A Link between Benzyl Isothiocyanate-Induced Cell Cycle Arrest and Apoptosis: Involvement of Mitogen-Activated Protein Kinases in the Bcl-2 Phosphorylation

Isothiocyanates (ITCs) are agents that occur as glucosinolates in a variety of cruciferous vegetables, such as Brussica species (1). Naturally occurring ITCs are effective chemoprotective agents against chemical carcinogenesis in experimental animals (1, 2, 3, 4, 5). ITCs inhibit rat lung, esophagus, mammary gland, liver, small intestine, colon, and bladder tumorigeneses (5, 6). Epidemiologic studies also indicate a significant correlation between the dietary intake of ITC-containing foods and the reduced risk for prostate cancer (7). Inhibition of phase I enzymes that are required for the bioactivation of carcinogens and enhancement of the carcinogen excretion or detoxification by the phase II enzyme, including glutathione S-transferase (GST) and NAD(P)H:(quinone-acceptor) oxidoreductase, is believed to play a significant role in the chemopreventive activity of ITCs against chemical-induced carcinogenesis. In our continuing studies, we have screened fruits for sources of GST inducers and described the isolation and identification of benzyl ITC (BITC) as a potent major inducer from papaya (8). The involvement of the redox regulation in the gene expression of GSTP1 isozyme induced by BITC also was suggested (9). In addition, BITC blocks the neoplastic effects of diethylnitrosamine or benzo(a)pyrene on the lung and forestomach (10, 11, 12).

Recent studies have demonstrated that ITCs inhibit cell growth by inducing apoptosis, which is suggested to be potentially involved in the anticarcinogenic action of ITCs. For instance, BITC and phenethyl ITC (PEITC) are shown to induce apoptotic cell death in cultured cells (13). Huang et al. (14) demonstrated that PEITC effectively blocks tumor-promoting phorbol ester-induced cell transformation in JB6 mouse epidermal cells by inducing apoptosis. Several pathways involved in ITC-triggered apoptosis have been postulated, e.g., a c-Jun N-terminal kinase (JNK)-dependent pathway found in Jurkat T-cells (15), a p53-dependent pathway found in JB6 cells (14), a Bax, cytochrome c-dependent, and p53-independent pathway found in HT29 colon cancer cells (16), and an extracellular signal-regulated kinase (ERK)-mediated pathway in p53-deficient PC-3 human prostate cancer cells (17). In the recent study, we demonstrated that one of the possibilities involved in the activation of a caspase-3-like protease by BITC is a mitochondrial death pathway (18). Although involvement of mitogen-activated protein kinases (MAPKs) in ITC-induced apoptosis has been established, involvements in the cell cycle arrest or the target molecules of ITC-activated MAPKs are not fully understood.

In the present study, we investigated the regulatory mechanism of cell death elicited by BITC in human T-cell leukemia Jurkat cells. We indicate for the first time the important roles of p38 MAPK in the BITC-induced G₂-M cell cycle arrest and apoptosis. We also demonstrate that BITC modifies synergistically the antiapoptotic Bcl-2 function by JNK-catalyzed phosphorylation, and discuss the differential and synergistic roles of the MAPK pathways.

BITC and Triton X-100 were obtained from Nacalai Tesque, Inc. (Kyoto, Japan). AlamarBlue was a product of Alamar BioScience (Sacramento, CA). Propidium iodide (PI) was obtained from Molecular Probes, Inc. (Eugene, OR). The antibodies against phospho-JNK (Thr183/Tyr185), phospho-c-jun (Ser73), phospho-p38 (Thr180/Tyr182), phospho-ATF-2 (Thr71), phospho-ERK (Thr202/Tyr204), phospho-Bcl-2 (Ser70), and phospho-Cdc2 (Tyr15) were purchased from Cell Signaling Technology, Inc. (Beverly, MA), and antibodies against anti-Bcl-2, anti-Bax, and anti-glyceraldehyde 3-phosphate dehydrogenase were from Santa Cruz Biotechnology (Santa Cruz, CA). Horseradish peroxidase-linked antirabbit and antimouse IgG immunoglobulins and enhanced chemiluminescence Western blotting detection reagents were obtained from Amersham Pharmacia Biotech (Buckinghamshire, United Kingdom). SB202190 and SP600125 were from BIOMOL Research Laboratories, Inc. (Plymouth Meeting, PA). Thymidine and nocodazole were purchased from Sigma (St. Louis, MO). All of the other chemicals were purchased from Wako Pure Chemical Industries (Osaka, Japan).

