, USA). The proteins on the blots were probed with antibodies against GAL-1, E-cadherin, Vimentin, TGF-β1, p-Smad2/3, and GAPDH(dilution = 1:2000). Goat anti-rabbit immunoglobulin conjugated with HRP comprised the secondary antibody (dilution= 1:2000). A West Pico chemiluminescent Substrate (Pierce, Carlsbad, CA, USA) was used to visualize the immunoreactive protein bands, which were quantified using densitometric image analysis software (Image Master VDS; Pharmacia Biotech, Little Chalfont, UK). As an internal reference, GAPDH level were determined. All experiments were carried out three times independently.

Histological examination and IHC evaluation

IHC was used to detect proteins in murine lung tissues, human GC tissues, and matching adjacent non-GC tissues according to our previous reports [14, 25]. Primary antibodies recognizing SMA-α (1:200), GAL-1 (1:200), E-cadherin (1:150), vimentin (1:150), TGF-β1 (1:150), and p-Smad2/3 (1:150) were incubated with the slides. Subsequent steps and the staining scores of the proteins of interest (GAL-1, SMA-α, p-Smad2/3, and TGF-β1) were the same as in our previous report [25]. Vimentin and E-cadherin staining was assessed as positive or negative according to the evaluation of the staining results as defined in our previous report [25].

Wound-healing assay

GC cells were cultured to a confluent monolayer (80–90 %) in a 6-well plate, and then a wound was scored across the cell surface using a sterile plastic tip. Phosphate-buffered saline (PBS) was used to wash the plates three times to remove cellular debris. The plates were incubated with serum-free medium containing mitomycin C at 10 μg/mL to block proliferation. At 0 and 48 h, the wound was photographed. The assays were performed three times independently.

Cell viability assay

The MTT assay was used to assess cell viability. Cells (5 × 10³ cells/well) were seeded into the well of a 96-well flat bottom plate and cultured overnight at 37° C in a 5% CO₂ humidified atmosphere. The cells were then treated with ITD1 at various concentrations (5, 10, 15, and 20 μM). After 48 h, MTT was added at 20 μL per well and incubated for 4 h. Thereafter, the supernatant was removed, and the formazan crystals were dissolved using 150 μL of DMSO. A microplate reader (Model 550; Bio-Rad) was used to measure the absorbance at 490 nm.

Cell invasion and migration assays

GC cell invasive and migratory abilities were assessed using 24-well Transwell units with polycarbonate filters (pore size, 8.0 μm; Corning, Corning, NY, USA). First, the upper Transwell inserts were coated with 100 μl of Matrigel basement membrane (BD Biosciences, San Diego, CA, USA) or left uncoated, and then added with 100 μL of serum-free RPMI medium. Then, 1× 10 ⁵ cells were seeded on the upper Transwell inserts. Medium (600 μL) containing 10% FBS as a chemoattractant was added to the lower chamber. For 24 h at 37° C, the cells were allowed to migrate or invade from the upper chamber, after which the non-migratory or non-invasive cells were removed. 4% paraformaldehyde was used to fix the filters, and a 0.05% crystal violet solution was used to stain the cells. In each sample, six fields were selected randomly to count the cells under a microscope (magnification = 100×). All assays were conducted three times independently.

Animal models

The Comparative Medicine Centre of Yangzhou University (Yang Zhou, JiangSu, China) provided 5-week-old male athymic mic, which were used to construct the lung metastasis model and subcutaneous GC implantation model of GC. The Ethics Committee of Yang Zhou University approved the animal experiments (YZU-EC-JS2352). The mice were bred under pathogen-free conditions in a laminar flow cabinet. LGALS1-overexpressing (OE-LGALS1) MGC-803 cells, OE-LGALS1 MGC-803 treated with ITD1, and MGC-803 cells (wild-type control) were inoculated separately into the right flank or the tail vein of the athymic mice at 2 × 10⁶ cells per mouse (n = 6). In the subcutaneous model group, the mice were sacrificed on day 21, whereas in the lung metastasis model group, the mice were sacrificed on day 50. We harvested the subcutaneous GC tumors or lung for hematoxylin and eosin (H&E) staining, WB, and immunohistochemical staining.

Statistical considerations

The statistical analyses were conducted using SPSS 20.0 (IBM Corp, Armonk, NY, USA). Means ± standard error were used to express the continuous variables. Based on the normal distribution of the data, comparisons between groups were performed using one-way analysis of variance (ANOVA) and Dunnett’s t test. To determine the relationship between two variables, Spearman’s or Pearson correlation coefficients were used. Residual analysis was used to determine whether the actual data were consistent with regression model assumptions or not. Values of P < 0.05 were regarded as statistically significant.

Ethics approval and consent to participate

This clinical study was approved by the Clinical Research Ethics Committee of Taizhou People’s Hospital (TZRY-EC-12-068). All patients consented to participate in our study. The animal experiments were approved by the Ethics Committee of Yang Zhou University (YZU-EC-JS2352).