The aerial part of Cimicifuga dahurica (Turcz) Maxim (synonyms: Actinospora dahurica Turczaninow ex Fischer & C. A. Meyer, Index Sem. Hort. Petrop. 1: 21. 1835; Actaea dahurica (Turczaninow ex Fischer & C. A. Meyer) Turczaninow ex Fischer & C. A. Meyer) was collected in Maojingba, Kalaqin Qi, Inner Mongolia Autonomous Region, China, in September 1999, and was identified by Prof. Ruile Pan of the Institute of Medicinal Plant Development, Chinese Academy of Medical Sciences and Peking Union Medical College. A voucher specimen has been deposited in the Herbarium of the Institute (XA99-09). The powdered aerial part of the plant (14.5 kg) was extracted exhaustively with 10 folds volume of 80% ethanol under refluxing for three times, one hour each time. Following combination and filtering, the solvent was evaporated under vacuum to obtain the crude extract (2.0 kg). Then the crude extract was mixed with siliceous earth and eluted with ethyl acetate. Removal of the solvent in vacuo, the TGA was (210 g) obtained.

Twenty nine triterpene glycosides (Table 1) including 25-anhydrocimigenol-3-O-β-D-xylopyranoside, 23-O-acetylcimigenol-3-O-β-D-xylopyranoside, 24-O-acetylcimigenol-3-O-β-D-xylopyranoside, and 25-O-acetylcimigenol-3-O-β-D-xylopyranoside, cimigenol xylopyranoside together with ferulic acid and isoferulic acid were isolated from TGA. The total triterpene glycosides content in TGA was estimated by a colorimetric method as we described previously 14, with slight modifications. A 50-μl aliquot of TGA methanol solution (1.50 mg/ml) was diluted with 1 ml water and then applied to a Waters Oasis HLB cartridge, which was preconditioned by rinsing with 1 ml methanol and followed by 1 ml water. The cartridge was washed with 2 ml water to remove carbohydrate compounds for interference, and then the triterpenes were eluted with 2 ml methanol from the cartridge to a clean glass tube. After drying by a gentle stream of nitrogen, a 0.2-ml aliquot of 5% vanillin acetic acid (w/v) and a 0.8-ml of aliquot of perchloric acid were added to the residue in the tube. Then the tube was kept in a 80°C water bath for 15 min. After cooling with water, the absorbance of the mixture was determined at 544 nm. The assay was conducted in triplicates. The total triterpene glycosides content was 72.67% ± 2.03 expressed as cimigenol xylopyranoside equivalents. Usually the total triterpene glycosides content in C. racemosa is calculated as 27-deoxyactein equivalent 15; However since 27-deoxyactein was not isolated from TGA, cimigenol xylopyranoside, one of the main components in TGA and many other Cimicifuga plants was used as a standard.

HepG2 (ATCC, Rockville, MD) cells were maintained in RPMI 1640 containing 10% FBS (Gibco, BRL, Carlsbad, CA), 2 mg/ml sodium bicarbonate, 100 μg/ml penicillin sodium salt and 100 μg/ml streptomycin sulfate. Cells were grown to 70% confluence, trypsinized with 0.25% trypsin-2 mM EDTA, and plated in 96 well plates. Mouse hepatocytes were isolated from normal CD-1 (ICR) mice (Beijing Vital Laboratory Animal Technology, Beijing, China) with enzymatic perfusion technique as we described previously 13. The viability of the mouse hepatocytes, tested with Trypan blue was about 80%. In all experiments, cells were grown in RPMI-1640 medium with 10% FBS for 24 h prior to treatment.

TGA was dissolved in DMSO at a concentration of 250 mg/ml, then diluted in tissue culture medium and filtered before use. The final concentration of DMSO (0.1%) did not affect the cell viability.

1.5 × 10⁴ HepG2 cells and 8 × 10³ mouse hepatocytes were seeded in 96 well plates and treated with TGA or vehicle (0.1% DMSO) at various concentrations and incubated for 48 h, followed by MTT (3- [4, 5-dimethylthiazol-2-yl]-2, 5-diphenyltetrazolium bromide) assay 16. Briefly, IC₅₀ of the TGA in HepG2 cells and normal mouse hepatocytes were derived from the dose-response curves.

