Molecular Mechanisms of c-Jun N-terminal Kinase-mediated Apoptosis Induced by Anticarcinogenic Isothiocyanates*

Isothiocyanates have strong chemopreventive properties against many carcinogen-induced cancers in experimental animal models. Here, we report that phenylmethyl isocyacyanate (PMITC) and phenylethyl isothio- cyanate (PEITC) induced sustained c-Jun N-terminal kinase (JNK) activation in a dose-dependent manner. The sustained JNK activation caused by isothiocyanates was associated with apoptosis induction in various cell types. An inhibitor of the caspase/interleukin-1β-converting enzyme blocked isothiocyanate-induced apoptosis without inhibiting the JNK activation, which suggests that JNK activation by isothiocyanates is an event that is independent or upstream of the activation of caspase/interleukin-1β-converting enzyme proteases. PEITC-induced apoptosis was suppressed by interfering with the JNK pathway with a dominant-negative mutant of JNK1 or MEKK1 (JNK1(APF) and MEKK1(KR), respectively), implying that the JNK pathway is required for apoptotic signaling. Isothiocyanate-induced JNK activation was blocked by the antioxidants 2-mercaptoethanol andN-acetyl-l-cysteine, suggesting that the death signaling was triggered by oxidative stress. Overexpression of Bcl-2 suppressed PEITC-induced JNK activation. In addition, Bcl-2 and Bcl-xL suppressed PEITC-induced apoptosis, but failed to protect cells from death induced by overexpression of activated JNK1. These results suggest that Bcl-2 and Bcl-xL are upstream of JNK. Taken together, our results indicate (i) that JNK mediates PMITC- and PEITC-induced apoptosis and (ii) that PMITC and PEITC may have chemotherapeutic functions besides their chemopreventive functions.

Apoptosis plays important roles in developmental processes, maintenance of homeostasis, and elimination of seriously damaged cells (1,2). The aberrant regulation of apoptosis has been observed in many disorders such as neuronal diseases, AIDS, autoimmune diseases, and cancers (2). In addition, many therapeutic agents eliminate tumor cells by inducing apoptotic cell death (2). Therefore, understanding the mechanism of apopto-sis has important implications in the prevention and treatment of many diseases.
Many isothiocyanates are effective chemopreventive agents against carcinogen-induced cancers in experimental animals. Isothiocyanates inhibit cancer formation in various tissues such as rat lung, esophagus, mammary gland, liver, small intestine, colon, and bladder cancers (29 -33). Isothiocyanates inhibit carcinogenesis caused by different compounds, including nitrosamines and polycyclic aromatic hydrocarbons (33). Previous studies suggested that isothiocyanates may inhibit enzymes (e.g. cytochrome P-450 isoforms) that are required for the bioactivation of carcinogens (34,35). In addition, isothiocyanates may increase the carcinogen excretion or detoxification by inducing the phase II detoxifying enzymes, including glutathione S-transferase (GST), quinone reductase, epoxide hydrolase, and UDP-glucuronosyltransferase (35)(36)(37). Here, we report that phenylmethyl isothiocyanate (PMITC; benzyl isothiocyanate) and phenylethyl isothiocyanate (PEITC) are capable of inducing persistent JNK activation in a dose-dependent manner. Our study indicates the involvement of JNK-mediated apoptosis in the anticarcinogenic functions of isothiocyanates.
DNA Fragmentation Assays-10 6 cells were lysed in 50 l of 100 mM NaCl, 40 mM Tris-Cl (pH 7.4), and 20 mM EDTA containing 0.5% SDS. The lysate was heated at 65°C for 10 min for inactivation of nucleases and digested with 0.5 mg/ml proteinase K at 50°C for 2 h. The lysate was then incubated with 0.2 mg/ml RNase A at 50°C for 2 h. The DNA fragmentation was analyzed on a 1.8% agarose gel in the presence of 0.5 g/ml ethidium bromide.
