Prostaglandin D(2) and J(2) induce apoptosis in human leukemia cells via activation of the caspase 3 cascade and production of reactive oxygen species.

(1)

Prostaglandin D

2

and J

2

induce apoptosis in human leukemia

cells via activation of the caspase 3 cascade and

production of reactive oxygen species

Yen-Chou Chen

a,

*, Shing-Chuan Shen

b,c

, Shu-Huei Tsai

d

a

Graduate Institute of Pharmacognosy, Taipei Medical University, 250 Wu-Hsing Street, Taipei 110, Taiwan

b

Department of Dermatology, School of Medicine, Taipei Medical University, 250 Wu-Hsing Street, Taipei 110, Taiwan

c_{Department of Dermatology, Taipei Medical University-Affiliated Taipei Municipal Wan-Fang Hospital, 250 Wu-Hsing Street, Taipei 110, Taiwan} d_{Department of Orthopedics and Traumatology, School of Medicine, Taipei Medical University, 250 Wu-Hsing Street, Taipei 110, Taiwan}

Received 7 July 2004; received in revised form 5 October 2004; accepted 6 October 2004 Available online 11 November 2004

Abstract

The presence of prostaglandins (PGs) has been demonstrated in the processes of carcinogenesis and inflammation. In the present study, we found that 12-o-tetradecanoylphorbol 13-acetate (TPA) induced cyclooxygenase 2 (COX-2), but not COX-1, protein expression in HL-60 cells, and the addition of arachidonic acid (AA) in the presence or absence of TPA significantly reduced the viability of HL-60 cells, an effect that was blocked by adding the COX inhibitors, NS398 and aspirin. The AA metabolites, PGD2and PGJ2, but not PGE2or PGF2a, reduced

the viability of the human HL60 and Jurkat leukemia cells according to the MTT assay and LDH release assay. Apoptotic characteristics including DNA fragmentation, apoptotic bodies, and hypodiploid cells were observed in PGD2- and PGJ2-treated leukemia cells. A dose- and

time-dependent induction of caspase 3 protein procession, and PARP and D4-GDI protein cleavage with activation of caspase 3, but not caspase 1, enzyme activity was detected in HL-60 cells treated with PGD2or PGJ2. Additionally, DNA ladders induced by PGD2and PGJ2

were significantly inhibited by the caspase 3 peptidyl inhibitor, Ac-DEVD-FMK, but not by the caspase 1 peptidyl inhibitor, Ac-YVAD-FMK, in accordance with the blocking of caspase 3, PARP, and D4-GDI protein procession. An increase in intracellular peroxide levels by PGD2and PGJ2was identified by the DCHF-DA assay, and anti-oxidant N-acetyl cysteine (NAC), mannitol (MAN), and tiron significantly

inhibited cell death induced by PGD2and PGJ2by reducing reactive oxygen species (ROS) production. The PGJ2metabolites,

15-deoxy-D12,14-PGJ2 and D12-PGJ2, exhibited effective apoptosis-inducing activity in HL-60 cells through ROS production via activation of the

caspase 3 cascade. The proliferator-activated receptor-g (PPAR-g) agonists, rosiglitazone (RO), troglitazone (TR), and ciglitazone (CI), induced apoptosis in cells which was blocked by the addition of the PPAR-g antagonists, GW9662 and BADGE, via blocking of caspase 3 and PARP cleavage. However, neither GW9662 nor BADGE showed any protective effect on PGD2- and PGJ2-induced apoptosis. A

Keywords: Apoptosis; Caspase 3; ROS; Prostaglandin; Cyclooxygenase; PPAR-g

1. Introduction

Prostaglandins (PGs) are a family of oxygenated metabolites of arachidonic acid (AA), and have a diverse range of actions depending on the PG type and cell target. PGs are divided into two groups, conventional PGs such as PGE2, PGF2a, and PGD2 and cyclopentenone PGs such as PGJ2, PGA1, and PGA2 [1,2]. AA is the precursor of PGs, and is primarily converted to PGH2 by

Abbreviations: MTT, (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxyMe-thoxy-phenyl)-2-(4-sulfophenyl)-2H-tetrazolium; TPA, 12-o-tetradecanoyl-phorbol 13-acetate; Ac-DEVD-FMK, acetyl-Asp-Glu-Val-Asp-fluoromethylketone; Ac-YVAD-FMK, acetyl-Tyr-Val-Ala-Asp-fluorome-thylketone; DCHF-DA, dichlorodihydrofluorescein diacetate; PARP, poly (ADP-ribose) polymerase; NAC, N-acetyl-cysteine; ALL, allopurinol; DPI, diphenylene iodonium; ROS, reactive oxygen species; PGs, prostaglandins; BADGE, biphenol A diglycidyl ether; GW9662, 2-chloro-5-nitrobenzanilide * Corresponding author. Tel.: +886 2 27361661x3224; fax: +886 2 23787139.

(2)

cyclooxygenases followed by conversion of PGH2 to several related PGs including PGD2, PGJ2, PGF2a, and PGE2 by tissue-specific isomerase. Several physiological effects of PGs have been identified. PGE2 production is increased in colon, gastric, and lung carcinomas with an increase in COX-2 protein levels [3,4]. Our previous data demonstrated that PGE2 is involved in 12-o-tetradeca-noylphorbol 13-acetate (TPA)- and epidermoid growth factor (EGF)-induced proliferation [5,6]. PGD2 is a major product in a variety of tissues or cells, and has significant effects including platelet aggregation and vasorelaxation

[7]. In vivo and in vitro studies have shown that PGD2 readily undergoes dehydration to yield active PGs of the J2series including PGJ2, D12–14 PGJ2, and 15-deoxy-D12– 14

PGJ2 [8,9]. Members of PGJ2 contain a reactive a,h-unsaturated ketone in the cyclopentenone ring that is important for their biological activities including anti-tumor, anti-inflammation, and antiviral replication effects

[10–12].

Many types of mammalian cells undergo apoptosis during normal development or in response to a variety of stimuli, including DNA damage, oxidative stress, growth factor deprivation, and chemical treatment. Apoptosis induced by these agents appears to be regulated by a set of downstream genes such as p53, p21, caspases, and Bcl-2 family genes

[13,14]. Human caspase-1 to -10 have been described, and previous studies demonstrated that activation of the caspase cascade is involved in chemical- and agent-induced apopto-sis[15,16]. Caspase 3 exists as an inactive pro-caspase 3 in the cytoplasm and is proteolytically converted to active caspase 3 by a single cleavage event in cells undergoing apoptosis. After caspase 3 activation, some specific sub-strates for caspase 3 such as PARP and D4-GDI proteins are cleaved, and these are important for the occurrence of apoptosis[17,18].

Several previous studies suggested that PGs might reduce cell viability via apoptosis induction, but their mechanisms of action are complex and not well defined. Our previous study demonstrated that apoptosis induced by chemicals was mediated by activation of the caspase 3 cascade through a distinct ROS-dependent or -independent pathway[19,20]. In the present study, we obtained evidence that PGD2and PGJ2, but not PGE2or PGF2a, exhibited effective apoptosis-inducing activities through ROS production and caspase 3 activation in human leukemia cells. ROS-dependent caspase 3 activation was identified in PGD2- and PGJ2-induced apoptosis.

