Portal de Periódicos da CAPES

Oxidative Stress-mediated Apoptosis

2002; Elsevier BV; Volume: 277; Issue: 41 Linguagem: Inglês

10.1074/jbc.m203842200

1083-351X

Jing Wen, Kyung-Ran You, So-Youn Lee, Changho Song, Dae‐Ghon Kim,

Plant Toxicity and Pharmacological Properties

The sesquiterpene lactone parthenolide, the principal active component in medicinal plants, has been used conventionally to treat migraines, inflammation, and tumors. However, the antitumor effects of parthenolide and the mechanism(s) involved are poorly understood. We found that parthenolide effectively inhibits hepatoma cell growth in a tumor cell-specific manner and triggers apoptosis of hepatoma cells. Parthenolide triggered apoptosis in invasive sarcomatoid hepatocellular carcinoma cells (SH-J1) as well as in other ordinary hepatoma cells at 5–10 μmconcentrations and arrested the cell growth (at G2/M) at sublethal concentrations (1–3 μm). During parthenolide-induced apoptosis, depletion of glutathione, generation of reactive oxygen species, reduction of mitochondrial transmembrane potential, activation of caspases (caspases-7, -8, and -9), overexpression of GADD153 (an oxidative stress or anticancer agent inducible gene), and subsequent apoptotic cell death was observed. This induced apoptosis could be effectively inhibited or abrogated by an antioxidantN-acetyl-l-cysteine, whereas l-buthionine-(S,R)-sulfoximine enhanced it. Furthermore, stable overexpression of GADD153sensitized the cells to apoptosis induced by parthenolide, and this susceptibility could be reversed by transfection with an antisense toGADD153. Parthenolide did not alter the expression of Bcl-2 or Bcl-XL proteins during apoptosis in hepatoma cells. Oxidative stress may contribute to parthenolide-induced apoptosis and to GADD153 overexpression in a glutathione-sensitive manner. The sensitivity of tumor cells to parthenolide appears to result from the low expression level of the multifunctional detoxification enzyme glutathione S-transferase π. Finally, parthenolide and its derivatives may be useful chemotherapeutic agents to treat these invasive cancers. The sesquiterpene lactone parthenolide, the principal active component in medicinal plants, has been used conventionally to treat migraines, inflammation, and tumors. However, the antitumor effects of parthenolide and the mechanism(s) involved are poorly understood. We found that parthenolide effectively inhibits hepatoma cell growth in a tumor cell-specific manner and triggers apoptosis of hepatoma cells. Parthenolide triggered apoptosis in invasive sarcomatoid hepatocellular carcinoma cells (SH-J1) as well as in other ordinary hepatoma cells at 5–10 μmconcentrations and arrested the cell growth (at G2/M) at sublethal concentrations (1–3 μm). During parthenolide-induced apoptosis, depletion of glutathione, generation of reactive oxygen species, reduction of mitochondrial transmembrane potential, activation of caspases (caspases-7, -8, and -9), overexpression of GADD153 (an oxidative stress or anticancer agent inducible gene), and subsequent apoptotic cell death was observed. This induced apoptosis could be effectively inhibited or abrogated by an antioxidantN-acetyl-l-cysteine, whereas l-buthionine-(S,R)-sulfoximine enhanced it. Furthermore, stable overexpression of GADD153sensitized the cells to apoptosis induced by parthenolide, and this susceptibility could be reversed by transfection with an antisense toGADD153. Parthenolide did not alter the expression of Bcl-2 or Bcl-XL proteins during apoptosis in hepatoma cells. Oxidative stress may contribute to parthenolide-induced apoptosis and to GADD153 overexpression in a glutathione-sensitive manner. The sensitivity of tumor cells to parthenolide appears to result