Human gastric cancer cell lines SGC-7901 and BCG823 were obtained from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). Cells were cultured in RPMI-1640 medium (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) with 10% fetal bovine serum (FBS) (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C in a 5% CO₂ incubator. GSK3β inhibitor IM-12 (10, 20 and 50 µM) and Rac1 inhibitor NSC23766 (25, 50 and 100 µM) were obtained from Merck KGaA (Darmstadt, Germany). Cells treated with inhibitors were cultured in a 37°C incubator at 5% CO₂.

Cell proliferation was assessed using an MTT (Sigma-Aldrich; Merck KGaA) assay. A total of 2×10³ cells in 100 µl culture medium were plated in a 96-well plate. MTT reagent (5 mg/ml) was added into each well for 48 h and the plate was returned to a 37°C incubator for 3 h. The culture medium was aspirated and 150 µl dimethyl sulfoxide (DMSO) was added into each well. The plate was shaken in an orbital shaker for 15 min. The absorbance at an optical density of 590 nm was measured using a microplate reader. GSK3β inhibitor IM-12 (10, 20 and 50 µM) and Rac1 inhibitor NSC23766 (25, 50 and 100 µM) were added 24 h after the cells (SGC-7901/vector and SGC-7901/sFRP1) were plated. Control groups were treated using DMSO.

sFRP1 vector was purchased from OriGene Technologies, Inc. (Rockville, MD, USA). The green fluorescent protein-fused wild-type (WT) Rac1, constitutively active (CA) mutant Rac1 (Q61L), dominant-negative (DN) mutant Rac1 (T17N) (27), Tag5Amyc-GSK3β WT (28), pCS2 Flag Smad3 S204A (29), pCMV5B-Flag-Smad3 (30) and pCMV5 Smad2-HA (31) were purchased from Addgene, Inc. (Cambridge, MA, USA). Top-flash luciferase plasmid (BPS Bioscience, San Diego, CA, USA) is a luciferase reporter plasmid that contains two sets of 3 copies of the wild-type T-cell factor (TCF) binding regions. If the canonical Wnt signaling is activated, β-catenin will translocate to the nucleus to associate with TCF/lymphoid enhancer factor transcription factors to activate transcription of Wnt target genes. pSV-β-Galactosidase control vector was used as an internal control for transfection and was purchased from Promega Corporation (Madison, WI, USA). For plasmid transfection, SGC-7901/vector cells were seeded into 6-well plate and allowed to grow 24 h prior to transfection. Cells in each well were transfected with 4 µg plasmid using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) for 6 h at 37°C and the medium was changed subsequent to transfection.

The transcriptional activity of β-catenin was measured by co-transfection with Top-flash luciferase plasmid and sFRP1 vector (OriGene Technologies, Inc.) or the control vector using Lipofectamine^® 2000 for 6 h at 37°C. TOPFlash encoding the LEF/TCF binding sites (insert gene) linked to firefly luciferase and reflecting Wnt/β-catenin signaling activity was used. After 24 h incubation, the luciferase activity was measured and normalized to β-galactosidase activity (Promega Corporation). The Luciferase Reporter Gene Detection kit (Promega Corporation) and GloMax^®-Multi+ Detection system (Promega Corporation) were used according to the manufacturer's protocol. The data presented were the mean value of three independent experiments.

The activation of Rac1 was measured using a Rac1 Activation Assay Biochem kit (Cytoskeleton, Inc., Denver, CO, USA) according to the manufacturer's protocol. Briefly, cell lysates were collected using the lysis buffer at 4°C for 30 min from the kit. The activated forms of Rac1 were combined by Rac1 activated kinase (PAK)-Rac/Cdc42 (p21) binding domain (PBD) affinity beads. The beads were centrifuged at 5,000 × g at 4°C for 1 min and activated GTPases were pulled-down into the bead pallets. Bound GTPases were eluted by SDS buffer and analyzed by 12% SDS-PAGE and western blotting. Rac1 levels were analyzed by the specific antibodies.

