Human astrocytoma CRT-MG cells were grown in RPMI 1640 medium that was supplemented with 10% heat-inactivated fetal bovine serum (FBS), penicillin G (100 U/ml), streptomycin (100 µg/ml) and L-glutamine (2 mmol/L) in a 5% CO₂ incubator at 37℃, as previously described (10). Other human astrocytoma cell lines, U251-MG and U87-MG cells, were maintained in Dulbecco's modified Eagle media (JBI, Korea) that was supplemented with 10% FBS and penicillin G (100 U/ml). Primary human fetal astrocytes were obtained from therapeutically aborted fetal brains and they were maintained in Dulbecco's modified Eagle media that was supplemented with 10% heat-inactivated fetal bovine serum (FBS), penicillin G (100 U/ml) and 1% nonessential amino acids (Gibco-BRL, Grand Island, NY), as previously described (11).

Ginseng saponin ginsenosides (F₁, R_o, R_c, R_e, R_d, R_f, C-K, Rh₂, Rg₁, Rg₂, Rg₃, Rb₁ and Rb₂) were obtained from KT&G (Daejeon, Korea). N-acetyl cysteine (NAC), 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) and diphenyl iodonium (DPI) were all purchased from Sigma (St. Louis, MO). Dichlorodihydrofluorescein diacetate (DCF-DA) and tetramethylrho-damine ethyl ester (TMRE) were purchased from Molecular Probe (Eugene, OR). An agonistic IgM type anti-Fas antibody (CH-11) was obtained from Upstate Biotechnology (Lake Placid, NY). Human recombinant TNF-α and Fas ligand were purchased from R&D Systems (Minneapolis, MN). Caffeic acid phenethyl ester (CAPE) and SB202190, SP100625 and U0126, which are pharmacological inhibitors of p38 MAPK, JNK and ERK, respectively, were obtained from Calbiochem (La Jolla, CA).

To detect intracellular ROS, an oxidation-sensitive probe 2, 7-dichlorofluorescein-diacetate (DCF-DA) was used as previously described (9). To study the time course of Fas-mediated ROS production, the cells were incubated with 2 µmol/L of DCF-DA for 10 min and then they were treated with CH-11 (500 µg/L) for varying time periods (0~60 min). To investigate the effect of NAC and CAPE on Fas-induced ROS production, the CRT-MG cells were maintained in serum-free media for 16 h, incubated in the absence or presence of these inhibitors for 1 h and then they were treated with CH-11 (500 µg/L) for an additional 15 min. The cellular fluorescence was measured using an inverted epifluorescence microscope (Zeiss, Germany).

Cell death was determined by staining with Annexin V (PharMingen), which is a 35.8-kDa protein that has a strong affinity for phosphatidylserine (12). The cells were washed with PBS, trypsinized, suspended in 200 µL of binding buffer and stained with 0.5 ng of Annexin V-fluorescein isothiocyanate (FITC) and 2.5 ng of propidium iodide (PI). Ten thousand cells were analyzed with using a FACStar (Becton Dickinson, Mountain View, CA) within 30 min after staining. Cell death was defined as those cells that were positive for Annexin V and/or PI. The mitochondrial transmembrane potential (Δψ_m) was assessed after staining with TMRE. The cell suspension was incubated in 2 µmol/L of TMRE for 30 min at room temperature and then it was analyzed with a FACStar.

The MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) assay was also used to determine cell viability, as described previously (13). Briefly, MTT is a yellow, water-soluble, tetrazolium salt. Metabolically active cells are able to convert this dye into a water-insoluble, dark-blue formazan by reductive cleavage of the tetrazolium ring. The formazan crystals can then be dissolved and quantified by measuring the absorbance of the solution at 595 nm, and the resultant value is related to the number of living cells.

The cells were washed with ice-cold PBS and then vortexed in the hypotonic buffer solution that contained 500 mmol/L sucrose, protease inhibitors and Na₃VO₄ (10 mmol/L Hepes, 2 mmol/L MgCl₂, 25 mmol/L KCl, 0.5 mmol/L EDTA, 0.5 mmol/L EGTA). Then NP-40 was added to the final concentration of 0.05%, and the mixture was centrifuged at 15,000g for 15 min. The supernatants were assayed for the cytosolic fractions, and the precipitants were analyzed for the mitochondrial fractions.

Immunoblot analysis for caspases was performed as previously described (14). The cell lysates (20 µg) were electrophoresed in 10% SDS-PAGE gels, the proteins were transferred to nitrocellulose membranes and these were probed with antibodies against human caspase-3, PARP, p38 MAPK, phosphor-specific p38 MAPK and β-actin (Cell Signaling, Danvers, MA). The blots were developed by chemiluminescence (AbFrontier, Korea).

The data is presented as means ±SDs. Comparisons between two different samples were analyzed by the Student t-test, and comparisons between more than three different samples were done by ANOVA (analysis of variance) with applying Tukey's post-hoc test to the significant main effects or interactions (SPSS 12.0K for Windows, SPSS, Chicago, IL).