from the low expression level of the multifunctional detoxification enzyme glutathione S-transferase π. Finally, parthenolide and its derivatives may be useful chemotherapeutic agents to treat these invasive cancers. reactive oxygen species mitochondrial transmembrane potential 5,6-carboxy-2′,7′-dichlorofluorescein diacetate 5,5′,6,6′-tetraethylbenzimidazocarbocyanine iodide glyceraldehyde-3-phosphate dehydrogenase glutathione N-acetyl-l-cysteine l-buthionine-(S,R)-sulfoximine poly(ADP-ribose)polymerase glutathione S-transferase 1,4-piperazinediethanesulfonic acid Sesquiterpene lactones are isolated from extracts of Mexican-Indian medicinal plants and have been widely used in indigenous medical practices, including treatment of migraines (1Johnson E.S. Kadam N.P. Hylands D.M. Hylands P.J. Br. Med. J. 1985; 31: 569-573Crossref Scopus (257) Google Scholar, 2Biggs M.J. Johnson E.S. Persaud N.P. Ratcliffe D.M. Lancet. 1982; 2: 776Abstract PubMed Scopus (29) Google Scholar), inflammation (3Hall I.H. Lee K.H. Starnes C.O. Sumida Y., Wu, R.Y. Waddell T.G. Cochran J.W. Gerhart K.G. J. Pharmacol. Sci. 1979; 68: 537-542Abstract Full Text PDF PubMed Scopus (162) Google Scholar), and tumors (4Douros J. Suffness M. Cancer Chemother. Pharmacol. 1978; 1: 91-100Crossref PubMed Scopus (41) Google Scholar, 5Lee K.H. Hall I.H. Mar E.C. Starnes C.O. ElGebaly S.A. Waddell T.G. Hadgraft R.I. Ruffner C.G. Weidner I. Science. 1977; 196: 533-536Crossref PubMed Scopus (234) Google Scholar). Parthenolide, the major sesquiterpene lactone found in medicinal plants such as feverfew (Tanacetum parthenium), is known to inhibit interleukin 1- and tumor necrosis factor α-mediated NF-κB activation, which is responsible for its anti-inflammatory activity (6Bork P.M. Schmitz M.L. Kuhnt M. Escher C. Heinrich M. FEBS Lett. 1997; 402: 85-90Crossref PubMed Scopus (302) Google Scholar,7Hehner S.P. Hofmann T.G. Droge W. Schmitz M.L. J. Immunol. 1999; 163: 5617-5623PubMed Google Scholar). Parthenolide has been reported to inhibit NF-κB by targeting the IκB kinase complex (7Hehner S.P. Hofmann T.G. Droge W. Schmitz M.L. J. Immunol. 1999; 163: 5617-5623PubMed Google Scholar). Parthenolide also exerts an anti-inflammatory activity by inhibiting the expression of inducible cyclooxygenase, proinflammatory cytokines (8Hwang D. Fischer N.H. Jang B.C. Tak H. Kim J.K. Lee W. Biochem. Biophys. Res. Commun. 1996; 226: 810-818Crossref PubMed Scopus (163) Google Scholar), and inducible nitric-oxide synthase (9Fukuda K. Hibiya Y. Mutoh M. Ohno Y. Yamashita K. Akao S. Fujiwara H. Biochem. Pharmacol. 2000; 60: 595-600Crossref PubMed Scopus (36) Google Scholar). In contrast, the cytotoxic and antitumor effects of sesquiterpene lactones have not been well studied because of their low potency. Recently it has been reported that parthenolide inhibits the in vitro growth of tumor cells in a cytostatic fashion, and it has been proposed that if their selectively cytotoxic or cytostatic actions against tumor cells can be established, sesquiterpene lactones may represent a new class of cancer chemotherapeutic drugs (10Ross J.J. Arnason J.T. Birnboim H.C. Planta Med. 1999; 65: 126-129Crossref PubMed Scopus (65) Google Scholar). Sarcomatoid changes of epithelial neoplasms (carcinomas with spindle-cell components) occur in many organs and histologic types. Sarcomatoid elements in liver cancers are derived from dedifferentiation of hepatocellular carcinoma or cholangiocarcinoma. Although the incidences of spindle-cell hepatocellular carcinoma and cholangiocarcinoma are less than 10% (11Haratake J. Horie A. Cancer. 1991; 68: 93-97Crossref PubMed Scopus (83) Google Scholar, 12Maeda T. Adachi E. Kajiyama K. Takenaka K. Sugimachi K. Tsuneyoshi M. Cancer. 