The following reagents were purchased and used according to the manufacturer’s instructions: glutathione S-transferase (GST)-TRAIL and anti-DR5 antibodies were from Koma Biotechnologies (Seoul, Korea); anti-caspase 3, anti-PARP and anti-CHOP antibodies were from Cell Signaling Technology; anti-DR4 antibody was from Rockland; anti-tubulin antibody was from Abcam; anti-actin antibody, thapsigargin (Tg), necrostatin-1, NAC and BHA were from Sigma; and zVAD was from R&D Systems. A formulated RGE was provided by the Korea Ginseng Corporation (Seoul, Korea).

The SK-Hep1, HepG2 and Hep3B cell lines were grown on culture plates in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. Huh-7 cells were grown on culture plates in RPMI-1640 medium supplemented with 10% FBS, 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. A commercially available normal liver cell line (HL-7702) was grown on culture plates in RPMI-1640 medium supplemented with 20% FBS, 2 mM glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin.

Upon treatment, cell lines were lysed in 20 mM Tris (pH 7.0), 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM DTT, 0.5% NP-40, 0.5 mM PMSF, 20 mM β-glycerol phosphate, 1 mM sodium vanadate and 1 μg/ml leupeptin. Lysates were loaded onto 10 or 12% SDS-PAGE gels. Following transfer and blotting, the proteins of interest were visualized by enhanced chemiluminescence (ECL, Pierce) and analyzed.

Viable cells were measured by the MTT assay (dual labeling method with 2 μM calcein-AM and 4 μM EthD-1), and MTT absorbance was read at 570 nm. Lactate dehydrogenase (LDH) leakage was quantified using a cytotoxicity detection kit (Promega). Images were recorded by microscope and the data presented are from at least three independent experiments.

shRNA-encoding plasmids were purchased from Sigma-Aldrich, and the targets were coding region sequences or the 3′ untranslated region of CHOP mRNA (NM: 004083). The 293T cell line (LV900A-1) was transfected with lentiviral plasmids using Lipofectamine 2000 transfection reagent (Invitrogen, 11668019). HepG2 cell lines were infected with pseudoviral particles collected from the 293T cell line 2 days after transfection in the presence of Polybrene (8 μg/ml). The lentivirus-infected HepG2 cell line was then isolated by puromycin (1 μg/ml) resistance 2 days after infection, and knockdown of CHOP was confirmed by western blot analysis. CHOP knockdown cell lines were seeded on 6-well plates and treated with the RGE or Tg for the times indicated for each experiment and were analyzed by western blot analysis.

A high-performance liquid chromatography/mass spectrometry (LC/MS) method was used for the targeted 14 ginsenosides to determine the ginsenoside content in the formulated RGE. Butanol extracts of white, red and sun ginseng were employed to isolate the standard ginsenosides, and each ginsenoside was isolated and purified by silica gel chromatography or preparative LC (21,22). Calibration curves for each ginsenoside standard were plotted (data not shown), and the ginsenoside contents of the formulated RGE were calculated (Fig. 1, Table I). Korean red ginseng has been reported to induce antitumor and immune-stimulating, antistress and antioxidant activities. The cytoprotective and chemoprotective properties of ginseng have been attributed to its ability to reduce oxidative or nitrosative stress (23–25). Previous studies have suggested that RGE, a major component of ginseng, inhibits cancer cell propagation and metastasis (24). Based on these observations, we tested the cytotoxicity of the RGE in the HepG2 cell lines. As shown in Fig. 2A, RGE concentrations up to 10 mg/ml showed no significant effects on cells (left panel) and no decrease in viability, as measured by LDH release (right panel). As a positive control, we treated the HepG2 cell lines with TRAIL (50 ng/ml) and observed the induction of cell death, as evidenced by the EthD-1-positive cell line. We found that TRAIL (25 ng/ml) induced poly (ADP-ribose) polymerase (PARP) cleavage at 12 h, but that the RGE alone had little effect on PARP and caspase-3 cleavage (Fig. 2B).

