Ursodeoxycholic acid switches oxaliplatin-induced necrosis to apoptosis by inhibiting reactive oxygen species production and activating p53-caspase 8 pathway in HepG2 hepatocellular carcinoma

Authors

  • Sung-Chul Lim,

    1. Research Center for Resistant Cells, College of Medicine, Chosun University, Gwangju
    2. Department of Pathology, College of Medicine, Chosun University, Gwangju
    Search for more papers by this author
  • Jeong Eun Choi,

    1. Research Center for Resistant Cells, College of Medicine, Chosun University, Gwangju
    Search for more papers by this author
  • Ho Sung Kang,

    1. Department of Molecular Biology, College of Natural Sciences, Pusan National University, Pusan, Korea
    2. Research Institute of Genetic Engineering, Pusan National University, Pusan, Korea
    Search for more papers by this author
  • Han SI

    Corresponding author
    1. Research Center for Resistant Cells, College of Medicine, Chosun University, Gwangju
    • Research Center for Resistant Cells, Chosun University, Gwangju 501-759, Korea
    Search for more papers by this author
    • Fax: +[82-62-233-6052]


Abstract

Hepatocellular carcinoma (HCC) is resistant to chemotherapy. Recently, however, several oxaliplatin-based combinatorial treatments have shown a promising anti-tumor activity in patients with HCC. Presently, we demonstrate that oxaliplatin triggers necrosis more than apoptosis in HepG2, SK-Hep1, SNU-423 and Hep3B HCC cells, while mainly inducing apoptosis in HCT116 and HT29 colon cancer cells. Interestingly, ursodeoxycholic acid (UDCA), a less hydrophobic bile acid that can suppress carcinogenesis, shifted oxaliplatin-induced necrosis to apoptosis in HepG2 cells. The same effect was produced by hydrophilic bile acids (tauroursodeoxycholic acid and taurohyodeoxycholic acid), but not by highly hydrophobic bile acids (deoxycholic acid and chenodeoxycholic acid). UDCA also triggered the necrosis-to-apoptosis switch when cotreated with other platinum-based chemotherapeutic drugs including cisplatin and carboplatin, suggesting that the cell death mode switching effect of UDCA is a general phenomenon when combined with platinum drugs. Oxaliplatin produced high level of reactive oxygen species (ROS) in HepG2 cells and UDCA significantly reduced oxaliplatin-induced ROS generation. In addition, N-acetyl-L-cysteine and the superoxide scavengers butylated hydroxyanisole and dihydroxybenzene-3,5-disulfonic acid attenuated necrosis, indicating a critical role(s) of ROS in occurrence of necrotic death. Apoptosis induced by combined treatment appeared to be mediated by p53-caspase 8-caspase 3 pathway. In conclusion, UDCA switches oxaliplatin-induced necrosis to apoptosis via inhibition of ROS production and activation of the p53-caspase 8 pathway in HepG2 cells. As necrosis and subsequent inflammation are implicated in tumor progression and malignancy, our results imply a potential improved efficacy of UDCA-combined chemotherapy in HCC by reducing inflammatory responses that may be triggered by oxaliplatin.

Hepatocellular carcinoma (HCC) is one of the most common malignancies and the leading causes of cancer-related mortality, and it generally represents highly resistant features to chemotherapy.1–3 Surgical resection and liver transplantation might be the most effective way to cure the cancer, but these approaches have limited applicability and advanced tumors frequently recur even after complete surgical resection.4, 5 Hence, chemotherapy remains an important option to treat HCC and developments of new chemotherapeutic strategies are necessary.

Oxaliplatin, a third generation platinum-based chemotherapeutic agent, displays a broader spectrum of antitumor activity than cisplatin and carboplatin and no cross-resistance against some cisplatin- and carboplatin-resistant tumors.6–9 Currently, it is widely used as a basis of combination regimens for various cancers including gastric, colorectal and advanced nasopharyngeal carcinoma.10–12 In addition, several oxaliplatin-combined regimens have been used for patients with advanced HCC and these have shown a moderate but promising anti-tumor effect.13, 14

In most cases, necrosis is closely associated with occurrence of inflammation. Due to the loss of cell membrane integrity, necrotic cells release cytosolic contents including high-mobility group box 1 (HMGB1) and other inflammatory molecules to the extracellular space, thereby causing a massive inflammatory response.15, 16 In addition, necrotizing cells are also involved in the induction of inflammation by nuclear factor kappa B (NF-κB)- and p38 mitogen activated protein kinase (MAPK)-mediated up-regulation and secretion of the pro-inflammatory cytokine interleukin (IL)-6, whereas this event is not observed in apoptotic cells.17 In cells dying by apoptosis, the cellular contents remain packed in the apoptotic bodies, which are removed by macrophages, limiting inflammation. Necrosis has been recently demonstrated to promote tumor growth through inflammation and angiogenesis, and to be correlated with malignancy and poor prognosis of tumor growth.18–21 Thus, the consequences of apoptosis and necrosis are quite different for a whole organism in terms of the effect on inflammatory responses and tumor prognosis.

Bile acids exert diverse physiological and pathological activities in hepatic and colorectal tissues. Hydrophobic bile acids include deoxycholic acid (DCA), chenodeoxycholic acid (CDCA) and lithocholic acid (LCA). Hydrophilic bile acids include ursodeoxycholic acid (UDCA), taurodeoxycholic acid (TUDCA), taurohyodeoxycholic acid (THDCA) and cholic acid (CA). Hydrophobic bile acids have been implicated in liver cholestasis and colorectal cancer development.22, 23 However, UDCA exerts cytoprotective, anti-apoptotic and immunomodulatory activities, and it is widely used for the treatment of certain cholestatic liver diseases such as primary biliary cirrhosis and cystic fibrosis-related cholestasis.24, 25 In addition, UDCA reduces carcinogenesis in several pre-malignant conditions of liver and colon in a rodent model.26–29 Furthermore, recurrence of colorectal adenomas with high-grade dysplasia in a phase III trial was shown to be decreased after treatment with UDCA.30 Although the tumor suppressive activity of UDCA is believed to be linked to the induction of apoptosis and cell cycle arrest, and the inhibition of oncogenic factors including Ras and COX-2, in human cancer cell lines,31–33 these mechanisms are not yet fully understood. In this study, we examined the effects of UDCA on platinum-induced cell death and demonstrate a novel activity of UDCA in switching platinum drug-induced cell death mode from necrosis to apoptosis, suggesting a possible beneficial role of UDCA in platinum-based anti-cancer chemotherapy.