1996; 77: 51-57Crossref PubMed Scopus (94) Google Scholar) and ∼5% (13Nakajima T. Tajima Y. Sugano I. Nagao K. Kondo Y. Wada K. Cancer. 1993; 72: 1872-1877Crossref PubMed Scopus (56) Google Scholar), respectively, the prognosis for patients with sarcomatoid liver carcinoma is worse than for those with ordinary carcinoma because of the aggressive intrahepatic spreading and the frequent metastasis of sarcomatous cells (12Maeda T. Adachi E. Kajiyama K. Takenaka K. Sugimachi K. Tsuneyoshi M. Cancer. 1996; 77: 51-57Crossref PubMed Scopus (94) Google Scholar). Thus, we have established a sarcomatoid hepatocellular carcinoma cell line (designated as SH-J1) and have searched for agents that arrest their growth and/or induce their apoptosis. Intriguingly, we have now observed that parthenolide effectively inhibits the proliferation and induces the apoptosis of those sarcomatoid hepatocellular carcinoma cells as well as other ordinary hepatoma cells. The underlying mechanism for the antitumor effects of parthenolide seems to be mediated by oxidative stress. This oxidative stress may subsequently be associated with overexpression ofGADD153, a growth arrest and DNA damage-inducible gene. Chang liver cells and human hepatoma cell lines, including Hep 3B, PLC/PRF/5, and SK-HEP-1, were obtained from the American Type Culture Collection (ATCC) (Manassas, VA). The sarcomatous SH-J1 cells were established in our laboratory (14Kim D.G. Park S.Y. Kim H. Chun Y.H. Moon W.S. Park S.H. Cancer Genet. Cytogenet. 2002; 132: 120-124Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar). To determine the effects on growth, cells were cultured on coverslips and treated with the indicated concentrations of parthenolide or vehicle (Me2SO) in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% fetal bovine serum for 48 h. The cells were fixed with 4% paraformaldehyde in 0.1m sodium phosphate buffer (pH 7.6) and were then stained with the Diff-Quick stain set (Dade Diagnostics of P. R. Inc., Aguada, Puerto Rico). To measure cytotoxic cell death, the cells were cultured in 6-cm culture dishes and exposed to the indicated concentrations of parthenolide. Cell death was determined by trypan blue dye assay, and the percentage of apoptotic cells was evaluated by Hoechst 33258 (Sigma) staining. At least 200 cells were counted for each time point, and all counting was done in a blinded fashion. SH-J1 cells were plated in 96-well flat-bottom microtiter plates (Nalge Nunc International, Naperville, IL) at a density of 103 cells/well and treated with 1–3 μm parthenolide in medium for 48 h. The cells were labeled with 1 μCi [methyl-3H]thymidine/well (specific activity, 5 Ci/mmol; Amersham Biosciences) for 16 h at 37 °C and then harvested on fiberglass filter paper strips with a multiple automated sample harvester (Inotech, Dottikon, Switzerland). Each sample was lysed hypotonically, and radioactivity was measured in a Tri-Carb 4530 liquid scintillation counter (Packard Instrument Company). Measurements of cell cycle distribution and apoptotic/hypodiploid cells were performed using a modification of the technique described previously (15Kim D.G., Jo, B.H. You K.R. Ahn D.S. Cancer Lett. 1996; 107: 149-159Crossref PubMed Scopus (37) Google Scholar). For synchronization, the cells were treated with 0.5 μg/ml nocodazole (Calbiochem) to induce M block for 16 h, washed, and incubated in the presence or absence of 3 μm parthenolide. Data were analyzed as single-parameter frequency histograms in an SFIT model. The sub-G1 fraction was estimated by gating hypodiploid cells in the DNA histogram using a C-30 program. SH-J1 cells (1 × 106) were seeded in 6-cm Petri dishes and allowed to settle and attach. Cells were then treated with the indicated concentrations of parthenolide for 