The human T-cell leukemia Jurkat cells (RIKEN Cell Bank; Tsukuba, Ibaraki, Japan) and acute promyelotic leukemia HL-60 cells (Health Science Research Resources Bank, Osaka, Japan) were maintained in RPMI 1640 (Sigma). The cervix carcinoma HeLa cells (provided Dr. H. Shibata of Nagoya University) were maintained in DMEM (Nissui Pharmaceutical Co. Ltd., Tokyo, Japan). These media were supplemented with 10% heat-inactivated FCS (Trace Scientific, Ltd., Melbourne, Australia), 50 units/ml penicillin, and 50 μg/ml streptomycin and grown in an atmosphere of 95% air and 5% CO₂ at 37°C. Jurkat cells treated with 1 mm thymidine (to arrest cells in G₁-S) or 150 ng/ml nocodazole (to arrest cells in G₂-M) for 16 h at 37°C were washed three times with fresh medium to release the cell cycle arrest and used in subsequent experiments (19).

Jurkat cells (2 × 10⁴) in 50 μl of culture medium were mixed with 50 μl of BITC in 96-well microculture plates. After culturing at 37°C for 24 h, the number of viable cells was determined. For quantitative analysis of cell viability, 100 μl of the culture medium and 10 μl of an AlamarBlue solution were added to each well, and the fluorescence was measured with excitation at 560 nm and emission at 590 nm according to the manufacturer’s direction after incubation at 37°C for 2 h in a humidified CO₂ incubator. The obtained values were compared with those of the control incubated with vehicle only.

Jurkat cells were incubated in culture medium in the presence or absence of BITC. For the DNA fragmentation analysis, 5 × 10⁵ cells were pelleted by centrifugation, and DNA was isolated from the cell pellets as described by Sellins and Cohen (20). The DNA then was subjected to electrophoresis in 2% agarose gels, stained with ethidium bromide, and imaged with a FluorImager (Molecular Dynamics, Tokyo, Japan), and the intensity was analyzed by ImageQuant (Molecular Dynamics).

Cell cycle analysis was performed on the harvested cell pellets treated with 0.2% Triton X-100 in PBS and PI solution (20 μg/ml) containing RNase A (100 μg/ml). The mixture was analyzed immediately by a flow cytometer. The cell cycle distribution was measured using the EPICS XL System II (Coulter, Tokyo, Japan). To evaluate the cell cycle distribution of apoptotic cells, terminal deoxynucleotidyl transferase-mediated (fluorescein-) nick end labeling (TUNEL)/PI double staining was done using MEBSTAIN Apoptosis Kit Direct (Medical and Biological Laboratories, Nagoya, Japan) according to the manufacturer’s instructions. Briefly, asynchronous or synchronized cells (1 × 10⁶) incubated with BITC were fixed by 4% paraformaldehyde at 4°C for 30 min, followed by washing with PBS. Cold 70% ethanol was added to the cell pellet. It then was incubated for 2 h at −20°C for permeabilization. After washing with PBS, terminal deoxynucleotidyl transferase reaction reagent was added to the cell pellet and incubated for 1 h at 37°C. The cell pellet was washed with PBS and suspended in PBS including PI (10 μg/ml) and RNase A (10 μg/ml) and analyzed by a flow cytometer. The cell cycle distribution of apoptotic cells was evaluated using the XL System II.

BITC-treated and untreated cells were harvested, washed twice with ice-cold PBS (pH 7.0), and lysed with lysis buffer [50 mm Tris-HCl (pH 7.5), 150 mm NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS 100 μg/ml phenylmethylsulfonyl fluoride, and 0.5 mm Na₃VO₄]. Whole cell lysates were incubated with the SDS sample buffer for 5 min at 100°C. The protein concentrations were determined using the BCA protein assay kit (Pierce, Rockford, IL). The same protein volume samples then were separated by 10–15% SDS-PAGE. The gel was transblotted onto a nitrocellulose membrane (Hybond enhanced chemiluminescence; Amersham), incubated with Block Ace (40 mg/ml; Dainippon Seiyaku, Osaka, Japan) for blocking, washed, and incubated with antibody. This procedure was followed by the addition of horseradish peroxidase conjugated to IgG and enhanced chemiluminescence reagents (Amersham). The bands were visualized using a Cool Saver AE-6955 (Atto, Tokyo, Japan).

