Two distinct modes of cell death induced by doxorubicin - Nature

Oncogene (2005) 24, 4765–4777

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Two distinct modes of cell death induced by doxorubicin: apoptosis and cell death through mitotic catastrophe accompanied by senescence-like phenotype Young-Woo Eom1, Mi Ae Kim1, Seok Soon Park1, Mi Jin Goo1, Hyuk Jae Kwon2, Seonghyang Sohn2, Wook-Hwan Kim3, Gyesoon Yoon4 and Kyeong Sook Choi*,1 1 Laboratory of Endocrinology, Ajou University School of Medicine, San 5, Wonchon-Dong, Youngtong-Gu, Suwon 442-749, South Korea; 2Laboratory of Cell Biology, Institute for Medical Sciences, Ajou University School of Medicine, Suwon 442-749, South Korea; 3Department of Surgery, Ajou University School of Medicine, Suwon 442-749, South Korea; 4Department of Biochemistry and Molecular Biology, Ajou University School of Medicine, Suwon 442-749, South Korea

Chronic exposure of many human hepatoma cell lines to a low dose (LD) of doxorubicin induced a senescence-like phenotype (SLP) accompanied by enlargement of cells and increased senescence-associated b-galactosidase activity. LD doxorubicin-induced SLP was preceded by multinucleation and downregulation of multiple proteins with mitotic checkpoint function, including CENP-A, Mad2, BubR1, and Chk1. LD doxorubicin-treated cells eventually underwent cell death through mitotic catastrophe. When we investigated whether LD doxorubicininduced cell death shares biochemical characteristics with high dose (HD) doxorubicin-induced apoptosis in Huh-7 cells, we observed that externalization of phosphatidyl serine and release of mitochondrial cytochrome c into the cytosol was associated with both types of cell death. However, propidium iodide exclusion assays showed that membrane integrity was lost in the initial phase of LD doxorubicin-induced cell death through mitotic catastrophe, whereas it was lost during the late phase of HD doxorubicin-induced apoptosis. Furthermore, HD doxorubicin-induced apoptosis but not LD doxorubicin-induced mitotic catastrophe led to transient activation of NF-jB and strong, sustained activations of p38, c-Jun N-terminal kinase, and caspases. Collectively, these results indicate that different doses of doxorubicin activate different regulatory mechanisms to induce either apoptosis or cell death through mitotic catastrophe. Oncogene (2005) 24, 4765–4777. doi:10.1038/sj.onc.1208627; published online 2 May 2005 Keywords: doxorubicin; mitotic catastrophe; apoptosis; senescence-like phenotype; hepatocellular carcinoma cells

*Correspondence: KS Choi; E-mail: kschoi@ ajou.ac.kr Received 2 July 2004; revised 2 February 2005; accepted 9 February 2005; published online 2 May 2005

Introduction The anthracycline antibiotic, doxorubicin, is one of the most important anticancer agents for solid tumors (Hortobagyi, 1997), and is a valuable component of intra-arterial infusion for the treatment of unresectable hepatocellular carcinoma (HCC) (Acunas and Rozanes, 1999). Free radical formation and DNA damages via inhibition of topoisomerase II may be primarily responsible for the cytotoxic effects of doxorubicin (Gewirtz, 1999). Doxorubicin induces apoptosis via the activation of caspases and disruption of mitochondrial membrane potential (Gamen et al., 2000). Despite the widespread clinical use of doxorubicin, its antiproliferative and death-inducing signal cascades are not yet fully understood. Several reports have demonstrated that various cancer cells treated with a low dose (LD) of doxorubicin show a senescence-like phenotype (SLP) that resembles the replicative senescence of normal cells at the morphological and enzymatic levels (Chang et al., 1999b; Wang et al., 1999). Senescent cells are generally characterized by a reduction in proliferative capacity, adoption of a flattened and enlarged cell shape, and the appearance of senescence-associated b-galactosidase (SA-b-gal) activity (Dimri et al., 1995). LDs of other DNA damaging agents, including various chemotherapeutic drugs and ionizing radiation (Chang et al., 1999a), and introduction of an activated ras oncogene (Serrano et al., 1997) have been reported to induce SLP in cancer cells in a similar manner. However, the molecular events that trigger SLP in cancer cells remain unclear. Other lines of evidence have indicated that LDs of chemotherapeutic drugs (Lock and Stribinskiene, 1996), g-radiation (Hendry and West, 1997), and activated Ras (Miranda et al., 1996) are capable of inducing mitotic catastrophe. Roninson et al. (2001) defined mitotic catastrophe as a type of cell death resulting from abnormal mitosis, usually ending in the formation of large cells with multiple micronuclei and decondensed chromatin. Cells undergoing mitotic catastrophe usually do not show DNA ladder formation (He et al., 2002) or

Apoptosis and cell death through mitotic catastrophe Y-W Eom et al

4766

DNA breaks detectable by terminal deoxynucelotidyl transferase-mediated deoxyuridine triphosphate nickend labeling (TUNEL) staining (Chang et al., 1999a), suggesting that this cell death is nonapoptotic. Castedo et al. (2004) recently proposed that mitotic catastrophe results from a combination of deficient cell-cycle checkpoints and cellular damage, and that failure of cell cycle arrest before or at mitosis triggers an attempt of aberrant chromosome segregation, which culminates in activation of an apoptotic default pathway and subsequent cellular demise. Thus, researchers have not yet reached a consensus as to whether apoptosis and cell death through mitotic catastrophe are fundamentally different death modes. More recently, some research groups have suggested that mitotic catastrophe should be regarded as an abnormal mitosis that leads to cell death (which can occur through necrosis or apoptosis) rather than an actual form of cell death (Chu et al., 2004; Nitta et al., 2004). The fact that subapoptotic doses of the same stimuli induce SLP or cell death by mitotic catastrophe prompted us to investigate the possible link between treatment-induced SLP and cell death through mitotic catastrophe in tumor cells. Here, we show that LD doxorubicin induces abnormal mitosis and SLP in Huh7 human hepatoma cells, and that treated cells finally die through mitotic catastrophe characterized by the formation of multiple micronuclei and loss of membrane integrity. Many human HCC cell lines show a similar cellular response, suggesting that induction of SLP and cell death by mitotic catastrophe may be a general response to LD doxorubicin. This work provides the first concrete evidence that the mode of LD doxorubicin-induced cell death through mitotic catastrophe is morphologically and biochemically distinct from apoptosis induced by high doses (HDs) of the same drug.