Abbreviations

IETD-pNA: IETD-p-nitroanilide; PARP: Poly(ADP-ribose)polymerase; PBS: phosphate-buffered saline; siRNA: small interference RNA; z-DEVD-FMK: z- Asp-Glu-Val-Asp-fluoromethylketone; z-LEHD-FMK: z-Leu-Glu-His-Asp-fluoromethylketone; z-IETD-FMK: z-Ile-Glu-Thr-Asp-fluoromethylketone; z-VAD-FMK: z-Val-Ala-Asp-fluoromethylketone

Material and Methods

Cell culture and drug treatment

HepG2, HT29 and HCT116 cells were obtained from American Type Culture Collection (Manassas, VA), and SK-Hep1, Hep3B and SNU-423 cells were provided by Korea Cell Line Bank (Seoul, Korea). Cells were grown in RPMI 1640 medium (Invitrogen, Carlsbad, CA) supplemented with 10% (v/v) fetal bovine serum (FBS; Invitrogen) and 1% penicillin-streptomycin (Welgene, Seoul, Korea) in a 37°C humidified incubator in an atmosphere of 5% CO2. Drug treatment of cells was performed by adding 50 μM oxaliplatin (L-OHP; Boryung Pharmaceutical, Seoul, Korea), 10–25 μM cisplatin (C-DDP; Sigma-Aldrich, St. Louis, MO), 10–50 μM carboplatin (CBDCA; Sigma-Aldrich), 10–25 μM 5fluoruracil (5-FU), or 2–5 μM doxorubicin to the culture medium and incubating for 24 hr. Cells were pretreated for 1 hr with 20 μM of a caspase inhibitor (z-IETD-FMK, z-LEHD-FMK, z-DEVD-FMK or z-VAD-FMK) or 30 μM pifithrin-α prior to simultaneous exposure to oxaliplatin and 100 μM UDCA (ICN Biomedicals, Irvine, CA), 30 μM DCA (Sigma-Aldrich), 100 μM TUDCA, 100 μM THDCA or 100 μM CDCA, or anti-oxidants including 10 mM N-acetyl-L-cysteine (NAC), 100 μM butylated hydroxyanisole (BHA), 10 mM 1,2-dihydroxybenzene-3,5-disulfonic acid (Tiron), or 1 mM 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) (Roche, Basel, Switzerland). Unless specified, drugs were purchased from Calbiochem (San Diego, CA).

Hoechst 33342 (HO)/propidium iodide double staining and annexin V/propidium iodide staining

Cells were incubated either with 1 μg/ml HO and 5 μg/ml propidium iodide (PI) at 37°C, 5% CO2 for 15 min at room temperature in the dark. Then, both floating and attached cells were collected. The pooled cell pellets were washed with ice-cold phosphate-buffered saline (PBS), fixed in 3.7% formaldehyde on ice, washed again with PBS, resuspended and a fraction of the suspension was centrifuged in a cytospinner (Thermo Shandon, Pittsburgh, PA). The slides were air dried, mounted in an anti-fade solution and examined using a DM5000 fluorescence microscope (Leica, Wetzlar, Germany) at respective excitation/emission wavelengths of 340/425 nm (HO) and 580/630 nm (PI). Morphological assessments of apoptosis and necrosis were performed; intact blue nuclei, condensed/fragmented blue nuclei, condensed/fragmented pink nuclei and intact pink nuclei were considered viable, early apoptotic, late apoptotic (secondary necrotic) and necrotic cells, respectively. A total of 500 cells from several randomly chosen fields were counted, and the number of apoptotic and necrotic cells were expressed as a percentage of the total number of cells scored.

To detect phosphatidylserine on the cell surface, cells cultured on 15 micron slide 8-well chamber dish (Ibidi GmbH, Munich, Germany) were treated with drugs for 18 hr and incubated with FITC-conjugated annexin V and PI (Annexin V-FITC apoptosis detection Kit from Calbiochem) for 20 min at room temperature in the dark. The cells were washed with ice-cold PBS and observed under a confocal microscope (×400, FV1000, Olympus).

Lactate dehydrogenase release and 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) viability assay

Lactate dehydrogenase (LDH) release was quantified using a LDH cytotoxicity assay kit II (BioVision, Mountain View, CA) according to the manufacturer's protocol. Briefly, cells were plated in wells of a 96-well plate at a density of 1 × 104 cells/well and incubated for 24 hr and treated with drugs for 24 hr. The plates were centrifuged at 600g for 10 min, and then the cell-free supernatant was transferred to 96-well plate, mixed with LDH Reaction Mix, incubated for 30 min and measured the absorbance at 450 nm. The percentage of specific LDH release was calculated by the following formula: % cytotoxicity = [(experimental LDH release)-(spontaneous LDH release by effector and target)/(maximum LDH release) − (spontaneous LDH release)] × 100. The spontaneous release of LDH activity from control cells was less than 2% of the maximal release of LDH activity, which was determined from the complete lysis by adding lysis buffer.

For the MTT assay, cells were prepared and treated as described earlier. MTT solution (0.5 mg/ml) was added and incubated for 4 hr. Cells were solubilized using dimethylsulfoxide (DMSO) and the solubilized formazan product was quantified using an enzyme-linked immunosorbant assay (ELISA) plate reader at 595 nm; absorbance of untreated cells was designated as 100% and cell survival expressed as a percentage of this value.