Whole cell lysates (SGC-7901/vector, SGC-7901/sFRP1, inhibitor treated cells and plasmid-transfected cells) were harvested using radio immunoprecipitation assay cell lysis buffer supplemented with a protease inhibitor cocktail (Sigma-Aldrich; Merck KGaA) at 4°C for 30 min. The nuclear and cytosol extracts were isolated using the nuclear extract kit (Active Motif, Carlsbad, CA, USA) according to the manufacturer's protocol. The concentration of the proteins were measured using a DC Protein assay kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA) according to the manufacturer's protocol. A total of 30 µg protein was loaded per lane and loaded into 10–12% SDS-PAGE gels. The proteins were then transferred to a polyvinylidene fluoride membranes followed by blocking with 5% bovine serum albumin (BSA) for 2 h at room temperature on a rocket shaker. Membranes were washed using tris-buffered saline with 0.1% Tween-20 (pH 8.0) three times for 10 min each. Primary antibodies (1:1,000) against sFRP1 (cat no. ab126613), phosphorylated (p-)Smad3L (cat no. ab63402), zinc finger E-box binding homeobox 2 (ZEB2; cat no. ab138222) and lamin A/C (cat no. ab108922) were purchased from Abcam (Cambridge, UK). Antibodies against Smad3 (cat no. 9523), Smad2 (cat no. 5339), Smad4, p-Smad3c (cat no. 9520), p-Smad2c (cat no. 3108), GSK3β (cat no. 12456), p-GSK3β Ser9 (cat no. 9323), p21 (cat no. 2947), β-catenin (cat no. 8480), p-Rac1/cell division cycle 42 S71 (cat no. 2461) (all obtained from Cell Signaling Technology, Inc., Danvers, MA, USA) and inhibitor of DNA binding 1 (ID1; cat no. 5559-1), Vav guanine nucleotide exchange factor 2 (VAV2; cat no. EP1067Y), plasminogen activator inhibitor 1 (PAI1; cat no. EPR21850-82) (all obtained from Epitomics, Burlingame, CA, USA) were used at 1:1,000 dilutions. Antibodies against GAPDH (cat no. G8795), Lamin A/C (cat no. SAB4200236) and Lamin C (cat no. MAB3540) (Sigma-Aldrich; Merck KGaA) were used at a 1:5,000 dilution. Primary antibodies were diluted in 5% BSA and incubated overnight at 4°C. Horseradish peroxidase (HRP)-conjugated goat anti-rabbit immunoglobulin G (IgG) H&L (cat no. ab6789) and HRP-conjugated goat anti-mouse IgG H&L (cat no. ab6721) secondary antibodies were purchased from Abcam and used at a 1:10,000 dilution in 5% BSA at room temperature for 2 h. Signals were visualized using enhanced chemiluminescence reagent (Pierce; Thermo Fisher Scientific, Inc.). Membranes were scanned using the ChemiDoc Touch Imaging system (Bio-Rad Laboratories, Inc.) and the images were captured using Image Lab Touch Software (version 1.0.0.15; Bio-Rad Laboratories, Inc.).

Cells were fixed with 4% formaldehyde at room temperature for 30 min and then permeabilized with PBS containing 0.2% Triton X-100 for 15 min at room temperature. Slides were blocked using 5% bovine serum albumin at room temperature for 1 h. For F-actin staining, cells were fixed with 4% formaldehyde at room temperature for 30 min and then incubated with Alexa Fluor 555^® phalloidin (Invitrogen; Thermo Fisher Scientific, Inc.) at room temperature for 30 min. The nucleus was counterstained using DAPI at room temperature for 5 min. Slides were washed using PBS, mounted and observed under a microscope. Immunofluorescence staining was visualized and captured using a Nikon Digital Sight DS-U2 (Nikon Corporation, Tokyo, Japan) and NIS elements F3.0 software was used (Nikon Corporation). Confocal images were obtained using an inverted ZEISS LSM710 confocal microscope (×40 oil lens; Carl Zeiss AG, Oberkochen, Germany). Zen 2009 Light Edition (Carl Zeiss AG) was used for measurement of the images.