2. Materials and methods 2.1. Cell culture

HL-60 and Jurkat human promyeloleukemic cells were obtained from ATCC (American Type Culture Collection; Rockville, MD). HL-60 and Jurkat cells were grown in

RPMI1640 containing 10% heat-inactivated fetal bovine serum (FBS) and maintained at 37 8C in a humidified incubator containing 5% CO2. Exponentially growing cells were exposed to drugs for the indicated time periods. All culture reagents were purchased from Invitrogen (Carlsbad, CA, USA).

2.2. Chemicals

The colorigenic synthetic peptide substrates, Ac-DEVD-pNA, Ac-DEVD-pNA, Ac-DEVD-FMK, and Ac-YVAD-FMK, were purchased from Calbiochem. Propidium, iodide, PGs, TPA, and AA were obtained from Sigma Chemical (St. Louis, MO). Rosiglitazone (RO), troglita-zone (TR), ciglitatroglita-zone (CI), GW9662, and BADGE were obtained from Cayman Chemical. Antibodies for PARP, caspase 3, and D4-GDI detection in Western blotting were obtained from IMGENEX. Antibodies for detecting Bcl-2 family proteins and a-tubulin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Dichlorodi-hydrofluorescein diacetate (DCHF-DA) was obtained from Molecular Probes.

2.3. Cell viability

Cell viability was assessed by MTT staining as described previously[21]. Briefly, cells were plated at a density of 105 cells/well in 24-well plates. After overnight growth, cells were treated under various conditions for 12 h. At the end of treatment, 30 Al of the tetrazolium compound, MTT (3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide), and 270 Al of fresh RPMI medium were added. The supernatant was removed, and formazan crystals were dissolved in DMSO. After incubation for 4 h at 37 8C, 200 Al of 0.1 N HCl in 2-propanol was placed in each well to dissolve the tetrazolium crystals. At the end, the absorbance at a wavelength of 600 nm was recorded using an ELISA plate reader.

2.4. Determination of ROS production

ROS production was monitored by flow cytometry using DCFH-DA[19]. This dye is a stable compound that readily diffuses into cells and is hydrolyzed by intracellular esterase to yield DCFH, which is trapped within cells. Hydrogen peroxide or low-molecular-weight peroxides produced by cells oxidize DCFH to highly fluorescent compound, 2V,7V-dichlorofluorescein (DCF). Thus, the fluorescence intensity is proportional to the amount of peroxide produced by the cells. In the present study, HL-60 cells were treated with each of the indicated compounds for 2 h and washed twice with PBS to remove the extracellular compounds. DCHF-DA (100 AM) was added for an additional hour. Green fluorescence was excited using an argon laser and was detected using a 525-nm band-pass filter by flow cytometric analysis.

(3)

2.5. Western blots

Total cellular extracts (30 Ag) were prepared and separated on 8% SDS-polyacrylamide mini gels for PARP detection and 12% SDS-polyacrylamide minigels for caspase 3, cleaved D4-GDI, Bcl-2 family, and a-tubulin detection, and then transferred to Immobilon polyvinylidene difluoride membranes (Millipore). The membrane was incubated at 4 8C with 1% bovine serum albumin at room temperature for 1 h and then incubated with the indicated antibodies for a further 3 h at room temperature followed by incubation with alkaline phosphatase-conjugated anti-mouse IgG antibody for 1 h. Protein was visualized by incubating with the colorimetric substrates, nitro blue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl-phosphate (BCIP), as described in our previous paper[20,21].

2.6. DNA gel electrophoresis

Cells (106/ml) under different treatments were collected, washed with PBS twice, and then lysed in 100 ml of lysis buffer [50 mM Tris, pH 8.0; 10 mM ethylenediaminetetra-acetic acid (EDTA); 0.5% sodium sarkosinate, and 1 mg/ml proteinase K] for 3 h at 56 8C and treated with 0.5 mg/ml RNase A for another hour at 56 8C. DNA was extracted with phenol/chloroform/isoamyl alcohol (25:24:1) before load-ing. Samples were mixed with loading buffer [50 mM Tris, 10 mM EDTA, 1% (w/w) low-melting-point agarose, and 0.025% (w/w) bromophenol blue] and loaded onto a pre-solidified 2% agarose gel containing 0.1 Ag/ml ethidium bromide. The agarose gels were run at 50 V for 90 min in TBE buffer, then observed and photographed under UV light[22].

Fig. 1. TPA and arachidonic acid (AA) induce apoptosis with increasing COX-2 protein expression in HL-60 cells. (A) Cells were treated with TPA (50, 100, and 200 ng/ml) for 12 h, and expression of COX-1 and COX-2 protein was detected by Western blotting using specific antibodies. The ratio COX-2/COX-1 was measured by densitometric analysis, and described as folds of control. (B) Effects of TPA (100 ng/ml) and AA (AA50, 50 AM; A100, 100 AM) on DNA fragmentation in HL-60 cells. Cells were treated with different compounds for 12 h, and the integrity of DNA in cells was analyzed by electrophoresis in 1.8% agarose gels. (C) Cells were treated as described in (B), and the cytotoxicity under different treatments was detected by LDH release. (D) As described in (C), HL-60 cells were treated with TPA (100 ng/ml) in the presence or absence of different doses of AA (25, 50, and 100 AM); the viability of cells was examined by the MTT assay. (E) The COX inhibitors, NS398 (NS10, 10 AM; NS20, 20 AM) and aspirin (ASP100, 100 AM; ASP200, 200 AM), inhibited TPA (100 ng/ ml) plus AA (50 AM)-induced cytotoxicity in HL-60 cells. HL-60 cells were treated with indicated doses of NS398 or aspirin for 1 h followed by TPA (100 ng/ ml) plus AA (50 AM) treatment. The cytotoxic effect of TPA plus AA in the presence or absence of NS398 or aspirin was evaluated by LDH release. Data are expressed as the meanFS.E. of three independent experiments. *pb0.05, **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.##

pb0.01, significantly differs between indicated groups, as analyzed by Student’s t-test. Western blotting and DNA ladder analyses were performed at least three times, and the results shown are representative of all of the data.

(4)

2.7. Analysis of respective caspase activities

Ac-DEVD-pNA for caspase 3 and Ac-YVAD-pNA for caspase 1 were used as colorimetric protease substrates. After different treatments, cells were collected and washed three times with PBS and resuspended in 50 mM Tris–HCl (pH 7.4), 1 mM EDTA, and 10 mM ethyleneglycoltetra-acetic acid (EGTA). Cell lysates were clarified by centrifu-gation at 15,000 rpm for 3 min, and clear lysates containing 50 Ag of protein were incubated with 100 AM of the indicated specific colorimetric substrates at 37 8C for 1 h. Activity of caspase 1 and 3 enzymes was described as the cleavage of colorimetric substrate by measuring the absorbance at 405 nm.

