Gastric cancer cell lines MKN45 and SGC7901 and the gastric epithelial cell GES-1 were purchased from the Chinese Committee of Type Culture Collection Cell Bank, Chinese Academy of Sciences (Shanghai, China). MKN45 and SGC7901 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS and 1% antibiotics. The GES-1 cells were cultured in RPMI-1640 medium containing 10% FBS and 1% antibiotics. LicA and specific signaling pathway inhibitors including LY294002, PD98059, Bay11-7082 and parthenolide were purchased from Selleck (Shanghai, China). Primary anti-HK2 (#2867), Glut1 (#12939), PKM2 (#4053), LDHA (#3582), p-Akt (#4060), p-ERK1/2 (#4370), p-p65 (#3031), p-IκBα (#9246), p-S6 (#4858), p-GSK3β (#12456), Akt1 (#75692), MCL-1 (#5453), BCL2 (#2870), BCL-XL (#2764), cleaved caspase-3 (#9664), cleaved PARP (#5625) antibodies and secondary antibodies anti-rabbit IgG HRP (#7074) and anti-mouse IgG HRP (#7076) were products of Cell Signaling Technology Inc. (Beverly, MA, USA). Anti-β-actin antibody (A5316) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Myr-Akt1 plasmid was purchased from Addgene (Cambridge, MA, USA). HK2 (ORF004940) construct was a product of Applied Biological Materials (ABM) Inc. (Richmond, BC, Canada). Lipofectamine^® 2000 was purchased from Invitrogen (Carlsbad, CA, USA). Annexin V-FITC and propidium iodide were products of Biolegend (San Diego, CA, USA).

MKN45 or SGC7901 cells in logarithmic growth phase (3×10³/well) were seeded into 96-well plate. Twenty-four hours later, the cells were treated with various concentrations of LicA. At different time points (24, 48 or 72 h), CellTiter96 Aqueous One Solution (Promega Corp., Madison, WI, USA) (20 µl/well) was added to a 96-well plate and the cell viability was assessed according to the manufacturer's protocol.

Cells cultured in 10-cm plastic dishes were treated with different concentrations of LicA for 24 h, and the cells were harvested and reseeded into a 6-well plate in duplicate at the appropriate density. The cells were cultured for 1–3 weeks until the colonies with substantially good size (minimum 50 cells per colony) were formed in the control group. After washing with PBS, the colonies were fixed with the fixation solution (methanol:acetic acid 3:1) and then stained with 0.5% crystal violet for 2 h at room temperature. The crystal violet was carefully removed by immersing the plates in tap water repeatedly and the plates were air-dried. The number of colonies was photographed and counted.

After the treatment of LicA, gastric cancer cells were lysed with RIPA lysis buffer containing protease cocktail (Roche, Mannheim, Germany) on ice for 30 min. The cell lysate was harvested and centrifugated at 12,000 × g for 5 min, and the supernatant was collected. The protein concentrations were determined by the Bradford assay (Bio-Rad, Philadelphia, PA, USA). Twenty micrograms per sample was subjected to SDS-PAGE and then transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). After blocking the non-specific binding site on the membrane with 5% non-fat milk solution, the membranes were incubated with specific primary antibodies at a dilution of 1:1,000 at 4°C overnight. After washing three times with TBS-Tween-20, the membranes were incubated with HRP-conjugated secondary antibody at a dilution of 1:2,000 at room temperature for 1 h, and then the bands on the membrane were visualized using an enhanced chemiluminescence reagent (Thermo Fisher Scientific, Inc., Waltham, MA, USA).

Gastric cancer cells (5×10⁵/well) were plated in 6-well tissue culture plate and cultured overnight. Then, the culture medium was discarded and cells were incubated with fresh medium containing different concentrations of LicA for 12 h. The Automatic Biochemical Analyzer (7170A, Hitachi, Tokyo, Japan) was used for the measurement of the levels of glucose and lactate in the culture medium. Protein concentration per sample acted as the control to normalize the relative glucose consumption and lactate production rate.