Densitometric measurement of immunoblot was performed using an Atto Lane Analyzer (Atto). Quantification of the protein levels was estimated by comparing the intensity of each specific protein band from the control or basal conditions with that of the BITC treatment conditions. All of the experimental data were compensated using each glyceraldehyde 3-phosphate dehydrogenase intensity as an internal standard. When applicable, the mean ± SE values are shown. The data were analyzed with an ANOVA when necessary, followed by Fisher’s exact test. Specific differences among treatments were examined using the Student’s t test (two sided), which assumed unequal variance.

BITC-treated and untreated cells were harvested, washed twice with ice-cold PBS (pH 7.0), resuspended in extraction buffer [20 mm Tris-HCl (pH 7.5), 137 mm NaCl, 2 mm EDTA, 1% Triton X-100, Complete Mini (Roche, Basel, Switzerland)], and centrifuged at 15,000 × g for 20 min at 4°C. The supernatants were precleaned using protein G-Sepharose (Amersham Pharmacia Biotech) and then immunoprecipitated using 1 μg anti-Bcl-2 antibody overnight at 4°C. Ten μl of protein-G suspension then were added to each of the immunocomplexes, and the mixtures were rotated for 1 h at 4°C. After four washes with extraction buffer, the immunocomplexes were eluted with SDS sample buffer. The immunoprecipitated proteins were subjected to SDS-PAGE and immunoblot analysis with the antibody.

We evaluated the effect of BITC on cellular viability using the modified 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide test, the AlamarBlue assay. When Jurkat cells were incubated with BITC, the viability was inhibited in a concentration-dependent manner up to 100 μm with an IC₅₀ value of ∼6 μm (Fig. 1,A). We also observed chromatin condensation (data not shown) and DNA fragmentation in a nucleosome unit (Fig. 1,B), both of which are the characteristic features of apoptosis, in the cells treated with 1–10 μm BITC for 20 h. The pretreatment with z-Val-Ala-Asp-fluoromethyl ketone significantly blocked the DNA fragmentation caused by BITC (data not shown). Although BITC-induced cytotoxicity was concentration dependent, DNA fragmentation was not detected at the higher concentration of BITC (>25 μm; Fig. 1,B). The increased population of necrotic cell death induced by the higher concentration of BITC was confirmed by fluorescence microscopy after staining with PI (data not shown). Flow cytometric analysis of Jurkat cells exposed to BITC for 15 h indicated that 2.5- and 5-μm concentrations of BITC arrested the cells in the G₂-M phase (Fig. 1 C). However, as the BITC concentration increased further, the dominant ratio of sub-G₁-degraded cells was observed in place of the G₂-M phase population after the 15-h incubation with BITC. Therefore, the cells were treated with 5 μm of BITC required for the G₂-M arrest and the apoptosis induction in subsequent experiments.

BITC shows the strongest inhibitory activity toward G₂-M cell cycle progression among the several ITC compounds (21). To analyze the relationship between the BITC-induced apoptosis and G₂-M cell cycle arrest, a flow cytometric analysis of the BITC-treated cells was performed using TUNEL and PI staining. As shown in Fig. 2,A, G₂-M phase-specific apoptosis induction was found in the Jurkat cells treated with BITC (5 μm) for 12 h. As the incubation period increased, greater amounts of G₂-M-TUNEL double-positive cells and the sub-G₁ population were observed (data not shown). Moreover, we examined the effect of BITC in the cells synchronized at the G₂-M phase by nocodazole treatment. Treatment of the G₂-M-synchronized cells with BITC only for 3 h led to the appearance of a G₂-M phase-TUNEL double-positive cell population (Fig. 2 B). These results indicated that G₂-M-arrested cells might be sensitive to undergoing apoptotic stimulation by BITC.