AO/EB (acridine orange/ethidium bromide) fluoresce staining method was used to observe the apoptosis morphological changes 17. Briefly, HepG2 cells were cultured in 3.5 cm dishes and treated with TGA at concentration of 50 μg/ml for 0, 12, 24 and 48 h respectively. After treatment, all the cultures were incubated at 37°C, 5% CO₂ for the indicated time. Photographs were taken under an inverted Leica fluorescence 40 × 10 microscope after staining.

Apoptosis was quantified by detecting surface exposure of phosphatidylserine in apoptotic cells using Annexin V-FITC/PI (propidium iodide) apoptosis detection kit (BD Biosciences Clontech). Cells were seeded in 3.5 cm dishes in 1 ml medium and incubated with TGA at the dose of 25, 50 and 100 μg/ml for 24 h, respectively. The adherent and floating cells were combined and treated according to the manufacturer's instruction and measured with FITC/PI staining using flow cytometry (Becton Dickinson, San Jose, CA). Apoptotic cells (annexin V⁺PI^-) were differentiated from necrotic cells (annexin V⁺PI⁺, including apoptotic cells at late stage).

HepG2 cells were treated with TGA at different concentrations (25, 50 and 100 μg/ml for 48 h) and time points (at 50 μg/ml for 12, 24 and 48 h). Then cells were collected and fixed in 70% cold ethanol (-20°C) overnight. After washing twice with PBS, cells were resuspended in PBS. RNase A (0.5 mg/ml) and PI (2.5 μg/ml) were added to the fixed cells for 30 min. The DNA content of cells was then analyzed with a FACSCalibur instrument (Becton Dickinson, San Jose, CA).

After treatments, cells were washed three times with ice-cold PBS and lysed with lysis buffer (50 mM Tris-HCl, pH 7.4, 10 mM EDTA, 1% Triton X-100, 26% urea, and 1 tablet/10 ml protease inhibitor cocktail tablets). Sticky DNA was removed from lysates with a sterile toothpick. The protein concentration of the supernatant was determined by the Bradford method. The lysates were subjected to electrophoresis on a 10 % SDS-polyacrylamide gel and then transferred to a nitrocellulose membrane 18. The nitrocellulose membrane was then incubated with mouse monoclonal anti-Bcl-2 and anti-Bax antibody (Santa Cruz Biotechnology, Santa Cruz, CA; sc-509 and sc-7480). Mouse monoclonal β-actin (Lab Vision, Fremont, CA) was used as an internal control. Secondary antibody to IgG conjugated to horseradish peroxidase was used. The blots were probed with the ECL Western blot detection system according to the manufacturer's instructions. The ratio of Bax/Bcl-2 was analyzed by pImage tool.

Male CD-1 (ICR) mice (Beijing Vital Laboratory Animal Technology, Beijing, China), weighing 20–22 g, were used for implantation of hepatoma H₂₂ cells (s.c.), which was maintained by weekly (i.p.) passage in CD-1 (ICR) mice. Ascites (0.2 ml of 1:6 dilution) from tumor-bearing mice 7 days after tumor inoculation were implanted (s.c.) into the armpit region of mice. Ten mice each were treated with either TGA (200,100 and 50 mg/kg b.w., i.g.) or vehicle, once a day for 10 days, 24 h after tumor inoculation. Cyclophosphamide (15 mg/kg b.w., i.p.) was used as a positive control. The tumor inhibition rate (TIR %) was calculated as we described previously 19. All the animal procedures were conducted in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

One-way ANOVA was used and followed by Dunnett's test. p < 0.05 was considered significant.

The cytotoxicity of TGA was evaluated by an MTT assay and the IC₅₀ values were derived from the dose-response curves (Fig 1). A 48 h exposure to TGA decreased the proliferation of HepG2 cells in a concentration-dependent manner with an IC₅₀ at 21 μg/ml; however, the effect on normal mouse hepatocytes showed bi-directional property: the proliferation was inhibited at higher concentrations, but promoted at lower concentrations, with an IC₅₀ at 105 μg/ml. This indicates that TGA may possess relative selective cytotoxicity to hepatoma cancer cells.