2.8. Flow cytometric analysis

Trypsinized cells were washed with ice-cold PBS and fixed in 70% ethanol at 20 8C for at least 1 h. After fixation, cells were washed twice, incubated in 0.5 ml of 0.5 ml of 0.5% Triton X-100/PBS at 37 8C for 30 min with 1 mg/ml of RNase A, and stained with 0.5 ml of 50 mg/ml propidium iodide for 10 min. Fluorescence emitted by the propidium-DNA complex was quantitated after excitation of

the fluorescent dye by FACScan flow cytometry (Becton Dickenson, San Jose, CA).

2.9. LDH release assay

Cells were treated under different conditions, and medium was collected for the LDH release assay. The amount of LDH in the medium was measured by the protocol suggested by the manufacturer (Roche Applied Science). The addition of 2% Triton X-100 to the cells was used as a positive control of the total amount of LDH in cells. The resulting cytotoxicity values were calculated by the following equation:

Cytotoxicity %ð Þ

¼ EXP: value CON=Triton X 100 value CONð Þ 100%:

2.10. Statistics

Values are expressed as the meanFS.E. The significance of the difference from the respective controls for each experimental test condition was assayed using Student’s

t-Fig. 2. Cytotoxic effect of PGD2and PGJ2, but not of PGE2or PGF2a, in human HL-60 and Jurkat leukemia cells. (A, B) Cells were treated with different doses

(2, 4, and 8 Ag/ml) of PGs for 12 h, and the viability of cells was detected by the MTT assay. (C) HL-60 cells were treated with different doses (4 and 8 Ag/ml) of PGD2and PGJ2for 12 h, and cytotoxicity was evaluated by LDH release. (D) HL-60 cells were treated with different doses (2, 4, and 8 Ag/ml) of PGD2or

PGJ2for 12 h (left panel), or PGD2 or PGJ2 (8 Ag/ml; right panel) for different time points (4, 8, and 12 h). The integrity of DNA was analyzed by

electrophoresis in 1.8% agarose gels. (E) Jurkat cells were treated with different doses (4 and 8 Ag/ml) of PGD2or PGJ2for 12 h, and the integrity of DNA was

(5)

test for each paired experiment. A P value of b0.01 or b0.05 was regarded as indicating a significant difference.

3. Results

3.1. TPA induction of COX-2 protein expression stimulates the occurrence of apoptosis induced in human HL-60 leukemia cells

Conversion of arachidonic acid (AA) to PGH2 is catalyzed by cyclooxygenases, and PGH2is isomerized into two groups of PGs including PGE2/PGF2aand PGD2/PGJ2. In the present study, we found that the addition of TPA at the doses of 100 and 200 ng/ml significantly induced COX-2, but not COX-1, protein expression in HL-60 cells by Western blot analysis (Fig. 1A). Addition of AA did not change the COX-2 protein expression induced by TPA (Data not shown). Results of the DNA fragmentation assay showed that TPA (100 ng/ml) exhibited a slight DNA

laddering effect, and the addition of AA at the doses of 50 or 100 AM significantly enhanced the occurrence of the intensity of DNA fragmentation in cells (Fig. 1B). We further analyzed the viability of cells under different treatments by MTT and LDH release assays. As elucidated inFig. 1C and D, cotreatment of HL-60 cells with AA/TPA showed a more significant decrease in the viability of cells, compared with the AA- or TPA-treated groups. Further-more, the COX enzyme inhibitors, NS398 and aspirin, significantly attenuated the cytotoxic effect of AA/TPA in cells (Fig. 1D). These data suggest that activation of COX-2 enzyme activity might be involved in cell death induced by AA and TPA.

3.2. Differential cytotoxic effect of PGs in the human HL-60 and Jurkat leukemia cells

Previous data indicated that AA potentiated cell death in the presence of TPA in HL-60 cells. Therefore, we inves-tigated the cytotoxic effects of four major AA metabolites

Fig. 3. Increased hypodiploid cells and apoptotic bodies in PGD2- or PGJ2-treated 60 cells by flow cytometric analysis and microscopic observations.

HL-60 cells were treated with PGD2or PGJ2(8 Ag/ml) for 12 h. (A) Apoptotic bodies in PGD2- and PGJ2-treated cells were detected by microscopic observations.

(B) The sub-G1 peak (hypodiploid cells) was examined by flow cytometric analysis, and the percentage of hypodiploid cells (the ratio of the sub-G1 peak) was quantitated from three independent experiments. **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.

(6)

PGE2, F2a, D2, and J2on the human leukemia cells, HL60 and Jurkat. As shown inFig. 2, PGD2and PGJ2, but not PGE2or PGF2a, dose-dependently reduced the viability of HL-60 and Jurkat cells by the MTT assay (Fig. 2A and B). Similarly, the LDH release assay showed that PGD2and PGJ2increased the release of LDH into the medium in HL-60 and Jurkat cells (Fig. 2C and data not shown). Furthermore, several apoptotic characteristics were investigated in the present study. Both PGD2and PGJ2induced DNA fragmentation in human HL-60 leukemia cells in a time- and dose-dependent manner (Fig. 2D). Induction of DNA fragmentation was also observed in Jurkat cells (Fig. 2E). Results of microscopic observations showed that apoptotic bodies were detected in HL-60 cells under PGD2 and PGJ2 treatment (Fig. 3A). Under flow cytometric analysis, we found an increase in hypodiploid

cells (sub-G1 peak) in PGD2- and PGJ2-treated HL-60 cells (Fig. 3B). These data indicate that reduction of cellular viability by PGD2and PGJ2in human leukemia cells occurs via apoptosis.

3.3. Induction of caspase 3 protein procession and PARP and D4-GDI protein cleavage by PGD2 and PGJ2

We investigated if activation of the caspase 3 cascade is involved in apoptosis induced by PGD2and PGJ2. Activation of caspase 3 (CPP32/Yama) causes it to recognize the sequence Asp-Glu-Val-Asp (DEVD) and to cleave a number of proteins, such as PARP and D4-GDI, another hallmark of apoptosis. With exposure to PGD2and PGJ2, degradation of 116-kDa PARP into 85-kDa fragments and production of

Fig. 4. Induction of the caspase 3 cascade was examined in PGD2- or PGJ2-induced apoptosis. (A) HL-60 cells were treated with different doses (2, 4, and 8 Ag/

ml) of PGD2or PGJ2for 12 h (left panel), and the expressions of caspase 3, PARP, and D4-GDI protein were detected by Western blotting using specific

antibodies. (B) Cells were treated with PGD2or PGJ2(8 Ag/ml; right panel) for different time points (4, 8, and 12 h), and the expressions of caspase 3, PARP,

and D4-GDI protein were detected by Western blotting using specific antibodies. (C) Jurkat cells were treated with PGD2or PGJ2(4 and 8 Ag/ml) for 12 h, and

expressions of caspase 3 and D4-GDI protein were analyzed. (D) HL-60 and Jurkat cells were treated with PGD2and PGJ2(8 Ag/ml) for 12 h, and the

expression of Bcl-2 family proteins was detected by Western blotting using specific antibodies. The intensity of indicated Bcl-2 family proteins was detected by densitometric analysis, and expressed as folds of control. Western blotting was performed at least three times, and the results shown are representative of all of the data.