Transient Transfection Cell Death/Protection Assays-The protection assay was performed as described (40) with modifications. Briefly, 293 cells were plated 24 h before transfection at a density of 1.5 ϫ 10 5 /35-mm well. Cells were cotransfected with the pCMV-␤gal plasmid encoding ␤-galactosidase and plasmids for control vector, Bcl-2 or Bcl-x L , or mutant kinases by a calcium phosphate precipitation protocol (Specialty Media), with duplicates in each transfection. After removing the transfection mixture, the cells were incubated in complete medium for 12 h for recovery and then treated with or without drugs. Cells were harvested 24 h post-treatment, washed, and fixed in 1% paraformaldehyde in phosphate-buffer saline (PBS). The fixed cells were washed once with PBS, resuspended in staining solution (PBS (pH 7.4), 1 mM MgCl 2 , 10 mM K 4 (Fe(CN) 4 ), 10 mM K 3 (Fe(CN) 4 ), 0.1% Triton X-100, and 1 mM X-gal) for 2-6 h, and then washed twice with PBS. The ratio of ␤-galactosidase-expressing cells (blue color) was examined with a hemocytometer. Cell survival was determined as follows: (% of blue cells in treated group/% of blue cells in untreated group) ϫ 100%.
To perform transient transfection/cell death assays, 293 cells were transfected with empty vectors or JNK1-plus ⌬MEKK1-expressing plasmids (in the absence or presence of Bcl-2 or Bcl-x L ). Cells were collected 48 h after transfection, fixed, and stained as described above. Cell survival was determined as follows: (% of blue cells in the experimental group/% of blue cells in the control) ϫ 100%.

Isothiocyanates Induce JNK Activation in a Dose-dependent
Manner-Human leukemia Jurkat T-cells were treated with different concentrations of PITC, PMITC, PEITC, and PPITC. Cells were collected 2 h after treatment, and endogenous JNK activity was determined by immunocomplex kinase assays. Among the tested isothiocyanates, PMITC and PEITC induced strong JNK activation at a concentration of 5 M, and PITC induced mild JNK induction at higher concentrations (50 -100 M). No apparent JNK activation was observed with PPITC treatment at all concentrations tested (Fig. 1A). In contrast to the JNK activation, decreases in JNK activity (in comparison with the basal levels in the untreated group) were observed in treatments with high concentrations (Ͼ50 M) of PMITC, PEITC, and PPITC (Fig. 1A). The decreases in JNK activity in treatments with high doses of isothiocyanates may be due to the acute cytotoxicity of these drugs. Jurkat cells swelled and lost the ability to exclude trypan blue (Fig. 1B), and the protein recovery in the treated cell lysate was significantly decreased after treatments with high doses of PMITC, PEITC, and PPITC. These phenomena indicated that the integrity of the cell membrane was lost and that the cells died by membrane disintegration and cytolysis, which is reminiscent of necrosis. However, Ͼ80% of the cells retained the ability to exclude trypan blue 24 h post-treatment with different isothiocyanates at 5 M (Fig. 1B).
PMITC and PEITC Induce Sustained JNK Activation and Apoptosis-To exclude the possibility that the differential regulation of JNK by isothiocyanates is due to the selected observation at the 2-h time point, we did a time course study of JNK activation with various isothiocyanates at either 5 M ( Fig. 2A) or 50 M (Fig. 2B) in Jurkat T-cells. At a concentration of 5 M, both PMITC and PEITC induced persistent JNK activation. JNK activity increased at the 1-h time point, peaked around 2-4 h post-treatment, and gradually decreased, but remained higher than basal levels even 12 h after treatment. In contrast, 5 M PITC or PPITC failed to induce any JNK activation. At 50 M, only PITC induced a slight JNK induction, whereas PMITC, PEITC, and PPITC decreased the basal levels of JNK activity. These results were consistent with the data in the dose-response experiment (Fig. 1A), showing that isothiocyanates induced JNK activation in a dose-dependent manner and also revealing that PMITC and PEITC were strong JNK activators, inducing sustained JNK activation at certain concentrations.
In our previous study, persistent JNK activation was associated with apoptosis induced by ␥-radiation and UV-C (26). Since most of the early apoptotic cells maintain an intact cytoplasmic membrane, trypan blue staining is not an appropriate criterion for apoptosis. Hence, a DNA fragmentation assay was used to examine if the persistent JNK activation induced by PMITC and PEITC was associated with apoptotic cell death. Cellular DNA was extracted from Jurkat cells treated with different isothiocyanates (5 M) and analyzed for DNA frag-mentation. We found that only PMITC and PEITC, which induced persistent JNK activation at 5 M, caused chromosomal DNA laddering at 12-24 h after treatment (Fig. 2C). This result shows the correlation between persistent activation of JNK and apoptosis induction by isothiocyanate treatments and the precedence of JNK activation to DNA fragmentation.
