c-Jun NH₂-Terminal Kinase Signaling Axis Regulates Diallyl Trisulfide–Induced Generation of Reactive Oxygen Species and Cell Cycle Arrest in Human Prostate Cancer Cells

Epidemiologic data continue to lend support to the premise that dietary intake of Allium vegetables, including garlic, may be protective against the risk of various types of malignancies, including cancer of the prostate (1–4). For example, the risk of prostate cancer was shown to be significantly lower in men consuming >10 g/d of total Allium vegetables than in men with total Allium vegetable intake of <2.2 g/d in a population-based case-control study (4). Anticarcinogenic effect of Allium vegetables is attributed to organosulfur compounds (OSC), which are released upon processing (e.g., cutting or chewing) of these vegetables (5, 6). Allium vegetable–derived OSCs, including diallyl sulfide, diallyl disulfide (DADS), and/or diallyl trisulfide (DATS), have been shown to inhibit cancer in animal models induced by a variety of carcinogens, including tobacco smoke–derived chemicals (7–11). For instance, the naturally occurring OSC analogues are highly effective in affording protection against benzo(a)pyrene-induced forestomach and pulmonary carcinogenesis in mice (8), N-nitrosomethylbenzylamine-induced esophageal cancer in rats (9), azoxymethane-induced colonic aberrant crypt foci formation in rats (10), and N-methyl-N-nitrosourea-induced mammary carcinogenesis in rats (11). The OSCs are believed to inhibit chemically induced cancers by increasing the expression of phase 2 carcinogen-inactivating enzymes, including glutathione (GSH) transferases and quinone reductase, and/or by inhibiting cytochrome P450–dependent monooxygenases (12–14). We have also shown previously that oral administration of DADS significantly inhibits growth of H-ras oncogene–transformed tumor xenografts in athymic mice without causing weight loss or any other side effects (15).

Evidence is accumulating to indicate that some naturally occurring OSC analogues, including DATS, can inhibit proliferation of cultured cancer cells by causing apoptosis and cell cycle arrest (16–24). An understanding of the mechanism by which OSCs cause apoptosis and cell cycle arrest is critical for their further development as clinically useful anticancer agents because this knowledge could lead to identification of mechanism-based biomarkers potentially useful in future clinical trials. The OSC-mediated apoptosis induction and cell cycle arrest was first documented by Milner (16, 17). The DADS-induced apoptosis in HCT-15 human colon cancer cells correlated positively with an increase in the level of intracellular free calcium (16). Our own work has revealed that the DATS-induced apoptosis in PC-3 and DU145 human prostate cancer cell lines is mediated by down-regulation as well as hyperphosphorylation of Bcl-2 (20). Despite considerable progress towards our understanding of the signaling pathways leading to OSC-mediated apoptosis (reviewed in ref. 21), the mechanism of cell cycle arrest caused by garlic-derived OSCs is not fully understood.

Recently, we investigated the mechanism of DATS-induced cell cycle arrest using DU145 and PC-3 cells as a model (22, 23). We found that the DATS-induced G₂-M phase cell cycle arrest in our model is independent of p21 but linked to reactive oxygen species (ROS)–dependent destruction and Ser²¹⁶ phosphorylation of cell division cycle 25C (Cdc25C) protein. The net result of these effects is accumulation of Tyr¹⁵-phosphorylated (inactive) cyclin-dependent kinase 1 (Cdk1), whose activation is critical for G₂-M transition. However, the mechanism by which DATS treatment causes ROS generation remains elusive. We now provide experimental evidence to indicate that the DATS-induced ROS generation in DU145 and PC-3 cells is caused by an increase in labile iron pool due to degradation of ferritin mediated by c-Jun NH₂-terminal kinase (JNK) signaling axis. The present study is the first published report to implicate JNK signaling axis, which is best known for its role in regulation of cell survival and cell death (25–28), in regulation of ferritin stability, iron homeostasis, and ROS generation.

Reagents. DATS was purchased from LKT Laboratories (St. Paul, MN). Tissue culture media, antibiotic mixture, and fetal bovine serum were from Life Technologies (Grand Island, NY). RNaseA was from Promega (Madison, WI). Propidium iodide and desferrioxamine mesylate were from Sigma (St. Louis, MO). 6-Carboxy-2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) and hydroethidine were from Molecular Probes (Eugene, OR). Combined superoxide dismutase and catalase mimetic EUK134 was a generous gift from Eukarion, Inc. (Bedford, MA). The antibodies against Cdc25C, phospho-(Ser²¹⁶)-Cdc25C, ferritin L, and ferritin H were from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody against total ferritin was from Sigma, and anti-actin antibody was from Oncogene Research Products (Boston, MA). The antibody against SEK1 was from Cell Signaling Technology (Beverly, MA). GSH level was determined using a kit from Cayman Chemical (Ann Arbor, MI). The JNK kinase activity was determined using a kit from Cell Signaling Technology.