Because our data suggested that the RGE alone has no obvious effect on cancer cell cytotoxicity, we then measured the sensitization effect of the RGE on TRAIL-derived cell death. When the RGE and TRAIL were added to cells consecutively, dramatic cell death occurred according to the results of the MTT colorimetric assay (Fig. 2C). Concentrations of TRAIL >50 ng/ml were required to induce significant cytotoxicity; however, concentrations of TRAIL <25 ng/ml showed minimal cytotoxicity. Viability measurements from concentration curves obtained from 1:2 serial dilutions demonstrated that 10 mg/ml of the RGE was necessary to achieve robust TRAIL-derived cell death (data not shown), but RGE alone showed a minimal cytotoxic effect. We further tested the sensitization effect in the Hep3B and Huh-7 cell lines. A similar degree of TRAIL (25 ng/ml) sensitization was observed, indicating that this effect was universal in HCC cell lines (Fig. 3A). In contrast, a combination of the RGE and TRAIL sensitize normal human liver cell lines, confirming that this sensitization was specific to the tumor cell lines (Fig. 3B).

One of the primary problems when clinically applying TRAIL as a cancer therapeutic agent is resistance. One effective strategy to surmount this issue would be to discover anticancer agents for use in combination therapy (26–29). Several natural products and other chemical compounds have been identified that sensitize cancer cell lines to TRAIL-derived apoptosis. In this study, we tested a formulated RGE that has been suggested to possess anticancer properties and the results provide substantial evidence that the RGE was capable of sensitizing HCC to TRAIL-derived apoptosis by upregulating the CHOP, which then induced DR5 expression.

We next examined the effects of the RGE on TRAIL activation of the caspase cascade in the SK-Hep1 cell line. As shown in Fig. 3C, treatment with a combination of RGE and TRAIL induced cleavage of caspase-3 and PARP, a well-established caspase substrate, whereas treatment with the RGE alone did not. In addition, cell death was apoptotic because necrostatin-1 failed to inhibit cell death, as shown in Fig. 3D. To further confirm RGE and TRAIL-derived apoptosis, cells were cultured on coverslips and treated with RGE, TRAIL or RGE plus TRAIL. The cells were fixed and stained with DAPI after 16–18 h, and examined for morphologic changes such as nuclear DNA fragmentation and/or condensation, which are indicative of apoptotic death. As shown in Fig. 3E, the RGE alone had no clear effect on nuclear DNA fragmentation and/or condensation, but RGE pretreatment significantly augmented TRAIL-derived nuclear DNA fragmentation and/or condensation in HepG2 cells.

TRAIL resistance is facilitated by the downregulation of DR4 and DR5 and upregulation of DcR1 and DcR2 (8,30–32). We examined TRAIL receptor protein levels in cells treated with the RGE to explore the molecular mechanisms underlying the sensitization effect of the RGE to TRAIL-derived cell death. Treatment with the RGE significantly increased the expression of DR5 in a dose-dependent manner in HepG2 cells (Fig. 4A), although no clear changes were found in DR4 expression (Fig. 4B). We compared the effects of the RGE in three different HCC cell lines to determine whether this effect was cell-type specific (Fig. 4C). All cell lines tested showed upregulated DR5 expression following the RGE treatment, suggesting that the effects of RGE on the upregulation of DR5 are not cell-type specific.