Previous studies have suggested that ITC-induced apoptosis may be mediated by activation of JNK 1/2 (16, 22). In contrast, the activation of ERK 1/2 by treatment with PEITC, one of the potent anticarcinogenic ITCs, was evident in prostate cancer cells as mentioned previously (17). To confirm whether BITC activates classical MAPKs in human Jurkat cells, the activity of MAPKs in the cells exposed to BITC for the indicated time periods was determined by immunoblot analysis using phosphorylated protein-specific antibodies. Phosphorylation of JNK 1/2 was detected 30 min after exposure to 5 μm BITC (data not shown) and increased up to 3 h, followed by a sustained activation to 9 h (Fig. 3 A). We found for the first time that exposure of Jurkat cells to 5 μm BITC resulted in a rapid increase and sustained activation of p38 MAPK phosphorylation. Moreover, BITC induced a transient increase in the phosphorylation of c-jun and activating transcription factor 2 (ATF-2), the representative substrates for JNK and p38 MAPK, respectively, with a peak at 3–6 h. Conversely, the ERK 1/2 phosphorylation level seems to be suppressed slightly rather than demonstrating no change, as in Jurkat cells treated with BITC.

The effect of SP600125 and SB202190, specific inhibitors of JNK (23) and p38 MAPK (24), respectively, on the kinase activity of the phosphorylated JNK and p38 MAPK (c-jun and ATF-2 activation) was determined. As shown in Fig. 3, B and C, the BITC-mediated phosphorylation of c-jun was reduced significantly in the presence of 10 μm SP600125, the JNK inhibitor, whereas the phosphorylation of ATF-2 was not altered significantly. Conversely, SB202190 efficiently inhibited ATF-2 phosphorylation at a 20-μm concentration, whereas the c-jun phosphorylation was inhibited slightly even at 50 μm. This partial inhibition of ATF-2 phosphorylation may be caused by activated JNK, which can phosphorylate ATF-2 (25). Thus, to examine the effect of MAPK inhibitors in the subsequent experiments, the cells were pretreated with SP600125 or SB202190 at a concentration required for specific inhibition of each kinase activity and BITC-induced apoptosis (10 μm or 20 μm, respectively).

To examine the possible role of JNK and p38 MAPK in BITC-induced apoptosis, we next examined the effect of these inhibitors on the BITC-induced apoptosis. As shown in Fig. 4,A, SP600125 and SB202190 significantly reduced the rate of TUNEL-positive cells. However, the ERK inhibitor PD98059 did not affect, even at 50 μm, the extent of apoptosis induced by 5 μm BITC. These results strongly suggested the possibility of p38 MAPK and JNK to regulate BITC-induced apoptosis. Moreover, the flow cytometric analysis of Jurkat cells exposed to BITC for 15 h showed the effect of SP600125 (10 μm) or SB202190 (20 μm) on the G₂-M progression (Fig. 4 B). Pretreatment of SB202190 for 30 min reduced the sub-G₁ and G₂-M populations, both of which were accumulated by the BITC treatment. Despite its reducing ability of the sub-G₁ population appearance, SP600125 had little effect on the G₂-M population.

It has been demonstrated previously that the accumulation of the inactive tyrosine 15-phosphorylated Cdc2 form is definite evidence of a cell division cycle arrest preventing entry into the G₂ and M phase (26, 27). Hence, we analyzed the amount of phosphorylated Cdc2 using a specific antibody. As shown in Fig. 5,A, 5 μm BITC time-dependently enhanced the level of tyrosine 15-phosphorylated Cdc2. SB202190, the p38 MAPK inhibitor, exhibited a marked suppression of Cdc2 phosphorylation to the control level (Fig. 5,B). In addition, SP600125 also slightly inhibited the Cdc2 inactivation, which may be caused by the nonspecific cross-reactions to the p38 MAPK (Fig. 3 B). The obtained data strongly suggested that the p38 MAPK pathway might play an important role in BITC-induced G₂-M arrest; UV-activated p38 also is capable of directly phosphorylating the serine 309 of Cdc25B and suppressing G₂-M progression (28). Additional mechanistic studies of the relationship between p38 activation and cell cycle progression signal are essential to provide supporting information.

