The HCC MHCC97 cell line was purchased from Shanghai Xiangf Biotechnology Co., Ltd. (Shanghai, China). The MHCC97 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Hangzhou Sijiqing Biological Engineering Materials Co., Ltd, Hangzhou, China), 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen Life Technologies, Carlsbad, CA, USA), and incubated in an atmosphere of 5% CO₂ at 37°C. Casticin was purchased from Chengdu Biopurify Phytochemicals Ltd. (Chengdu, China, dissolved in dimethyl sulfoxide (DMSO) as a 10-mmol/l stock solution, and diluted in a medium to the indicated concentration. MTT and lithium chloride were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trypsin and DMSO were purchased from Amersco Company (Solon, OH, USA). Mouse anti-human β-catenin, cyclin D1, β-actin antibodies and horseradish peroxidase-conjugated rabbit anti-mouse secondary antibody were all purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA).

Cell sorting was performed with MHCC97 cells using the cell surface phenotype, CD133⁺, through magnetic activated cell sorting (MACS) separation columns (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. Cells were trypsinized and washed with phosphate-buffered saline (PBS) and suspended in PBS containing 0.5% bull serum albumin (BSA). FcR Blocking Reagent (100 μl; anti-CD133 antibody), 100 μl CD133-conjugated MicroBeads (AC133 Cell Isolation Kit, Miltenyi Biotec), and 10⁸ cells were subsequently added to the sample and incubated in parallel for 30 min on ice. After washing the cells, CD133-positive and -negative fractions were isolated through MACS separation columns. The CD133⁺ and parental cells were collected and washed to remove serum, and suspended in serum-free DMEM/F12, which was supplemented with 20 ng/ml human recombinant epidermal growth factor (EGFR), 20 ng/ml human recombinant basic fibroblast growth factor, 2% B27 supplement without vitamin A, 0.4% BSA, 4 ng/ml insulin, 100 IU/ml penicillin and 100 μg/ml streptomycin. The single-cell suspensions were suspended at a density of 2,000 cells/ml in stem cell-conditioned medium and seeded into ultra-low attachment six-well plates (Corning, Inc., Corning, NY, USA). When the spheroid diameter reached 50 μm, the suspension cultures were passaged every six days. Colonies were counted in 10 different views under a microscope (IX71, Olympus, Tokyo, Japan). The volume of the spheroids (μm³) was estimated using the following formula: V=(4/3) πR³, where R denotes radius. The experiments were repeated three times in duplicate.

The parental cells, and sorted CD133⁺ and CD133⁻ cells were resuspended in PBS, sub-packaged in Eppendorf tubes (density, 1×10⁵ cells/ml) and incubated directly with the conjugated monoclonal antibodies, mouse anti-human CD133-R-phycoerythrin (PE) and mouse IgG2b isotype control-PE for 30 min at 4°C in the dark. The fluorescence value was measured by FCM with 10,000 cells per tube.

The CD133⁺ SFCs of the MHCC97 cell line were collected by gentle centrifugation at 80 × g (TL-5-A, Jintan Shenglan Instrument Manufacturing Co., Ltd., Jintan, China), dissociated with trypsin-EDTA and mechanically disrupted using a pipette. The resulting single cells were centrifuged to remove the enzyme, resuspended in a stem cell-conditioned culture medium and allowed to reform spheres. The tumorspheres were passaged every six days until reaching a diameter of 50 μm. The dissociated single SFCs were diluted to a density of 500 cells/ml, the diluted cell suspension was plated onto an ultra-low attachment 96-well plate (Corning Inc.) with 2 μl/well of serum-free medium (150 μl). The wells containing only one cell were marked, observed and photographed with an inverted microscope (IX71, Olympus) daily for approximately nine days.

To examine the effects of casticin on sphere formation, the resulting single-cell suspension, with a density of 2×10³ cells/ml, was plated onto an ultra-low attachment six-well plate supplemented with serum-free medium at the same volume as the primary tumorsphere formation experiment (primary experiment). As a second experimental process the density was altered to ~1×10³ cells/ml. In the primary experiment the medium was supplemented with various concentrations of casticin, however, this was not the case in the second experiment.