Morphologic analysis on semi-thin sections

Cells were collected and fixed in 2.5% glutaraldehyde with 0.1 M cacodylate buffer (pH 7.2) for 2 hr at 4°C and washed twice with cold PBS, post-fixed in OsO4, dehydrated in graded ethanol and embedded in epon mixture. Sections were prepared with ultra-microtome (MT-7000), stained with toluidine blue and observed by light microscopy (400×).

Western blot analysis

For Western blot analysis, an equal amount of proteins was electrophoretically separated by SDS-PAGE and transferred to a nitrocellulose membrane. After blocking with 5% skim milk/TBST, the membrane was incubated overnight at 4°C with primary antibodies to active caspase-8, active caspase-3, phospho-p53 (Cell Signaling Technology, Danvers, MA), cytochrome C, p53, p65, c-Jun (Santa Cruz Biotechnology, Santa Cruz, CA), PARP (BD Biosciences, San Jose, CA) and α-tubulin (BioGenex Laboratories, San Ramon, CA). After washing 3 times in Tris-buffered saline (TBS) with 0.1% Tween-20, the membrane was incubated with horseradish peroxidase-conjugated secondary antibody and visualized with the Super Signal West Pico chemiluminescence kit (Pierce Biotechnology, Rockford, IL) and signals were acquired by image analyzer (image station 4000MM, Kodak, NY).

Assay for reactive oxygen species

To determine production of intracellular reactive oxygen species (ROS), cells were plated in wells of a 96-well plate at a density of 1 × 104 cells/well and cultured for 24 hr and treated with oxaliplatin in the presence or absence of UDCA for indicated times, then loaded with 50 μM 2, 7-dichlorofluorescin diacetate (DCFDA, Molecular Probes, Eugene, OR) to measure ROS generation and 0.5 μg/ml HO to quantify cell number for 30 min. After rinsing, fluorescence measurements were obtained with a Fluorocount model MQX200 plate reader (Packard Instrument, Meriden, CT) at respective excitation/emission wavelengths of 490/530 nm (DCFDA) and 340/425 nm (HO), and values of ROS production were obtained by determining the ratio of DCFDA/HO signals per well.

Caspase 8 activity assay

Caspase 8 activity assay was carried out using a FADD-like IL-1β-converting enzyme (FLICE)/Caspase-8 Colorimetric assay kit (BioVision) according to the manufacturer's protocol. Briefly, after proper treatment, cells were washed, lysed and centrifuged. The supernatant was analyzed for protein concentration and 200 μg of protein in a 50 μl volume was mixed with 50 μl of 2× reaction buffer, and 5 μl of 4 mM IETD-pNA substrate was added to produce a final substrate concentration of 200 μM. After 90 min incubation, absorbance was measured at 405 nm in a plate reader. Fold increase in FLICE activity was determined by comparing the results of treated samples with the level of the untreated control.

RNA interference

For the RNA interference (RNAi) experiment, short interfering RNA (siRNA) of p53 5′-CAC UAC AAC UAC AUG UGU A-3′ (S) 5′-UAC ACA UGU AGU UGU AGU G(dTdT)-5′ (AS), c-Jun 5′-ACU GUA GAU UGC UUC UGU A(dTdT)-3′ (S) 5′-UAC AGA AGC AAU CUA CAG U(dTdT)-3′ (AS), NF-kB (p65) 5′-CCU GAG CAC CAU CAA CUA U(dTdT)-3′ (S) 5′-AUA GUU GAU GGU GCU CAG G(dTdT)-3′ (AS) and scrambled RNA sequence 5′-CCU ACG CCA CCA AUU UCG U(dTdT)-3′ (S) 5′-ACG AAA UUG GUG GCG UAG G(dTdT)-3′ (AS) were designed and synthesized by BIONEER (Daejeon, Korea). HepG2 cells at 70% confluence in 60 mm dishes were transfected with 5 μg siRNA using the JET-PEI reagent (1:3.2; Polyplus-Transfection, France) according to the manufacturer's protocol. The transfected cells were incubated for 36 hr then exposed to drugs and whole experimental processes were completed in 72 hr post-transfection.

Electrophoretic mobility shift assay

The transcription factor (TF) binding activity was measured using electrophoretic mobility shift assay “Gel Shift” kit (Panomics, Fremont, CA) according to the manufacturer's protocol. Briefly, the assay was done in 10 μl of reaction buffer in the presence of 10 μg nucleic extract and 1 μg poly(dI-dC), then incubated for 30 min at 15°C after the addition of the Avidin-labeled probe DNA. For competition assay, excess cold probe was added to the reaction mixture with the labeled probe. The TF-DNA complexes were then run in 6% nondenaturing gel, transferred to nylon membrane and fixed using a UV cross-linker. The membrane is incubated in Streptavidin-HRP solution, rinsed with PBS, developed by ECL solution and signals were acquired by image analyzer.

Statistical analysis

All numerical data are reported as mean ± SE unless otherwise specified. All data represent the results of at least 3 independent experiments. Groups were compared by means of Student's t-test.