48 h. For analysis of genomic DNA, cells were harvested and collected together with nonattached cells in the supernatant. Cells were resuspended in 0.5 ml of lysis buffer (50 mm Tris-HCl, 100 mm EDTA, 0.5% sodium dodecyl sulfate (pH 8.0) containing 0.1 mg/ml RNase A. After incubation at 37 °C for 30 min, extracts were treated with 1 mg/ml proteinase K for an additional 16 h at 37 °C. DNA was extracted with phenol/chloroform and then with chloroform and precipitated with ethanol and sodium acetate, and 5 μg of DNA was separated on a 2% agarose gel. DNA in the gel was stained with ethidium bromide, visualized under UV light, and photographed. The intracellular generation of ROS was measured using the oxidation-sensitive fluorescein 5,6-carboxy-2′,7′-dichlorofluorescein diacetate (DCFH-DA) (Molecular Probes, Eugene, OR). For measurement of ΔΨm, cells were treated with 10 μm parthenolide for the indicated times. 5,5′,6,6′-Tetraethylbenzimidazocarbocyanine iodide (JC-1) (Molecular Probes) was added to the medium at 5 μg/ml, and incubation was continued in the dark for 15 min (16You K.R. Wen J. Lee S.T. Kim D.G. J. Biol. Chem. 2002; 277: 3870-3877Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). Stained cells were harvested, washed once in phosphate-buffered saline, and analyzed by flow cytometry. A BD Biosciences FACscan was used to analyze a minimum of 10,000 cells/sample. Forward and side scatter were used to gate the viable population of cells. At relatively high ΔΨm the dye JC-1 forms J-aggregates, which emit at 590 nm, in the red/orange range of visible light (FL-2 channel). However, in the absence of or at low ΔΨm, JC-1 exists as a monomer, remaining in the cell but emitting at 527 nm, in the green range (FL-1 channel). Cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum until they were 60% confluent at which time they were treated with 5 μmparthenolide for 48 h. Total RNA was extracted from treated or untreated cells using phenol and guanidine thiocyanate solution (Tri Reagent; Molecular Research Center Inc., Cincinnati, OH) following the manufacturer's instructions. RNAs were then fractionated by electrophoresis on 1.0% agarose gels containing formaldehyde and transferred to membranes. Northern blot analysis was performed as described previously (16You K.R. Wen J. Lee S.T. Kim D.G. J. Biol. Chem. 2002; 277: 3870-3877Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). The blots were stripped and then rehybridized with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe to normalize the amount of RNA loaded. Transfection of the GADD153 gene into SH-J1 cells was performed using an expression plasmid vector encoding human GADD153 cDNA or control pcDNA3. The construct of the GADD153 expression vector was made by ligating humanGADD153 (a gift from Dr. Nikki J. Holbrook, NIA, National Institutes of Health, Baltimore, MD) with pcDNA3 at eachBamHI/XhoI site in the sense orientation. To generate the antisense GADD153 expression vector, humanGADD153 was PCR-amplified with the forward primer containing the XbaI restriction enzyme site (5′-GCTCTAGAGGGCTGCAGAGATGGC-3′) and the reverse primer containing theEcoRI restriction enzyme site (5′-GGAATTCGGGACTGATGCTCCCA-3′). The PCR-amplified GADD153double strand DNA was ligated to pcDNA3 at theXbaI/EcoRI site in the antisense orientation. Transfections were carried out using Lipofectin (Invitrogen) according to the manufacturer's protocol. GADD153-transfected and neo-transfected cells were selected in the presence of 600 μg/ml G418 for 2–3 weeks. Finally, individual colonies were isolated using cloning rings, expanded, and assayed for expression of the transfected gene by Northern analysis and by Western analysis. pGADD-LUC, a hamsterGADD153 promoter-driven