(7)

cleaved (23-kDa) D4-GDI proteins were found to be dose-and time-dependent in HL60 cells, dose-and these were associated with activation of caspase 3 brought about by its cleavage, represented here as the production of cleaved fragments (Fig. 4A and B). Induction of caspase 3 protein procession, as well as PARP and D4-GDI protein cleavage, was also identified in Jurkat cells (Fig. 4C). Bcl-2 family proteins are important apoptotic regulators and are located upstream of caspase activation. In results of Western blotting, decreases in the antiapoptotic 2 protein level in HL-60 cells and the Bcl-XL protein level in Jurkat cells were detected with PGD2and PGJ2treatment. A decrease in Bad protein expression was found in HL-60, but not in Jurkat, cells; however, neither PGD2nor PGJ2induced the phosphorylation of Bad protein in both cells by Western blotting using a specific anti-body for phosphorylated Bad protein (Fig. 4D and data not shown). No

significant change in Bax or Mcl-1 protein was observed. Furthermore, two colorimetric substrates, Ac-DEVD-pNA and Ac-YVAD-pNA, were used to detect caspase 1 and caspase 3 enzyme activities in HL-60 cells under PGD2and PGJ2treatment, respectively. As illustrated inFig. 5A and B, PGD2and PGJ2dose-dependently induced DEVD-specific, but not YVAD-specific, caspase activity in HL-60 cells. These data suggest that activation of caspase 3, but not of caspase 1, was exhibited in PGD2- and PGJ2-induced apoptosis in human leukemia cells.

3.4. The caspase 3 peptidyl inhibitor, Ac-DEVE-FMK, inhibits apoptosis induced by PGD2 and PGJ2

In order to confirm if activation of caspase 3 is essential for the apoptosis induced by PGD2 and PGJ2, peptidyl

Fig. 5. Activation of caspase 3 enzyme plays an important role in PGD2- and PGJ2-induced apoptosis in HL-60 cells. (A, B) Activation of caspase 3, but not of

caspase 1, enzyme was detected in PGD2- and PGJ2-treated HL-60 cells. Ac-DEVD-pNA for caspase 3 and Ac-YVAD-pNA for caspase 1 were used to detect

respective enzyme activity. HL-60 cells were treated with different doses (2, 4, and 8 Ag/ml) of PGD2or PGJ2for 12 h, and the enzyme activity was measured

as described in Materials and methods. DEVD, addition of Ac-DEVD-FMK in the reaction; YVAD, addition of Ac-YVAD-FMK in the reaction. (C) The caspase 3 peptidyl inhibitor, Ac-DEVD-FMK, but not the caspase 1 peptidyl inhibitor, Ac-YVAD-FMK, inhibited PGD2- or PGJ2-induced apoptosis in HL-60

cells. HL-60 cells were treated with Ac-DEVD-FMK (3; 20 AM) or Ac-YVAD-FMK (1; 20 AM) for 1 h followed by PGD2or PGJ2(8 Ag/ml) treatment. The

integrity of DNA was analyzed by electrophoresis in 1.8% agarose gels. (D) Cells were treated as described in (C), and the viability of cells was detected by the MTT assay. (E) Cells were treated as described in (C), and expression of indicated proteins was evaluated by Western blotting using specific antibodies. **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.##

pb0.01, significantly differs between indicated groups, as analyzed by Student’s t-test. Western blotting and DNA fragmentation assay were performed at least three times, and the results shown are representative of all of the data.

(8)

inhibitors including DEVD-FMK for caspase 3 and Ac-YVAD-FMK for caspase 1 were used. InFig. 5C, PGD2and PGJ2 induced DNA fragmentation in HL-60 cells, which was blocked by adding DEVD-FMK, but not Ac-YVAD-FMK, to HL-60 cells. Results of the MTT assay showed that the addition of DEVD-FMK, but not of Ac-YVAD-FMK, protected HL-60 cells from PGD2- and PGJ2 -induced cell death (Fig. 5D). Furthermore, results of Western blotting showed that caspase 3 protein procession, as well as PARP and D4-GDI protein cleavage induced by PGD2and PGJ2, was significantly reduced by the addition

of Ac-DEVD-FMK, but not Ac-YVAD-FMK (Fig. 5E).

These data demonstrate that activation of caspase 3 is an essential event in apoptosis induced by PGD2and PGJ2. 3.5. Production of reactive oxygen species (ROS) is involved in PGD2- and PGJ2-induced apoptosis and is located upstream of caspase 3 activation

ROS are important apoptosis mediators, therefore we investigated if ROS production involves PGD2- and

PGJ2-induced apoptosis. The intracellular peroxide level in HL-60 cells under PGD2 and PGJ2 treatment was detected using DCHF-DA as a fluorescent ROS indicator. In results of flow cytometry analysis, increases in intracellular peroxide levels were found in PGD2 -and PGJ2-treated HL-60 cells. The addition of N-acetyl cysteine (NAC), mannitol (MAN), and tiron significantly reduced the intracellular peroxide level induced by PGD2 and PGJ2 (Fig. 6A and B). The ratio of hypodiploid

cells induced by PGD2 and PGJ2 was reduced by

the addition of NAC, MAN, and tiron to HL-60 cells (Fig. 6C). In addition, NAC addition significantly inhibited PGD2- and PGJ2-induced apoptosis by DNA

fragmentation and MTT assays (Fig. 7A and B).

Activation of caspase 3 enzyme activity and protein procession and PARP and D4-GDI protein cleavage induced by PGD2 and PGJ2 were significantly blocked by the addition of NAC (Fig. 7C and D). These data indicated that ROS production is involved in PGD2- and PGJ2-induced apoptosis and is located upstream of caspase 3 activation.

Fig. 6. Increased intracellular peroxide levels by PGD2and PGJ2in HL-60 cells. (A) HL-60 cells were pretreated with NAC (10 mM), mannitol (MAN; 20

AM), or tiron (TIR; 20 AM), followed by the addition of PGD2or PGJ2(8 Ag/ml) for a further hour. DCHF-DA (100 AM) was added at the end of reaction for

an additional hour. The level of intracellular peroxide was detected by flow cytometric analysis as described in Materials and methods. (B) Quantification of the fluorescent intensity from three independent experiments as described in (A) was performed, and results are expressed as the meanFS.E. (C) NAC, MAN, and TIR attenuated PGD2- and PGJ2-induced hypodiploid cells by flow cytometric analysis. Cells were treated with NAC (10 mM), NAM (20 AM), or TIR (20 AM)

for 1 h followed by incubation with PGD2or PGJ2(8 Ag/ml) for a further 12 h. The percentage of hypodiploid cells was quantitated as described inFig. 4.

**pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.##

pb0.01, significantly differs between indicated groups, as analyzed by Student’s t-test.

