Flavonoid compounds, luteolin (3′,4′,5,7-tetra-hydoxyflavone) and apigenin (4′,5,7-trihydroxyflavone) were purchased from Sigma-Aldrich Co. LLC (St. Louis, MO, USA). Each compound was dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich Co. LLC) to 20–200 μg/ml stock solutions for further experiments.

Human HCC cell line HepG2 and Hep3B, and human hepatocyte-derived Chang liver cells were obtained from American Type Culture Collection (Manassas, VA, USA). HepG2 and Hep3B were maintained in minimum essential medium (MEM, Invitrogen Life Technologies, Carlsbad, CA, USA) with 10% fetal bovine serum (FBS, Invitrogen Life Technologies) and antibiotics (100 U/ml of penicillin and 100 μg/ml streptomycin, Invitrogen Life Technologies). Chang liver cells were maintained with RPMI-1640 (Invitrogen Life Technologies) supplemented with 10% FBS with antibiotics in humidified atmosphere of 37°C, 5% CO₂.

Each cell type (0.96×10⁴ cells/well) was seeded in Falcon 96-well plate for MTT assay, which measures mitochondrial activity in viable cells. This method is based on the conversion of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich Co. LLC) to MTT-formazan crystals by mitochondrial enzyme. Cells were grown overnight, and the media were replaced with fresh media, treated with each compound at various concentrations, and incubated for 48 h. Control groups were treated with DMSO, equal to the highest percentage (<0.1%) of solvent used in experimental conditions for growth inhibition and MTT assay. After 48 h the media were replaced with serum-free media. MTT was freshly prepared at 5 mg/ml in phosphate-buffered saline (PBS, Sigma-Aldrich Co. LLC) and passed through a 0.2-μm pore-size filter. An aliquot of 100 μl of MTT stock solution was added to each well, and the plate was incubated at 37°C for 4 h in humidified 5% CO₂ incubator. After 4 h media were removed. Ethanol-DMSO (1:1 mixture solution) (200 μl) was added per well in order to solubilize the formazan. In addition, after 10 min the optical density of each well was measured with a spectrophotometer equipped with a 560-nm filter. Proliferation rate was calculated from 4 wells using percentage of control.

The treated cells were detached using trypsin/EDTA (Invitrogen Life Technologies), washed with PBS and fixed in 75% ethanol at 4°C for 30 min. Prior to analyses, cells were washed again with PBS, resuspended in cold PI solution (PI in PBS, 50 μg/ml) and incubated at room temperature in the dark for 30 min. Before analysis cell suspensions were filtered with 40-μm pore nylon mesh for removing debris. Flow cytometry analyses were performed on a FACScan (Becton-Dickinson, San Jose, CA, USA).

Cells were harvested and washed twice in PBS at 4°C. Total cell lysates were lysed in lysis buffer [40 mM Tris (pH 8.0), 120 mM NaCl, 0.5% NP-40, 0.1 mM sodium orthovanadate, 2 μg/ml aprotinin]. The supernatant was collected and protein concentrations were then measured with protein assay reagents (Pierce, Rockford, IL, USA). Equal amount of proteins were boiled for 3 min and chilled on ice, subjected to 10–12.5% SDS-PAGE, and electrophoretically transferred to a nitrocellulose membrane. The blotting membrane was blocked with PBS/0.1% Tween-20 containing 10% skim milk for 1 h. Antibodies specific for p21^WAF1/CIP1, p53, p27^KIP1, B-cell lymphoma 2 (Bcl-2), cyclin E, CDK2, caspase-3, Fas, Fas-ligand (FasL), c-Myc, Bcl-2 associated X protein (Bax), poly(ADP-ribose) polymerase (PARP), and Smad 4 were purchased from Santa Cruz Biotechnology Inc. (Dallas, TX, USA). Antibody against c-Jun was obtained from BD Biosciences Pharmingen (San Diego, CA, USA). Monoclonal antibody to β-actin (Sigma-Aldrich Co. LLC) was used as an internal control. Horseradish peroxidase (HRP)-labeled donkey anti-rabbit immunoglobulin and HRP-labeled donkey anti-goat immunoglobulin were purchased from Santa Cruz Biotechnology. The HRP-labeled sheep anti-mouse immunoglobulin was from GE Healthcare Life Sciences (Piscataway, NJ, USA). The proteins were visualized with the enhanced chemiluminescence (ECL) detection system (GE Healthcare Life Sciences).