Results

Oxaliplatin induces mainly necrotic cell death in HCC

To investigate whether oxaliplatin could be used as a chemotherapeutic agent in HCC, as a first step, its cytotoxic effect was analyzed using MTT assay in 4 different HCC cells including HepG2, SK-Hep1, SNU-423 and Hep3B, and in HCT116 and HT29 colon cancer cells. MTT assay demonstrated that oxaliplatin exerted dose-dependent cytotoxic effects in a similar pattern in both HCCs and colon cancer cells (Fig. 1a). The reduction in MTT values represents a growth arrest or mitochondrial or other damage associated with both apoptosis and necrosis. Then, we used HO/PI double staining to identify oxaliplatin-induced cell damage in terms of cell death (Figs. 1b and 1c). HO stains the nuclei of all cells with blue fluorescence, while PI only penetrates and fluoresces in cells with damaged membranes. Thus, intact blue nuclei, condensed/fragmented blue nuclei, condensed/fragmented pink nuclei and intact pink nuclei were considered viable, early apoptotic, late apoptotic (secondary necrotic) and necrotic cells, respectively. In colon cancer cells, oxaliplatin induced mostly apoptotic cell death but less necrotic cell death. In contrast, massive necrotic cell death than apoptotic cell death was observed in oxaliplatin-treated HepG2, SK-Hep1, SNU-423 (Figs. 1b and 1c) and Hep3B HCC cells (see Fig. 7d) with an increasing effect depending on the concentration. The ratio of necrotic and apoptotic death was slightly different depending on the cell types and the most prominent necrotic cell death in response to oxaliplatin was observed in HepG2. Oxaliplatin-induced necrotic cell death in HCC but not in colon cancer cells was also evident by release of LDH into the medium (Fig. 1d). LDH release at early stage of cell death is indicative of cell membrane rupture, a typical feature of necrotic death. However, LDH could be released from secondary necrosis in cell culture, and thus we measured LDH release after 24 hr treatment of oxaliplatin, at this time point secondary necrosis was not observed in HO/PI staining.

Figure 1.

Oxaliplatin induces mainly necrosis in HCC cells, but apoptotic death in HCT116 or HT29 cells. (a) HepG2, SK-Hep1 and SNU423 HCC cells, and HCT116 and HT29 colon cancer cells were incubated in control medium or with 25, 50, 100, 150 or 200 μM oxaliplatin for 24 hr, and viability was assessed by a MTT assay. (b, c) Cells were exposed to the indicated concentrations of oxaliplatin for 24 hr, stained with HO/PI, and cells with condensed or fragmented nuclei were counted as apoptotic cell (b), and cells stained with PI and those having no condensed or fragmented nuclei were scored as necrotic cells (c) by fluorescence microscopy. (d) After incubation in oxaliplatin-containing medium for 24 hr, released LDH activities were measured from the medium.

Co-administration of UDCA with oxaliplatin leads to switch of cell death mode from necrosis to apoptosis

Necrotic death is accompanied by inflammation, an aggravating factor of tumor malignancy and poor prognosis, and thus we tried to explore a possible approach to reduce necrotic death during chemotherapy. We examined the effects of UDCA on oxaliplatin-induced necrosis in HepG2 cells. Combined treatment of oxaliplatin with 100 μM UDCA, a concentration that is itself not cytotoxic, did not cause any significant alterations in oxaliplatin-induced cytotoxicity (Fig. 2a). However, the addition of UDCA markedly reduced LDH release as compared with oxaliplatin alone (Fig. 2b). Meanwhile, as bile acids especially DCA could directly inhibit LDH activity at high concentrations,34 we examined whether UDCA itself affects LDH activity using lysed cells after exposure to UDCA or DCA and found that the presence of UDCA showed no reduction in LDH activity at all, while 20 mM of DCA decreased LDH activity (Supporting Information Fig. 1). These results demonstrate that UDCA may switch oxaliplatin-induced necrosis to other types of cell death such as apoptosis without altering cytotoxicity in addition to preventing necrosis. We observed that combined treatment of oxaliplatin and UDCA markedly decreased cells with intact pink nuclei and increased the number of those with fragmented/condensed blue nuclei, as revealed by HO/PI double staining (Figs. 2c and 2d). These results suggest that UDCA may switch the oxaliplatin-induced necrotic death to apoptotic death. The necrosis-to-apoptosis switching effect of UDCA was further supported by annexin V-FITC staining and semi-thin epon section technique. Externalized phosphatidylserine that was detected by annexin V-FITC was observed in UDCA/oxaliplatin-treated cells but not in oxaliplatin treatment alone (Fig. 2e). Furthermore, in semi-thin epon section staining, oxaliplatin-treated cells had necrotic features, such as disintegrated membrane and loose cytoplasm and cell debris as a result of necrosis, while UDCA/oxaliplatin-treated cells had compacted and fragmented nuclei with intact membrane, which is a typical feature of apoptosis (Fig. 2f). Taking these results together, UDCA appeared to switch the oxaliplatin-induced necrotic death to apoptotic death. The cell death mode switching effect of UDCA in oxaliplatin-induced cell death increased in a dose-dependent manner at concentrations of 10–100 μM, with a maximal effect at 100–200 μM (data not shown).

Figure 2.

UDCA switches oxaliplatin-induced necrosis to apoptosis in HepG2 cells. (a) Cell viability was measured by a MTT assay after 24 hr incubation with oxaliplatin in the presence (+) or absence (−) of 100 μM UDCA. (b) HepG2 cells were incubated with vehicle, 100 μM UDCA, 50 μM oxaliplatin, or 100 μM UDCA plus 50 μM oxaliplatin for 12, 24, 36 or 48 hr. LDH release into the medium was assessed by a LDH activity assay. (c, d). Cells were incubated with the indicated concentrations of oxaliplatin for 24 hr in the presence (+) or absence (−) of 100 μM UDCA, stained with HO/PI and apoptotic and necrotic cells were scored by fluorescence microscopy (c). The images of 50 μM oxaliplatin, 100 μM UDCA and 50 μM oxaliplatin plus 100 μM UDCA treatments were captured by confocal microscopy (×400, FV1000, Olympus) (d). (e) Cells treated with vehicle, 100 μM UDCA, 50 μM oxaliplatin or 100 μM UDCA plus 50 μM oxaliplatin for 15 hr were stained with annexin V-FITC/PI, and observed by confocal microscopy (×400, FV1000, Olympus). (f) Cells, incubated with vehicle (control), 100 μM UDCA, 50 μM oxaliplatin, or 100 μM UDCA plus 50 μM oxaliplatin were prepared for semi-thin epon sections, stained with toluidine blue dye and observed by light microscopy (DM5000, Leica, Germany).