Cell lines and cell culture. Monolayer cultures of PC-3 cells were maintained in F-12K Nutrient Mixture (Kaighn's modification) supplemented with 7% non–heat-inactivated fetal bovine serum and antibiotics. The DU145 cells were cultured in DMEM supplemented with 10% fetal bovine serum, 0.1 mmol/L nonessential amino acids, 1 mmol/L sodium pyruvate, and antibiotics. Each cell line was maintained at 37°C in an atmosphere of 95% air and 5% CO₂.

Measurement of ROS. Intracellular ROS generation was measured by flow cytometric monitoring of oxidation of H₂DCFDA, which is cleaved by nonspecific cellular esterases and oxidized in the presence of peroxides, including hydrogen peroxide, to yield fluorescent 2′,7′-dichlorofluorescein (DCF; refs. 29, 30). Briefly, 2 × 10⁵ cells were plated, allowed to attach overnight, and exposed to desired concentration of DATS for specified time period. Stock solution of DATS was prepared in DMSO, and an equal volume of DMSO was added to the controls. Subsequently, the cells were counterstained with 2 μmol/L hydroethidine and 5 μmol/L H₂DCFDA for 30 minutes at 37°C. The cells were collected, and the fluorescence was analyzed using a Coulter Epics XL Flow Cytometer. In some experiments, cells were treated with DATS in the absence or presence of 20 μmol/L EUK134, 25 μmol/L desferrioxamine, or 25 μmol iron-saturated desferrioxamine (2-hour pretreatment) before analysis of ROS generation.

Analysis of cell cycle distribution. The effect of DATS treatment on cell cycle distribution was determined by flow cytometry following staining with propidium iodide. Briefly, DU145 or PC-3 cells (1 × 10⁶) were seeded in T75 flasks and allowed to attach by overnight incubation. The cells were then exposed to desired concentrations of DATS for specified time periods at 37°C. Both floating and adherent cells were collected, washed with PBS, and fixed with 70% ethanol. The fixed cells were treated with RNaseA and propidium iodide, and the stained cells were analyzed using a Coulter Epics XL Flow Cytometer as described by us previously (22, 23, 31). In some experiments, the cells were treated with DATS in the absence or presence of 20 μmol/L EUK134, 25 μmol/L desferrioxamine, 25 μmol/L iron-saturated desferrioxamine, or 1 μmol/L MG132 (2-hour pretreatment) before analysis of cell cycle distribution.

Immunoblotting. The cells were treated with DATS and/or EUK134 or MG132 as described above. Both floating and attached cells were collected; washed with PBS; resuspended in a lysis solution containing 50 mmol/L Tris-HCl (pH 8), 150 mmol/L NaCl, 0.1% SDS, 1% Triton X-100, 10 μg/mL phenanthroline, 10 μg/mL aprotinin, 10 μg/mL leupeptin, 10 μg/mL pepstatin A, 1 mmol/L phenylmethylsulfonyl fluoride; and incubated for 40 minutes on ice with gentle shaking. The cell lysate was cleared by centrifugation at 18,000 × g for 20 minutes. Lysate proteins were resolved by 10% to 12% SDS-PAGE and subjected to immunoblotting as described previously (32, 33). Change in protein level was assessed by densitometric scanning of the bands and corrected for actin loading control.

Determination of GSH level. The effect of DATS treatment on intracellular GSH level was determined using a kit from Cayman Chemical according to the manufacturer's instructions. Briefly, PC-3 or DU145 cells (1 × 10⁶) were plated, allowed to attach, and exposed to DMSO (control) or 40 μmol/L DATS for specified time intervals at 37°C. The cells were collected by scraping, washed with PBS, resuspended in 0.5 mL of PBS, and counted. Equal number of cells from each treatment group was processed for analysis of GSH.

Determination of labile iron. The effect of DATS treatment on level of labile iron was determined by (a) a modified high-pressure liquid chromatography (HPLC) procedure described by Gower et al. (34) and (b) analysis of calcein fluorescence as described by Breuer et al. (35). The HPLC method detects desferrioxamine chelatable iron in cellular extracts. Briefly, cells (6 × 10⁶) were treated with 40 μmol/L DATS for specified time periods, collected by trypsinization, and washed twice with PBS. Cells were lysed by five cycles of freeze-thaw, and cellular extracts were suspended in 1 mL of 50 mmol/L Tris-HCl (pH 7.4) containing 2 mmol/L desferrioxamine. The samples were incubated for 1 hour at 37°C and centrifuged for 15 minutes at 15,000 × g. The desferrioxamine and ferrioxamine in the supernatant fractions were extracted using a Sep-Pak C₁₈ cartridges (Waters, Boston, MA) presoaked with methanol, and conditioned with distilled water. The supernatants were then applied to the Sep-Pak C₁₈ cartridge and washed with water, and desferrioxamine and ferrioxamine were eluted with methanol. The eluted material was dried under vacuum, reconstituted, and subjected to reversed-phase HPLC using a Waters 600E controller equipped with a 717 plus autosampler, 996 photodiode array detector, and Nova-Pak C₁₈ reversed-phase column (150 mm × 3.9 mm; particle size = 4 μm). The mobile phase consisted of 20 μmol/L sodium phosphate buffer (pH 6.6) containing 4 mmol/L EDTA, 1 mol/L ammonium acetate, and 10% (v/v) acetonitrile. Elution was carried out isocratically at a column flow rate of 0.75 mL/min. Eluates were monitored at 240 nm. For determination of calcein fluorescence, cells were seeded at a density of 1 × 10⁵ per well, allowed to attach, and exposed to 40 μmol/L DATS in the absence or presence of 25 μmol/L desferrioxamine or 1 μmol/L MG132 (2-hour pretreatment). Cells were then stained with 5 μmol/L calcein for 30 minutes and rinsed with PBS, and calcein fluorescence was examined using a Leica fluorescence microscope.