Cell migration was analyzed using a Transwell chamber assay. A 24-well plate with 8 µm pore size inserts was used. SGC-7901/vector and SGC-7901/sFRP1 cells treated with vehicle (DMSO), NSC23766 (25 µM) and IM-12 (10 µM) for 24 h were used. A total of 1×10⁴ cells were mixed in 100 µl RPMI-1640 medium and added to the upper chamber of the Transwell insert. A total of 600 µl RPMI 1640-medium with 10% FBS was added to the lower chamber. The Transwell inserts were then placed into the wells of a 24-well plate. After 12 h incubation at 37°C, the cells were fixed using 4% formaldehyde at room temperature for 30 min and stained by 0.01% (v/v in methanol) crystal violet at room temperature for 10 min. Subsequent to being washed three times by H₂O, the cells on the inner side of the chamber were removed with a cotton swab and the cells on the outer side of the chamber were counted under a light microscope at a magnification of ×100. Cells were visualized using an Olympus BX50 microscope, images were captured using Nikon Digital Sight DS-U2 and NIS elements F3.0 software was used for analysis.

Total RNA was isolated from cultured SGC7901/vector and SGC-7901/sFRP1 cells using the RNeasy mini kit (Qiagen, Inc., Valencia, CA, USA) according to the manufacturer's protocols, and cDNA was synthesized with oligo (dT) primers by using of a SuperScript first-strand cDNA synthesis kit (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. A total of 1 µg RNA was used to synthesize cDNA. Gene expression was assessed by RT-qPCR using an Applied Biosystems 7500 Fast Sequence Detection System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The PCR reaction mixture consisted of QuantiTect SYBR Green PCR master mix (2× QuantiTect SYBR Green kit, containing HotStart Taq^® DNA polymerase, QuantiTect SYGB Green PCR buffer, dNTP mix, SYGB I, Rox passive reference dye and 5 mM MgCl₂; Qiagen, Inc.), 0.5 µmol/l of each primer and cDNA. The thermocycling conditions were as follows: 95°C for 30 sec, 40 cycles at 95°C for 5 sec, 60°C for 30 sec; and the dissociation stage at 95°C for 15 sec, 60°C for 1 min and 95°C for 15 sec. The transcript of the housekeeping gene, GAPDH was used as an endogenous control to normalize the expression data. The comparative Cq method was used to calculate the relative changes in gene expression. Expression fold change was calculated using the equation 2 ^{− (Cq gene - Cq GAPDH)} (32). The primers used were as follows: PAI1 forward, 5′-TGGCACGGTGGCCTCCTCAT-3′ and reverse, 5′-ACTGTTCCTGTGGGGTTGTGCC-3′; ID1 forward, 5′-CGAGATCAGCGCCCTGACGG-3′ and reverse, 5′-GGCCGCCGATCGGTCTTGTT-3′; Smad3 forward, 5′-GGAGAAATGGTGCGAGAAGG-3′ and reverse, 5′-GAAGGCGAACTCACACAGC-3′; p21 forward, 5′-GCCGAAGTCAGTTCCTTGTG-3′ and reverse, 5′-TTCTGACATGGCGCCTCCT-3′; activating transcription factor 3 (ATF3) forward, 5′-GAGGTGGGGTTAGCTTCAGT-3′ and reverse, 5′-TTGATTTTGGGGCAAGGTGC-3′; GAPDH forward, 5′-GGGCTCTCTGCTCCTCCCTGTTCT-3′ and reverse, 5′-CAGGCGTCCGATACGGCCAAA-3′.

SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. Values are presented as the mean ± standard deviation of samples measured in triplicate. P<0.05 was considered to indicate a statistically significant difference. Each experiment was repeated three times, unless otherwise indicated. The significance of differences between experimental groups compared to the vehicle control group was analyzed using a paired Student's t-test and two-tailed distribution. Multiple comparisons were analyzed using a one-way analysis of variance (ANOVA) test. Newman-Keuls test was used following ANOVA.