In order to assess the effect of the Wnt/β-catenin pathway on the formation of sub-tumorspheres, dissociated MHCC97 CD133⁺ SFCs were treated with a culture medium containing casticin-LiCl (0 μmol/l, 20 mmol/l), casticin-LiCl (1 μmol/l, 0 mmol/l), casticin-LiCl (1 μmol/l, 10 mmol/l), casticin-LiCl (1 μmol/l, 20 mmol/l) or control, 0.1% DMSO respectively, for 24 h and the formation of sub-tumorspheres was observed.

Twenty pathogen-free male Balb/c-nu mice (age, 5–6 weeks) were purchased from the Animal Institute of the Chinese Academy of Medical Science. The animal studies were performed in accordance with standard protocols approved by the Ethics Committee of Hunan Normal University and the Committee of Experimental Animal Feeding and Management (Changsha, China). The mice were randomly divided into five groups (n=4 per group) and maintained under standard conditions, according to typical protocols. The cells were suspended in a serum free-DMEM/Matrigel (BD Biosciences, Franklin Lakes, NJ, USA) mixture (1:1 volume). The mice were inoculated with different quantities of CD133⁺ SFCs (5×10², 1×10³, 5×10³, 1×10⁴ and 5×10⁴ cells) in one flank, and unsorted MHCC97 cells (5×10⁴, 1×10⁵, 2×10⁵, 5×10⁵ and 1×10⁶ cells) in the other. Tumorigenicity experiments were terminated two months after cell inoculation. Tumor size was measured using a caliper and the volume was calculated as follows: V (mm³) =L × W² × 0.5, where L denotes length and W denotes width. The harvested tumors were photographed and weighed immediately. Specimens from tumor tissue samples were fixed in 10% neutral-buffered formalin, processed in paraffin blocks and sectioned. The sections were stained with hematoxylin and eosin (H&E) and examined under an inverted microscope (IX71, Olympus).

CD133⁺ SFCs or parental MHCC97 cells were seeded in 96-well plates pre-coated with 0.6% agarose at a density of 5,000 cells/well as described previously (19). One day after plating, 8-bromo-7-methoxychrysin of different concentrations was added to each well and cultured for 48 h at 37°C. Following removal from the medium, the cells were incubated with 5 mg/ml MTT for 4 h. The cells were extracted with acidic isopropanol and the absorbance at a 570-nm wavelength (A570) was measured using an enzyme-labeling instrument (ELx800 Absorbance Microplate Reader type, BioTek Instruments, Inc., Winooski, VT, USA). The relative cell proliferation inhibition rate was calculated as follows: Average A570 of the experimental group/average A570 of the control group × 100%.

The preparation of whole cell lysates and western blot analysis were performed as previously described (19). Mouse anti-human β-catenin, cyclin D1 and β-actin antibodies served as primary antibodies. The signals were visualized using a chemiluminescent substrate (enhanced chemiluminescence; Amersham Life Science, Arlington Heights, IL, USA) and β-actin served as an internal control. Images were scanned and densitometry analysis was performed with a UN-SCAN-IT graph digitizer (Silk Scientific, Inc., Orem, UT, USA).

In order to assess the effect of LiCl attenuated during the casticin-induced downregulation of β-catenin or cyclin D1 protein expression, dissociated MHCC97 CD133⁺ SFCs were treated with a culture medium containing casticin-LiCl (0 μmol/l, 20 mmol/l), casticin-LiCl (1 μmol/l, 0 mmol/l), casticin-LiCl (1 μmol/l, 10 mmol/l), casticin-LiCl (1 μmol/l, 20 mmol/l) or control, 0.1% DMSO, respectively, for 24 h and the expression of β-catenin or cyclin D1 was observed.

The data are expressed as means ± SD and the data were analyzed with SPSS software, version 15.0 (SPSS, Inc., Chicago, IL, USA). In addition, one-way analysis of variance was performed. After the equal check of variance, two-two comparisons of the means between the test and control groups were performed using the least-significant difference method; or Dunnett’s test was used. P<0.05 was considered to indicate a statistically significant difference.