The Caspase/ICE Protease Inhibitor Fails to Inhibit JNK Activation by Isothiocyanates-The caspase/ICE family of proteases are known to be important apoptosis mediators (42); hence, we determined if caspases/ICE-like proteases are involved in JNK-mediated apoptosis induced by isothiocyanates. PMITC-and PEITC-induced DNA fragmentation was completely inhibited by cotreatment with a caspase/ICE protease inhibitor, Z-VAD-FK (Fig. 3A). Z-VAD-FK blocked the cleavage of CPP32 (caspase 3) caused by anti-Fas treatment (Fig. 3B), indicating that it is an effective caspase inhibitor. In contrast to the inhibition of DNA fragmentation, Z-VAD-FK did not abolish JNK activation induced by PMITC (Fig. 3C) or PEITC (data not shown). The slight enhancement of JNK activation by Z-VAD-FK was not reproducible in repeated experiments, and therefore, was fluctuation of kinase assays. These data indicate the requirement of caspases/ICE-like proteases in isothiocyanate-induced apoptosis; however, JNK activation can occur in the absence of caspase activity. Also, in light of the fact that JNK activation occurred hours before the onset of DNA fragmentation during the isothiocyanate treatments (Fig. 2), JNK activation is unlikely to be a secondary effect of the cellular damage during apoptosis. These data suggest that JNK activation initiates apoptotic signaling.
PEITC Induces JNK Activation and Apoptosis in Various Cell Types-We next examined if the isothiocyanate induces persistent JNK activation and apoptosis in different cell types. We found that different concentrations of PEITC were needed to induce JNK activation in distinct cell types; the effective concentration varied from 5 to 50 M (data not shown). 20 M PEITC induced persistent JNK activation in HeLa and 293 cells (Fig. 4A). At the concentration that induced sustained JNK induction, PEITC also caused apoptosis (Fig. 4B). On examination after Hoechst 33258 staining, the cell nuclei were condensed and fragmented after the PEITC treatment, in comparison with the homogeneous nuclear staining of the untreated cells (Fig. 4B). These data demonstrate that the induction of JNK activation and apoptosis by PEITC can occur in various cell types.
Interfering with the JNK Pathway Suppresses PEITC-induced Apoptosis-Previously, we have shown that the JNK pathway is involved in and required for radiation-induced apoptosis (26). Other investigators also showed that the JNK cascade is required for apoptosis induced by growth factor withdrawal or ceramide treatment (24,27). We then tested if interfering with the JNK pathway had a suppressive effect on isothiocyanate-induced apoptosis. 293 cells were transfected with pCMV-␤gal with or without plasmids encoding a dominant-negative kinase mutant, and each transfection was duplicated for treatments with or without PEITC (20 M). The cells were harvested 24 h after treatment and stained with X-gal to examine the ␤-galactosidase-expressing cells (blue in color). The survival rate of transfected cells after the drug treatment was determined as the percentage of blue cells in the treated group divided by the percentage of blue cells in the untreated group. The dominant-negative mutants of MEKK1 and JNK1 (MEKK1(KR) and JNK1(APF), respectively) blocked PEITCinduced cell death in transfected 293 cells (Fig. 5). In contrast, transfection of wild-type JNK1, a dominant-negative Raf1 mutant (Raf-BXB301), or dominant-negative p38 mitogenactivated protein kinase (p38(AGF)) did not significantly affect apoptosis induced by PEITC. This result indicates that interfering with the JNK pathway prevents the PEITC-induced apoptosis, thereby suggesting that the JNK pathway is required for isothiocyanate-induced apoptosis.