The death mode switching effect of UDCA is also effective in combination with cisplatin and carboplatin

Next, to clarify whether the UDCA effect was specific to oxaliplatin, we tested the action of UDCA on cell death mode induced by 2 other platinum-based drugs, cisplatin and carboplatin, and other widely-used types of anti-cancer drugs including the antitumor antibiotic, doxorubicin and the pyrimidine analogue 5-fluorouracil (5-FU). In HepG2 cells, cisplatin, carboplatin and doxorubicin showed similar cytotoxic effects to those of oxaliplatin and induced massive LDH release and necrotic cell death in a concentration-dependent fashion (Figs. 3a and 3b). In contrast, 5-FU appeared to induce apoptotic cell death as revealed by a low level of LDH release and appearance of HO/PI stained cells having condensed/fragmented blue nuclei (Figs. 3b and 3c). As shown in Figures 3b and 3c, UDCA significantly suppressed LDH release triggered by cisplatin or carboplatin, but not by doxorubicin, and switched cisplatin- or carboplatin-induced necrosis to apoptosis without affecting doxorubicin-induced necrosis. UDCA did not exert any significant effects on 5-FU-induced cytotoxicity and apoptotic cell death. These results suggest that UDCA has a general necrosis-to-apoptosis switching effect when combined with platinum-based chemotherapeutic drugs but not with other drugs.

Figure 3.

UDCA switches necrosis induced by platinum-based anti-cancer drugs to apoptosis. Cell viability and LDH release were measured by a MTT assay (a) and LDH assay (b) after exposure to various anti-cancer drugs (cisplatin, carboplatin, 5-FU and doxorubicin) in the presence (+) or absence (−) of 100 μM UDCA. (c) Cells were incubated with the indicated concentrations of anticancer drugs in the presence or absence of 100 μM UDCA for 24 hr, stained with HO/PI, and apoptotic and necrotic cells were scored by fluorescence microscopy.

Hydrophilic, but not hydrophobic, bile acids switch oxaliplatin-induced necrosis to apoptosis

As hydrophobicity of bile acids is postulated to be a critical feature in their physiological functions, we examined whether the cell death mode switching effect of UDCA is related to its hydrophilic characteristics. TUDCA (100 μM), THDCA (100 μM), DCA (30 μM) and CDCA (100 μM) were co-administrated with oxaliplatin, and their effects on oxaliplatin-induced necrotic cell death were examined. As DCA was cytotoxic by itself, the highest concentration (30 μM) showing no cytotoxicity for 24 hr was used. As shown in Figure 4, co-administration of oxaliplatin with TUDCA or THDCA reduced the necrotic cell population and increased the apoptotic cell population, as was observed in combination with UDCA, whereas DCA and CDCA showed no significant effects on necrotic population. These observations suggest that the necrosis-to-apoptosis switching effect of UDCA is closely associated with its hydrophilic features.

Figure 4.

Hydrophilic, but not hydrophobic, bile acids switch oxaliplatin-induced necrosis to apoptosis. Cells were exposed to 50 μM oxaliplatin alone or in combination with 100 μM UDCA, 100 μM TUDCA, 100 μM THDCA, 30 μM DCA and 100 μM CDCA. After 24 hr incubation, cells were stained with HO/PI, and apoptotic and necrotic cells were scored by fluorescence microscopy.

UDCA decreases oxaliplatin-induced ROS production that may be responsible for necrosis

ROS has been suggested as a critical factor in the determination of cell death mode. Therefore, we explored whether ROS could participate in the regulation of cell death mode in UDCA/oxaliplatin-induced cell death. The level of ROS production was measured by DCFH-DA in oxaliplatin- or UDCA/oxaliplatin-treated HepG2 cells. Exposure to oxaliplatin gradually increased ROS production and UDCA significantly suppressed oxaliplatin-induced ROS production after 6 hr of treatment, without affecting the slight ROS production observed at an early time point (1 hr) (Fig. 5a). To investigate whether ROS production was associated to occurrence of oxaliplatin-induced necrosis, we examined the effects of various antioxidants on oxaliplatin-induced necrotic cell death. Addition of antioxidants represented differential effects on oxaliplatin-based cell death mode. NAC suppressed both apoptosis and necrosis, whereas the superoxide scavengers BHA and Tiron and the lipid peroxidation inhibitor Trolox only reduced oxaliplatin-induced necrosis, without affecting apoptosis (Fig. 5b). These results suggest that ROS plays a critical role(s) not only in execution of cell death but also in determination of cell death mode by oxaliplatin.

Figure 5.

UDCA reduces oxaliplatin-induced ROS production that is responsible for necrosis. (a) ROS production was measured by 50 μM DCFH-DA after incubation with 50 μM oxaliplatin in the presence (+) or absence (−) of 100 μM UDCA. *p < 0.05 and represent the significance of the difference from control. (b) Effects of various antioxidants on oxaliplatin-induced cell death mode were determined. Cells were incubated with oxaliplatin alone or in combination with 10 mM NAC, 100 μM BHA, 10 mM Tiron and 1 mM Trolox. After 24 hr incubation, cells were stained with HO/PI, and apoptotic and necrotic cells were scored by fluorescence microscopy.