Bcl-2 family members are well known to be important proapoptotic or antiapoptotic regulators. Enhanced expression of the Bcl-2 has been reported to block ITC-induced apoptosis (15). However, the involvement of Bcl-2 protein expression and/or modification and its relation to the MAPK pathways in BITC-induced apoptosis are not fully understood. Hence, we determined their expression or phosphorylation of Bcl-2 in the untreated and BITC-treated cells by immunoblot analysis of the total cell lysates. As shown in Fig. 6, A and C, although slightly inducing the total Bcl-2 protein (1.6-fold) for 12 h in Jurkat cells, BITC enhanced dramatically the phosphorylated Bcl-2 (∼9.5-fold for 12 h) in a time-dependent manner (Fig. 6, B and C). Moreover, the Bcl-2 phosphorylation was abrogated completely in the presence of the JNK-specific inhibitor SP600125, whereas SB202190 also slightly inhibited it (Fig. 6, B and D). These results obviously implied that BITC-activated JNK might mediate the inactivation of Bcl-2 by phosphorylation corresponding to Bax induction (∼1.8-fold; data not shown) because the equilibrium between the Bcl-2 and Bax expressions is suggested to be an important regulator of the apoptotic signaling pathway (29). An immunoprecipitation experiment demonstrated that BITC reduced the Bcl-2/Bax interaction (Fig. 6,E), strongly supporting this assumption. In addition, involvement of the MAPK activation in the Bcl-2 phosphorylation and apoptosis induction also was observed in HL-60 and HeLa cells. In brief, BITC significantly enhanced the phosphorylated Bcl-2 in a time-dependent manner in both cells (Fig. 7, A and D). This phosphorylation was abrogated in the presence of the JNK-specific inhibitor SP600125, whereas SB202190 also slightly inhibited it (Fig. 7, B and E). Moreover, SP600125 and SB202190 significantly reduced the rate of TUNEL-positive cells in both cell lines (Fig. 7, C and F). However, the ERK inhibitor PD98059 did not affect, even at 50 μm, the extent of apoptosis induced by BITC (data not shown).

Partial inhibition of the Bcl-2 phosphorylation by SB202190 led us to an assumption that the p38 MAPK-dependent G₂-M arrest may be involved in the accumulation of phosphorylated Bcl-2. To support this, the effect of the apoptosis-inducing compounds on the Bcl-2 modification was examined compared with that of BITC. It has been widely known that etoposide and H₂O₂ induce apoptotic cell death through the mitochondrial death pathway accompanied by G₀-G₁ or S phase arrest in the cell cycle (30, 31). As shown in Fig. 8,A, treatment of Jurkat cells with etoposide or H₂O₂ at the concentrations (25 μm) required for significant induction of G₀-G₁ or S phase arrest (data not shown) did not show any dramatic change in the total and phosphorylated protein level of Bcl-2. A correlation between the phosphorylated Bcl-2 amount and the ratio of G₂-M-arrested cells was observed (Fig. 8 B). Therefore, BITC-induced cell cycle regulation seems to contribute partly to the Bcl-2 phosphorylation.

To determine the effect of the BITC-induced cell cycle regulation on the Bcl-2 phosphorylation, the synchronized Jurkat cells were prepared by the release of thymidine block, by which the cells were arrested at the G₁-S phase. The synchronized cells mainly entered into the S phase 6 h after release from the thymidine block and then into the G₂-M phase 9 h after release (Fig. 9,A). To examine the effect of BITC in the cells at each phase, the synchronized cells untreated (dotted line interval) or treated with BITC for the past 3 h during the indicated periods (solid line interval) after the thymidine block release were prepared (Fig. 9,B). As shown in Fig. 9,C, the enhanced phosphorylation was observed in the cells only released from thymidine block for 9–12 h (G₂-M phase-dominant; entries d and e) but not during 0–6 h (G₁-S phase-dominant; entries a through c). Furthermore, BITC stimulation for 3 h at the G₂-M phase (the cells were treated with BITC 9 h after the thymidine block release and then incubated for 3 h; entry i) synergistically increased in the phosphorylated Bcl-2 protein by 4.3-fold compared with the spontaneous phosphorylation in the G₂-M-dominant Jurkat cells (synchronized by the release of thymidine block for 12 h; entry e; Fig. 9,C). In the other phase cells, BITC-enhanced Bcl-2 phosphorylation was not prominent, strongly suggesting that BITC is likely to specifically enhance the Bcl-2 phosphorylation in the G₂-M phase. Moreover, the BITC-induced enhancement was suppressed to the control level in the presence of the JNK inhibitor SP600125 but not by SB202190 (Fig. 9,D). In addition, the treatment of these inhibitors did not change the spontaneous (G₂-M cell cycle arrest-dependent) Bcl-2 phosphorylation level. Although completely inhibiting the Bcl-2 phosphorylation induced by the treatment of the normal (asynchronous) cells with BITC for 12 h (Fig. 6,D), SP600125 showed no effect on the spontaneous phosphorylation in the G₂-M synchronized cells (Fig. 9 D). These results implied the existence of a distinct kinase from JNK, regulating the Bcl-2 phosphorylation universally at the G₂-M phase (32).