Results LD doxorubicin induces a SLP and subsequent cell death through mitotic catastrophe While various cytotoxic agents have been reported to induce senescence in cancer cells at LDs (Chang et al., 1999a, b; Wang et al., 1999), the biochemical changes associated with treatment-induced senescence are not clearly understood. Here, we sought to investigate these changes in cells treated with LD doxorubicin, a widely used anticancer drug. We first determined the optimal condition of doxorubicin for induction of SA-b-gal activity, a biomarker for cellular senescence, in Huh-7 human hepatoma cells. We found that chronic exposure to 50 ng/ml doxorubicin for 6 days induced SA-b-gal activity in over 82% of Huh-7 cells, but not in cells incubated in doxorubicin-free media for the same period (Figure 1a and b). Huh-7 cells treated with 50 ng/ml doxorubicin for 6 days demonstrated a characteristic SLP, including enlargement of cell volume, flattened cell morphology, and the appearance of multinucleated and vacuolated cells (Figure 1a). Moreover, RT–PCR Oncogene

Figure 1 Induction of a SLP by LD doxorubicin. (a) Expression of senescence-associated b-gal in Huh-7 cells following chronic exposure to LD doxorubicin. Huh-7 cells incubated with or without 50 ng/ml doxorubicin for 6 days were fixed and subjected to SA-b-gal assay as described in Materials and methods. Representative pictures are shown (magnification, 200). Note the substantial increase in cell volume and the blue staining of cells treated with doxorubicin. (b) Quantitation of b-gal expression in Huh-7 cells treated with or without doxorubicin. Data are presented as the mean and s.d. (bars) of four independent experiments based on three random fields with 100 cells counted per field. (c) RT–PCR analysis of gene products associated with cellular senescence. RNA was prepared from Huh-7 cells treated with 50 ng/ml doxorubicin for the indicated time points and RT– PCR was performed as described in Materials and methods. (d) Changes in cellular viability following treatment with LD doxorubicin were analysed by staining with trypan blue

analysis of gene products commonly associated with cellular senescence showed that the mRNA levels of osteonectin, SM22, TGase II, and PAI-1 (Dumont et al., 2000) were significantly increased at 6 days and were maintained up to 9 days (Figure 1c). The percentage of SA-b-gal-positive cells decreased after 9 days of doxorubicin exposure (Figure 1b), and the giant cells gradually floated from the plate. Trypan blue exclusion assays showed that cellular viability was maintained for the first 6 days of doxorubicin exposure and then gradually decreased to 54% of the original level by day 9 (Figure 1d). The changes in nuclear morphology following exposure to 50 ng/ml doxorubicin were further assessed by


4767

Figure 2 Changes in the cellular morphologies of Huh-7 cells treated with 50 ng/ml doxorubicin. (a) Changes in the nuclear morphologies. Representative pictures of cells stained with Hoechst 33258 and visualized under a fluorescence microscope (magnification, 200). (b) Electron microscopic observation of Huh-7 cells treated with 50 ng/ml doxorubicin for the indicated time periods. Bar is 2.5 mm. Electron micrographs are the same magnification. White arrowheads, black arrowheads, and white arrows denote the electron-dense lysosomes, vacuoles and the convolution of nuclear membranes, respectively. A higher magnification of the nuclear membrane collapse at 6 day is shown on the left

Hoechst 33258 staining (Figure 2a). After 3 days of LD doxorubicin treatment, the nuclei became significantly larger and some cells contained several nuclei of unequal sizes. After 6 days, we observed an increased number of micronuclei, but did not observe the condensed or fragmented nuclei characteristic of apoptotic cells. These phenomena were further clearly demonstrated by more detailed electron microscopic observation (Figure 2b). After 6 days, we observed considerable increases in the numbers of vacuoles and electron-dense lysosomes, and we additionally observed the convolution or collapse of some nuclear membranes. After 9 days of LD doxorubicin treatment, most of the cells that had floated from the culture plate demonstrated necrosis-like characteristics such as diffuse chromatin, disruption of intact cellular boundaries, and losses of distinctive nuclear membrane structures. We did not detect the DNA fragmentation characteristic of apop-

Figure 3 Characterization of the time course of multinucleation and SA-b-gal expression in Huh-7 cells treated with 50 ng/ml doxorubicin. (a) Microscopic analysis of multinucleation and SAb-gal expression. Huh-7 cells were treated with 50 ng/ml doxorubicin for the indicated time points and SA-b-gal expression was assayed. Cells were counterstained with Giemsa solution and photographed at 400 magnification with phase contrast. b þ , b, SN, or MN denotes SA-b-gal positive, SA-b-gal negative, single nucleus, and multiple nuclei, respectively. Representative pictures are shown. (b) Measurement of the percentages of cells with SA-b-gal negativity and single nucleus (b-gal () and SN), cells with SA-b-gal positivity and multiple nuclei (b-gal ( þ ) and MN), cells with SA-b-gal positivity and single nucleus (b-gal and SN), and cells with SA-b-gal negativity and multiple nuclei (b-gal () and MN)

tosis, as examined by agarose gel electrophoresis and TUNEL assay (data not shown). Next, we investigated whether abnormal nuclear morphologies were observed before or after the increase in SA-b-gal-positive cells by SA-b-gal assay and subsequent nuclear staining with Giemsa solution. Both the percentages of multinucleated and those of SA-bgal-positive cells increased gradually following treatment with 50 ng/ml doxorubicin (Figure 3a and b). Over 60% of the cells demonstrated both multinucleation and SA-b-gal-positivity on day 6, suggesting that multinucleation and SLP are induced in most cells treated with 50 ng/ml doxorubicin. However, the increase of the SA-b-gal () and multinucleated cells peaked at day 3 and it was followed by the increase of SA-b-gal ( þ ) and multinucleated cells, suggesting that multinucleation might precede the acquisition of SLP in cells treated with 50 ng/ml doxorubicin. In addition, we further investigated whether multinucleation by LD doxorubicin was associated with abnormal mitosis. We examined the expression of atubulin (a major component of microtubules) by staining with anti-a-tubulin antibody, and observed Oncogene


4768

Figure 4 Abnormal spindle formation in cells treated with 50 ng/ ml doxorubicin. (a) Representative pictures of Huh-7 cells treated with or without 50 ng/ml doxorubicin for 3 days and then immunostained with anti-a-tubulin antibody and DAPI. Images from confocal laser scanning microscopy are shown (magnification, 1200). (b) The percentages of cells with abnormal spindles or cells with multiple nuclei following treatment with 50 ng/ml doxorubicin for the indicated time periods

nuclear morphologies by DAPI staining. While the percentage of multinucleated cells gradually increased following treatment with 50 ng/ml doxorubicin, abnormal formation of spindles (tripolar, quadripolar, or asymmetrical mircotubules) peaked at 3 days of doxorubicin treatment (Figure 4a and b). These results suggest that abnormal spindle formation may contribute to the observed LD doxorubicin-induced multinucleation. FACS analysis was then employed to monitor the changes of DNA contents in cells treated with LD doxorubicin (Figure 5). At day 1.5 of doxorubicin treatment, the S-phase-cell population was dramatically increased from 23.7 to 50.2%, and there was a concomitant decrease in the G0/G1-phase-cell population, from 52.5 to 25.3%. At day 3, about 37.7% of cells were detected at G2/M phase and the hyperploid cell population (>4N DNA cells) had increased by 21.4%. Thereafter, the hyperploid cell population increased further and the DNA contents of the doxorubicintreated cells became very heterogeneous, possibly due to the formation of multiple micronuclei in the absence of cytokinesis. Interestingly, there was only a slight increase (7.5%) in the sub-G1 population after 9 days of LD doxorubicin treatment, whereas trypan blue exclusion assay revealed that cellular viability had decreased to 54% of the original level (Figure 1d). Taken together, these results suggest that apoptosis might not be the major mode of the cell death in Huh-7 Oncogene