luciferase reporter construct, was a gift from Dr. Nikki J. Holbrook (17Fawcett T.W. Eastman H.B. Martindale J.L. Holbrook N.J. J. Biol. Chem. 1996; 271: 14285-14289Abstract Full Text Full Text PDF PubMed Scopus (104) Google Scholar). Cells were transfected with the pGADD-LUC construct using Lipofectin (Invitrogen) as noted above. Cells were plated at 2 × 104 cells/well in 24-well plates, and 18 h later, the cells were incubated at 37 °C with 500 ng of GADD-LUC plasmid and 50 ng of pRL-TK plasmid (Promega, Madison, WI) in the presence of Lipofectin for 16 h. The cells were then replenished with complete medium and treated with equitoxic levels of parthenolide. The cells were lysed in 120 μl of lysis buffer at the indicated time intervals and stored at −20 °C until measurement. Luciferase activity was measured using the dual luciferase reporter Assay system (Promega) as instructed by the manufacturer and normalized by Renilla luciferase activity. Reduced glutathione was measured using a colorimetric assay kit (OXIS International, Inc., Portland, OR). Cells were treated with either parthenolide alone or in combination with the glutathione precursorN-acetyl-l-cysteine (NAC) for 48 h or with the glutathione depleting agentl-buthionine-(S,R)-sulfoximine (BSO). Cells treated with BSO were incubated overnight and then treated with parthenolide after washing for 48 h. The homogenate was centrifuged at 3000 × g, 4 °C for 10 min, and its supernatant was used for GSH measurement according to the manufacturer's instruction. The cell pellet dissolved in 1m NaOH was used for the measurement of protein content (18You K.R. Shin M.N. Park R.K. Lee S.O. Kim D.G. Hepatology. 2001; 34: 1119-1127Crossref PubMed Scopus (41) Google Scholar). The GSH content was expressed as nmol/mg of protein or percent of the control. Cell lysis and immunoblotting were performed as described previously (18You K.R. Shin M.N. Park R.K. Lee S.O. Kim D.G. Hepatology. 2001; 34: 1119-1127Crossref PubMed Scopus (41) Google Scholar). Cytoplasmic extracts were prepared from cells in lysis buffer (10 mm PIPES (pH 7.4), 10 mm KCl, 2 mm MgCl2, 1 mm dithiothreitol, 5 mm ethylene glycol-bis(β-aminoethyl ether)N,N,N′,N′-tetraacetic acid, 25 μg/ml leupeptin, 40 mm β-glycerophosphate, and 1 mm phenylmethylsulfonyl fluoride). Cells were homogenized with a Dounce homogenizer, and lysis was confirmed by microscopy. Nuclei were collected by centrifugation through 30% sucrose (800 × g, 10 min at 4 °C), mitochondria were collected at 10,000 × g for 20 min at 4 °C, and the remaining cell membranes were removed by centrifugation at 100,000 ×g for 45 min at 4 °C, as described previously. The supernatant (S-100 fraction) was used as the cytosolic extract. After adjustment for protein concentration, 30 μg of each cell lysate was boiled in Laemmli buffer and resolved by 15% SDS-PAGE before immunoblot analysis with antibody against cytochrome c(Pharmingen). Signals were detected using an ECL Western blotting kit (Amersham Biosciences). The mouse anti-poly(ADP-ribose)polymerase (PARP) monoclonal antibody (C2–10), rabbit anti-caspase-3 polyclonal antibody, mouse anti-caspase-6 monoclonal antibody (B93-4), mouse anti-caspase-7 monoclonal antibody (B94-1), rabbit anti-caspase-8 polyclonal antibody (B9-2), and mouse anti-caspase-9 monoclonal antibody (B40) were purchased from Pharmingen. The mouse anti-Bcl-2 monoclonal antibody (Ab-1) and mouse anti-Bcl-XL monoclonal antibody (H-5) were obtained from Calbiochem and from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively. All data were entered into Microsoft Excel 5.0. GraphPad Software was used to perform paired ttests. All p values of less than 0.05 were considered to be statistically significant. To determine whether