(9)

3.6. Involvement of ROS production and caspase 3 protein procession in 15-deoxy-D1214 PGJ2- and D12PGJ2-induced apoptosis

D12PGJ2is a decomposition product of PGD2 and may be converted to 15-deoxy-D12–14 PGJ2 in an acidic condition. 15-Deoxy-D12–14 PGJ2 is a metabolite of PGJ2 and can be formed from PGD2 by the elimination of two molecules of H2O. Therefore, we investigated the apop-totic mechanism of 15-deoxy-D12–14 PGJ2and D12PGJ2in human HL-60 leukemia cells. As illustrated inFig. 8, both 15-deoxy-D12–14 PGJ2 and D12PGJ2 showed effective cytotoxicity in HL-60 cells by the MTT assay in accordance with the induction of DNA fragmentation in cells (Fig. 8A and B). The IC50 values of 15-deoxy-D12–14 PGJ2 and D12PGJ2 were 2.8F0.2 and 2.3F0.4 Ag/ml, much lower than those of PGD2(4.5F0.3 Ag/ml) and PGJ2 (4.1F0.5 Ag/ml) in HL-60 cells. An increase in intra-cellular peroxide levels was detected in 15-deoxy-D12–14 PGJ2- and D12PGJ2-treated HL-60 cells. Furthermore, NAC significantly inhibited 15-deoxy-D12–14 PGJ2- and D12PGJ2-induced apoptosis with a reduction in intracellular peroxide levels (Fig. 8D and E). An increase in

hypo-diploid cells by 15-deoxy-D12–14 PGJ2 and D12PGJ2 was identified by flow cytometric analysis, which was blocked by the addition of NAC (Fig. 8C). Furthermore, induction of caspase 3 and PARP protein procession was detected in 15-deoxy-D12–14 PGJ2- and D12PGJ2-treated HL-60 cells, and those events were blocked by the addition of NAC (Fig. 8F).

3.7. The peroxisome proliferator-activated receptor-c (PPAR-c) agonists, RO, TR, and CI, induce apoptosis in human HL-60 leukemia cells

J2series cyPGs have been identified as ligands of PPAR-g, and cyPGs form from PGD2via dehydration. Therefore, we investigated if apoptosis induced by PGD2and PGJ2in human HL-60 leukemia cells occurs through activation of PPAR-g. Three well-known PPAR-g agonists, RO, TR, and CI, and two PPAR-g antagonists, GW9662 and BADGE, were used in the study. Results in Fig. 9A show that RO, TR, and CI at doses of 10 and 20 AM induced DNA laddering in human HL-60 leukemia cells. DNA ladders induced by RO, TR, and CI were significantly reduced by the addition of the PPAR-g antagonists, GW9662 (G) and

Fig. 7. NAC protects HL-60 cells from PGD2- or PGJ2-induced apoptosis via blocking caspase 3 activation. (A) NAC inhibits PGD2- and PGJ2-induced DNA

fragmentation in HL-60 cells. Cells were treated with N-acetyl cysteine (NAC; 2.5, 5, and 10 mM) for 1 h followed by PGD2or PGJ2(8 Ag/ml) treatment. The

integrity of DNA was analyzed by electrophoresis. (B) Cells were treated with PGD2or PGJ2in the presence or absence of NAC pretreatment for 12 h, and the

caspase 3 enzyme activity was measured as described inFig. 6. (C) Cells were treated with PGD2or PGJ2in the presence or absence of NAC (5 and 10 mM)

pretreatment for 12 h, and the viability of cells was detected by the MTT assay. (D) Cells were treated with PGD2or PGJ2in the presence or absence of NAC (5

and 10 mM) pretreatment for 12 h, and the expressions of caspase 3, PARP, and D4-GDI protein were detected. **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.##

pb0.01, significantly differs from the PGD2- or PGJ2-treated group, as analyzed by Student’s t-test. Western

(10)

BADGE (B). However, neither GW9662 nor BADGE inhibited the occurrence of DNA ladders induced by PGD2 and PGJ2 (Fig. 9A, right panel). Results of the MTT assay indicated that GW9662 addition at doses of 10 and 20 AM protected HL-60 cells from RO-, TR-, or CI- but not PGD2- or PGJ2-induced cell death (Fig. 9B). Results of Western blotting showed that RO, TR, and CI induced caspase 3 and PARP protein procession, described herein as an increase in cleaved fragments (cleaved) and a decrease in intact protein (caspase 3 and PARP) (Fig. 9C). The addition of GW9662 or BADGE significantly reduced caspase 3 and PARP protein procession induced by RO, TR, and CI.

4. Discussion

TPA induction of COX-2 protein expression with reduction in the viability with or without AA was observed in human leukemia cells, and the cytotoxic effect induced by TPA and AA was attenuated by the COX inhibitors, NS398 and aspirin. PGD2 and PGJ2, but not PGE2 or PGF2a, exhibited apoptosis-inducing activity through elevation of the intracellular ROS level, an event upstream of caspase 3 activation. Metabolites of PGD2 and PGJ2, including 15-deoxy-D12–14 PGJ2 and D12PGJ2, showed more effective apoptotic activity than PGJ2 and PGJ2, and induction of

Fig. 8. PGJ2metabolites, 15-deoxy-D12–14PGJ2and D12PGJ2, induced apoptosis in HL-60 cells by ROS production and activation of the caspase 3 cascade.

(A) HL-60 cells were treated with different doses of 15-deoxy-D12–14_PGJ

2(2, 4, and 8 Ag/ml) or D12PGJ2(4 Ag/ml) for 12 h, and the integrity of DNA was

analyzed by DNA electrophoresis. In the presence of NAC pretreatment for 1 h, cells were treated with 15-deoxy-D12–14_PGJ

2or D12PGJ2(4 Ag/ml) for 12 h,

and the integrity of DNA was analyzed. (B) As described in A, the viability of cells under different treatments was measured by the MTT assay. (C) Similarly, hypodiploid cells induced by 15-deoxy-D12–14_PGJ

2and D12PGJ2were detected by flow cytometric analysis, and that induction was prevented by the addition

of NAC. (D) Increases in intracellular peroxide production were examined in 15-deoxy-D12–14_PGJ

2- and D12PGJ2-treated HL-60 cells by the DCHF-DA assay.

Addition of NAC (N10; 10 mM) significantly attenuated the peroxide production induced by 15-deoxy-D12–14PGJ2and D 12

PGJ2(4 Ag/ml). (E) Intracellular

peroxide levels under different treatments were quantified, and results are presented as the meanFS.E. from three independent experiments. (F) Activation of the caspase 3 cascade was involved in 15-deoxy-D12–14PGJ2- and D

12

PGJ2-induced apoptosis, which was blocked by the addition of NAC. Cells were treated

with NAC (N5, 5 mM; N10, 10 mM) for 1 h followed by the addition of 15-deoxy-D12–14PGJ2and D 12

PGJ2(4 Ag/ml) for 12 h. Expression of PARP and

caspase 3 protein was examined by Western blotting as described previously. **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.##pb0.01, significantly differs from indicated group, as analyzed by Student’s t-test. Western blotting and DNA fragmentation assay were performed at least three times, and the results shown are representative of all of the data.

(11)

ROS production and activation of the caspase 3 cascade were involved. PPAR-g antagonists showed no protective effect on PGD2- and PGJ2-induced apoptosis. These data suggest the possibility that COX-2 up-regulation during inflammation is involved in the apoptotic process by enhancing the production of PGs such as PGD2, PGJ2, 15-deoxy-D12–14PGJ2, and D12PGJ2, and that ROS production, upstream of the caspase 3 cascade, is involved in the apoptotic mechanisms.

Apoptosis induced by PGs, particularly J-series PGs, has been documented, and it has been shown that activation of PPAR-g by J2 series cyPGs such as 15d-PGJ2 exhibits antiproliferative, anti-apoptotic, and anti-inflammatory properties in several types of cancer cells. Kondo et al.