We screened thirteen ginsenosides for their anti-cancer effect on human malignant astrocytoma cells (CRT-MG, U87-MG and U251-MG cells). Among the 13 ginsenosides, compound K (C-K) and Rh₂ induced marked time-dependent cytotoxicity even at a concentration as low as 25 mg/L, while the other ginsenosides showed marginal cytotoxic effects at a concentration as high as 50 mg/L (Fig. 1A). The cytotoxic effect of C-K and Rh₂ was only observed for the transformed tumor cells, but not for the primary cultured human astrocytes even at higher concentrations of C-K and Rh₂ (Fig. 1B). Interestingly, the cells treated with C-K showed vacuolar changes even with using concentrations that had little toxicity. To test what type of cell death was induced by ginsenosides, the cells treated with the ginsenosides were stained for Annexin V and propidium iodide (Fig. 1C). As expected, treatment with C-K and Rh₂ induced typical apoptotic cell death, which was confirmed by the positive staining with Annexin V. Pretreatment with a broad-spectrum caspase inhibitor z-VAD-Fmk markedly reduced the ginsenoside induced cell death, confirming that ginsenosides induced caspase-dependent apoptotic cell death in the malignant astrocytoma cells. Since it has been reported that the mitochondrial intermembrane potential (Δψ_m) is selectively decreased during the activation of the apoptotic cascades (15), we further investigated whether treatment with ginsenosides affected the Δψ_m by staining with TMRE (Fig. 1D). In concordance with the Annexin V staining results, treatment with C-K and Rh₂ induced a rapid depolarization of the Δψ_m, and this was nearly completely reversed by pretreatment with z-VAD-fmk. These results clearly indicate that the ginsenosides C-K and Rh₂ induced caspase-dependent apoptotic cell death in human malignant astrocytoma cells.

We next investigated the signaling pathways that are responsible for the ginsenoside-induced apoptotic cell death. Since we have observed that inhibition of caspases significantly suppressed the ginsenoside-induced cell death, we first tested the involvement of specific caspases by performing western blot analysis (Fig. 2A, B). Caspase-3 is known as the common executioner of apoptotic cell death, and it was cleaved into active fragments (17/19 kDa) in a time-dependent manner by treatment with C-K or Rh₂. We observed the proteolytic cleavage of caspase-3 as early as 12 h after treatment with ginsenosides, while the maximal cleavage was observed 24 h after treatment. The proteolytic cleavage of PARP (poly-ADP ribose polymerase), a well known cellular target of activated caspase-3, was also observed in a time-dependent manner after treatment with C-K (Fig. 2B). To further investigate the molecular mechanisms of ginsenoside-induced apoptosis, we further examined the cytosolic translocation of cytochrome-C (Fig. 2C). Treatment with C-K induced a marked translocation of mitochondrial cytochrome-C into the cytosolic fractions. These results collectively suggest that ginsenosides induce apoptotic cell death via the intrinsic mitochondrial pathways.

We further tried to investigate the intracellular signaling pathways that are responsible for the activation of the intrinsic apoptotic pathways. We examined the involvement of the mitogen activated kinase (MAPK) pathways for controlling the ginsenoside-induced apoptosis by using various specific pharmacological inhibitors for ERK1/2, p38 and JNK. Pretreatment with the p38-specific inhibitor significantly suppressed C-K-induced cell death, while the pretreatment with ERK or JNK inhibitors had no effect on C-K-induced cell death (Fig. 3A). The protective effect of p38 inhibitor was not observed for the Rh₂-induced cell death. The protective effect of p38 inhibitor on the C-K-induced cell death was also confirmed by Annexin V staining (data not shown). We next tested the involvement of p38 MAPK in the C-K-induced cell death by performing western blot analysis. As expected, treatment with C-K induced a time-dependent phosphorylation of p38 MAPK, which was sustained up to 18 h after treatment (Fig. 3B).

We next investigated whether pro-apoptotic ginsenosides can sensitize Fas-resistant astrocytoma cells to Fas-induced cell death by performing MTT assay. To test the effect of C-K on Fas-mediated cell death, the CRT-MG cells were incubated in the absence or the presence of varying concentrations of C-K for 1 h and then the cells were treated with increasing doses of CH-11 anti-Fas antibody for 24 h. As CRT-MG cells are known to be resistant to Fas-induced cell death (16), the cells did not undergo cell death upon Fas ligation in the absence of C-K (Fig. 4A). Yet pretreatment with increasing doses of C-K dramatically increased the sensitivity to Fas-mediated cytotoxicity in a dose-dependent manner. A similar synergistic effect on Fas-mediated cell death was observed by pretreatment with Rh₂ (data not shown). The synergistic effect of ginsenosides with fas ligation was also confirmed by Annexin V staining and the mitochondrial depolarization (Fig. 4B, C). We also observed that the inhibition of caspases suppressed the synergistic cytotoxicity induced by ginsenosides (data not shown).

We have previously reported that ROS scavengers such as NAC and CAPE enhanced Fas-induced cell death in a caspase-dependent manner (9). Since various ginsenosides are known for their effective ROS scavenging activities, we tested the possibility that C-K and Rh₂ can enhance Fas-induced apoptosis in a ROS-dependent manner. As reported previously (9), we were able to confirm that Fas ligation induced a time-dependent increase of the intracellular ROS levels (Fig. 5A). Because the maximal induction of the intracellular ROS levels was observed after 15 min of Fas ligation, we used that time point for the following experiments. In agreement with our hypothesis, the pretreatment with C-K or Rh₂ significantly reduced the intracellular ROS levels induced by Fas ligation (Fig. 5B). To confirm whether inhibition of ROS generation can enhance Fas-induced cell death via caspase-3 activation, we further tested the in vitro caspase-3 enzymatic activity of the lysates of the cells treated with CH-11 in the absence or presence of ginsenosides (Fig. 5C). Treatment with the ginsenosides alone had little effect on the caspase-3 enzymatic activity, while the same treatment significantly enhanced the Fas-induced enzymatic activity of caspse-3.