parthenolide inhibits cell proliferation, SH-J1 cells were cultured on coverslips and exposed to parthenolide at different concentrations (0–10 μm) for 48 h. The cells were fixed and stained with the Diff-Quick stain set, and cell numbers were histocytologically counted. Parthenolide effectively decreased cell numbers in a dose-dependent manner (Fig. 1 A). To examine whether these cytotoxic effects of parthenolide are tumor cell-specific, SH-J1 cells and Chang liver cells (which were derived from normal liver tissue) were treated with 10 μmparthenolide for 48 h. Histocytological examination revealed that the nonmalignant Chang liver cells were resistant to parthenolide-mediated cell growth inhibition, as compared with SH-J1 cells. The cell numbers of Chang liver cells and SH-J1 cells are about 80 and 20% that of the untreated controls, respectively (Fig.1 B). The parthenolide-induced reduction of cell number may result from cell cycle inhibition or cytotoxic cell death. Thus, to measure quantitative cell death, SH-J1 cells, Chang liver cells, and ordinary hepatoma cells (Hep 3B, PLC/PRF/5, and SK-HEP-1) were cultured in 6-cm dishes and exposed to parthenolide at different concentrations (0–10 μm) for 48 h. The cells were harvested, and dead cells stained with trypan blue dye were counted. Parthenolide decreased SH-J1 cell survival in a dose-dependent manner, and the IC50 for cell survival was about 7.5 μm (Fig. 1 C). The quantitation of cell survival of SH-J1 and Chang liver cells was consistent with the histocytological examination (24.7 ± 2.2% versus80.0 ± 7.4%, p < 0.01). These results suggest that parthenolide exerts its cytotoxic effects in a tumor cell-specific manner. However, cytotoxic cell death was rarely detectable at lower concentrations (1–2.5 μm) of parthenolide. Hep 3B and PLC/PRF/5 cells were more susceptible to the drug-induced cell death than were SK-HEP-1 cells. To demonstrate whether parthenolide induces apoptotic cell death, DNA fragmentation was analyzed by gel electrophoresis. SH-J1 cells treated with parthenolide showed increased fragmentation of lower molecular weight DNAs at higher concentrations of parthenolide (between 5 and 10 μm) (Fig.1 D). Lower concentrations of parthenolide (0–2.5 μm) did not result in any DNA fragmentation. Cells with condensed or fragmented chromosomes and with membrane blebbing, both characteristics indicative of apoptotic morphologies, were examined by staining with Hoechst 33258 after treatment with parthenolide (5–10 μm) (Fig. 1 E). These results indicate that parthenolide effectively induces apoptosis at the relatively higher concentrations but not at the lower concentrations. Lower concentrations of parthenolide seem to reduce cell number by cell cycle inhibition. This was supported by the previous report that parthenolide inhibited cell growth in a cytostatic fashion and that the effect was reversible at lower concentrations (10Ross J.J. Arnason J.T. Birnboim H.C. Planta Med. 1999; 65: 126-129Crossref PubMed Scopus (65) Google Scholar). Thus, we quantitatively determined the effect of parthenolide on cell proliferation at sublethal concentrations (1–3 μm). Using a thymidine uptake assay, DNA synthesis was significantly inhibited by 8–51% in cells treated with 1–3 μmparthenolide (Fig. 2 A), which is consistent with a previous report (1Johnson E.S. Kadam N.P. Hylands D.M. Hylands P.J. Br. Med. J. 1985; 31: 569-573Crossref Scopus (257) Google Scholar). We next analyzed the impact of parthenolide on the cell cycle of SH-J1 cells. In asynchronized cells, parthenolide increased the fraction of cells in G2/M from 15 to 30% compared with untreated control cells at 12 h of culture (Fig. 2 B). Thus, we further studied the cell cycle inhibition in synchronized