[10] demonstrated that 15d-PGJ2 induces apoptosis via accumulation and phosphorylation of p53 and results in activation of the caspase cascade. Ward et al.[11]suggested that PGD2 and 15d-PGJ2 selectively induce apoptosis in eosinophils via inhibiting I kappa B degradation in a PPAR-g-independent manner. In neuroblastoma cells, 15d-PGJ2 -induced apoptosis occurs through activation of the ERK pathway in a PPAR-g-dependent manner[23]. Okano et al.

[24] indicated that 15d-PGJ2 induces apoptosis with

suppression of NF-kappa B activation in a independent manner in SH-Hep1 cells, but in a PPAR-g-dependent manner in Hep G2 cells, through a caspase-3-independent pathway. Additionally, Castrillo et al. [25]

provided evidence suggesting that 15d-PGJ2 promotes apoptosis in activated macrophages via sustained activation of PKC zeta and JNK and inhibition of NF-kappa B activity. Nakamura et al. [26] indicated that inhibitors of the AA cascade modulate TPA-induced oxidative stress in mouse skin. These data indicate that apoptosis induced by PGs, especially 15d-PGJ2, is very complicated, and dependent on the types of cells tested. However, the apoptotic mechanism induced by PGD2and PGJ2in HL-60 leukemia cells is still unclear. Constitutive expression of the PPAR-g receptor has been identified in human HL-60 leukemia cells[27]. In the present study, we identified that PGD2- and PGJ2-induced apoptosis occurs through a caspase 3-dependent pathway in human leukemia cells. The PPAR-g agonists, RO, TR, and CI, induced apoptosis in cells in accordance with caspase 3 and PARP, and this was blocked by the PPAR-g antagonists, GW9662 and BADGE. However, neither GW9662 nor BADGE showed any preventive effect against PGD2- or PGJ2-induced apoptosis. This suggests that apoptosis

Fig. 9. The PPAR-g agonists, RO, TR, and CI, induced DNA laddering with activation of caspase 3 and PARP protein procession. (A) Left panel: HL-60 cells were treated with different doses (10 or 20 AM) of RO, TR, or CI for 12 h, and the integrity of DNA was analyzed by agarose electrophoresis. Right panel: HL-60 cells were treated with GW9662 (G; 20 AM) or BADGE (B; 20 AM) for 1 h followed by incubation with RO (20 AM), PGD2(8 Ag/ml), or PGJ2(8 Ag/ml)

for a further 12 h. The integrity of DNA was analyzed by agarose electrophoresis. (B) The viability of cells under different treatments was examined by the MTT assay. HL-60 cells were treated with the PPAR-g, antagonist GW9662 (10 or 20 AM), for 1 h followed by the addition of PGD2(8 Ag/ml), PGJ2(8 Ag/ml),

or RO (20 AM) for a further 12 h, and the viability of cells was detected by the MTT assay. (C) PARP and caspase 3 protein procession induced by the PPAR-g agonists, RO, CI, and TR, was attenuated by the addition of GW9662 and BADGE (20 AM). Cells were treated with GW9662 or BADGE (20 AM) for 1 h followed by incubation with RO, TR, or CI (20 AM) for a further 12 h. The occurrence of caspase 3 and PARP protein procession were examined by Western blotting using specific antibodies as described previously. **pb0.01, significantly differs from the control (CON) group, as analyzed by Student’s t-test.

##

pb0.01, significantly differs from the indicated group, as analyzed by Student’s t-test. Western blotting was performed at least three times, and the results shown are representative of all of the data.

(12)

induced by PGD2 and PGJ2 might occur in a PPAR-g-independent manner.

Oxidative stress is seen as an upstream event in the signaling cascade in many cellular functions such as proliferation, inflammatory responses, and apoptosis, and ROS produced during oxidative stress may play critical roles in these effects by damaging cellular components such as membrane lipids [28–30]. However, both ROS-dependent and -inROS-dependent apoptosis has been found in a variety of chemical treatments [18,28]. Kondo et al. [31]

reported that cyclopentenone PGs including PGA2, PGJ2, D12-PGJ2, and 15-deoxy-D12–14-PGJ2showed a potent oxidant effect in human neuroblastoma cells. A pro-oxidative characteristic of cyclopentenone PGs is that they contain a,h-unsaturated ketones for nucleophilic addition reactions with thiols. Although PGs have been reported to cause oxidative stress in neuroblastoma cells, the molecular mechanism underlying this has not been delineated. Based on the observations that (a) PGD2and PGJ2(but not PGE2 or PGF2a) treatment resulted in apoptosis induction in human leukemia cells, (b) increases in intracellular per-oxide levels in PGD2- and PGJ2-treated cells, and (c) NAC reduction of ROS production and apoptosis by PGD2 or PGJ2, these data suggest that PGD2and PGJ2possess the ability to induce apoptosis in human leukemia cells via their pro-oxidant activity. DPI and ALL are inhibitors of

NADPH oxidase and xanthine oxidase, respectively. Our previous study demonstrated that DPI protected NIH3T3 cells from arsenic-induced apoptosis by decreasing ROS production [30]. In the present study, neither diphenyl iodine (DPI) nor allpurinol (ALL) showed an inhibitory effect on PGD2- or PGJ2-induced ROS production and apoptosis (data not shown). This suggests that increasing intracellular ROS levels by PGD2 and PGJ2did not occur through activation of NADPH oxidase and xanthine oxidase in cells.

Activation of the caspase cascades has been shown in the process of apoptosis, and caspase 3 is an executioner caspase, and is extensively activated by proteolytical cleavage of its 32-kDa precursor to generate 17- or 15-kDa active forms in the apoptotic process. Cleavage of PARP (116 kDa) into 85- and 31-kDa fragments, and of D4-GDI (28 kDa) into 23- and 5-kDa fragments by activated caspase 3, has been identified in apoptosis in response to several stimuli such as chemicals, growth factor deprivation, and UV irradiation [32]. Its relation with the effect of PGs on apoptosis induction is still controversial. Ishaque et al. [33] reported that PGE2 and PGI2 protected renal glomerular mesangial cells against TNFa-mediated apoptosis, and down-regulation of PGs by the COX-2 inhibitors, SC-236 and NS398, may induce the occurrence of apoptosis [34]. However, Ragolia et al.[35]

Fig. 10. A tentative model for PGs, proliferation, and apoptosis as proposed in the present study. TPA, 12-o-tetradecanoylphorbol 13-acetate; PGD2,

(13)

reported that PGD2 synthase induced apoptosis in PC12 neuronal cells which was inhibited by the addition of PGE1, E2, and F2a. Based on the observations that (a) PGD2 and PGJ2 (but not PGE2 or PGF2a) treatment resulted in caspase 3 activation in the presence of PARP and D4-GDI cleavage in leukemia cells, (b) the caspase 3 peptidyl inhibitor, Ac-DEVD-FMK, protected cells from PGD2- and PGJ2-induced apoptosis, and (c) NAC treat-ment significantly inhibited PGD2- or PGJ2-induced caspase 3 activation and the occurrence of apoptosis, these data indicate that apoptosis induced by PGD2and PGJ2 is mediated by activation of the caspase 3 cascade located downstream of ROS production.