Involvement of p53 but not AP-1 or NF-kB in oxaliplatin/ UDCA-induced apoptotic cell death

We investigated the possible involvement of AP-1, NF-kB and p53 in the cell death mode switching effect of UDCA. These transcription factors are known to play a crucial role(s) in sensitivity or resistance to apoptotic death in many transformed cells.35, 36 In addition, p53 is known to mediate oxaliplatin-induced apoptosis in colon cancer cells.37 Oxaliplatin caused an induction in total and nuclear protein levels of c-Jun without affecting total and nuclear protein levels of NF-kB-p65 (Fig. 6a) and increased the DNA binding activities of AP-1 and NF-kB (Fig. 6b). Addition of 100 μM UDCA did not cause any differences in these protein patterns (Fig. 6a) and did not affect DNA-binding activities induced by oxaliplatin (Fig. 6b). Further, knockdown of c-Jun or p65 using siRNAs demonstrated no alterations in cell death mode by oxaliplatin in the absence or presence of UDCA, indicating that AP-1 or NF-kB activity may not be related with the death mode regulation activities of UDCA (Fig. 6c). In case of p53, oxaliplatin up-regulated p53 phosphorylation and protein level in a time-dependent manner, even though the cell death mode was necrosis rather than apoptosis in HepG2 cells (Fig. 7a). UDCA did not alter its phosphorylation and protein level (Fig. 7b). Thus, oxaliplatin appears to increase p53 activity and protein level independent of the presence of UDCA. In a continuous study, we found that the inhibition of p53 activity by either 30 μM pifithrin-α, an inhibitor of p53 transcriptional activity, or RNAi significantly suppressed apoptosis induced by oxaliplatin/UDCA and augmented cell survival due to the reduction of apoptosis (Fig. 7c), whereas it did not exert any significant effects on necrosis induced by oxaliplatin alone (data not shown). These results are consistent with the suggestion that oxaliplatin/UDCA-induced apoptotic cell death is mediated through p53. Furthermore, although addition of UDCA decreased oxaliplatin-induced LDH release and increased apoptotic body formation in p53 wild type HepG2 and SK-Hep1 HCC cells (even though the effect of UDCA in SK-Hep1 cells was milder than that in HepG2 cells), UDCA did not affect oxaliplatin-induced cell death mode in p53 mutant SNU-423 or p53 null Hep3B cells, indicating an important role(s) of p53 in UDCA-mediated cell death mode switch (Fig. 7d).

Figure 6.

UDCA-induced cell death mode switch is not associated with the activity of c-Jun or NF-kB. (a) Total and nuclear protein levels of p65 and c-Jun were measured after 18 hr of treatment with 50 μM oxaliplatin in the presence (+) or absence (−) of 100 μM UDCA. (b) DNA binding activity of AP1 and NF-kB were measured after 4 hr of treatment. Lane 1, control; Lane 2, UDCA; Lane 3, L-OHP; Lane 4, UDCA+L-OHP; Lane 5, cold probe to lane 4 sample. (c) Cells transiently transfected with scrambled RNAi (Ctl RNAi), c-jun or p65 RNAi were cultured for 36 hr, then the cells were exposed to 50 μM oxaliplatin or oxaliplatin plus 100 μM UDCA. After 24 hr incubation, apoptotic cells were determined by counting cells with fragmented or condensed nuclei, and protein levels of cleaved caspase 3, c-Jun, p65, and tubulin were measured by Western blot analysis.

Figure 7.

Implication of p53 in the oxaliplatin/UDCA-induced apoptotic cell death. Cells were incubated with 50 μM oxaliplatin (a) or incubated with 50 μM oxaliplatin in the presence (+) or absence (−) of 100 μM UDCA (b) for the indicated time and protein levels were assessed by Western blotting analysis using anti-phospho p53 (Ser-15), and anti-p53 antibody. (c) Cells were treated with 50 μM oxaliplatin plus 100 μM UDCA in the presence (+) or absence (−) of pifithrin-α for 24 hr or cells were transiently transfected with scrambled RNAi (ctl RNAi) or p53 RNAi and cultured for 36 hr, then the treated cells and the transfected cells were exposed to 50 μM oxaliplatin plus 100 μM UDCA. After 24 hr incubation, apoptotic cells were determined by counting cells with fragmented or condensed nuclei. *p < 0.05, **p < 0.01. (d) Death pattern of oxaliplatin-treated HCCs in the absence or presence of UDCA. HepG2, SK-Hep1, SNU-423 and Hep3B cells were incubated in control medium or medium containing 50, 100, 200 or 300 μM oxaliplatin for 24 hr, and cell viability was assessed by a MTT assay and LDH activity released into medium were measured. For identification of apoptotic cells, cells were stained with HO, collected, and centrifuged using cytospinner, and then condensed or fragmented nuclei were counted under fluorescence microscope. (e) NAC, BHA, Tiron, or Trolox were treated to cells alone or co-exposed with oxaliplatin for 12 hr and protein level of p53 was assessed by Western blotting analysis.

We also observed that NAC, which could prevent both necrosis and apoptosis, completely prevented p53 induction, while the antioxidants BHA, Tiron and Trolox, which could prevent necrosis but not apoptosis, did not affect oxaliplatin-induced p53 induction (Fig 7e). These results further support the notion that p53 plays a critical role(s) in oxaliplatin/UDCA-induced apoptotic cell death. However, in spite of its up-regulation, p53 activity was independent from oxaliplatin-induced necrosis. A possible explanation could be that oxaliplatin activates p53 at an early stage, but certain late-onset signals such as excessive ROS production keep the downstream path of p53 suppressed during necrosis, with UDCA acting to release this suppression.

UDCA/oxaliplatin-induced apoptotic cell death is regulated through caspase 8 and caspase 3 activation

To further understand the underlying molecular mechanism of apoptosis by the combined treatment, we investigated the possible alterations of pro-apoptotic molecules. Combined treatment of UDCA and oxaliplatin induced cytochrome C release, caspase 3 activation and poly(ADP-ribose)polymerase (PARP) cleavage (Fig. 8a). Next, the caspase(s) responsible for the apoptosis were examined using specific caspase inhibitors including z-DEVD-FMK (caspase 3 inhibitor), z-IETD-FMK (caspase 8 inhibitor), z-LEHD-FMK (caspase 9 inhibitor) and z-VAD-FMK (pan-caspase inhibitor). UDCA/oxaliplatin-induced apoptosis was completely suppressed by z-VAD-FMK and z-IETD-FMK and also significantly reduced by z-DEVD-FMK. Inhibitory effects of z-LEHD-FMK were detected, but were lower than those of the other inhibitors (Fig. 8b). Activation of caspase 8 by the combined treatment was reassessed by Western blot and caspase 8-like activity assay. As shown in Figures 8c and 8d, caspase 8 cleavage and activity were clearly up-regulated upon the combined treatment with UDCA, but not with oxaliplatin treatment alone. Moreover, z-IETD-FMK suppressed the UDCA/oxaliplatin-triggered cytochrome C release and caspase 3 activation without affecting p53 induction (Fig. 8e). Inhibition of p53 activity by either 30 μM pifithrin-α or silencing significantly reduced UDCA/oxaliplatin-induced caspase 8 and caspase 3 activation (Fig. 9), indicating that caspase 8 activation is a downstream event of p53 activation. Taken together, these results suggest that oxaliplatin triggers apoptosis via p53-caspase 8 pathway when combined with UDCA in HepG2 cells.