cells. For synchronization, SH-J1 cells were arrested in M phase with 0.5 μg/ml nocodazole. Cells were then analyzed at 3-h interval after release from the M block in the presence or absence of parthenolide. Cells treated with parthenolide clearly showed delayed cell cycling (G2/M arrest). In control cells, a normal cell cycle distribution appeared 12 h after release, whereas it appeared only after 36 h after release in cells treated with parthenolide. Treatment of SH-J1 cells with 10 μmparthenolide resulted in ∼80% apoptotic death within 48 h. Remarkably, in the presence of non-cytotoxic doses of general antioxidants, including NAC (800 μm), pyrrolidine dithiocarbamate (0.25 μm), and nordihydroguaiaretic acid (0.1 μm), the killing activity of parthenolide is effectively inhibited or abrogated (Fig.3 A). NAC was able to completely abrogate parthenolide-induced apoptosis (95 ± 5.5% inhibition, p < 0.01). Vitamin E (200 μm) and vitamin C (200 μm) also effectively inhibited parthenolide-induced apoptosis by 40 ± 3.5 and by 18 ± 2%, respectively (p < 0.05) (data not shown). Thus, the protection afforded by antioxidants against the induction of cell death by parthenolide suggests that free radicals may be involved in this phenomenon. To examine this possibility, we measured the levels of intracellular free radicals before and after exposure to parthenolide using the cell permeant dye DCFH-DA. Measurements of cellular fluorescence revealed that parthenolide generated a 3-fold increase of intracellular ROS in SH-J1 cells. A noncytotoxic dose of NAC completely suppressed parthenolide-induced ROS generation (Fig. 3 B). Reduction of ΔΨm is believed to be mediated by ROS. An assay to detect these changes of ΔΨm in mitochondrial function employs a fluorescent cation, JC-1, which emits a red color when sequestered in mitochondria of healthy cells but emits a green color when it is in the cytoplasmic compartment of apoptotic cells. Treatment of cells with 10 μm parthenolide for 48 h caused a disruption of ΔΨm, as evidenced by a shift in the fluorescence of the probe JC-1, from 3.3 ± 0.5% in the right lower quadrant fluorescence fraction in control cells to 78.3 ± 5.7% in parthenolide-treated cells (Fig. 3 C). The parthenolide-induced reduction in ΔΨm was completely abrogated by NAC (5.2 ± 0.4%). The sum of these results suggests that parthenolide induces ΔΨm dissipation in an antioxidant-sensitive pathway and indicates that parthenolide-mediated generation of ROS causes the reduction in ΔΨm. The ΔΨm disruption appears to be critical for the apoptosis cascade. To determine whether the loss of ΔΨm precedes cytochrome c release, we analyzed the time course of ΔΨm reduction and cytochrome c release. Flow cytometric measurements from 0 to 48 h after parthenolide treatment revealed that the percentage of cells with low ΔΨm in the right lower quadrant began to increase 12 h after the treatment started and reached a maximum at the end of the experimental period, in a time-dependent manner (Fig. 4 A). About 80% of the cells showed a collapsed ΔΨm 48 h after treatment. The time course of cytochrome release analyzed by Western immunoblot revealed that cytosolic cytochrome c appeared 12 h after parthenolide treatment (Fig. 4 B). Thus, our data suggest that the collapse in ΔΨm correlates with mitochondrial cytochrome c release. Because caspase activation plays a central role in the induction of apoptosis (19Grutter M.G. Curr. Opin. Struct. Biol. 2000; 10: 649-655Crossref PubMed Scopus (437) Google Scholar), to understand the mechanism of parthenolide-induced apoptosis we examined its effect on the activation of initiator and effector caspases and cleavage of their substrate, PARP. The cleavage and/or the decreased level of the