Figure 5 Changes in DNA contents following treatment with 50 ng/ml doxorubicin. Cells were fixed with ethanol and their DNA contents were measured by FACS analysis. Representative histograms of three independent experiments are shown. Percentages of G0/G1-, S-, G2/M-, and sub-G1-phase cells were calculated by deconvolution of the DNA content histograms. Values are presented as the mean and s.d. of three independent experiments

cells treated with LD doxorubicin. Instead, these cells likely died through mitotic catastrophe. Several mitosis-associated proteins are downregulated during LD doxorubicin-induced SLP and cell death through mitotic catastrophe There is a broad consensus that p53-deficient cells are able to aberrantly reenter the cell cycle and undergo unchecked DNA reduplication, leading to increased nuclear content and subsequent chromosomal instability (Lanni and Jacks, 1998). In this context, p21 appears to be a major target of p53 (Mantel et al., 1999). In Huh-7 cells harboring mutant p53 (Hsu et al., 1993), neither p53 nor p21 protein levels were altered following treatment with 50 ng/ml doxorubicin (Figure 6). This is contrary to previous reports that p53 and p21 acted as positive regulators of doxorubicin-mediated senescence (Chang et al., 1999b; Wang et al., 1999; Elmore et al., 2002). When we further analysed the protein levels of mitosis-associated proteins following treatment with 50 ng/ml doxorubicin, we observed downregulation of


4769

Figure 6 Changes in the expression of p53, p21, and some mitosiscontrolling proteins following treatment with 50 ng/ml doxorubicin. Huh-7 cells were treated with 50 ng/ml doxorubicin for the indicated time periods, and Western blotting was performed to detect expression changes of the listed proteins. To confirm that anti-p21 antibody used in this experiment was immunologically active, Huh-7 cells were treated with 5 ng/ml TGF-b1 for 9 days and Western blotting of p21 was performed

the mitosis initiator Cdc2, the centromere protein CENP-A (Kalitsis et al., 1998), spindle checkpoint control proteins Mad2 and BubR1 (Li and Benezra, 1996), and the DNA damage-induced checkpoint kinase Chk1 (Chan et al., 1999). In contrast, there were no changes in the protein levels of the DNA damageinduced checkpoint kinase Chk2 (Hirao et al., 2000) or the mitotic checkpoint protein Plk1 (Seong et al., 2002). Our results suggest that LD doxorubicin may induce depletion of some proteins controlling mitosis, thus contributing to abnormal nuclear division and subsequent cell death in Huh-7 cells. The SLP and mitotic catastrophe is observed in many human HCC cell lines following treatment with LD doxorubicin To analyse whether LD doxorubicin-induced SLP and cell death through mitotic catastrophe is restricted to a particular cell line, we examined the effect of LD doxorubicin on the human HCC cell lines, SNU-354, -398, -449, and -475. Although there were cell typespecific variations in the optimal concentrations of doxorubicin for induction of SLP and the incubation time to reach cell death (data not shown), all of the

Figure 7 Induction of SLP and mitotic catastrophe by doxorubicin in other human HCC cell lines. (a) SA-b-gal assay. After treatment of human HCC cells with or without doxorubicin at the indicated concentrations (SNU-354, 60 ng/ml for 8 days; SNU-398, 15 ng/ml for 10 days; SNU-449, 120 ng/ml for 8 days; SNU-475, 45 ng/ml for 8 days), SA-b-gal assays were performed and cells were counterstained with Giemsa solution. Representative pictures are shown (magnification, 200). (b) Quantitation of b-gal expression in HCC cells treated with LD doxorubicin. Data are presented as the mean and s.d. (bars) of four independent experiments based on three random fields with 100 cell counts per field. (c) Occurrence of multinucleation. After treatment of HCC cells with or without doxorubicin at the indicated experimental conditions (SNU-354, 60 ng/ml for 10 days; SNU-398, 15 ng/ml for 12 days; SNU-449, 120 ng/ml for 10 days; SNU-475, 45 ng/ml for 10 days), nuclei were stained with Hoechst 33258 and cells were observed under a fluorescence microscope

tested cell lines demonstrated similar induction of SLP and mitotic catastrophe in response to doxorubicin treatment (Figure 7). Differences in cellular morphologies and kinetics between LD doxorubicin-induced cell death by mitotic catastrophe and HD doxorubicin-induced apoptosis Although the biochemical characteristics of apoptotic cell death have been well defined, those of cell death through mitotic catastrophe are not fully understood, prompting us to compare the two cell death modes. After 24 h treatment with 10 mg/ml of doxorubicin, Huh7 cells showed a typical apoptosis characterized by reduction of cell volume, apoptotic blebbing, and an Oncogene


4770

ymethyl ester (calcein-AM) and ethidium homodimer-1 (Etd-1), which detect live and dead cells, respectively (Figure 8c). Apoptosis induced by 10 mg/ml doxorubicin was a rapid process (completed within 5 days), whereas cell death induced by 50 ng/ml doxorubicin progressed more slowly, beginning after 6 days and continuing until day 15. In terms of DNA contents, Huh-7 cells treated with 10 mg/ml doxorubicin demonstrated an obvious gradual increase in sub-G1 populations (Figure 8d), as compared with the hyperploid and very heterogeneous DNA contents observed in Huh-7 cells undergoing LD doxorubicin-induced cell death through mitotic catastrophe (Figure 5). These results suggest that HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death through mitotic catastrophe may differ both morphologically and biochemically. Plasma membrane integrity is lost during the initial phase of LD doxorubicin-induced cell death by mitotic catastrophe

Figure 8 Comparison of morphologies and viabilities between cells undergoing HD doxorubicin-induced apoptosis and LD doxorubicin-induced cell death through mitotic catastrophe. (a) Cellular morphologies. Cell death by mitotic catastrophe or apoptosis in Huh-7 cells was respectively induced by treatment with 50 ng/ml doxorubicin for 9 days or 10 mg/ml doxorubicin for 24 h. Representative phase contrast microscopic pictures of cells are shown (magnification, 200). (b) Nuclear and cytoskeletal morphologies of cells undergoing doxorubicin-induced cell death by mitotic catastrophe and apoptosis. Nuclei and actin were, respectively, stained with PI- and FITC-conjugated phalloidin in cells treated as described above. Representative fluorescence microscopic pictures are shown (magnification, 1200). (c) Cellular viability after exposure to doxorubicin. Huh-7 cells were plated in 24-well plates at 1 104 cells/well and exposed to 50 ng/ml or 10 mg/ml doxorubicin for different time periods. The viable cells were determined according to their ability to hydrolyse calcein-AM and exclude Etd-1, as described in Materials and methods. (d) Changes in DNA contents following treatment with 10 mg/ml doxorubicin. Cells were fixed with ethanol and their DNA contents were measured by FACS analysis. Representative histograms of three independent experiments are shown

increase in the nucleus-to-cytoplasm ratio (Figure 8a and b). In contrast, treatment with 50 ng/ml doxorubicin for 9 days induced significant increases in cell volume and formation of micronuclei with indistinct nuclear boundaries. Cellular viability in the two treatment groups was compared by staining with calcein-acetoxOncogene

To further understand the mode of cell death through mitotic catastrophe, cells treated with 50 ng/ml or 10 mg/ ml doxorubicin for different time periods were stained with FITC-conjugated Annexin-V (a Ca2 þ -dependent phospholipid-binding protein with high affinity for phosphatidyl serine) and propidium iodide (PI; a membrane-impermeable DNA stain). While most of the Annexin V ( þ ) cells were PI () at 24 h of 10 mg/ml doxorubicin treatment, the percentage of Annexin V ( þ ) and PI ( þ ) cells were increased by 20% at 48 h (Figure 9). Interestingly, almost all of Annexin V ( þ ) cells were PI ( þ ) following treatment with 50 ng/ml doxorubicin. These results suggest that phosphatidyl serine externalization occurred during both doxorubicin-induced apoptosis and cell death through mitotic catastrophe. However, the integrity of the plasma membrane was lost during the initial phase of LD doxorubicin-induced cell death through mitotic catastrophe, whereas this loss was detected during the later stages of HD doxorubicin-induced apoptosis.