Figure 8.

Caspase 8 activation is responsible for oxaliplatin/UDCA-induced apoptotic cell death. (a) Cells were incubated in control medium or in medium containing 25 μM or 50 μM oxaliplatin with (+) or without (−) UDCA for 24 hr and released cytochrome C (cyto C) into the cytosolic fraction, active caspase 3 (p20), and cleaved PARP were analyzed by Western blot. (b) Cells were untreated or pretreated with caspase inhibitors (z-DEVD-FMK, z-IETD-FMK, z-LEHD-FMK, and z-VAD-FMK) for 1 hr followed by 50 μM oxaliplatin plus 100 μM oxaliplatin treatment. After 24 hr incubation, apoptotic cells were counted under fluorescence microscopy after nuclear HO staining. (c, d) Cells were incubated with 50 μM oxaliplatin, or 50 μM oxaliplatin plus 100 μM UDCA for the indicated time and caspase 8 activation was assessed by detecting the active form of caspase 8 (c) or measuring IETD-pNA substrate cleavage (d). Cells were untreated or treated with oxaliplatin, or UDCA plus oxaliplatin without or with z-IETD for 24 hr, then harvested and analyzed by Western blot, in which the protein levels of p53, active caspase 3, cytoplasmic cytochrome C, or tubulin (loading control) were detected (e).

Figure 9.

p53 is an upstream regulator of oxaliplatin/UDCA-induced caspase 8 activation. Cells were untreated or treated with oxaliplatin, or UDCA plus oxaliplatin without or with pifithrin-α (a, c), or cells were transfected with scrambled siRNA (ctl RNAi) or p53 siRNA and then exposed to UDCA plus oxaliplatin (+) or not (−) for 24 hr (b, c). The treated cells were collected and analyzed by Western blot for p53, active caspase 8, active caspase 3, cytoplasmic cytochrome C, or tubulin (loading control) (a, b), or subjected for caspase 8 activity assay (c). **p < 0.01.

Discussion

Necrosis is distinguished from apoptosis by several morphological features including loss of cellular membrane integrity, which consequently leads to the release of cytoplasmic contents into the extracellular space, and results in inflammatory responses in the surrounding tissues.15, 16 Necrosis and the subsequent inflammatory response have been suggested to exert potent tumor promoting activities through aggravation of tumor growth, angiogenesis and invasion,38, 39 and are, thus, considered to be risk factors for the patient harboring a tumor. Especially, the liver is an organ where massive necrosis is observed upon exposure to drugs and under various stress conditions such as ischemia and viral infections.40 We observed that oxaliplatin, a widely used anticancer drug, induced relatively high rate of necrosis in HepG2, SK-Hep1, SNU423 and Hep3B HCC cells in a fashion that was different from the predominant apoptosis that occurred in HCT116 or HT29 colon cancer cells. This highlights the increased susceptibility of hepatic cells to necrosis.

In this study, we investigated the effects of the hydrophilic bile acid UDCA on oxaliplatin-induced cytotoxicity and necrotic cell death in HepG2 cells. UDCA is well-utilized not only in the treatment of cholestatic liver disease but also in tumor prevention.24, 29, 41 The activity of UDCA to suppress cholestatic liver disease has been intensively studied by many investigators and is attributed to the compounds' cytoprotective roles that include the prevention of pro-apoptotic gene expression, collapse of the mitochondrial membrane potential, or death receptor-mediated cell death events.31, 33, 42–44 On the other hand, UDCA has been shown to exert tumor suppressing activity by pro-apoptotic mechanisms32, 33 and enhance CPT-11 induced-apoptosis in colon cancer cells and promote an apoptotic response of tumor cells to photosensitizers and Bcl-2/Bcl-XL antagonist HA14-1.45, 46 Pro-apoptotic activity of UDCA was also demonstrated in hepatocytes.47 Although the mechanisms driving tumor suppressive activity are not fully understood yet, the tumor suppressive activity of UDCA is explained by its potential to activate pro-apoptotic mechanisms, inhibit the cell cycle, or prevent oncogenic factors. In addition, the anti-inflammatory action of UDCA has been suggested to be a crucial mechanism underlying treatment of various disorders, including cancer, primary sclerosing cholangitis and inflammation in symptomatic gallbladders with cholesterol gallstones.26, 48, 49 In this study, we report a novel activity of UDCA as a death mode regulator in oxaliplatin and other platinum anticancer drug-induced cell death. We also observed that other types of hydrophilic bile acids (TUDCA and THDCA) but not a highly hydrophobic cytotoxic bile acid (DCA) could switch oxaliplatin-induced necrosis to apoptosis. These results suggest that the cell death mode switching effect of UDCA is a general phenomenon that is shared with other platinum drugs and is linked to its hydrophilic feature. Our finding may provide another explanation that UDCA may exert tumor suppressive activity through the modulation of inflammatory responses via necrosis-to-apoptosis switching activity.

