Berberine was from Sigma (St. Louis, MO); PD98059 (MEK inhibitor) and LY294002 (PI3K inhibitor) were obtained from Merk. Cell culture reagents and DCFH-DA were obtained from Invitrogen. ERK 1/2 and AKT, the total and phosphorylated protein antibodies, MMP-9 antibody and horseradish peroxidase (HRP)-labeled anti-rabbit secondary antibody were purchased from Cell Signaling Technology (Boston, MA). All other reagents were from Sigma (St. Louis, MO) unless stated otherwise.

The HepG2 cell line was originally obtained from the American Tissue Culture Collection (ATCC, USA). Cells were cultured at 37°C and 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (Gibco), 2 mM Glutamine, 1% non essential amino acids (NEAA) and 1% antibiotics (100 U/mL of penicillin and 100 µg/mL of streptomycin).

Cell viability was determined by the MTT quantitative colorimetric assay, as previously reported ^[17]. 5×10⁴ cells in 100 µL of serum-free DMEM were seeded in 96-well and incubated with berberine at various concentration (0–80 µM) for 24 hours. Thereafter the medium was aspirated, and the cells were fixed with 0.2 mL of 10% cold TCA/per well at 4°C for 30 min, washed with deionized water, dried at room temperature overnight and incubated with MTT (0.5 mg/mL) for 4 hours. The viable cell number was directly proportional to the production of formazan solubilized with isopropanol, which could be measured spectrophotometrically at 563 nm.

The intracellular generation of ROS was measured using DCFH-DA. The cell-permeable non-fluorescent dye penetrates into the cells and is hydrolyzed to DCFH by the cellular esterases. The probe (DCFH) is rapidly oxidized by ROS to the highly fluorescent compound 2′,7′-dichlorofluorescein (DCF). Cells were seeded in 6-well plates at 2×10⁵ cells/well and treated with or without berberine followed by incubated with 5 µM of DCFH-DA at 37°C for 15 min. After the preincubation, the cells were then washed twice with PBS, trypsinized, and resuspended in phosphate-buffered saline (PBS). At least 20,000 cells were acquired for each sample. The mean fluorescence intensity at 530 nm was assayed using a flow cytometer (Beckman Coulter, CA).

The effect of berberine on the invasiveness of HepG2 cells was determined using modified Boyden chambers consisting of Transwell (Corning Costar Corp, Cambridge, MA) with 8 µm pore size polycarbonate membrane filters precoated with 50 µL of Matrigel (1.25 mg/mL). During MTT assay, equal HepG2 cells (5×10⁴ cells) of the second group suspended in the serum-free DMEM of 100 µL in the presence or absence of berberine were seeded onto the upper chamber of Matrigel-coated filter inserts. Serum-containing DMEM (500 µL) was added to the lower chamber. After 24-hour incubation, filter inserts were removed from the wells. The cells on the upper surface of the filter were wiped with a cotton swab. Filters were fixed 4% paraformaldehyde for 30 minutes and stained with 0.1% crystal violet for 30 minutes, and then the invaded cells were determined as eight high-power fields of cells were counted in each well under an inverted microscope at 200× magnification. Invasion was calculated as the relative invasive score of treated group (invaded cell number/total cell number assayed by MTT represented by OD₅₇₀) divided by that of control.

Western blotting protocols were as in previous studies ^[18]. Briefly, cell lysates were separated by SDS/PAGE in 10% Tris-glycine gels and transferred to a NC membrane. For Western blot analysis of ERK1/2, phospho-ERK1/2, AKT and phosphor-AKT, blots were probed with their specific antibodies (diluted with 5% BSA to 1: 1000). Nonphosphorylated total ERK1/2 and AKT bands were chosen as loading control for MAPKs activation. For Western blot analysis of MMP-9, blots were probed with MMP-9 specific antibody (diluted with 5% BSA to 1: 500). Membranes were probed with horseradish peroxidase (HRP)–labeled anti-rabbit secondary antibody (diluted with 5% BSA to 1: 1000; all antibodies from Cell Signaling). Antibody binding was detected by enhanced enhanced chemiluminescence detection kit (ECL) (UK Amersham International plc). All Western blot exposures were in the linear range of detection, and the intensities of the resulting bands were quantified by Quantity One software on a GS-800 densitometer (Bio-Rad).