After LicA treatment, MKN45 cells and the cell culture medium were collected and centrifugated. After washing with PBS, the cell pellets were resuspended and stained with Annexin V-FITC and propidium iodide according to the manufacturer's instructions, and then the stained cells were subjected to flow cytometry. The results were analyzed and quantified by a flow cytometer (BD Biosciences, San Jose, CA, USA).

For Myr-Akt1 or pORF-HK2 transfection, HCC cells were seeded into a 6-well culture plate and grown to 70–90% confluence. Before transfection, the culture medium was replaced with 1.5 ml fresh medium without fetal bovine serum. Plasmid DNA (2.5 µg) and Lipofectamine 2000 reagent (7.5 µl) were diluted with 150 µl Opti-MEM medium respectively and incubated at room temperature for 5 min. The diluted DNA and transfection reagent were mixed and incubated for 10 mins avoiding light, then 250 µl mixture was added per well and the culture medium was replaced with fresh medium with fetal bovine serum after 4–6 h. Forty-eight hours later, the transfected cells were used for further studies.

In accordance with the protocol approved by the Institutional Animal Care and Use Committee, BALB/ca nude mice maintained under specific pathogen-free (SPF) conditions were subcutaneously injected with MKN45 cells (2×10⁶ cells/mice). Once the tumor was formed and the volume was ~50 mm³, the mice were randomly assigned into a vehicle and experimental group. The vehicle group received 0.2 ml sterile PBS solution, and the experimental group was treated with 10 mg/kg LicA by i.p. injection daily. The tumor volume (V) was measured twice per week with microcalipers and was calculated as V= (length × width²)/2. At the end of the experiment, the mice were sacrificed and the tumors were removed.

Tumor tissues isolated from the mice were embedded in paraffin and cut into 5-µm sections. After dewaxing in xylene and hydration in serial ethanol, the slides were boiled in sodium citrate buffer (10 mmol/l, pH 6.0) for 10 min to expose antigens. To block endogenous peroxidase, the slides were treated with 3% H₂O₂ for 10 min. After incubation with goat serum at room temperature for 1 h, the slides were incubated with the anti-HK2 (1:100) or anti-Ki-67 (1:200) antibody respectively at 4°C in a humidified chamber overnight. Following washing with PBS three times, the slides were hybridized with biotinylated goat anti-rabbit secondary antibody at room temperature for 30 min. After washing with PBS, the sections were probed with HRP-conjugated streptavidin and then visualized with DAB solution. After counterstaining with Harris' hematoxylin, the slides were dehydrated and mounted. The intensity was analyzed by Image-Pro PLUS (v.6) and ImageJ (NIH) software programs.

SPSS software (version 13.0; IBM SPSS, Armonk, NY, USA) was used for statistical analysis. Student's t-test or one-way ANOVA was adopted to evaluate statistical significance. P<0.05 indicated a statistically significant difference.

Firstly, we tested the effect of LicA on normal gastric epithelium cell line GES-1. The results showed that LicA had no obvious cytotoxicity at concentrations ≤180 µM (Fig. 1B). We next examined the anti-proliferative activity of LicA in gastric cancer cells in vitro. As shown in Fig. 1C, at the low concentration (15 µM), no obvious inhibitory effect was observed. With the increase in concentration (30–60 µM) and the extension of treatment time (48–72 h), cell proliferation in the MKN45 and SGC7901 cells was significantly inhibited. The IC₅₀ of LicA in MKN45 and SGC7901 cells was 63.57 and 55.56 µM respectively (data not shown). Beyond that, the effect of LicA on clonogenic survival was also investigated. The incubation of LicA resulted in the decrease in colonies formed in the agar, and the number of clones was substantially decreased in a dose-dependent manner. At 60 µM, little colonies were formed after exposure to LicA (Fig. 1D). All these data demonstrated that LicA exerted a profound antitumor activity in gastric cancer cells in vitro.