Figure 9 Membrane alterations in HD doxorubicin-induced apoptosis and LD doxorubicin-induced cell death through mitotic catastrophe. After treatment of Huh-7 cells with 50 ng/ml or 10 mg/ ml doxorubicin for the indicated time periods, cells were incubated with FITC-Annexin V and PI as described in Materials and methods. Annexin V- and/or PI-positive cells were counted under a fluorescence microscope and expressed as a percentage of the total cell number


4771

Figure 10 Patterns of nuclear lamin B distribution in LD doxorubicin-induced cell death through mitotic catastrophe and HD doxorubicin-induced apoptosis. Huh-7 cells were treated with 50 ng/ml doxorubicin for 9 days or 10 mg/ml doxorubicin for 24 h. Cells were immunostained using antilamin B antibody (green). Nuclei were counterstained with PI (red). Cells were visualized by confocal laser scanning microscopy (magnification, 1200) and representative pictures are shown. Arrows indicate the frequent disruption of nuclear envelopes in cells undergoing mitotic catastrophe

Nuclear lamin B distribution differs in cells undergoing HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death via mitotic catastrophe Next, we investigated the structure of the nuclear envelope in cells undergoing LD doxorubicin-induced cell death through mitotic catastrophe. Lamin B is a major component of the nuclear lamina, a fibrous structure adjacent to the nucleoplasmic face of the nuclear membrane (Aebi et al., 1986). Nuclear envelopes were strongly labeled by lamin B antibody in control interphase cells (Figure 10). In contrast, the ring-like structure was frequently disrupted in cells treated with LD doxorubicin for 9 days, suggesting that the integrity of the nuclear envelope was lost. However, in apoptotic cells treated with 10 mg/ml doxorubicin for 24 h, the lamin B-positive peripheral ring shape was almost absent and the intranuclear staining of numerous spots was intense. These results demonstrate that lamin B distribution differed in cells undergoing doxorubicininduced apoptosis and cell death through mitotic catastrophe. Mitochondrial cytochrome c is released during both HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death through mitotic catastrophe Disruption of the mitochondrial membrane function and release of mitochondrial cytochrome c into the cytosol play a critical role in apoptosis induced by many chemotherapeutic agents (Decaudin et al., 1998). As expected, cells treated with 10 mg/ml doxorubicin

Figure 11 Release of mitochondrial cytochorme c in HD doxorubicin-induced apoptosis and LD doxorubicin-induced cell death through mitotic catastrophe. (a) Subcellular fractionation and Western blot analysis of the release of mitochondrial cytochrome c. Cytosolic and membrane fractions were separated from cells treated with 50 ng/ml or 10 mg/ml doxorubicin at the indicated time points as described in Materials and methods. In all, 10 mg of cytosolic and mitochondrial proteins per lane were resolved by 12% SDS–PAGE, and the contents of cytochrome c in the respective fractions were analysed by Western blotting using an anticytochrome c antibody. Successful fractionation of the cytosolic and mitochondrial fractions was confirmed by Western blotting of a-tubulin as a specific marker for cytosolic proteins and VDAC as a specific marker for mitochondrial proteins. (b) Analysis of the release of mitochondrial cytochrome c by immunocytochemistry. Huh-7 cells treated with or without 50 ng/ml doxorubicin for 9 days or 10 mg/ml doxorubicin for 24 h were fixed and subjected to immunocytochemical analysis of cytochrome c (green). Nuclei (red) were counterstained with PI and cells were observed under a confocal microscope (magnification, 1200)

showed a gradual increase in the cytochrome c content of cytosolic fractions (Figure 11a). Interestingly, cells treated with 50 ng/ml doxorubicin also showed progressive release of cytochrome c into the cytosol, beginning at day 3, and reaching significant levels by day 9. The Oncogene


4772

release of cytochrome c into the cytosol in cells treated with HD or LD doxorubicin was further confirmed by immunofluorescence experiments using confocal laser scanning microscopy (Figure 11b). While punctate staining patterns of cytochrome c were observed in untreated Huh-7 cells, both LD and HD doxorubicin treatments were associated with diffuse cytochrome c staining patterns after 9 days and 24 h, respectively. p38, c-Jun N-terminal kinase (JNK) mitogen-activated protein (MAP) kinases, and NF-kB are significantly activated during HD doxorubicin-induced apoptosis but not LD doxorubicin-induced cell death through mitotic catastrophe Both p38 and JNK have been implicated in apoptosis induced by various cytotoxic agents (Davis, 2000; Ono and Han, 2000). Therefore, we investigated whether these stress-activated MAP kinases are also involved in LD doxorubicin-induced SLP and cell death through mitotic catastrophe. The activities of p38 and JNK were analysed with antisera specific to the actively phosphorylated proteins (Figure 12a). As early as 1 h after exposure to 10 mg/ml doxorubicin, phosphorylation of both p38 and JNK was substantially enhanced; this activity was sustained up to 24 h. In contrast, treatment

with 50 ng/ml doxorubicin did not induce any significant activation of these proteins over 12 days. MKK3/6 (Raingeaud et al., 1996) and SEK1 (Yan et al., 1994), which are the upstream kinases of p38 and JNK, respectively, demonstrated similar activation profiles under LD and HD doxorubicin conditions. The total protein levels of these MAP kinases were not significantly altered in response to either 10 mg/ml or 50 ng/ml doxorubicin (data not shown). Activation of the NF-kB has been linked to apoptosis, with this transcription factor playing either an antiapoptotic or a proapoptotic role depending on the cell type (Baichwal and Baeuerle, 1997). We analysed changes in the expression levels of IkB-a, an inhibitor protein of NF-kB, during the progression of HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death through mitotic catastrophe (Figure 12b). When cells were treated with 10 mg/ml doxorubicin, IkB-a protein was markedly diminished at 1 h and completely disappeared thereafter. We then analysed whether IkB-a degradation in the apoptotic cells was directly associated with nuclear translocation of NF-kB. Interestingly, the appearance of NF-kB p65, p50, and cRel in the nuclear fractions was significantly increased 30 min after doxorubicin treatment, peaked at 0.5 h and then declined (Figure 12c). These results demonstrate that NF-kB is only transiently activated during the initial phase of doxorubicin-induced apoptosis. In contrast, IkB-a protein levels were not significantly decreased at any time point in cells treated with 50 ng/ml doxorubicin (Figure 12b). Furthermore, there was no detectable nuclear translocation of NF-kB in cells undergoing LD doxorubicin-induced cell death through mitotic catastrophe (data not shown). Taken together, these results demonstrate that the stress-activated signaling molecules are involved in HD doxorubicininduced apoptosis but not in doxorubicin-induced cell death through mitotic catastrophe. Caspases are significantly activated during HD doxorubicin-induced apoptosis but not LD doxorubicininduced cell death through mitotic catastrophe

