The following antibodies were used: Anti-phosphorylated(p)-EGFR (Tyr-1173) immunoglobulin rabbit polyclonal IgG, anti-p-EGFR (Tyr-845) rabbit polyclonal IgG, anti-p-EGFR (Tyr-1086) rabbit polyclonal IgG, anti-p-EGFR (Tyr-1068) rabbit polyclonal IgG, anti-p-HFGR (Met; Tyr1313) rabbit polyclonal IgG, ant-p-Met (Tyr-1349) rabbit polyclonal IgG, ant-p-Met (Tyr-1365) rabbit polyclonal IgG, anti-Akt1/2/3 rabbit polyclonal IgG, anti-P-Akt1/2/3 (Ser473) rabbit polyclonal IgG, anti-P-Akt1/2/3 (Thr308) rabbit polyclonal IgG, anti-phospholipaseCγ (PLCγ)1 rabbit polyclonal IgG, anti-p-PLCγ1 (Tyr-783) rabbit polyclonal IgG, anti-p-p44/42 mitogen-activated protein kinase (MAPK) rabbit polyclonal IgG (Erk1/2; Thr-202/Tyr-204), anti-β-actin mouse polyclonal IgG, goat anti-rabbit polyclonal horseradish peroxidase-conjugated IgG, rabbit anti-goat polyclonal horseradish peroxidase-conjugated IgG, goat anti-mouse polyclonal horseradish peroxidase-conjugated IgG and horseradish peroxidase-labeled streptavidin (all from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Anti-GM3 IgM was obtained from Wako Pure Chemical Industries Ltd (Osaka, Japan) The dilution used was 1:500 for all antibodies.

LY294002 was obtained from Sigma-Aldrich (St. Louis, MO, USA), EGF was obtained from ProSpec-Tany TechnoGene, Ltd. (East Brunswick, NJ, USA), Lipofectamine™2000 and opti-MEM™ were obtained from Invitrogen Life Technologies, Carlsbad, CA, USA, Protease inhibitor cocktail was obtained from Sigma-Aldrich and GM3 and GM2 were purchased from Sigma-Aldrich.

Mouse hepatocellular carcinoma Hepa1–6 cells were grown in 10-cm cell culture dishes or in multi-well plates in RPMI-1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin at 37°C in a 5% CO₂ atmosphere for 48 h.

A negative control and four types of GM3 synthase short hairpin (sh)RNA plasmids were constructed, as follows: PGPU6/green fluorescent protein (GFP)/Neo-shNC (negative control), PGPU6/GFP/Neo-St3ga15-mus-469, PGPU6/GFP/Neo-St3 ga15-mus-613, PGPU6/GFP/Neo-St3ga15-mus-772 and PGPU6/GFP/Neo-St3ga15-mus-832. To select an effective shRNA plasmid that inhibited GM3 synthase, the Hepa1–6 cells were transfected with the GM3 synthase shRNA plasmids using Lipofectamine™ 2000 in opti-MEM™. The ratio of DNA (μg) to lipofectamine 2000 (μl) was 1:3. The effective GM3 synthase shRNA plasmid was selected using reverse transcription polymerase chain reaction (RT-PCR).

Total RNA was isolated from cells using Trizol (Invitrogen Life Technologies), and RT-PCR was conducted with Taq polymerase, according to the manufacturer’s instructions (Boehringer-Mannheim, Indianapolis, IN, USA). The cycling parameters were implemented as follows: denaturation at 94°C for 30 sec, annealing at 58–61°C for 1 min (temperature dependant on primer specifications), and elongation at 72°C for 1 min. The number of cycles varied between 25 and 35, depending on the quantity of mRNA. The identity of the PCR products was confirmed through sequencing. The relative quantitative analysis was normalized to that of the endogenous reference gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The mouse GM3 synthase forward primer used was 5′-AGCTGCCAGAGGTGATGAGT-3′ and its reverse primer was 5′-TCAAGTGGCTTCAAGCAATG-3′. The mouse GAPDH forward primer used was 5′-TACTTATGCCGATGTCGTTGT-3′ and its reverse primer was 5′-CCAGCCTCGTCCCGTAGA-3′. The PGPU6/GFP/Neo-St3ga15-mus-772 exhibited effective inhibition and was used in the subsequent investigation (GenePharma Co., Ltd., Shanghai, China).