In an attempt to identify the mechanisms underlying the necrosis-to-apoptosis switching activity of UDCA, we examined the role of ROS in the cell death mode determination because ROS is suggested to play critical roles in necrotic cell death.50, 51 For instance, excess ROS could trigger necrosis in response to metabolic stress and alleviation of ROS production by antioxidants switches the death mode from necrosis to apoptosis.52, 53 In addition, the superoxide generator menadione prevents apoptosis by inactivation of caspases, thus leading to necrosis in HepG2 cells.54 We observed that oxaliplatin gradually increased ROS level and UDCA could suppress ROS production. UDCA reduces oxidative stress by preventing mitochondrial ROS production or lipid peroxidation.43, 55 As oxaliplatin-induced necrosis was suppressed by the antioxidants NAC, BHA, Tiron and Trolox, ROS seems to be responsible for triggering the necrotic cell death. However, in addition to necrosis, apoptosis observed in oxaliplatin-treated cells could be also prevented by NAC, indicating a dual role of ROS in the induction of necrosis as well as apoptosis. At present, it is not fully understood how ROS trigger both necrosis and apoptosis, but the ROS quantity seems to be critical for determining the cell death mode. Interestingly, UDCA did not affect the initially slight ROS production observed at an early time point (1 hr) (Fig. 6a). This result suggests that this initially low level of ROS production may be responsible for the initiation of apoptosis, and that UDCA can switch the cell death mode to apoptosis without affecting the initial ROS production. However, because BHA, Tiron and Trolox, which are anti-oxidants that exhibit narrow specificities for ROS compared to NAC, could reduce oxaliplatin-induced necrosis but not switch the cell death mode to apoptosis, the cell death mode switching effect of UDCA is not explained only by its ROS reducing activity. The details of UDCA activity remain to be clarified.

Next, we investigated the possible involvement of AP-1, NF-kB and p53 as a death mode regulator. DCA is reported to activate AP-1 and NF-kB, whereas UDCA inhibit those induced by interleukin 1 beta and DCA in colon cancer cells.56 In this study, we showed that UDCA did not affect the changes in protein level or binding activity of these factors that occurred by oxaliplatin, and silencing of c-Jun or p65 did not affect the UDCA-induced apoptotic switch. Thus, AP-1 or NF-kB appeared not to be involved in the UDCA-mediated death mode regulation. Another factor p53 is a key regulator of apoptotic pathway and plays a crucial role(s) in platinum drug-induced apoptosis in many different types of cancer cells, including colon cancer, ovarian cancer and renal epithelial cells.57–60 Either combined treatment of UDCA and oxaliplatin or treatment of oxaliplatin alone could increase p53 phosphorylation and protein levels in HepG2 cells. Inhibition of p53 activity by pifithrin-α or RNA silencing markedly reduced UDCA/oxaliplatin-induced apoptosis without affecting oxaliplatin-induced necrosis. Further, UDCA did not induce oxaliplatin-induced cell death mode switch to apoptosis in p53 mutant SNU-423 or p53 null Hep3B cells. In addition, treatment with NAC (that suppresses apoptosis) but not BHA, Tiron and Trolox (that do not prevent apoptosis) prevented oxaliplatin-induced p53 activation. These results suggest that p53 may be an essential factor regulating apoptotic program potentiated by oxaliplatin. Our observation that p53 phosphorylation and protein level could be also upregulated by oxaliplatin alone indicate that the early signals elicited by oxaliplatin may be pro-apoptotic, as evidenced by p53 activation, but that certain late-onset signals such as excessive ROS production block a further link to the downstream effectors. Thus, the necrotic path may be chosen instead of the apoptotic path in HepG2 cells. We also showed that caspase 8 was activated dependently of p53 and inhibition of caspase 8 by z-IETD-FMK suppressed the apoptosis induced by the combined treatment, indicating a critical role(s) of p53-caspase 8 pathway in apoptosis. On the basis of these results, we suggest that the action point of UDCA that switches necrosis to apoptosis may exist between p53 activation and caspase 8 activation (Fig. 10). In general, caspase 8 is known to be activated by an extrinsic pathway including death receptor activation. For instance, cisplatin induces apoptosis in Jurkat cells through ROS-dependent Fas aggregation and activation of the FADD-caspase 8 cascade.61 In addition, UDCA as a membrane stabilizer has been described, suggesting the possible relationship of death mode regulatory mechanisms with membrane-associated events.62, 63 On the other hand, caspase 8 can also be activated by a FADD-independent mechanism including the mitochondrial pathway.64, 65 Therefore, the molecular mechanism for caspase 8 activation remains to be elucidated.

Figure 10.

Model for the cell death mode switching effect of UDCA on oxalipatin-induced necrosis.

So far, successful treatment of HCC with anti-cancer drugs has not been demonstrated. Even though certain drugs such as paclitaxel induce cytotoxicity in HCC cells such as HuH7 and HepG2, they do not show a very promising result in patients with unresectable HCC.66 Recently, administration of platinum drugs such as cisplatin and oxaliplatin has been shown to be effective in the treatment of HCC,13, 14, 67 but their efficacy is still moderate. In this study, we show that necrotic cell death is induced by treatment of oxaliplatin and other platinum drugs, and demonstrate that UDCA can switch platinum drug-induced necrosis to apoptosis. Combined treatment of UDCA does not affect 5-FU-induced apoptotic cell death or doxorubicin-induced necrotic cell death. Platinum-based chemotherapeutic agents alkylate and cause cross-linking of DNA, whereas doxorubicin intercalates DNA. Recently, cell death in response to DNA alkylators such as N-methyl-N′-nitro-N-nitrosoguanidine have been reported to be a PARP-mediated programmed necrosis. Thus, the cell death mode switching effect of UDCA may be linked to alkylating agent-triggered programmed necrosis. From this study, our results may provide new insight to develop a UDCA combination strategy in platinum-based chemotherapeutic treatment for patients with HCC, although clinical effects of combined treatment of UDCA and oxaliplatin should be further evaluated.

Acknowledgements

The authors wish to thank Prof. Tae-Hyoung Kim, Department of Biochemistry, Chosun University School of Medicine for the excellent discussion.

Ancillary