Many mechanisms have been elucidated for inducing DR5, including ER stress, p53 induction, reactive oxygen species (ROS) generation and NF-κB/MAPK activation (33–35). It has been suggested that C/EBP homologous protein/GADD153, a transcription factor of the C/EBP family that has a role in ER stress, is involved in activating DR5 expression and thus contributes to the effectiveness of the RGE and TRAIL combined therapy (36–38). The RGE treatment increased CHOP levels in the HepG2 cell lines corresponding to its sensitization effect on TRAIL-derived cell death (Fig. 5A–C). As shown in Fig. 5D, RGE-induced CHOP upregulation preceded the increase in DR5 levels. Such a temporal pattern further supports the notion that CHOP plays a role in the activation of DR5 expression by the RGE. As CHOP is a major transcription factor induced by ER stress, we examined whether RGE treatment would induce ER stress in HCC. Thus, we probed for two markers of ER stress: Ser-51 phosphorylation of eIF2α (P-eIF2α) and induction of the ER chaperone protein GRP78. As shown in Fig. 5E, the RGE upregulated C/EBP homologous protein but did not induce GRP78 or eIF2α phosphorylation in the HepG2 cell lines, indicating that the RGE was capable of over-expressed the CHOP without substantively affecting ER stress. However, the RGE did not influence CHOP expression in the normal liver cell lines, which strongly agrees with the observation that the RGE did not boost the effect of TRAIL in the normal cell lines (Fig. 3B). Additionally, the RGE was unable to induce expression of the C/EBP homologous protein or other ER stress markers in the normal cell lines (Fig. 5F).

When investigating the induction of the TRAIL receptors DR4 and DR5 in tumors it is very important to assess susceptibility of the tumor to TRAIL treatment (39). Several studies have shown that high expression of the TRAIL receptors DR4 and DR5 facilitates sensitization of cancer cells to TRAIL-derived cell death (1,40). One therapeutic approach being tested is to induce the expression of death receptors, including DR4 and particularly DR5, by small molecules, which results in TRAIL-derived tumor cell death or sensitization of TRAIL-resistant cell lines to cell death (33). Furthermore, DR5 expression levels are highly correlated with TRAIL, which binds to its receptors, DR4 and DR5, and activates the extrinsic apoptosis pathway (34,41,42). In the present study, we found that the improved effects of the combination therapy were primarily dependent upon control of DR5 induction. We also found that the combination therapy induced apoptosis in the presence of ROS (Fig. 6), suggesting that ROS do not contribute to RGE-facilitated sensitization of TRAIL-derived cell death.

Tg treatment was compared with RGE treatment to confirm that upregulation of CHOP plays a key role in the RGE sensitization effect, as shown in Fig. 7A(43,44). Next, we established stable cell lines for knockdown of the C/EBP homologous protein using a lentiviral shRNA system. As shown in Fig. 7B, we confirmed knockdown of CHOP. In addition, knockdown of CHOP inhibited the robust upregulation of CHOP induced by Tg. We then examined the effect of CHOP knockdown on RGE-induced sensitization to TRAIL-derived cell death. CHOP knockdown inhibited the effect of the RGE plus TRAIL treatment in HepG2 cell lines, suggesting that augmentation of cell death by the RGE combined treatment in HCC was mediated by CHOP expression (Fig. 7C).

Activation of NF-κB in cancer cell lines contributes to the induction of DR5 as well as resistance to TRAIL-derived apoptosis. Therefore, we examined whether the RGE might regulate NF-κB to facilitate DR5 upregulation and, consequently, sensitize cell lines to TRAIL-derived apoptosis. As shown in Fig. 7D, degradation and phosphorylation of IκBα did not occur in response to the RGE. In addition, the RGE did not affect TRAIL-derived NF-κB activation, suggesting that RGE-induced sensitization to cell death is not mediated by NF-κB. Although TRAIL is capable of activating NF-κB in some cancer cell lines and the induction of DR5 is linked to NF-κB activation, activated NF-κB also upregulates the anti-apoptotic gene Bcl-xL, effectively blocking TRAIL-derived apoptosis (10,27,34). We failed to detect upregulation of Bcl-xL in response to RGE treatment (data not shown), which further indicates that NF-κB was activated by the RGE. Additionally, RGE-induced CHOP upregulation was not caused by ER stress, which can lead to apoptosis in HCC cell lines. Inducing CHOP expression or TRAIL sensitization was unsuccessful when RGE was applied to the normal cell line HL7702, suggesting that CHOP protein upregulation by the RGE is specific to the tumor cell lines.

Although we have not tested the antitumor activity of the RGE in vivo, the formulated RGE examined could potentially be further developed as an important chemosensitizer to be utilized as a dietary supplement with anticancer drugs.