To decrease the expression of GM3 in the Hepa1–6 cell, the cells were transfected with the GM3 synthase shRNA plasmid at 37°C in a 5% CO₂ atmosphere. After 48 h incubation, the cells were harvested and the inhibition of ganglioside synthesis was monitored by high performance thin layer chromatography (HPTLC). For the GM3 synthase shRNA, the cells were transfected with PGPU6/GFP/Neo-St3ga15-mus-772 using lipofectamine™ 2000 in opti-MEM medium. After 48 h transfection, the cells were harvested and the expression of GM3 synthase was monitored by RT-qPCR. To increase the expression of ganglioside in the Hepa1–6 cell, the cells were treated with 50 μM GM3 or 50 μM GM2 for 48 h prior to harvesting.

HPTLC analysis was performed, as described previously (18). Briefly, the cells were grown in 10 cm dishes until ~90% confluent. The cells were then trypsinized and washed three times with phosphate-buffered saline (PBS; Sangon Biotech, Shanghai, China). The total lipids were extracted twice from the cell pellet using chloroform/methanol (1:1). The extracts were then combined and dried under a stream of N₂. The gangliosides were purified by partitioning the dried total lipid in di-isopropyl ether/1-butanol/17 mM aqueous NaCl followed by sephadex G-50 gel filtration and lyophilization. The individual gangliosides were separated onto silica gel 60 HPTLC plates (Merck KGaA, Darmstadt) with a solvent system of chloroform/methanol/0.25% aqueous CaCl2·2H2O (60:40:9; Sangon Biotech). The gangliosides were visualized as purple bands by spraying them with resorcinol-HCI reagent and heating the plate at 120°C.

The cell migration assays were performed using a Boyden chamber (Corning Costar, Cambridge, MA, USA) with 8 μm pore polycarbonate filters (BD Biosciences, Franklin Lakes, NJ, USA). The Hepa1–6 cells were treated with GM3 and transfected with the GM3 synthase shRNA plasmid, as described above. To inhibit the phosphatidylinositol 3-kinase (PI3K)/Akt pathway with LY294002, the Hepa1–6 cells treated, as above, were exposed to 15 μM LY294002 for 4 h. The cells were then harvested and used for migration assays.

Following incubation of the cells in serum-free medium overnight, the cells (1×10⁵) were resuspended in 300 μl serum-free medium containing 50 ng/ml EGF or 100 ng/ml HGF and the desired agents and placed in the top compartment of the chamber. Subsequently, 250 μl 10% FBS medium was placed in the bottom chamber. After 36 h incubation at 37°C in a 5% CO₂ incubator, the cells on the top membrane surface were mechanically removed. The cells that migrated to the lower surface of the membrane were fixed and stained with 0.1% DAPI. Images of the cells were captured and the stained cells from five randomly selected fields were counted under a microscope (Nikon Corporation, Shanghai, China).

The cells treated, as described above, were seeded into 12-well plates (Falcon, BD Biosciences) and incubated overnight in serum-free medium containing the desired agents, as described in the figure legends. The medium was changed and the cells were stimulated with 50 ng/ml EGF or 100 ng/ml HGF for 10 min at room temperature or left unstimulated as a control. The cells were harvested and lysed in 200 μl radioimmunoprecipitation assay buffer containing 1% Triton X-100, 150 mM NaCl, 25 mM Tris (pH 7.5), 0.5% sodium deoxycholate, 0.1% SDS, 5 mM pyrophosphate and 50 mM NaF with 1 mM Na₃VO₄, 1 mM DTT, 1% protease inhibitor cocktail and 1% phosphatase inhibitor cocktail per well. The lysate was subjected to SDS-PAGE using 10% gel. Following electrophoresis, the proteins were transferred onto polyvinylidene difluoride (PVDF) membranes at 100 mA for 2 h. The non-specific binding sites on the PVDF membranes were inhibited using 3% (w/v) bovine serum albumin. The target protein bands in the PVDF membranes were revealed by immunoblotting with the primary antibodies given in the ‘antibodies’ section of ‘Materials and methods’ and detected using enhanced chemiluminescence reagents (Amersham Pharmacia Biotech, Amersham, UK).