The structure–activity relationship (SAR) of PGs with apoptosis induction is still unclear. Vosseler et al. [36]

indicated that 15-deoxy-D12,14-PGJ2, but not others, mark-edly reduced cell viability in HUVECs, and that both 15-deoxy-D12,14-PGJ2 and PGD2 treatment induced apoptosis by the annexin V assay. However, apoptosis induced by PGD2seems to occur to a lesser degree than that induced by 15-deoxy-D12,14-PGJ2. Results of the present study showed that both PGD2and PGJ2were effective apoptotic inducers in human leukemia cells, and that IC50values for PGD2and PGJ2 were 4.5F0.3 and 4.1F0.5 Ag/ml in HL-60 cells, respectively. PGJ2 has a reactive a,h-unsaturated carbonyl group in its cyclopentenone ring, and has been shown to exhibit several biological effects, including antiviral and antitumor activities. Interestingly, PGD2without a reactive a,h-unsaturated carbonyl group in its structure also showed apoptosis-inducing activity in cells in the present study. Shibata et al. [37] reported that PGD2 might be non-enzymatically converted to PGJ2 in the culture medium. Results of the present study demonstrated that the PGJ2 metabolites, 15-deoxy-D12–14PGJ2and D12PGJ2, exhibited an effective apoptotic effect in cells, and the IC50values of 15-deoxy-D12–14 PGJ2 (2.8F0.2 Ag/ml) and D12PGJ2 (2.3F0.4 Ag/ml) were much lower than those of PGD2 and PGJ2. This suggests that PGD2induction of apoptosis in human leukemia cells may be attributed to the conversion of PGD2to more reactive cyclopentenone PGs such as PGJ2, 15-deoxy-D12–14 PGJ2, and D12PGJ2 which in turn induce DNA fragmentation in cells.

In summary, the findings described in the present study suggest that PGD2and PGJ products, including PGJ2, 15-deoxy-D12–14 PGJ2, and D12PGJ2, possess apoptosis-inducing activity in human leukemia cells via activation of the caspase 3 cascade, and that production of ROS participates upstream of the caspase 3 cascade in the apoptotic mechanisms, which might be independent of activation of PPAR-g. Therefore, involvement of PGs in apoptosis during inflammation through ROS production is proposed in the present study (Fig. 10). Understanding the roles of ROS in other biological functions of PGs such as regulation of cell growth, differentiation, and inflamma-tion is an important topic and is reserved for future studies.

Acknowledgements

This study was supported by grants (NSC 92-2320-B-038-021 and 93-2321-B-038-009) from the National Sci-ence Council of the R.O.C.

References

[1] D.S. Straus, C.K. Glass, Cyclopentenone prostaglandins: new insights on biological activities and cellular targets, Med. Res. Rev. 21 (2001) 185 – 210.

[2] H.K. Na, Y.J. Surh, Peroxisome proliferator-activated receptor gamma (PPARg) ligands as bifunctional regulators of cell proliferation, Biochem. Pharmacol. 66 (2003) 1381 – 1391.

[3] R. Pai, B. Soreghan, I.L. Szabo, M. Pavelka, D. Baatar, A.S. Tarnawski, Prostaglandin E2 transactivates EGF receptor: a novel

mechanism for promoting colon cancer growth and gastrointestinal hypertrophy, Nat. Med. 8 (2002) 289 – 293.

[4] I. Nagatsuka, N. Yamada, S. Shimizu, M. Ohira, H. Nishino, S. Seki, K. Hirakawa, Inhibitory effect of a selective cyclooxygenase-2 inhibitor on liver metastasis of colon cancer, Int. J. Cancer 100 (2002) 515 – 519.

[5] S.C. Shen, C.H. Ko, K.C. Hsu, Y.C. Chen, 3-OH flavone inhibition of epidermal growth factor-induced proliferation through blocking prostaglandin E2production, Int. J. Cancer 108 (2004) 502 – 510.

[6] C.H. Ko, S.C. Shen, H.Y. Lin, W.C. Hou, W.R. Lee, L.L. Yang, Y.C. Chen, Flavanones structure related inhibition on TPA-induced tumor promotion through suppression of extracellular signal-regulated protein kinases: involvement of prostaglandin E2in anti-promotive

process, J. Cell. Physiol. 193 (2002) 93 – 102.

[7] H. Giles, P. Leff, The biology and pharmacology of PGD2,

Prostaglandins 5 (1988) 277 – 300.

[8] Y. Kikawa, S. Narumiya, M. Fukushima, H. Wakatsuka, O. Hayaishi, 9-Deoxy-delta 9, delta 12-13,14-dihydroprostaglandin D2, a

metabo-lite of prostaglandin D2formed in human plasma, Proc. Natl. Acad.

Sci. U. S. A. 81 (1984) 1317 – 1321.

[9] Y. Hirata, H. Hayashi, S. Ito, Y. Kikawa, M. Ishibashi, M. Sudo, H. Miyazaki, M. Fukushima, S. Narumiya, O. Hayaishi, Occurrence of 9-deoxy-delta 9,delta 12-13,14-dihydroprostaglandin D2 in human

urine, J. Biol. Chem. 263 (1988) 16619 – 16625.

[10] M. Kondo, T. Shibata, T. Kumagai, T. Osawa, N. Shibata, M. Kobayashi, S. Sasaki, M. Iwata, N. Noguchi, K. Uchida, 15-Deoxy-Delta(12,14)-prostaglandin J2: the endogenous electrophile that

induces neuronal apoptosis, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 7367 – 7372.

[11] C. Ward, I. Dransfield, J. Murray, S.N. Farrow, C. Haslett, A.G. Rossi, Prostaglandin D2 and its metabolites induce caspase-dependent

granulocyte apoptosis that is mediated via inhibition of I kappa B alpha degradation using a peroxisome proliferator-activated receptor-gamma-independent mechanism, J. Immunol. 168 (2002) 6232 – 6243. [12] S. Laurora, S. Pizzimenti, F. Briatore, A. Fraioli, M. Maggio, P. Reffo, C. Ferretti, M.U. Dianzani, G. Barrera, Peroxisome prolif-erator-activated receptor ligands affect growth-related gene expres-sion in human leukemic cells, J. Pharmacol. Exp. Ther. 305 (2003) 932 – 942.

[13] Y.C. Chen, S.C. Shen, W.R. Lee, H.Y. Lin, C.H. Ko, C.M. Shih, Wogonin and fisetin induction of apoptosis through activation of caspase 3 cascade and alternative expression of p21 protein in hepatocellular carcinoma cells SK-HEP-1, Arch. Toxicol. 76 (2002) 351 – 359.

[14] D.A. Hildeman, T. Mitchell, B. Aronow, S. Wojciechowski, J. Kappler, P. Marrack, Control of Bcl-2 expression by reactive oxygen species, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 15035 – 15040.