ene silencing by RNA interference (siRNA) was used to down-regulate MMP-9 expression in HepG2 cells. The siRNA specific for human MMP-9 was synthesized (Qiagen, Valencia, CA) against the target sequences for MMP-9: [gene: 5′-AACATCACCTATTGGATCCAAACTAC-3′], nucleotides 377 to 403 ^[19]. Nonsilencing siRNA ([gene: 5′-AATTCTCCGAACGTGTCACGT-3′], Qiagen) with no homology to mammalian genes was used as the negative control ^[20]. Transfection of 7×10⁵ HepG2 cells with 0.1 µM of siRNA was done in a 66-mm Petri dish using 15 µL Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.

Data were statistically analyzed using Unpaired Student's t test at a significance level P value of <0.05 and are presented as means±SD, using Sigma Plot software (Jandel Scientific).

To determine the antiproliferative/cytotoxic effect of berberine on hepatoma cell line, in comparison with Chang liver cells, a non-tumor liver cell line, HepG2 cells (5×10⁴ cells) and Chang liver cells (5×10⁴ cells) were respectively suspended in 100 µL of DMEM and simultaneously seeded in 96-well plates. Then these cells were incubated in the absence or presence of increasing concentrations of berberine for 24 hours with the cytotoxicity of berberine measured by a standard MTT assay. As indicated in Figure 1A, the survival curve showed the dose-dependent cytotoxicity of berberine in HepG2 cells. After 24-hour of berberine (40 µM) treatment, cell viability was reduced by 40% approximately. In contrast, no marked antiproliferative/cytotoxic effects were seen in Chang liver cells under the exposure of same concentrations of berberine for 24 hours (Figure 1B).

Excessive ROS production has been proposed as a vital role in induction of cell death by various agents ^[21]. To test the hypothesis that berberine-induced cytotoxicity in HepG2 cells is also initiated through upregulation of ROS level, we first used the DCFH-DA flow cytometry system to detect the effect of berberine on ROS production ^[22]. As shown in Figure 2A, treatment of HepG2 cells with berberine for 24 hours resulted in a dose-dependent increase in ROS generation compared with non-berberine-treated cells, which was demonstrated by the increase in intensity of DCF fluorescence. Pretreatment with DPI (10 µM), an inhibitor of NADPH oxidase ^[23], blocked the berberine-increased ROS production (Figure 2B). Next, we found that elimination of increased ROS by pretreatment with DPI (10 µM) reversed berberine-induced cytotoxic effect on HepG2 cells (Figure 2C). These results suggested that berberine exerted cytotoxic effect on HepG2 cells through enhancement of ROS production. On the contrary, berberine had no effect on ROS production in Chang liver cells (Figure 2D).

Berberine has been shown to exert inhibitory effect on invasion of multiple human cancer cells. To test whether berberine also has the same effect in hepatoma cells, HepG2 cell invasion with berberine treatment was analyzed by Matrigel invasion assay. Figure 3A showed that berberine administration led to concentration-dependent decrease in cell invasion after 24-hour incubation. Berberine (40 µM) diminished the invasive ability of HepG2 cells substantially up to 32.8% of the control.

To rule out the possibility that decreased cell invasion after berberine treatment might be caused by decreased total cell number, in fact, we plated cells at the same density and culture medium volume as in the MTT assay into transwell chambers for invasion assay (see the section‘Cell invasion assay’ of methods). At the same time, moreover, equal cells of berberine-treated and -untreated groups were plated to 96-well plates for total cell number assay (MTT) represented by OD₅₇₀. The invasiveness of HepG2 cells was expressed by the invasive score (number of invaded cells/total cell number).

MMP-9 has been described to be closely participated in capsular infiltration in hepatoma cells ^[24]. The present study further investigated the effect of berberine on MMP-9 expression in HepG2 cells for determination of the mechanism for berberine-induced suppressive effect on cell invasion. Figure 3B showed that treatment of cells with berberine (40 µM) significantly suppressed MMP-9 expression and the decrease in MMP-9 level relative to that of non-berberine-treated group was approximately 40% after a 24-hour incubation period.