Firstly, the cell morphological changes induced by sFRP1 overexpression were investigated. SGC-7901/sFRP1 cells exhibited extended and protruded lamellipodia composed of F-actin fibro stained by phalloidin (Fig. 1A, upper); however, SGC-7901/vector cells exhibited a polygon-like shape with less lamellar extensions around the entire cell periphery (Fig. 1A, lower). This phenomenon indicated that Rac1, which is well known to be involved in filopodia and lamellipodia formation and thus control cell movement (33), was activated by sFRP1 overexpression. Therefore, Rac1 activity was measured in control and sFRP1-overexpressing cells by kinase activity assays. In agreement with a previous study (33), sFRP1-overexpressing cells exhibited increased Rac1 activity (Fig. 1B, left); however, the loss of Rac1 activity/activation was observed in sFRP1-knockdown BGC823 cells, compared with vector only control cells (Fig. 1B, right). Immunoblotting also demonstrated a lower level of its inactivated form p-Rac1 Ser71 in sFRP1-overexpressing cells (Fig. 1C). These data indicated that sFRP1 activates Rac1 activity.

In addition, it was reported previously that sFRP1 abrogates GSK3β inactivation by preventing its phosphorylation at the Ser9 residue (34). The present study also demonstrated a lower level of p-GSK3β Ser9 in sFRP1-overexpressing cells compared with the control cells (Fig. 2A). In agreement with the notion that sFRP1 is an inhibitor of Wnt signaling, it was determined that TCF-responsive luciferase activity was significantly repressed by sFRP1 overexpression compared with the control cells (P<0.05; Fig. 2B) and the nuclear accumulation of β-catenin was attenuated (Fig. 2C). Consistent with other data, the present cell model also demonstrated that sFRP1 overexpression restored GSK3β activity and inhibited the Wnt/canonical pathway.

Due to sFRP1 overexpression activating Rac1 and GSK3β, and GSK3β being previously reported to modulate Rac1 activity (35), the present study investigated whether GSK3β regulated Rac1 activity in SGC-7901/sFRP1 cells. Decreased lamellipodia formation, a feature of Rac1 inactivation, was observed in SGC-7901/sFRP1 cells treated with GSK3β inhibitor IM-12 or Rac1 inhibitor NSC23766 compared with vehicle cells (Fig. 3A). As depicted in Fig. 3B, a reduced amount of Rac1 bound to PAK-PBD compared with vehicle cells, which indicated reduced Rac1 activity. Levels of VAV2, a guanine nucleotide exchange factor (GEF) and activator of Rac1 (36), were lower in NSC23766 and IM-12 treated cells that were precipitated by PAK-PBD compared with vehicle cells, indicating that GSK3β or Rac1 inhibition suppressed Rac1 activity. Notably, GSK3β was also one of the components that was precipitated by PAK-PBD beads, and its level was decreased upon Rac1 or GSK3β inhibition compared with the vehicle control cells (Fig. 3B, left). The total levels of Rac1, GSK3β, and VAV2 remained consistent in cells with different treatments (Fig. 3B, right). Due to GSK3β being precipitated by PAK-PBD, which bound the activated form of Rac1, this indicated that GSK3β may directly or indirectly interact with Rac1; therefore, the levels of precipitated GSK3β were decreased in a similar pattern to the levels of the activated-Rac1, indicating that GSK3β may regulate Rac1 activity. Subsequently, a GSK3β overexpression model was used to investigate whether GSK3β was able to regulate Rac1 activity. As expected, a low level of the inactivated form of Rac1 (p-Rac1 Ser71) was observed in GSK3β-overexpressing cells compared with the vector cells (Fig. 3C). Due to NSC23766 inhibiting Rac1-GEF interaction (37) and IM-12 directly suppressed GSK3β activity (38), GSK3β activity may be necessary for regulating Rac1 activity.

Activated Rac1 signaling has been determined to be important in gastric cancer tumorigenesis (39) and induces the high mobility cell phenotype (40). Although GSK3β is a classic inhibitor of the Wnt/canonical pathway, it is able to activate other signaling pathways and promote tumorigenesis (41,42). Subsequently, whether Rac1 or GSK3β mediated the tumor-promoting effects of sFRP1 was investigated. To address this question, a specific Rac1 inhibitor NSC23766 (37) and a small molecule GSK3β inhibitor IM-12 (38) were used, which were demonstrated to inhibit Rac1-GEF interaction and GSK3β kinase activity, respectively. Cell proliferation and migratory ability were then investigated. The significant inhibition of SGC-7901/sFRP1 cell growth by NSC23766 or IM-12 was depicted in Fig. 4A (P<0.05). sFRP1-overexpressing cells exhibited significantly inhibited migration following NSC23766 or IM-12 treatment, compared with control cells (P<0.05; Fig. 4B, upper). NSC23766 or IM-12 treatment abolished the formation of lamellipodia and membrane ruffles in SGC-7901/sFRP1 cells (Fig. 4B, lower). These data indicated that Rac1 and GSK3β serve essential functions in regulating the growth and migration of sFRP1-overexpressing cells.