Total RNA was prepared using RNAzol (Teltest, Finewood, TX, USA). The reverse transcriptase (RT) reaction was carried out with SuperScript II reverse transcriptase (Life Technologies). The RT reaction mixture containing 1 μg total RNA, 100 pmol oligo (dT)₁₈, 0.1 mM dNTP mixture, 40 U RNasin ribonuclease inhibitor (Progmega, Madison, WI, USA), 5× First-strand buffer and 200 U SuperScript II reverse transcriptase (Life Technologies). The synthesis of cDNA was performed at 42°C for 1 h, followed by 15 min of heating at 70°C for inactivating enzyme. A final volume of 20 μl of reaction containing 1 μl of template cDNA from RT reaction, 20 pmol of sense and antisense primers, 2 μl of 10× PCR buffer, 0.5 mM of dNTP mixtur, and 1 U of AmpliTaq polymerase (PE Biosystem, Waltham, MA, USA) was carried out on a GeneAmp PCR System 2400 (PE Biosystem). The primer sequence is as follows (33); transforming growth factor β1 (TGF-β1) sense 5′-GCCCTGGACACCAACTATTGCT-3′, TGF-β1 antisense 5′-AGGCTCCAAATGTAGGGGCAGG-3′. PCR reaction performed was denature at 94°C for 40 sec, annealing at 64°C for 45 sec, and extension at 72°C for 45 sec, and 35 cycles were used for amplification. The amplified PCR products were electrophoresed on 2.5% agarose gels and visualized by ethidium bromide staining.

To examine the growth inhibitory potency of luteolin on hepatocellular cells, cell proliferation was determined by MTT assay. We also employed apigenin to examine cell proliferation since luteolin and apigenin were identified as active components in Ixeris sonchifolia Hance, and luteolin showed a significant potent effect when comparing their ability on growth inhibition (32). Luteolin was more potent than apigenin in all the cell lines tested in the present study (Fig. 1). In addition, HepG2 was the most sensitive to luteolin (Fig. 1A). IC₅₀ values of luteolin on HepG2 and Hep3B cells were ~9 and 55 μg/ml, respectively (Fig. 1A and B). On Chang liver cells, no significant antiproliferative effect was observed in either luteolin or apigenin treatment (Fig. 1C).

To determine whether luteolin’s growth inhibitory effect was caused by specifically perturbing cell cycle-related events, a set of experiment was performed to measure DNA content and the cell cycle distribution by flow cytometry analysis after staining with PI. Fig. 2 shows the relative percentages of HepG2 and Hep3B cells in each phase of the cell cycle, following a 12-h treatment with varying luteolin concentrations. In HepG2 cells, luteolin induced the accumulation of G1 phase of cell cycle in a concentration-dependent manner (Fig. 2, left panel). However, Hep3B cells showed increase of subG1 population at the same time-point (Fig. 2, right panel). These results suggest that the growth inhibition of HepG2 cells was the result of a G1 phase arrest and that of Hep3B cells occurred mainly through apoptosis.

Since luteolin arrested HepG2 cells in the G1 phase of the cell cycle, we determined the expression levels of cell cycle regulating factors involved in G1 boundary, such as cyclin E and Cdk2, by western blot analysis. As shown in Fig. 3A, the protein levels of cdk2 and cyclin E were decreased in a concentration-dependent manner. These data indicate that the growth inhibitory effect of luteolin in HepG2 cells are caused by downregulating cdk2 and cyclin E expression.

Because it has been reported that p53, a tumor suppressor, regulates a DNA damage-triggered G1 checkpoint by upregulation of CDK inhibitor p21^WAF1/CIP1 (34), we examined the expression patterns of p53 and p21^WAF1/CIP1 by luteolin-treatment. As shown in Fig. 3B, HepG2 cells treated with luteolin increased the expression of p53 in a concentration-dependent manner. Unlike p53, the concentrations of luteolin ranging from 7 to 12.5 μg/ml markedly increased the protein level of p21^WAF1/CIP1; however, this induction of p21^WAF1/CIP1 was decreased at 25 μg/ml luteolin (Fig. 3B). This result is consistent with our previous study which was performed by using luteolin from Ixeris sonchifolia Hance (32). Because p27^KIP1, another CDK inhibitor, is reported to arrest cells only at the G1 phase (35,36), we next determined the effect of luteolin on p27^KIP1 expression. HepG2 cells treated with luteolin showed similar expression pattern of p27^KIP1 compared with that of p21^WAF1/CIP1 although there was no prominent increase at 12.5 μg/ml luteolin (Fig. 3B). These results suggest that other factors besides p53 may be involved in luteolin-induced G1 arrest as well as growth inhibition in HepG2 cells.