(14)

[15] V.S. Marsden, L. O’Connor, L.A. O’Reilly, J. Silke, D. Metcalf, P.G. Ekert, D.C. Huang, F. Cecconi, K. Kuida, K.J. Tomaselli, S. Roy, D.W. Nicholson, D.L. Vaux, P. Bouillet, J.M. Adams, A. Strasser, Apoptosis initiated by Bcl-2-regulated caspase activation independ-ently of the cytochrome c/Apaf-1/caspase-9 apoptosome, Nature 419 (2002) 634 – 637.

[16] M. Whyte, G. Evan, Apoptosis. The last cut is the deepest, Nature 376 (1995) 17 – 18.

[17] C. Soldani, A.I. Scovassi, Poly(ADP-ribose) polymerase-1 cleavage during apoptosis: an update, Apoptosis 7 (2002) 321 – 328. [18] F. Essmann, T. Wieder, A. Otto, E.C. Muller, B. Dorken, P.T. Daniel,

GDP dissociation inhibitor D4-GDI (Rho-GDI 2), but not the homologous rho-GDI 1, is cleaved by caspase-3 during drug-induced apoptosis, Biochem. J. 346 (2000) 777 – 783.

[19] Y.C. Chen, S.C. Shen, W.R. Lee, F.L. Hsu, H.Y. Lin, C.H. Ko, S.W. Tseng, Emodin induces apoptosis in human promyeloleukemic HL-60 cells accompanied by activation of caspase 3 cascade but independent of reactive oxygen species production, Biochem. Pharmacol. 64 (2002) 1713 – 1724.

[20] S.C. Shen, C.H. Ko, S.W. Tseng, Y.C. Chen, Structurally-related anti-tumor effects of flavanones in vitro and in vivo: involvement of caspase 3 activation, p21 gene expression and reactive oxygen species production, Toxicol. Appl. Pharmacol. 197 (2004) 84 – 95.

[21] Y.C. Chen, S.C. Shen, H.Y. Lin, Rutinoside at C7 attenuates the apoptosis-inducing activity of flavonoids, Biochem. Pharmacol. 66 (2003) 1139 – 1150.

[22] S.C. Shen, Y.C. Chen, F.L. Hsu, W.R. Lee, Differential apoptosis-inducing effect of quercetin and its glycosides in human promyelo-leukemic HL-60 cells by alternative activation of the caspase 3 cascade, J. Cell. Biochem. 89 (2003) 1044 – 1055.

[23] E.J. Kim, K.S. Park, S.Y. Chung, Y.Y. Sheen, D.C. Moon, Y.S. Song, K.S. Kim, S. Song, Y.P. Yun, M.K. Lee, K.W. Oh, Y. Yoon do, J.T. Hong, Peroxisome proliferator-activated receptor-gamma activator 15-deoxy-Delta12,14-prostaglandin J2inhibits neuroblastoma cell growth

through induction of apoptosis: association with extracellular signal-regulated kinase signal pathway, J. Pharmacol. Exp. Ther. 307 (2003) 505 – 517.

[24] H. Okano, K. Shiraki, H. Inoue, Y. Yamanaka, T. Kawakita, Y. Saitou, Y. Yamaguchi, N. Enokimura, N. Yamamoto, K. Sugimoto, K. Murata, T. Nakano, 15-deoxy-delta-12-14-PGJ2 regulates apoptosis induction and nuclear factor-kappaB activation via a peroxisome proliferator-activated receptor-gamma-independent mechanism in hepatocellular carcinoma, Lab. Invest. 83 (2003) 1529 – 1539. [25] A. Castrillo, P.G. Traves, P. Martin-Sanz, S. Parkinson, P.J. Parker, L.

Bosca, Potentiation of protein kinase C zeta activity by 15-deoxy-delta(12,14)-prostaglandin J(2) induces an imbalance between

mito-gen-activated protein kinases and NF-kappa B that promotes apoptosis in macrophages, Mol. Cell. Biol. 23 (2003) 1196 – 1208.

[26] Y. Nakamura, M. Kozuka, K. Naniwa, S. Takabayashi, K. Torikai, R. Hayashi, T. Sato, H. Ohigashi, T. Osawa, Arachidonic acid cascade inhibitors modulate phorbol ester-induced oxidative stress in female ICR mouse skin: differential roles of 5-lipoxygenase and cyclo-oxygenase-2 in leukocyte infiltration and activation, Free Radic. Biol. Med. 35 (2003) 997 – 1007.

[27] N. Yamakawa-Karakida, K. Sugita, T. Inukai, K. Goi, M. Nakamura, K. Uno, H. Sato, K. Kagami, N. Barker, S. Nakazawa, Ligand activation of peroxisome proliferator-activated receptor gamma induces apoptosis of leukemia cells by down-regulating the c-myc gene expression via blockade of the Tcf-4 activity, Cell Death Differ. 9 (2002) 513 – 526.

[28] N. Maulik, D.K. Das, Redox signaling in vascular angiogenesis, Free Radic. Biol. Med. 33 (2002) 1047 – 1060.

[29] C.H. Ko, S.C. Shen, Y.C. Chen, Hydroxylation at c4V or c6 is essential for apoptosis-inducing activity of flavanone through activation of the caspase 3 cascade and production of reactive oxygen species, Free Radic. Biol. Med. 36 (2004) 897 – 910.

[30] P. Fortini, B. Pascucci, E. Parlanti, M. D’Errico, V. Simonelli, E. Dogliotti, 8-Oxoguanine DNA damage: at the crossroad of alternative repair pathways, Mutat. Res. 531 (2003) 127 – 139.

[31] M. Kondo, T. Oya-Ito, T. Kumagai, T. Osawa, K. Uchida, Cyclo-pentenone prostaglandins as potential inducers of intracellular oxidative stress, J. Biol. Chem. 276 (2001) 12076 – 12083. [32] Y.C. Chen, S.Y. Lin-Shiau, J.K. Lin, Involvement of reactive oxygen

species and caspase 3 activation in arsenite-induced apoptosis, J. Cell. Physiol. 177 (1998) 324 – 333.

[33] A. Ishaque, M.J. Dunn, A. Sorokin, Cyclooxygenase-2 inhibits tumor necrosis factor alpha-mediated apoptosis in renal glomerular mesan-gial cells, J. Biol. Chem. 278 (2003) 10629 – 10640.

[34] H. Kusuhara, H. Matsuyuki, M. Matsuura, T. Imayoshi, T. Okumoto, H. Matsui, Induction of apoptotic DNA fragmentation by nonsteroidal anti-inflammatory drugs in cultured rat gastric mucosal cells, Eur. J. Pharmacol. 360 (1998) 273 – 280.

[35] L. Ragolia, T. Palaia, L. Frese, S. Fishbane, J.K. Maesaka, Prostaglandin D2synthase induces apoptosis in PC12 neuronal cells,

Neuroreport 12 (2001) 2623 – 2628.

[36] C.A. Vosseler, W. Erl, P.C. Weber, Structural requirements of cyclopentenone prostaglandins to induce endothelial cell apoptosis, Biochem. Biophys. Res. Commun. 307 (2003) 322 – 326.

[37] T. Shibata, M. Kondo, T. Osawa, N. Shibata, M. Kobayashi, K. Uchida, 15-deoxy-delta 12,14-prostaglandin J2. A prostaglandin D2

metabolite generated during inflammatory processes, J. Biol. Chem. 277 (2002) 10459 – 10466.