Figure 12 Differential regulation of p38, JNK cascades, and NFkB in HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death via mitotic catastrophe. (a) Changes in p38 and JNK activity. Cells treated with 50 ng/ml or 10 mg/ml doxorubicin for the indicated time periods were examined for kinase phosphorylation by Western blotting. (b) Expression of IkB-a during the progression of HD doxorubicin-induced apoptosis and LD doxorubicin-induced cell death through mitotic catastrophe. (c) Transient activation of NF-kB via its translocation into the nucleus during HD doxorubicin-induced apoptosis. Huh-7 cells were treated with 10 mg/ml doxorubicin for the indicated time periods and total extracts and nuclear extracts were Western blotted for expression of p65, c-rel, and p50 Oncogene

Apoptosis triggered by various cytotoxic agents depends on activation of caspases, which play pivotal roles in the proteolysis of specific targets (Gupta, 2001). To examine whether caspases are involved in doxorubicin-induced cell death through mitotic catastrophe, we used Western blotting to compare the proteolytic processing patterns of caspases in cells undergoing LD doxorubicin-induced cell death with those of cells undergoing HD doxorubicin-induced apoptosis (Figure 13a). First, we examined cells treated with 10 mg/ml doxorubicin. In these cells, processing of caspase-3, -8, and -9 was first detected 4, 8, and 12 h post-treatment, respectively. By 48 h, caspase processing was extensive; we noted substantial losses of the proforms (32 kDa for procaspase-3, 55 kDa for procasape-8, and 46 kDa for procaspase-9) and the concomitant appearances of their processed subunits (20 and 17 kDa for caspase-3, 20 kDa for caspase-8, and 37 kDa for caspase-9).


4773

Discussion

Figure 13 Differential regulation of caspases in HD doxorubicininduced apoptosis and LD doxorubicin-induced cell death through mitotic catastrophe. (a) Changes in the proteolytic processing of caspases in doxorubicin-induced apoptosis and cell death through mitotic catastrophe. Huh-7 cells were treated with 50 ng/ml or 10 mg/ml doxorubicin for the indicated time periods and cell lysates were prepared for Western blotting analysis of caspases. (b) Changes in the expressions of caspase substrate proteins during the progression of the two modes of cell death

Furthermore, procaspase-6 and -7 were proteolytically processed, as shown by analysis of substrate cleavage (Figure 13b). We observed processing of FAK and PARP (substrates of caspase-3) into their respective discrete degradation products 8 h after treatment with 10 mg/ml doxorubicin. Bid (a substrate of caspase-8) was proteolytically degraded into its 15 kDa product beginning at 4 h, and degradation of Lamin B (a substrate protein of caspase-6) was detected at 24 h. Next, we analysed the proteolytic processing of caspases during LD doxorubicin-induced cell death through mitotic catastrophe (Figure 13a). Interestingly, while viable cell numbers had decreased by 54% after 9 days of treatment with 50 ng/ml doxorubicin (Figure 8c), small amounts of procaspase-3 and -9 were degraded into their active forms at this time. Procaspase-8 was not processed into its active form during cell death through mitotic catastrophe, whereas the protein levels of procaspase-6 and -7 were slightly decreased at 12 days after treatment with 50 ng/ml doxorubicin. To examine this further, we monitored the levels of several possible caspase substrates at a variety of time points (Figure 13b). While the levels of intact proteins were significantly decreased at 12 days, the discrete degradation products of PARP, FAK, Bid, and lamin B were not observed in cells undergoing death through mitotic catastrophe. These results seem to indicate that caspases may not play a critical role in LD doxorubicin-induced cell death through mitotic catastrophe. Taken together, our results suggest that different doses of doxorubicin induce different modes of cell death with distinctive morphological characteristics and signaling pathways.

Although there are no currently established chemotherapeutic regimens for the treatment of unresectable HCC (Patt et al., 1988), doxorubicin is one of the most active anti-HCC agents, having single-agent response rates of 0–15% in advanced HCC (Sciarrino et al., 1985). In contrast to the well-studied HD doxorubicin-induced apoptosis seen in various cancer cells, the antiproliferative effects induced by LD doxorubicin are not well understood. In this study, we demonstrate that LD doxorubicin induces a SLP (evidenced by increased SAb-gal activity) and cell death through mitotic catastrophe in Huh-7 hepatoma cells. Treatment of many other HCC cell lines with LD doxorubicin (15–120 ng/ ml) induced similar responses, suggesting that induction of SLP and cell death through mitotic catastrophe is a general response to LD doxorubicin in HCC cells. Similar to our results, SLP plus mitotic catastrophe have been observed in several cell lines treated with LD anticancer drugs (Chang et al., 1999a, b; Ruth and Roninson, 2000). The term ‘mitotic catastrophe’ has been widely used to describe a type of cell death that occurs during mitosis, but there is still no broadly accepted definition of this term. The term ‘mitotic catastrophe’ is sometimes restrictively used for abnormal mitosis that leads to cell death (which can occur through necrosis or apoptosis) rather than to the cell death itself. Nitta et al. (2004) proposed that ‘mitotic catastrophe’ is a mechanism for the induction of cell death in cancer cells by antineoplastic agents that damage DNA. Moreover, Chu et al. (2004) suggested that mitotic catastrophe is an event that results from a cell either failing to complete division or fusing after division. Thus, these groups have argued that mitotic catastrophe is not a mode of cell death such as apoptosis or necrosis. Here, we examined this controversy by comparing Huh-7 cells undergoing LD doxorubicin-induced cell death and HD doxorubicininduced apoptosis. In our study, abnormal nuclear morphologies (likely arising through abnormal mitosis) were observed slightly earlier than the principal increase in SA-b-galpositive cells following LD doxorubicin treatment. Cells with multiple nuclei increased gradually beginning on day 2 (Figure 4b) and the numbers of micronuclei within the cells were further increased thereafter (Figure 2). In contrast, SA-b-gal activity was significantly increased beginning on day 5 (Figure 3b). The extensively micronucleated cells remain viable for some time, but eventually die through loss of integrity of both the cytoplasmic and nuclear membranes. These results suggest that LD doxorubicin-treated cells may first undergo mitotic catastrophe and then enter a temporary senescence-like arrest before eventually dying. Inhibition of mitotic checkpoint genes has been described as a characteristic of senescence in several studies, including a report on colon carcinoma cell senescence induced by LD doxorubicin (Chang et al., 2002). In cells capable of developing a reversibly senescent phenotype through regulated expression of p21, inhibition of mitotic genes Oncogene