To decrease the expression of GM3 in the Hepa1–6 cells, the cells were transfected with GM3 synthase shRNA plasmids. The RT-qPCR results revealed that, among the four GM3 synthase shRNA plasmids, the PGPU6/GFP/Neo-St3ga15-mus-772 plasmid was the most effective at inhibiting the expression of GM3 synthase (Fig. 1A). The inhibitory effect was monitored by HPTLC (Fig. 1B).

The effects of GM3 on the motility and migration of the EGF- or HGF-stimulated Hepa1–6 cells were investigated. A decrease in GM3 content by the GM3 synthase shRNA plasmid promoted EGF-stimulated cell migration (Fig. 2a and b) and inhibited HGF-stimulated migration (Fig. 2e and f). Conversely, an increase in GM3 content, caused by the addition of exogenous GM3 in the culture medium, inhibited EGF-stimulated cell migration (Fig. 2a and c) and promoted HGF-stimulated migration (Fig. 2e and g). The addition of exogenous GM2 had no effect on the migration of the Hepa1–6cells treated with EGF and HGF (Fig. 2a, d and e, h). These results indicated that GM3, but not GM2, inhibited EGF-stimulated cell migration and promoted the HGF-stimulated migration in vitro.

To elucidate the precise molecular mechanism by which GM3 affected EGF- and HGF-stimulated migration of the Hepa1–6 cells, the present study examined the effect of GM3 on the phosphorylation of EGFR and HGFR. Under control conditions, no background phosphorylation at either Tyr-1173 or Tyr-845 was observed. However, a decrease in GM3 content, by treating cells with the GM3 synthase shRNA plasmid, significantly elevated the EGF-stimulated phosphorylation of EGFR at Tyr-1173 and reduced the EGF-stimulated phosphorylation of Tyr-845 (Fig. 3Aa). An increase in GM3 content, by treating cells with exogenous GM3, markedly inhibited the EGF-stimulated phosphorylation of EGFR at the Tyr-1173 and elevated the phosphorylation of Tyr-845 (Fig. 3Ab). The phosphorylation of EGFR at Tyr-1068 and Tyr-1086 were not detected (data not shown). The results demonstrated that GM3 suppressed the EGF-stimulated phosphorylation of EGFR at Tyr-1173, but promoted the phosphorylation of Tyr-845. Tyr-1173 is located at the c-terminus of EGFR and is a major EGFR autophosphorylation site. Phosphorylated Tyr-1173 serves as a docking site for SH2-domain containing signaling molecules and leads to the activation of the downstream signaling pathways. Tyr-845 is distinct from others located at the EGFR c-terminus. It is located in the kinase domain of the EGF receptor and can be transphosphorylated by src kinase (19). The significance of Tyr-845 phosphorylation remains to be elucidated, however, it may be involved in the modulation of tyrosine kinase activity of the receptor. These results suggested that the mechanism for the GM3-suppressed phosphorylation of EGFR in Hepa1–6 cells is complicated.

By contrast, a decrease in GM3 content, by treating cells with a GM3 synthase shRNA plasmid, significantly reduced the HGF-stimulated phosphorylation at Tyr-1313 and Tyr-1365 (Fig. 3Ba). An increase in GM3 content, by treating cells with exogenous GM3, markedly promoted the HGF-stimulated phosphorylation of cMet at Tyr-1313 and Tyr-1365 (Fig. 3Bb). The phosphorylation of Tyr-1349 was not detected (data not shown). GM3 promoted the HGF-stimulated phosphorylation of HGFR at Tyr-1313 and Tyr-1365.