To further confirm our hypothesis that other mechanisms are involved in luteolin-induced HCC cell growth inhibition, we examined the effect of luteolin on p53, p21^WAF1/CIP1, and p27^KIP1 expressions using P53-deleted HCC Hep3B cells. As expected, no p53 expression was observed in Hep3B cells (Fig. 3C). However, treatment of Hep3B cells with luteolin resulted in a concentration-dependent increase in the expression of p21^WAF1/CIP1 and p27^KIP1 (Fig. 3C). These results from HepG2 and Hep3B cells suggest that the mechanism(s) other than p53 may be involved in the upregulation of p21^WAF1/CIP1 by luteolin.

Transforming growth factor β1 (TGF-β1) is an essential regulator of cellular processes including proliferation, differentiation, migration, cell survival and angiogenesis. TGF-β1 has been reported to exert its function via specific receptors and intracellular Smad transcription factors. Phosphorylation of receptor-activated Smads, such as Smad2 or Smad3, leads to formation of complexes with the common mediator Smad (Smad4), which are imported to the nucleus, induce cycle-dependent kinase inhibitors, and then lead to G1 arrest (37). Therefore, TGF-β1 is known as an upstream G1 arrest signal (34,36). Whether luteolin affects the expression of TGF-β1 and Smad4 in HCC cells was examined. The mRNA level of TGF-β1 was gradually increased by luteolin treatment in HepG2 cells (Fig. 4A). Treatment with various concentrations of luteolin also induced the mRNA level of TGF-β1 in Hep3B cells (Fig. 4B).

We next tested the effect of luteolin on Smad4 expression in HCC cells. Results show that Smad4 expression level in HepG2 cells slightly increased at 12.5 μg/ml and decreased at 25 μg/ml concentration of luteolin (Fig. 4C). In Hep3B cells, luteolin treatment increased Smad4 expression concentration-dependently (Fig. 4D).

Because TGF-β1 signaling is reported to activate Fas-mediated apoptotic pathways (38), we investigated whether luteolin affects the expression of Fas and its ligand FasL in HCC cells. Fig. 4C shows that luteolin upregulated FasL in a concentration-dependent manner, but Fas increased up to 12.5 μg/ml, then decreased at 25 μg/ml in HepG2 cells. We also observed similar expression pattern of FasL in Hep3B cells although Fas was not significantly altered by luteolin (Fig. 4D).

To confirm the contribution of Fas/FasL and p53 on luteolin-induced apoptosis, we investigated whether the expression of apoptotic protein Bax, Bcl-2, caspase-3 and PARP, were modulated by luteolin. Luteolin treatment increased Bax and slightly decreased Bcl-2 expression in a concentration-dependent manner in HepG2 cells (Fig. 5A). Luteolin also increased Bax expression in Hep3B cells (Fig. 5B). In contrast, Bcl-2 level was increased simultaneously in Hep3B cells by luteolin (Fig. 5B). To elucidate the mechanism of Bcl-2 upregulation by luteolin in Hep3B cells, we determined the effect of luteolin on oncogene expression, such as c-Jun and c-Myc. An increase of c-Myc and c-Jun expression was observed in HepG2 cells (Fig. 5A); however, the reduction of these proteins was found in Hep3B cells (Fig. 5B).

Finally, the pro-caspase-3 level and PARP cleavage were measured in luteolin-treated HCC cells. In the presence of lutein, the increase of cleavage forms of caspase-3 and PARP were observed in HepG2 cells (Fig. 5A). Similarly, PARP cleavage was detected in Hep3B cells in a concentration-dependent manner and pro-form of caspase-3 decreased slightly (Fig. 5B).