Individual apoptosis in the cell population in HepG2 cells treated by TGA was studied by fluorescence staining method. Morphological alteration, such as chromatin aggregation, nuclear and cytoplasmic condensation, and partition of cytoplasm and nucleus into membrane-bound vesicles (apoptotic bodies) were observed in HepG2 cells treated by TGA at 50 μg/ml for 12, 24 and 48 h (Fig 2).

To further confirm that TGA induces HepG2 cell apoptosis, cells were stained with Annexin V-FITC and PI, and then subsequently analyzed by flow cytometry. This assay is based on the translocation of phosphatidylserine from the inner leaflet of the plasma membrane to the cell surface in the early apoptotic cells. HepG2 cells were treated with TGA at 25, 50 and 100 μg/ml for 24 h. The dual parameter fluorescent dot plots showed the viable cell population in the lower left quadrant (annexin V^- PI^-), the cells at the early apoptosis are in the lower right quadrant (annexin V⁺PI^-), and the ones at the late apoptosis are in the upper right quadrant (annexin V⁺I⁺). As indicated in Fig. 3, in untreated cells, 0.29% of cells were Annexin V-positive/PI-negative, whereas 1.02% of cells were Annexin V/PI double positive. After treatment with TGA at 25, 50 and 100 μg/ml for 24 h, the corresponding quantities were 0.75 and 1.76%; 9.05 and 0.93%; 13.23 and 6.09% respectively.

The ability of a substance to affect specific phases of the cell cycle may provide clues to its mechanism of action. To determine the effects of TGA on the cell cycle, HepG2 cells were treated with TGA at different concentrations (25, 50 and 100 μg/ml) and time points (at 50 μg/ml for 12, 24 and 48 h) respectively. The cells were then stained with PI and analyzed DNA content by flow cytometry. After exposure to 25 μg/ml of the TGA, there was an increase of cells in G₀/G₁ when compared to the DMSO solvent control and a concomitant decrease of cells in S and G₂/M phases. After treatment with 50 and 100 μg/ml of TGA, there was a decrease of cells in G₀/G₁ and an increase of cells in G₂/M. Moreover, the sub-G₁ apoptotic peak was induced by TGA in dose- and time- dependent manners (Fig 4). This indicates that TGA contains more than one component with the more active or abundant component inducing G₀/G₁ arrest and the less active component inducing G₂/M arrest and/or individual component(s) in TGA exert different effects at different concentrations.

The raise of the ratio of Bax/Bcl-2 is of benefit to apoptosis. In light of our study, following treatment with TGA at 50 μg/ml, pro-apoptotic protein Bax expression was up regulated in a time-dependent manner; the anti-apoptotic protein Bcl-2 was down regulated at 12 h and 24 h time points, but slightly up regulated at 48 h time point. At all events, the ratio of Bax/Bcl-2 was increased during all time points compared with control. Although the ratio of Bax/Bcl-2 at 48 h time point was decreased in some degree than 24 h time point, it was still far higher than that of control (Fig 5). The reason of this might be due to more apoptosis at 48 h in HepG2 cells negatively feedback to inhibit the ratio of Bax/Bcl-2 via enhancing the expression of anti-apoptotic protein.

After tumor implantation for 24 h, administration of TGA (200, 100 or 50 mg/kg b.w., i.g) and cyclophosphamide (15 mg/kg B.W., i.p) once a day for 10 days, could significantly suppress the growth of H₂₂ tumor and TGA at 200 mg/kg was more effective than the lower dosages. One-way ANOVA followed by Dunnett's test was used for statistic analysis and significant differences were found for treatment groups vs. control (Table 2). In addition, marked body weight loss was observed in cyclophosphamide-treated group compared to the control group, whereas only slight body weight loss was observed in TGA-treated groups. This implies that TGA might be a promising antitumor agent with low toxicity.