Isothiocyanate-induced JNK Activation Is Inhibited by Antioxidants-JNK activity is regulated by the upstream kinase cascade (MEKK1 3 MKK4 3 JNK, 3); however, it is not clear which upstream signals regulate the JNK module after cells receive apoptotic stimuli. Isothiocyanates react with GSH and form dithiocarbamates (R-NH-C(ϭS)-SG) in the presence or absence of GST (43). Since apoptosis is activated by many oxidative agents (44), the isothiocyanates may induce oxidative stress by reacting with and depleting the intracellular GSH pool, which may then induce JNK activation and apoptosis. We decided to test the roles of oxidation in isothiocyanate-induced JNK activation using antioxidants. JNK activation by PMITC and PEITC was inhibited by preincubation of the Jurkat cells with antioxidants, 2-mercaptoethanol (10 mM, 1 h) or N-acetylcysteine (20 mM, 2 h) (Fig. 6). The antioxidant treatments also blocked ␥-radiation-induced JNK activation, but had no effect on JNK induction by phorbol 12-myristate 13-acetate plus ionomycin or by anisomycin (Fig. 6). The inhibition of JNK activation is not due to the loss of viability of Jurkat cells after antioxidant treatment because Ͼ95% of the cells retained the ability to exclude trypan blue 6 h after antioxidant treatment (data not shown). These data indicate that the JNK activation in cells exposed to PMITC, PEITC, or ␥-radiation may be due to the induction of oxidative stresses. In contrast, the induction of JNK by phorbol 12-myristate 13-acetate plus ionomycin or by anisomycin may not be mediated through the intracellular oxidative changes.
Bcl-2 Suppresses PEITC-induced JNK Activation and Apoptosis-Bcl-2 family members are known to be important apoptosis regulators (45). We used a transient transfection/cell death protection assay to test if Bcl-2 or Bcl-x L can block apoptosis induced by isothiocyanates. The levels of the Bcl-2 (Bcl-x L ) protein were examined by Western blot analysis (Fig.  7A), revealing the production of transfected genes. The empty vector control group lost 40% of cells after the drug treatment. In comparison, transfection of Bcl-2 or Bcl-x L prevented PEITC-induced apoptosis in most of the transfected cells (Fig.  7A), indicating that Bcl-2 and Bcl-x L were capable of protecting cells from PEITC-induced apoptosis. We also examined the PEITC-induced JNK activation in 293 cells transfected with empty vector or Bcl-2-encoding plasmids. Although only 40 -50% of the cells were transfected, the endogenous JNK activity induced by PEITC was evidently decreased in Bcl-2-transfected cells (Fig. 7B). This result implicates Bcl-2 as an upstream suppressor for JNK activation by apoptotic stimuli.
To further examine the molecular ordering between Bcl-2 or Bcl-x L and the JNK pathway, we cotransfected Bcl-2 or Bcl-x L with JNK1 plus constitutively active MEKK1 (⌬MEKK1) into 293 cells. The transfection of JNK1 plus ⌬MEKK1 led to persistent JNK activation and caused apoptosis independent of the upstream signals (Fig. 7C). Neither Bcl-2 nor Bcl-x L significantly affected the cell death caused by JNK1 plus ⌬MEKK1, although the levels of Bcl-2 and Bcl-x L were increased in the transfected cells (Fig. 7C). In comparison to the Bcl-2-or Bcl-x L -transfected cells, the lower levels of Bcl-2 and Bcl-x L detected in the cells cotransfected with JNK1 plus ⌬MEKK may be due to the loss of transfected cells caused by apoptosis (Fig. 7C). This result indicates that Bcl-2 and Bcl-x L failed to prevent cell death induced by JNK activation; therefore, they may not be downstream of JNK in the apoptotic signaling pathway. Taken together, our results suggest that Bcl-2 and Bcl-x L are upstream, but not downstream, of JNK in apoptotic signaling induced by isothiocyanates (Fig. 8). DISCUSSION Activation of the JNK pathway has been shown to be a common phenomenon in apoptotic cell death (24 -27, 46); however, the importance of this activation seems to vary in apoptosis caused by different agents. The JNK pathway is required for apoptosis induction by growth factor withdrawal, heat shock, radiation, and ceramide (24 -27). In contrast, JNK may not be essential for receptor-mediated apoptosis (e.g. Fas-and tumor necrosis factor-mediated apoptosis) (47,48). In this study, we demonstrate the involvement of JNK in isothiocyanate-induced apoptosis by proving that interfering with the JNK pathway suppressed isothiocyanate-induced cell death. We observed that high doses of PMITC, PEITC, and PPITC (Ͼ50 M) caused acute cell death in the absence of JNK activation; however, the cell death resembled necrosis rather than apoptosis. This result indicates that induction of JNK activity is not a general event caused by stress during cell death, but rather it is a specific phenomenon associated with apoptotic cell death. In addition, the failure of a caspase/ICE inhibitor to block JNK activation caused by isothiocyanates indicates that JNK activation may initiate apoptotic signaling and that it is not a secondary effect of cellular damage from apoptotic cell death. However, we (26) and others (48) have shown that JNK activation in Fas-mediated apoptosis can be suppressed by a caspase/ICE inhibitor, which suggests that JNK induction could be augmented by activation of caspases/ICE-like proteases. A recent report indicates that CPP32, a caspase/ICElike protease, is capable of cleaving D4-GDI, a GDP dissociation inhibitor of the Ras-related Rho family GTPase (49). The process of D4-GDI may irreversibly activate the Rho family G proteins, which are activators of the JNK pathway (50,51). Therefore, the activation of CPP32 may lead to the enhancement of JNK activation. If this regulatory mechanism exists, it may work as a signaling circuit to amplify the apoptotic signal, but it may not be essential for Fas-mediated apoptosis.