4774

by p21 was shown to produce mitotic catastrophe once the cells reentered the cell cycle after release from the enforced p21 induction (Chang et al., 2000). In our experimental conditions, downregulation of Cdc2, CENP-A. Mad2, BubR1, and Chk1 (but not Chk2 and Plk1) was noted following treatment with LD doxorubicin (Figure 6). Moreover, the changes in DNA ploidy observed in this study (Figure 5) are similar to those in doxorubicin-treated fibrosarcoma cells, and are associated in part with senescence following abnormal mitosis (Chang et al., 1999b). Several recent studies have shown that drug-induced senescence is dependent on p53. MCF-7 breast cancer cells with functional p53 enter senescence following doxorubicin treatment, whereas those lacking functional p53 undergo apoptosis (Elmore et al., 2002). Similarly, glioblastoma cells with intact p53 show SLP in response to inhibition of topoisomerase I, whereas cells with disrupted p53 undergo apoptosis (Wang et al., 2004). In contrast, p53 inhibition in HT1080 fibrosarcoma cells was associated with decreased drug-induced SLP and increased cell death through mitotic catastrophe (Chang et al., 1999b). In our study, SLP was effectively induced by LD doxorubicin in human HCC cell lines harboring mutant p53, including Huh-7 (Hsu et al., 1993), SNU-354, -398, -449, and -475 (Kang et al., 1996). However, these SLPs were only transiently maintained, and were followed by cell death through mitotic catastrophe in most cases. Therefore, it is interesting to speculate whether reintroduction of functional wild-type p53 into these cells could attenuate LD doxorubicin-induced cell death through mitotic catastrophe, leading to a prolonged senescence-like arrest. In an attempt to better understand the biochemical differences between cell death through mitotic catastrophe and apoptosis, we compared various aspects of the two cell death modes induced by different doses of doxorubicin. Our molecular analysis showed that LD doxorubicin-induced cell death through mitotic catastrophe shares some common characteristics with HD doxorubicin-induced apoptotic cell death, such as the release of mitochondrial cytochrome c into the cytosol (Figure 11) and externalization of phosphatidyl serine in the plasma membrane (Figure 9). However, our novel results suggest that LD doxorubicin-induced cell death through mitotic catastrophe differs from HD doxorubicin-induced apoptosis as follows: (a) cell death by mitotic catastrophe is induced after 6 days, while apoptosis can occur within 12 h (Figure 8c); (b) cells undergoing death by mitotic catastrophe become multinucleated and enlarged, whereas apoptotic cells shrink (Figure 8b); (c) plasma membrane integrity is lost during the initial phase of cell death through mitotic catastrophe, but at a later phase of apoptosis (Figure 9); (d) the DNA contents in cells undergoing cell death through mitotic catastrophe are heterogeneously hyperploid, while cells undergoing apoptosis are mostly subdiploid (Figures 5 and 8d); (e) expression of lamin B at the nuclear periphery is often disrupted during cell death by mitotic catastrophe, whereas lamin B is relocalized from Oncogene

the nuclear periphery into intranuclear structures during apoptosis (Figure 10); and (f) caspases and stressactivated signaling pathways such as JNK, p38, and NF-kB are significantly activated during apoptosis but not during cell death through mitotic catastrophe (Figures 12 and 13). Although anticancer drugs have traditionally been valued based on their ability to cause apoptosis of cancer cells, recent studies have shown that apoptosis may not be the primary death mode in solid tumors. Both senescence and cell death through mitotic catastrophe were efficiently induced by subapoptotic concentrations of anticancer agents in vitro and in vivo (te Poele et al., 2002). Induction of senescence or cell death through mitotic catastrophe by LD anticancer drugs may have several advantages for patients. As high plasma concentrations of doxorubicin tend to cause severe toxicity to normal tissues, including cardiotoxicity (Speth et al., 1987), administration of LDs of doxorubicin for extensive periods of time rather than HDs for short periods of time may benefit patient survival. Furthermore, since cell death through mitotic catastrophe is similar to necrosis in relation to the early loss of plasma membrane integrity, it is possible that induction of cell death though mitotic catastrophe may lead to inflammation and mobilization of the immune system against the tumor. In contrast, apoptosis induced by HDs of anticancer drugs is noninflammatory and does not recruit the resources of the immune system. Therefore, cell death through mitotic catastrophe may create a more profound, longer-lasting effect in vivo. Supporting these ideas, the cellular immune function of HCC patients was significantly impaired by conventional doses of anticancer drugs through transcatheter arterial chemoembolization, while it remains possible that LD of the drugs might enhance the cellular immune functions (Lu et al., 2002). In conclusion, our study demonstrates that HD doxorubicin-induced apoptosis and LD doxorubicininduced cell death through mitotic catastrophe are largely independent, and that the activation of distinct signaling pathways may be responsible for the induction of these morphologically distinct cell death modes. Elucidation of the factors regulating the various aspects of treatment-induced apoptosis and cell death through mitotic catastrophe should greatly benefit the development of new strategies for improving the efficacy of cancer therapy.

Materials and methods Cells and culture conditions To explore the mode of cell death after induction of SLP by doxorubicin, we first determined the optimal conditions to induce SLP. Of the tested experimental protocols (including different doses of doxorubicin, pulse treatment with doxorubicin for different periods, intermittent or chronic treatment, etc.), we obtained the best results with chronic exposure of HCC cells to an LD of doxorubicin (50 ng/ml for Huh-7 cells). Briefly, Huh-7 cells were first plated in 10 cm dishes with 6 ml


4775 of DMEM containing 10% fetal bovine serum (FBS). After overnight culture, this medium was replaced with 6 ml of fresh DMEM containing 10% FBS and 50 ng/ml doxorubicin. After 3 days, 3 ml of fresh DMEM containing 10% FBS and 50 ng/ml doxorubicin was added to the pre-existing medium (to avoid possible nutritional depletion in long-term culture and to maintain the long-term concentration of doxorubicin at 50 ng/ ml). Thereafter, 3 ml of doxorubicin-containing fresh medium was added every 3 days in the same manner, until day 12. Welldefined human HCC cells (SNU-354, -398, -449, and -475) were obtained from the Korean Cell Line Bank (Seoul, Korea) (Park et al., 1995) and grown in RPMI 1640 (Gibco-BRL, Grand Island, NY, USA) supplemented with 10% FBS. In the respective HCC cell lines, the optimal concentrations of doxorubicin for induction of SLP were determined to be 60 ng/ml for SNU-354, 15 ng/ml for SNU-398, 120 ng/ml for SNU-449, and 45 ng/ml for SNU-475. Doxorubicin-induced apoptosis in Huh-7 cells was produced by treatment with 10 mg/ml doxorubicin.

Annexin V/PI staining Huh-7 cells grown on coverslips were treated with 50 ng/ml or 10 mg/ml doxorubicin at different time periods. Adherent cells on coverslips were washed with PBS, and externalized phosphatidyl serine and DNA were dually labeled for 15 min with Annexin-V-fluorescein and PI (1 mg/ml) in HEPES buffer (10 mmol/l HEPES/NaOH, pH 7.4, 140 mmol/l NaCl, 5 mmol/l CaCl2). The stained cells were washed, the coverslips were mounted on glass slides, and samples were observed under a fluorescence microscope. Floated dead cells were separately collected by centrifugation, washed with PBS, and then incubated for 15 min in HEPES buffer containing Annexin V-FITC and PI. The cells were then centrifuged, resuspended in HEPES buffer, dropped onto a slide, covered with a coverslip, and observed under a fluorescence microscope. Total cell numbers were measured after counting of the numbers of adherent cells per coverslip and the numbers of floated dead cells. Percentages of cells stained with Annexin V-FITC and/or PI relative to total cell numbers were assessed.