These results indicated that GM3 suppressed the EGF-stimulated phosphorylation and activation of EGFR, but promoted the HGF-stimulated phosphorylation and activation of cMet, which may explain why GM3 exerted opposing affects on the motility and migration of the EGF- and HGF-stimulated Hepa1–6 cells.

It is well known that the autophosphorylation of tyrosine kinase receptors is an essential step for their activation, which results in phosphorylation at specific tyrosine residues on the intracellular domain of the receptors. Certain cytoplasmic SH2-domain containing signaling molecules, including PI3K, PLC-γ, growth factor receptor-bound protein 2 and protein tyrosine phosphatases, bind to specific phosphotyrosine sequences on the receptor, thereby initiating the intracellular corresponding signaling pathways (20). To further investigate the mechanism underlying the different effects of GM3 on EGF- and HGF-stimulated motility and migration in vitro, the present study investigated the effects of GM3 on the activity of intracellular signaling pathways, which are essential for the modulation of tumor cell invasion and migration. Three of the main downstream signal transduction pathways activated by GFRs are the MAPK, PLCγ and PI3K/Akt pathways. The present study examined the effect of GM3 on the activity of these signaling pathways. Fig. 4 shows that an alteration in GM3 content had no significant affect on the phosphorylation of MAPK (p44/42) in the Hepa1–6 cells when stimulated by either EGF (Fig. 4Aa and b) or HGF (Fig. 4Ba and b). This result suggested that GM3 did not affect the activity of MAPK signaling in the Hepa1–6 cells. The effect of GM3 on the activity of the PLCγ signaling pathway was also assessed. It is understood that the phosphorylation and translocation of PLCγ from the cytosol to the plasma membrane are necessary for its activation and function (13). The present study attempted to examine the effect of GM3 on the phosphorylation of PLCγ1, however phosphorylated PLCγ1 was undetected in the control cells and in the cells treated with EGF or HGF (data not shown). The localization of PLCγ1 in the cells was further examined, however, membrane-binding PLCγ1 was not detected in the cells stimulated with either EGF or HGF (data not shown). As shown in Fig. 4, GM3 promoted the expression of PLCγ1 in the cells stimulated with EGF (Fig. 4Ac and d) and inhibited the expression of PLCγ1 in the cells stimulated with HGF (Fig. 4Bc and d). These data indicated that GM3 did not affect the activity of the PLCγ1 signaling pathway. The phosphorylation of Akt was also examined. In the EGF-treated cells, suppression of GM3 synthesis by treatment of the cells with a GM3 synthase shRNA plasmid, elevated the phosphorylation of Akt at Ser 473 (Fig. 4Ae), whereas increasing the GM3 content by treating the cells with exogenous GM3 inhibited the phosphorylation of Akt at Ser 473 (Fig. 4Af). By contrast, in the HGF-treated cells, suppression of GM3 synthesis by treatment of the cells with a GM3 synthase shRNA plasmid inhibited the phosphorylation of Akt at Ser 473 (Fig. 4Be). An increase in the GM3 content following treatment with exogenous GM3 promoted the phosphorylation of Akt at Ser 473 (Fig. 4Bf). These results suggested that GM3 suppressed the EGF-stimulated activation of the PI3K/Akt signaling pathway and promoted the HGF-stimulated activation of the PI3K/Akt signaling pathway, indicating that the PI3K/Akt signaling pathway may be important in GM3-regulated cell migration.

To confirm whether Hepa1–6 cell motility and migration were mainly regulated by the PI3K/Akt signaling pathway, the activity of PI3K in cells was inhibited using LY294002, an inhibitor of PI3K. Inhibiting the PI3K/Akt pathway with LY294002 reversed the positive effect following transfection with a GM3 synthase shRNA plasmid (Fig. 5a–c) on EGF-stimulated migration. The positive effect on HGF-stimulated migration, caused by the addition of exogenous GM3, was also reduced by inhibiting the PI3K/Akt pathway (Fig. 5d–f). This finding confirmed that the PI3K/Akt pathway was involved in the modulation of Hepa1–6 cell motility and migration in vitro.