apoptotic substrate PARP and procaspase-7 (which is indicative of apoptotic cell death) were demonstrated along the time course, implying that cytochrome c release leads to apoptotic cell death through oxidative stress as previously described (20Yang J.C. Cortopassi G. Free Radic. Biol. Med. 1998; 24: 624-631Crossref PubMed Scopus (197) Google Scholar). Interestingly, parthenolide caused a decrease in the procaspase-7 level and its cleavage without changing the level of procaspases-3 or -6, implying that caspase-7 activation is the major effector pathway involved in parthenolide-induced apoptosis. Activation of initiator caspases-8 and -9 were observed in a time-dependent manner, as evidenced by decreased level or cleavage of their pro-caspases. To identify the gene(s) responsible for parthenolide-induced apoptosis, we screened genes known to be involved in drug-induced apoptosis through generation of ROS. We focused on GADD153, the growth arrest and DNA damage gene, because parthenolide triggered suddenGADD153 mRNA overexpression during exposure to the drug (Fig. 5 A). To determine the effect of parthenolide on the activation of the GADD153promoter, SH-J1 cells were transiently transfected with pGADD-LUC, which contains the hamster GADD153 promoter coupled to the luciferase reporter gene. The transfected cells were exposed to 5 μm parthenolide for 48 h (Fig. 5 B). The change in luciferase activity, expressed relative to the level in untreated control cells, indicated that parthenolide increasedGADD153 promoter activation maximally after 24 h of exposure, which preceded the increase of endogenous GADD153mRNA after 36 h of treatment with parthenolide. These results suggest that parthenolide transcriptionally regulatesGADD153 mRNA expression in SH-J1 cells. To further investigate the transcriptional regulation by parthenolide, the decline of GADD153 mRNA levels in parthenolide-treated cells was examined after parthenolide withdrawal and/or the addition of an inhibitor of transcription, actinomycin D. Subconfluent SH-J1 cells were treated with 5 μm parthenolide for 48 h to bring about the accumulation of GADD153 mRNA levels. Cells treated with parthenolide were then fed parthenolide-free medium to which actinomycin D (5 μg/ml) or vehicle was added (21Eymin B. Dubrez L. Allouche M. Solary E. Cancer Res. 1997; 57: 686-695PubMed Google Scholar). At various times thereafter, relative mRNA levels were calculated by comparing GAPDH-normalized values with the levels observed in cells at time 0 (Fig. 5 C). Parthenolide sustainedGADD153 mRNA after treatment for 9 h in the cells. In the presence of actinomycin D, parthenolide decreased theGADD153 mRNA levels as if parthenolide had been removed from the culture medium. These results imply that parthenolide regulates GADD153 mRNA levels transcriptionally in SH-J1 cells, in accordance with results from the in vitrotransfection experiment. Parthenolide-induced apoptosis was completely blocked by NAC, a glutathione precursor, but other general antioxidants, including pyrrolidine dithiocarbamate, nordihydroguaiaretic acid, vitamin C, and vitamin E, did not block this apoptosis to the same extent, suggesting that the induction of oxidative stress by parthenolide may be related to glutathione depletion. Thus, we measured the cellular glutathione level during parthenolide-induced apoptosis. Parthenolide exposure resulted in a time-dependent decrease in cellular glutathione, which was about 75% that of the control level within 48 h (Fig.6 A). NAC (800 μm) abrogated the parthenolide-mediated glutathione depletion but increased the cellular glutathione level (171% ± 15.2%, p < 0.05), implying that parthenolide-induced glutathione depletion activates the glutathione rescue syste