SA-b-gal activity Cells were stained for b-gal activity as described by Dimri et al. (1995). Briefly, 1 104 cells were seeded in 24-well plate. After appropriate exposure, the cells were washed twice with phosphate-buffered saline (PBS), fixed with 2% formaldehyde and 0.2% glutaraldehyde in PBS, and washed twice in PBS. Cells were stained for 12 h in X-gal staining solution (1 mg/ml X-gal, 40 mmol/l citric acid/sodium phosphate (pH 6.0), 5 mmol/l potassium ferricyanide, 5 mmol/l potassium ferrocyanide, 150 mmol/l NaCl, 2 mmol/l MgCl2). Cells were then counterstained with 0.1% hematoxylin solution or Giemsa stain solution. Measurement of cellular viability Huh-7 cells were treated with doxorubicin at the indicated concentrations for the fixed time points. Cellular viability was assessed by double labeling of cells with 2 mmol/l calcein-AM and 4 mmol/l Etd-1. Calcein-positive live cells were counted under a fluorescence microscope (Nikon Diaphot 300, Japan), since Etd-1-positive dead cells were floated from the culture plate following treatment of doxorubicin. Alternatively, cellular viability was assessed by trypan blue exclusion assay. Transmission electron microscopic examination Huh-7 cells were treated with 50 ng/ml doxorubicin for the indicated time periods. The cells were prefixed in Karnovsky’s solution (1% paraformaldehyde, 2% glutaraldehyde, 2 mmol/l calcium chloride, 0.1 mol/l cacodylate buffer, pH 7.4) for 2 h and washed with cacodylate buffer. Postfixing was carried out in 1% osmium tetroxide and 1.5% potassium ferrocyanide for 1 h. After dehydration with 50–100% alcohol, the cells were embedded in Poly/Bed 812 resin (Pelco, Redding, CA, USA) and polymerized, and observed under electron microscope (EM 902A, Zeiss, Oberkochen, Germany). Cell cycle analysis Trypsinized and floating cells were pooled, washed with PBSEDTA, and fixed in 70% (v/v) ethanol. For assessment of DNA contents, cells were stained with PI and monitored by FACScan. Cell cycle distribution was determined with the ModFit LT program (Verity Software House Inc.).

Immunoblotting Immunoblotting was performed using standard procedures. The membrane was incubated with antibodies against p53, p21 (each diluted 1:500, purchased from Calbiochem, San Diego, CA, USA), a-tubulin, VDAC, Bid, phospho-specific-p38, -MKK3/6, -SEK1, -JNK, -p38, -IkB-a, (each diluted 1 : 1000, purchased from Cell Signaling Technology, Beverly, MA, USA), caspase-9, FAK, Lamin B, Cdc2, BubR1, Chk1, Chk2, (each diluted 1:500, purchased from Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), caspase-3, -6, -7, -8 (1 : 500, Stressgen Biotechnologies Co., Victoria, BC, Canada), CENPA, PARP (1 : 500, Upstate Biotechnology, Lake Placid, NY, USA), actin, Plk1, and cytochrome c (1 : 500, BD Transduction Lab., San Diego, CA, USA) in blocking buffer. Bound primary antibodies were detected with HRP-conjugated secondary antibodies and enhanced chemiluminescence (Amersham, Arlington heights, IL, USA). Immunocytochemistry Cells were washed twice with PBS and fixed in 4% formaldehyde for 10 min at room temperature, and then washed three times with PBS. To examine the expression of atubulin or cytochrome c, fixed cells were permeabilized in 0.1% Triton X-100/2% BSA, and stained with mouse anti-atubulin antibody (1 : 200, Calbiochem, San Diego, CA, USA) or mouse anticytochrome c antibody (1 : 50, BD Transduction Lab., San Diego, CA, USA). Cells were further incubated with FITC-conjugated anti-mouse antibody (1 : 50, Molecular Probes Inc., Eugene, OR, USA). To examine the expression of lamin B, cells were stained with goat antilamin B antibody (1 : 100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) and further incubated with FITC-conjugated anti-goat antibody (1 : 50, Sigma, St Louis, MO, USA). Nuclei were further stained with Hoechst 33258, DAPI, or propium iodide (1 mg/ ml, Sigma, St Louis, MO, USA). To examine the expression of actin, cells were incubated with FITC-conjugated phalloidin (1 : 50, Sigma, St Louis, MO, USA). Stained cells were examined by confocal microscopy (Olympus, Shinjuku-ku, Tokyo, Japan). RT–PCR analysis Total RNA was isolated from cells treated with 50 ng/ml doxorubicin for the indicated time points using RNAZolB Oncogene


4776 Table 1 Gene Osteonectin Forward Reverse SM22 Forward Reverse TGase II Forward Reverse PAI-1 Forward Reverse GAPDH Forward Reverse

Primer used for the analysis of the gene expression by RT–CR Primer sequence

Product (bp)

50 -CTGTGGGAGCTAATCCTG-30 50 -GGGTGCTGGTCCAGCTGG-30

602

50 -TGGCGTGATTCTGAGCAA-30 50 -CTGCCAAGCTGCCCAAGG-30

534

50 -CTCGTGGAGCCAGTTATCAACAGCTAC-30 50 -TCTCGAAGTTCACCACCAGCTTGTG-30

310

50 -GTGTTTCAGCAGGTGGCGC-30 50 -CCGGAACAGCCTGAAGAAGTG-30

300

50 -CGTCTTCACCATGGAGA-30 50 -CGGCCATCACGCCCACAGTTT-30

310

(Tel-Test, Friendswood, TX, USA). Total RNA (2 mg) from each cell culture was reverse transcribed using oligo-dT primers and AMV reverse transcriptase (TaKaRa, Otsu, Shiga, Japan). The cDNAs were amplified by PCR (941C for 30 s, 601C for 30 s and 721C for 1 min) with Taq DNA polymerase. To ensure exponential amplification, four aliquots were removed from each PCR assay at cycles 20, 25, 30, or 35 cycles (which were determined in preliminary experiments to produce the weakest detectable PCR product for each gene) and every two to four cycles thereafter. Amplified products were analysed by agarose gel electrophoresis at cycles within the linear range of mRNA amplification. The sequences of oligonucleotide primers used for RT–PCR and the expected transcript sizes are listed in Table 1. Assessment of nuclear translocation of NF-kB in response to doxorubicin After treatments, nuclear extracts of Huh-7 cells were prepared using an established protocol as described previously (Liu et al., 1998). The expression of the translocated NF-kB was assessed by Western blotting using ani-p65, c-rel, and p50 antibody (1 : 500, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA). Subcellular fractionation for analysis of the release of mitochondrial cytochrome c After treatments, Huh-7 cells were collected and washed twice in ice-cold PBS, resuspended in S-100 buffer (20 mM HEPES, pH 7.5, 10 mM KCl, 1.9 mM MgCl2, 1 mM EGTA, 1 mM

EDTA, mixture of protease inhibitors), and incubated on ice for 20 min. After 20 min incubation on ice, the cells were homogenized with a Dounce glass homogenizer and a loose pestle (Wheaton, Millville, NJ, USA) for 70 strokes. Cell homogenates were spun at 1000 g for 10 min to remove unbroken cells, nuclei, and heavy membranes. The supernatant was spun again at 14 000 g for 30 min to collect the mitochondria-rich (pellet) and cytosolic (supernatant) fractions. The mitochondria-rich fraction was washed once with extraction buffer, followed by a final resuspension in lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EGTA) containing protease inhibitors for Western blot analysis.