To confirm a causal link between berberine-mediated downregulation of MMP-9 expression and decreased invasion, the expression of MMP-9 was blocked by transfecting cells with MMP-9 siRNA. Figure 3C showed that siRNA to MMP-9 at the concentrations of 0.1 µM decreased MMP-9 expression by 89% as compared with control. The nonsilencing siRNA had no effect on MMP-9 expression. Both MMP-9 siRNA and the nonsilencing siRNA exerted no cytotoxic effect on HepG2 cells (data not shown). Knockdown of MMP-9 expression with MMP-9 siRNA resulted in a significant reduction of HepG2 cell invasion. Berberine significantly inhibited HepG2 cell invasion, but this effect was not seen when cells were pretreated with MMP-9 siRNA (Figure 3D). Altogether, these results suggest that berberine inhibits HepG2 cell invasion through suppression of MMP-9 expression.

To investigate the mechanism for anti-invasive effect of berberine on HepG2 cells, we determined whether interfering with the PI3K-AKT and ERK pathways affected the inhibition of cell invasion and MMP-9 expression by berberine.

We first determined the specific effect of the PI3K-AKT inhibitor LY294002 and ERK inhibitor PD98059. After 24-hour incubation, both LY294002 and PD98059 at the indicated concentrations had no impact on HepG2 cell growth as assessed by MTT assay (Figure 4A). Pretreatment with LY294002 (10 µM) and PD98059 (25 µM) for 1 hour significantly decreased the invasive ability of the cells, as the relative invasive score was reduced by near 50% and 58%, respectively (Figure 4B). Figure 4C showed that pretreatment with two inhibitors substantially downregulated MMP-9 expression by 31% and 56% respectively after 24-hour incubation, which suggested the critical contribution of PI3K-AKT and ERK pathways-dependent MMP-9 expression to the invasive phenotype of HepG2 cells.

To exclude the possibility that the PI3K-AKT and ERK pathways-dependent MMP-9 expression is restricted to hepatoma cells, experiments were also performed in normal Chang liver cells. Like HepG2 cells, PI3K-AKT and ERK pathways-dependent regulation of MMP-9 expression may also exist in normal Chang liver cells (See Text S1 and Figure S1, available online).

Figure 5A illustrated that treatment of HepG2 cells with berberine (40 µM), LY294002 (10 µM) and PD98059 (25 µM) at the indicated time of incubation resulted in significant decrease in invasive ability of HepG2 cells, as the relative invasive score was reduced by approximately 67%, 50% and 42%, respectively. The combination treatment of LY294002 or PD98059 with berberine could even reduce the cell invasion by 83% or 85%, as compared with control. Furthermore, combination treatment of LY294002 or PD98059 with berberine could further downregulate the MMP-9 level compared with that of berberine treatment only (Figure 5B). Next, we determined the effect of berberine on AKT and ERK activities represented as the levels of phosphorylated forms of AKT and ERK by western blotting. Berberine (40 µM) significantly decreased the levels of p-AKT (Figure 5C) and p-ERK1/2 (Figure 5D) with no changes on total AKT and ERK levels under 4-, 12-, and 24-hour exposure, suggesting the suppressive effect of berberine on the PI3K-AKT and ERK pathways.

Thus, these results demonstrate that berberine inhibits the invasive ability of HepG2 cells through downregulation of MMP-9 expression by concomitant inactivation of the PI3K-AKT and ERK pathways.

To explore whether the cytotoxic effect and the anti-invasive effect of berberine on HepG2 cells are independently exerted or have some crosslinks, we determined if enhancement of ROS production by berberine had some influence on its anti-invasive effect.

As shown in Figure 6, berberine administration significantly suppressed MMP-9 expression and invasion of HepG2 cells. Pretreatment with DPI had no influence on the suppressive effects of berberine on the activity of PI3K-AKT (Figure 6A) and ERK (Figure 6B) pathways, as well as MMP-9 expression (Figure 6C) and HepG2 cell invasion (Figure 6D). Thus, the cytotoxic effect and the anti-invasive effect of berberine on HepG2 cells seem to be independently exerted.