CD133 is classified as a CSC marker in HCC (10). Therefore, the CD133⁺ subpopulation was sorted from the MHCC97 cell line using MACS and cultured in vitro. The expression of the stem cell marker, CD133, was examined by FCM. The subpopulation of CD133⁺ cells showed a high purity of 56.26±2.34% compared with the purity of CD133⁻ (1.04±0.27%) and parental cells (3.32±0.38%; Fig. 1A). To establish long-term cultures enriched in stem cells from sorted CD133⁺ cells, the tumorsphere formation assay in a stem cell-conditioned medium was performed. The spheroids from CD133⁺ and parental cells were obtained after six days of culture (Fig. 1B). The CD133⁺ subpopulation exhibited a greater quantity of tumorsphere formation and increased size compared with the parental cells (Fig. 1B and C). These findings indicate the existence of LCSCs in sorted CD133⁺ cells and that LCSCs are highly enriched in CD133⁺ tumor-forming cells.

To further investigate stem-cell properties and the function of CD133⁺ SFCs, self-renewal capacity and tumorigenic potential were analyzed. The capacity of single cells (obtained from CD133⁺ dissociated spheres) to form secondary tumorspheres was measured. Within nine days of culture, new spheroids of growing undifferentiated CD133⁺ cells were observed (Fig. 1D). Thus, the in vitro CD133⁺ SFCs from the MHCC97 cell line demonstrated a self-renewing capacity. Furthermore, the tumorigenic potential of CD133⁺ SFCs of the MHCC97 cell line was investigated in Balb/c-nu mice. Our findings demonstrated that ≤2×10⁵ parental cells were required to initiate stable tumor formation 37 days after inoculation. By contrast, only 1×10³ CD133⁺ SFCs were sufficient to generate visible tumors 27 days after inoculation (Table I). These data indicate that CD133⁺ SFCs of the MHCC97 cell line have a greater tumerogenic capacity compared with parental cells in vivo. Additionally, H&E staining revealed histological characteristics in tumor xenografts, which were derived from CD133⁺ SFCs, were similar to those of the parental cells (Fig. 1E). Collectively, these data demonstrate that CD133⁺ SFCs possess an ability to self-renew in vitro and initiate tumor growth in vivo, indicating that the CD133⁺ SFCs may provide a true representation of LCSCs in the HCC MHCC97 cell line.

CD133⁺ SFCs and parental cells were treated with different concentrations of casticin (0.1, 0.3, 1.0, 3.0 and 10.0 μmol/l) to examine its effect on cell viability of LCSCs using the MTT assay. Casticin preferentially inhibited cell viability of CD133⁺ SFCs derived from MHCC97 cells in a dose-dependent manner (Fig. 2A). The half maximal inhibitory concentration of the parental cells and the CD133⁺ SFCs was 17.9 and 0.5 μmol/l, respectively (Fig. 2A).

In order to evaluate whether casticin suppresses the self-renewal of LCSCs derived from the MHCC97 cell line in vitro, the primary tumorspheres were treated with various concentrations of casticin, followed by drug removal and culturing in another passage to form secondary spheres. Treatment with casticin resulted in a decreased number of tumorspheres in LCSCs (Fig. 2B) and a decreased number of secondary tumorspheres; these findings are consistent with the reduced self-renewal capacity of LCSCs by casticin treatment (Fig. 2C).

The Wnt/β-catenin signaling pathway is a well-established and significant regulator of stem cell self-renewal. Wnt/β-catenin signaling has been implicated in the maintenance of CSCs that are present in liver cancer (21). The expression level of the stem cell signal molecule, β-catenin, and its downstream target molecule, cyclin D1, were measured following casticin treatment in LCSCs and parental cells. Western blot analysis revealed that β-catenin and cyclin D1 were highly expressed in LCSCs compared with the parental cells. Additionally, casticin treatment (0.1, 0.5 and 1.0 μM) resulted in a significant decrease in β-catenin and cyclin D1 expression in LCSCs (Fig. 3).

The role of β-catenin in maintaining the self-renewal characteristics of LCSCs was investigated. LCSCs were treated with lithium chloride, an agonist known to activate the Wnt/β-catenin pathway. The addition of lithium chloride resulted in the upregulation of β-catenin and cyclin D1 in LCSCs. In addition, lithium chloride antagonized the inhibitory effects of casticin on the self-renewal of LCSCs and attenuated the casticin-induced downregulation of β-catenin and cyclin D1 expression in LCSCs (Fig. 4).