In order to study the effect of LicA on tumor glycolysis, the amount of glucose consumption in gastric cancer cells was determined. As shown in Fig. 2A, after LicA treatment, the glucose consumed by MKN45 and SGC7901 cells was significantly decreased dose-dependently. At the high concentration of 60 µM, ~50% inhibition was observed. Accompanied by the reduction of glucose absorption, the amount of lactate secreted by gastric cancer cells was also decreased significantly. Furthermore, we tested the effect of LicA on key glycolytic enzymes and the results demonstrated that the expression of HK2 was decreased by LicA. The expression of other glycolytic enzymes, such as GLUT1, PKM2 and LDHA, had no obvious change (Fig. 2B). All these data imply that HK2 is involved in the suppression of glycolysis mediated by LicA.

To confirm which signaling pathway is engaged in the regulation of HK2 expression, the effect of LicA on the main signaling pathways in gastric cancer cells was investigated. As demonstrated, several signaling pathways in gastric cancer, such as ERK, Akt and NF-κB were blocked by LicA dose-dependently (Fig. 3A). To further validate the inhibition of these pathways, we examined the effect of LicA on EGF-induced activation of ERK and Akt, as well as TNF-α-induced NF-κB activation. As shown in Fig. 3B and C, the activation of ERK, Akt and NF-κB signaling was suppressed by LicA dose and time-dependently. By using different specific signaling pathway inhibitors to treat gastric cancer cells and observe the change in HK2 expression, we identified that LY294002, a specific PI3-K inhibitor, demonstrated a similar effect on HK2 as LicA, implying that the Akt signaling pathway may be involved in the mediation of HK2 expression (Fig. 3D).

Based on the results shown above, the HK2 decrease caused by LicA may be attributed to the blockade of Akt activation, thus, we tested the effect of LicA on the Akt signaling pathway. After LicA treatment, along with the inactivation of Akt, the phosphorylation of S6 and GSK3β, which are the main downstream signaling of Akt, were inhibited dose-dependently (Fig. 4A and B). To illustrate the importance of Akt in the regulation of HK2 expression, constitutively activated Akt1 (myr-Akt1) was transfected into MKN45 and SGC7901 cells. As expected, with the recovery of Akt activity in gastric cancer cells, HK2 reduction caused by LicA was notably reversed (Fig. 4C and D). Moreover, glucose consumption and lactate production in myr-Akt1 transfected cells were also significantly increased (Fig. 4E and F). These results demonstrated that Akt-mediated HK2 expression plays a pivotal role in the effectd of LicA on tumor glycolysis.

Except glycolysis suppression, LicA also induced apoptosis in gastric cancer cells. The expression of cleaved-caspase-3 and PARP, which are signals of cells undergoing apoptosis, were significantly increased (Fig. 5A and B). The results of FACS analysis also confirmed that MKN45 cells exhibited induced apoptosis dose-dependently; 25% of tumor cells were subjected to apoptosis at 60 µM (Fig. 5C). Furthermore, the anti-apoptotic proteins such as Bcl-2, and Mcl-1 were found to be decreased after LicA treatment, whereas the expression of pro-apoptotic proteins including Bad and Bak had no significant change. In addition to mediating tumor glycolysis, HK2 is also involved in apoptosis regulation. Given the decrease in HK2 expression by LicA, we speculated that HK2 was also involved in LicA-induced cell apoptosis. After exogenous HK2 overexpression in gastric cancer, the apoptosis induced by LicA was substantially attenuated and the expression of cleaved-PARP and caspase-3 was significantly decreased in contrast with the control group, suggesting that HK2 was involved in LicA-induced apoptosis (Fig. 5D-F).

The in vivo antitumor activity of LicA was investigated in an MKN45 xenograft model. In contrast with the vehicle group, tumor growth in the LicA-treated group was significantly inhibited (Fig. 6A-C). The tumor volume of the vehicle group had reached about 900 mm³, whereas the average volume of the LicA-treated group was ~400 mm³. The tumor weight of the LicA-treated group was also significantly smaller than the vehicle group (0.52 vs. 0.85 g). Based on the change in body weight, no obvious toxicity was observed (Fig. 6D). As shown in Fig. 6E, immunohistochemistry staining of LicA-treated tumor tissues demonstrated that the expression of HK2 was substantially decreased. Meanwhile, the expression of Ki-67, a marker of cell proliferative potential, was also decreased, indicating that LicA inhibited tumor growth in vivo by suppressing tumor glycolysis.