We next analyzed the cell cycle specificity of the apoptosis induction by comparing various phase synchronized Jurkat cells. The synchronized Jurkat cells were prepared by the release of thymidine block. The DNA fragmentation level, one of the hallmarks of apoptosis, was increased together with the cell cycle progression from G₁-S to G₂-M (Fig. 9,E). We observed a positive correlation between inactivation of antiapoptotic Bcl-2 by the BITC-enhanced phosphorylation and DNA fragmentation content in the Jurkat cells (Fig. 9 F). No DNA fragmentation was detected in the synchronized cells without BITC treatment (data not shown).

Diverse types of cancer chemopreventive agents, including naturally occurring and pharmaceutical compounds, have been studied for efficacy in vitro and in vivo. One of the most important groups of dietary compounds is the ITCs, which occur naturally in a variety of cruciferous vegetables. BITC is an ITC that has been isolated from papaya fruit as a major phase II enzyme inducer present in the organic solvent extracts of this fruit (8). We have reported more recently that BITC induces apoptosis in rat liver epithelial cells dependent on the mitochondrial death pathway-activating caspase-9/-3 activities (18). Our interest in the molecular mechanism of the apoptosis-inducing effect of BITC stemmed from the following observations: (a) BITC is quickly and continuously accumulated into cells, and the intracellular concentration of BITC increased up to 300 μm (33); (b) BITC modifies the mitochondrial function (18); (c) reactive oxygen species production and the resulting cellular redox change can be part of the signal transduction pathway during apoptosis (18); and (d) MAPK cascade is involved partially in apoptosis in several ITC compounds (16, 17, 22). We additionally confirmed that BITC induces MAPK activation (JNK and p38 MAPK) and apoptosis in cultured human T lymphocytes (Figs. 1 and 3). Because apoptosis is the end of the cell differentiation pathways of several types of cells in some tissues, a process that results in apoptosis should minimize the proliferative signal, an important determinant of the tumor development process.

We demonstrated herein that two cooperative antiproliferative pathways are activated by BITC in Jurkat cells; one is a p38 MAPK-dependent pathway, first demonstrated in the present study, resulting in cell cycle arrest accompanied with the inactive Cdc2 accumulation, and the other is a JNK-dependent pathway leading to the Bcl-2 phosphorylation. The cooperation between cell cycle regulation and apoptosis was strongly suggested by the observation that the G₂-M-arrested cells were sensitive to apoptosis induction by the BITC treatment (Fig. 2). We showed a strong correlation between each MAPK (p38 MAPK and JNK) inhibitor concentration required to inhibit the corresponding MAPK activities, cell cycle arrest, and apoptosis. It also was demonstrated that the p38 MAPK pathway could be operative in cell cycle arrest and apoptosis induced by BITC, whereas the JNK pathway plays a role in the apoptosis but not in the cell cycle regulation (Figs. 4 and 5). Therefore, we proposed the existence of a key molecule linking the p38-dependent cell cycle arrest with the apoptosis induction by BITC. In addition, these phenomena might not be the result of the Jurkat cell-specific effects because involvement of the MAPK activation in the Bcl-2 phosphorylation and apoptosis induction also was observed in the additional cell lines such as HL-60 and HeLa (Fig. 7, A through F).