As demonstrated previously (25), sFRP1-overexpressing cells retained nuclear Smad2 levels but exhibited notably reduced Smad3 levels, compared with vector control cells, indicating unbalanced Smad2 and Smad3 activity (Fig. 5A). PAI1, ID1 and ZEB2, downstream targets of the TGFβ signaling pathway (43,44), were also upregulated in SGC-7901/sFRP1 cells (Fig. 5A). Rac1 was determined to selectively antagonize TGFβ/Smad3 mediated growth inhibition via its ability to promote Smad2 activation (45). GSK3β was previously reported to be responsible for the linker region of Smad3 and inhibited its transcriptional activity on molecules that mediated the growth inhibition activity of TGFβ signaling (46). Subsequently, whether Rac1 and GSK3β participated in the regulation of TGFβ signaling through sFRP1 was investigated; therefore, immunoblotting was performed using the nuclear extracts from SGC-7901/sFRP1 cells treated with Rac1 or GSK3β inhibitors. As depicted in Fig. 5B, nuclear Smad2 expression levels were not altered, and Smad3 and Smad4 expression levels were decreased following NSC23766 or IM-12 treatment. Inhibition of Rac1 or GSK3β activity also decreased ID1 and ZEB2 levels, which explained why Rac1 or GSK3β inhibition suppressed cell growth and migration.

Elevated mRNA levels of Smad3-responsive genes (47), including p21, ATF3, PAI1 and Smad3, were significantly elevated by Rac1 or GSK3β inhibition, whereas the ID1 mRNA level was significantly inhibited (P<0.05; Fig. 5C). To further observe the different gene responses to Smad2 and Smad3 signaling, HA-Smad2 or Flag-Smad3 constructs were transfected into SGC-7901 cells (Fig. 5D and E). Notably, cells overexpressing Smad3 exhibited high levels of pSmad3C, PAI1, pSmad3L and p21. Transfection of the Smad3-S204 mutant, a mutant form of WT Smad3 with the Ser204 mutant that cannot be phosphorylated by GSK3β, construct into SGC-7901 cells resulted in even higher levels of p21, without exhibiting a pSmad3L band. Additionally, Smad2 overexpression also resulted in elevated PAI1, which was potentially caused by increased pSmad2C levels, as pSmad3C levels were unaltered. These observations supported the observation that sustained Smad2 activity was able to compensate some of the Smad3-responsive functions in sFRP1-overexpressing cells. These data strongly indicated that the Rac1 and GSK3β were able to suppress Smad3 function, whilst retaining the expression of genes that were critical in mediating TGFβ-induced survival and the EMT phenotype.

To further examine the function of Rac1 or GSK3β in suppressing Smad3 activity, Rac1 or GSK3β were ectopically overexpressed in SGC-7901/vector cells. It is known that GSK3β phosphorylates the linker region (Ser204) of Smad3 and inhibited its transcriptional activity (48). In the present study, higher levels of pSmad3L (Ser204) and unaltered pSmad2 levels (Fig. 6A) were also observed in GSK3β-overexpressing cells compared with the vector cells. Subsequently, the function of Rac1 overexpression on Smad3 activity was investigated by transfecting Rac1-WT and Rac1-CA plasmids into SGC-7901/vector cells. Rac1-WT overexpressing cells exhibited lower pSmad3C expression levels compared with the vector cells (Fig. 6B, left). Rac1-CA overexpressing cells exhibited a more notable decrease in pSmad3C expression, compared with Rac1-WT overexpressing cells (Fig. 6B, left); however, Rac1-DN overexpressing cells exhibited an elevated pSmad3C expression level (Fig. 6B, right). Collectively, these data indicated that GSK3β and Rac1 are responsible for modulating TGFβ/Smad3 signaling in sFRP1-overexpressing cells (Fig. 6C).