Our result indicates that oxidative stress may initiate JNK activation (Fig. 6). The anti-apoptotic regulator Bcl-2 may inhibit apoptosis by suppressing the formation or the damaging effects of reactive oxygen species (ROS) (52,53), which are generated by many apoptotic agents (44). In this report, we showed that Bcl-2 suppressed PEITC-induced JNK activation. Bcl-2 or Bcl-x L blocked apoptosis caused by PEITC, but failed to suppress apoptosis caused by overexpression of activated JNK1. In addition, a recent report showed that Bcl-2 blocks JNK activation induced by serum depletion or nerve growth factor withdrawal in PC-12 cells (54). Taken together, Bcl-2 may be an upstream suppressor of the JNK pathway acting to relieve the oxidative stress and to prevent JNK induction and the initiation of apoptotic signals (Fig. 8). The failure of Bcl-2 or Bcl-x L to block cell death induced by JNK1 plus ⌬MEKK1 suggests that Bcl-2 may not be downstream of JNK. However, these data do not exclude the possibility that JNK may directly or indirectly down-regulate the function of Bcl-2, therefore causing cell death. The tumor suppressor p53, which is essen-Bcl-2 and Bcl-x L was determined by Western blot analysis. B, 293 cells were transfected with control vector or Bcl-2-encoding plasmid (3 g), and transfected cells were treated for the indicated times with 20 M PEITC 6 h after removing the transfection mixture. Endogenous JNK activity was examined by immunocomplex kinase assays. C, 293 cells were transfected with pCMV-␤gal (1 g) plus different combinations of plasmids as indicated (JNK1, 1.5 g; ⌬MEKK1, 1.5 g; and Bcl-2 or Bcl-x L , 3 g). Empty vector was added to normalize the total DNA amount. Cells were collected 48 h after transfection, fixed, stained, and examined as described under "Materials and Methods." The cell survival data presented are the means of three experiments. Western blotting for Bcl-2 and Bcl-x L was performed as described under "Materials and Methods." tial for ionizing radiation-induced apoptosis (55), has been shown to be a substrate of JNK (56). p53 is a positive regulator for the expression of Bax (57), a counteracting molecule of Bcl-2 and a potent apoptosis inducer (58). Although the exact physiological function of this p53 phosphorylation is still unclear, p53 may be a mediator for JNK-induced apoptosis.
Isothiocyanates are well known for their chemopreventive effects on various carcinogens (29 -33). Previous studies attributed this anticarcinogenic property to their ability to affect the bioactivation, detoxification, and excretion of carcinogens (34 -37). Here, we show that isothiocyanates induced apoptosis by activation of the JNK pathway. We propose that isothiocyanate-mediated apoptosis may be one possible mechanism to achieve the anticarcinogenic function. Carcinogens usually cause genomic damages in the exposed cells. If the damages are limited, the cells can repair those damages and maintain normal functions. The cells that fail to repair the damages are normally eliminated by apoptosis, which prevents the propagation of genomic damages to progenitor cells. The cells with genomic damages that escape cell death are prone to develop into cancerous cells. In the presence of isothiocyanates, the cells will start the apoptosis process because of an increase in oxidative stress and JNK activity. The simultaneous or subsequent exposure to a carcinogen will trigger the cell death progression, and the damaged cells will be eliminated by apoptosis; therefore, fewer cells can survive and become cancer cells. In addition, cancer cells usually have a higher metabolic rate and generate higher levels of intracellular oxidants than normal cells. The ability of isothiocyanates alone to generate oxidative stress, activate the JNK pathway, and induce apoptosis suggests that they may have a therapeutic function in addition to their chemopreventive functions.