We observed no significant change in the expression or degradation of Bcl-2 (Fig. 6,A) and Bax (unpublished data), even though previous studies indicated the importance of the Bcl-2 family protein expression during ITC-related compound-induced apoptosis (15). Contrarily, we revealed for the first time that BITC stimulates JNK-dependent Bcl-2 phosphorylation (Fig. 6, B and D). Additionally, we observed that the Bcl-2/Bax interaction was decreased exactly by BITC treatment (Fig. 8,A). Therefore, phosphorylated Bcl-2 appears to lose a binding ability with Bax, leading to the enhanced susceptibility of the cells to apoptosis (34, 35). It has been shown recently that Bcl-2 phosphorylation universally occurs at the G₂-M phase of the cell cycle (36, 37, 38) and that a JNK-dependent pathway may be involved (38). In fact, Bcl-2 phosphorylation occurs not only in untreated synchronized cells but also in response to BITC accompanied with the G₂-M transition (Fig. 9,C). These findings suggest that BITC-induced Bcl-2 phosphorylation at the G₂-M phase may dissociate the interaction between the Bcl-2 and Bax, resulting in the apoptosis induction triggered by the mitochondrial death signal. Conversely, treatment of the G₂-M phase-synchronized cells with BITC dramatically enhanced the JNK-dependent phosphorylation of Bcl-2 (Fig. 9, C and D), suggesting that the BITC-induced Bcl-2 phosphorylation is not merely a consequence of the G₂-M arrest but a synergistic event caused by the activation of the p38 and JNK pathways (Fig. 10). Bcl-2 phosphorylation can occur in response to various cytokines and microtubule-damaging agents (34). In some systems, phosphorylation of Bcl-2 at residues Thr-56, Thr-74, and Ser-87 can prevent proteasome targeting and can confer resistance against the induction of apoptosis (39). We have observed more recently that BITC did not induce the Bcl-2 phosphorylation at sites other than Ser-70, including Thr-56, Thr-74, and Ser-87 (supplementary data).¹ These findings suggested that Bcl-2 phosphorylation at Thr-56, Thr-74, and Ser-87 could be ruled out in the mechanism, and thus a main target of apoptosis induction by BITC is Ser-70 of Bcl-2.

Lack of selectivity in the killing of tumor and normal cells is a major obstacle in the apoptosis study and cancer therapy. In principle, chemotherapic drugs and chemopreventive agents possessing a cytotoxic activity poorly discriminate between normal and cancer cells. Interestingly, a recent study demonstrated that cucurbitacin E, a microtubule active agent inducing G₂-M arrest, potently inhibits proliferating human endothelial cells as compared with quiescent cells in vitro, suggesting it as a potential anticancer agent (40). Blagosklonny et al. (41) envisioned the protection of normal cells that is mainly based on four principles: (a) cancer cells are less dependent on growth factors; (b) it is possible to selectively arrest normal cells in the interphase; (c) G₂-M-arrest inducers are toxic in cycling agents; and (d) cells arrested interphase may be insensitive to G₂-M-arrest inducers. Therefore, we propose that BITC, inducing the G₂-M cell cycle arrest-dependent apoptosis, in combination with reversible and selective inhibitors of the cycle of proliferating epithelial and hematopoietic cells may be able to induce apoptosis selectively in cancer cells. In preliminary experiments, we have compared the quiescent and log-phase human normal epithelial cells and found that the latter are significantly more sensitive to BITC-inducing cytotoxicity. In addition, the narrow threshold of BITC to exhibit apoptosis should be considered. Treatment with an excessive concentration of BITC resulted in severe cytotoxicity without DNA ladder formation (Fig. 1 B). We have demonstrated recently that BITC produced intracellular reactive oxygen species in a dose-dependent manner (9, 18). These oxidative phenomena may lead to necrotic cell death and thus damage to the surrounding cells. It is likely that the intracellular reactive oxygen species generation plays an important role in the unfavorable side effects of BITC on carcinogenesis. Additional mechanistic studies on the involvement and/or disturbance of intracellular oxidative stress in cell cycle arrest and apoptosis induction by BITC are necessary.