Abbreviations PBS, phosphate-buffered saline; TUNEL, terminal deoxynucelotidyl transferase-mediated deoxyuridine triphosphate nickend labeling; SDS–PAGE, sodium dodecyl sulfate/polyacrylamide gel electorphoresis; MAP, mitogen-activated protein; JNK, c-Jun N-terminal kinase; calcein-AM, calcein-acetoxymethyl ester; Etd-1, ethidium homodimer-1. Acknowledgements This study was supported by a grant of the Korea Health 21 R&D Project, Ministry of Health and Welfare, Republic of Korea (HMP-02-PJ1-PG3-20708-0003) and a grant from the KOSEF (R05-2004-000-10740-0) (to KSC).

References Acunas B and Rozanes I. (1999). Eur. J. Radiol., 32, 86–89. Aebi U, Cohn J, Buhle L and Gerace L. (1986). Nature, 323, 560–564. Baichwal VR and Baeuerle PA. (1997). Curr. Biol., 7, R94–R96. Castedo M, Perfettini JL, Roumier T, Andreau K, Medema R and Kroemer G. (2004). Oncogene, 23, 2825–2837. Chan GK, Jablonski SA, Sudakin V, Hittle JC and Yen TJ. (1999). J. Cell. Biol., 146, 941–954. Chang BD, Broude EV, Dokmanovic M, Zhu H, Ruth A, Xuan Y, Kandel ES, Lausch E, Christov K and Roninson IB. (1999a). Cancer Res., 59, 3761–3767. Oncogene

Chang BD, Broude EV, Fang J, Kalinichenko TV, Abdryashitov R, Poole JC and Roninson IB. (2000). Oncogene, 19, 2165–2170. Chang BD, Swift ME, Shen M, Fang J, Broude EV and Roninson IB. (2002). Proc. Natl. Acad. Sci. USA, 99, 389–394. Chang BD, Xuan Y, Broude EV, Zhu H, Schott B, Fang J and Roninson IB. (1999b). Oncogene, 18, 4808–4818. Chu K, Teele N, Dewey MW, Albright N and Dewey WC. (2004). Radiat. Res., 162, 270–286. Davis RJ. (2000). Cell, 103, 239–252. Decaudin D, Marzo I, Brenner C and Kroemer G. (1998). Int. J. Oncol., 12, 141–152.


4777 Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, Medrano EE, Linskens M, Rubelj I, Pereira-Smith O, Peacocke M and Campisi J. (1995). Proc. Natl. Acad. Sci. USA, 92, 9363–9367. Dumont P, Burton M, Chen QM, Gonos ES, Frippiat C, Mazarati JB, Eliaers F, Remacle J and Toussaint O. (2000). Free Radic. Biol. Med., 28, 361–373. Elmore LW, Rehder CW, Di X, McChesney PA, JacksonCook CK, Gewirtz DA and Holt SE. (2002). J. Biol. Chem., 277, 35509–35515. Gamen S, Anel A, Perez-Galan P, Lasierra P, Johnson D, Pineiro A and Naval J. (2000). Exp. Cell Res., 258, 223–235. Gewirtz DA. (1999). Biochem. Pharmacol., 57, 727–741. Gupta S. (2001). Life Sci., 69, 2957–2964. He QY, Liang YY, Wang DS and Li DD. (2002). Int. J. Oncol., 20, 261–266. Hendry JH and West CM. (1997). Int. J. Radiat. Biol., 71, 709–719. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, Yoshida H, Liu D, Elledge SJ and Mak TW. (2000). Science, 287, 1824–1827. Hortobagyi GN. (1997). Drugs, 54, 1–7. Hsu IC, Tokiwa T, Bennett W, Metcalf RA, Welsh JA, Sun T and Harris CC. (1993). Carcinogenesis, 14, 987–992. Kalitsis P, MacDonald AC, Newson AJ, Hudson DF and Choo KH. (1998). Genomics, 47, 108–114. Kang MS, Lee HJ, Lee JH, Ku JL, Lee KP, Kelley MJ, Won YJ, Kim ST and Park JG. (1996). Int. J. Cancer, 67, 898–902. Lanni JS and Jacks T. (1998). Mol. Cell. Biol., 18, 1055–1064. Li Y and Benezra R. (1996). Science, 274, 246–248. Liu L, Kwak YT, Bex F, Garcia-Martinez LF, Li XH, Meek K, Lane WS and Gaynor RB. (1998). Mol. Cell. Biol., 18, 4221–4234. Lock RB and Stribinskiene L. (1996). Cancer Res., 56, 4006– 4012. Lu W, Li YH, He XF, Chen Y, Zeng QL and Qiu YR. (2002). Di. Yi. Jun. Yi. Da. Xue. Xue. Bao., 22, 524–526.

Mantel C, Braun SE, Reid S, Henegariu O, Liu L, Hangoc G and Broxmeyer HE. (1999). Blood, 93, 1390–1398. Miranda EI, Santana C, Rojas E, Hernandez S, OstroskyWegman P and Garcia-Carranca A. (1996). Mutat. Res., 349, 173–182. Nitta M, Kobayashi O, Honda S, Hirota T, Kuninaka S, Marumoto T, Ushio Y and Saya H. (2004). Oncogene, 23, 6548–6558. Ono K and Han J. (2000). Cell Signal., 12, 1–13. Park JG, Lee JH, Kang MS, Park KJ, Jeon YM, Lee HJ, Kwon HS, Park HS, Yeo KS, Lee KU, Kim ST, Chung JK, Hwang YJ, Lee HS, Kim CY, Lee YI, Chen TR, Hay RJ, Song SS, Kim WH, Kim CW and Kim YI. (1995). Int. J. Cancer, 62, 276–282. Patt YZ, Claghorn L, Charnsangavej C, Soski M, Cleary K and Mavligit GM. (1988). Cancer, 61, 1884–1888. Raingeaud J, Whitmarsh AJ, Barrett T, Derijard B and Davis RJ. (1996). Mol. Cell. Biol., 16, 1247–1255. Roninson IB, Broude EV and Chang BD. (2001). Drug Resist. Updat., 4, 303–313. Ruth AC and Roninson IB. (2000). Cancer Res., 60, 2576– 2578. Sciarrino E, Simonetti RG, Le Moli S and Pagliaro L. (1985). Cancer, 56, 2751–2755. Seong YS, Kamijo K, Lee JS, Fernandez E, Kuriyama R, Miki T and Lee KS. (2002). J. Biol. Chem., 277, 32282–32293. Serrano M, Lin AW, McCurrach ME, Beach D and Lowe SW. (1997). Cell, 88, 593–602. Speth PA, Linssen PC, Boezeman JB, Wessels HM and Haanen C. (1987). Cancer Chemother. Pharmacol., 20, 305–310. te Poele RH, Okorokov AL, Jardine L, Cummings J and Joel SP. (2002). Cancer Res., 62, 1876–1883. Wang Y, Blandino G and Givol D. (1999). Oncogene, 18, 2643–2649. Wang Y, Zhu S, Cloughesy TF, Liau LM and Mischel PS. (2004). Oncogene, 23, 1283–1290. Yan M, Dai T, Deak JC, Kyriakis JM, Zon LI, Woodgett JR and Templeton DJ. (1994). Nature, 372, 798–800.

Oncogene