Transient transfection. Plasmid expressing catalytically inactive mutant of JNK kinase 2 [JNKK2(AA)] was kindly provided by Dr. Michael Karin (36). DU145 cells were transfected with the plasmid encoding inactive JNKK2(AA) or empty pcDNA3.1 vector at 50% to 60% confluency using FuGENE 6 (Roche, Indianapolis, IN). After 24 hours of transfection, the cells were treated with DATS and processed for different assays. The DU145 cells stably transfected with pSilencer-SEK1 small interfering RNA (siRNA) or empty vector were generously provided by Dr. Yong J. Lee (37).

Immunoprecipitation-immunoblotting to determine ferritin ubiquitination. Cells were treated with DATS for specified time periods, washed twice with ice-cold PBS, and lysed as described above. Aliquots containing 800 μg of lysate protein were incubated with 3 μg of anti-ferritin antibody overnight at 4°C. Protein A/G plus-agarose (30 μL; Santa Cruz Biotechnology) was then added to each sample, and the incubation was continued for an additional 3 hours at 4°C. The immunoprecipitated complexes were washed five times with lysis buffer and subjected to SDS-PAGE followed by immunoblotting using anti-ubiquitin antibody.

ROS dependence of DATS-induced cell cycle arrest. We have shown previously that the DATS-induced G₂-M phase cell cycle arrest in DU145 and PC-3 cells is associated with ROS-dependent destruction and Ser²¹⁶ phosphorylation of Cdc25C (22). In the present study, we used a combined mimetic of superoxide dismutase and catalase (EUK134) to firmly establish the association between DATS-induced ROS production and cell cycle arrest. Figure 1A summarizes data on effect of EUK134 on DATS-mediated oxidation of H₂DCFDA, a measure of ROS production. Treatment of DU145 cells with 40 μmol/L DATS for 4 hours resulted in an ∼2.5-fold increase in DCF fluorescence compared with DMSO-treated control (Fig. 1A). The DATS-mediated oxidation of H₂DCFDA was evident as early as 1 hour after treatment and observed even at lower drug concentrations (data not shown). The DATS-mediated oxidation of H₂DCFDA was nearly completely blocked in the presence of 20 μmol/L EUK134 (Fig. 1A). As can be seen in Fig. 1B, treatment of DU145 cells with 40 μmol/L DATS for 8 hours resulted in an ∼2.3-fold enrichment of G₂-M fraction compared with DMSO-treated control. The DATS-induced G₂-M phase cell cycle arrest was partially but statistically significantly attenuated in the presence of EUK134 (P < 0.05, DATS alone group versus DATS plus EUK134 group by one-way ANOVA followed by Bonferroni's multiple comparison test; Fig. 1B). Moreover, EUK134 offered a marked protection against DATS-induced degradation and Ser²¹⁶ phosphorylation of Cdc25C (Fig. 1C). Collectively, these results indicated that ROS generation was a critical event in DATS-induced G₂-M phase cell cycle arrest.

Iron chelation prevented DATS-induced ROS generation. Although a role for oxidative stress in cellular responses to DADS and DATS has been suggested in other cellular systems (19, 24), the mechanism by which OSCs cause ROS production remains elusive. Initially, we considered the possibility that DATS-induced ROS production might be due to depletion of intracellular GSH, which is the predominant non-protein thiol in cells. We explored this possibility by measuring GSH levels in PC-3 and DU145 cells following treatment with DMSO (control) or 40 μmol/L DATS for 1, 4, or 8 hours. DATS treatment did not affect intracellular GSH level in either DU145 or PC-3 cell line (data not shown). These results indicated that ROS production in DATS-treated DU145/PC-3 cells was not caused by depletion of GSH levels.

Labile iron is considered an important determinant of ROS generation in cells because even trace amounts of labile iron can cause oxidative stress through Fenton/Haber-Weiss reactions (38, 39). Therefore, we tested the possibility whether ROS production in our model was iron-dependent by using an iron chelator desferrioxamine. As can be seen in Fig. 2A, the DATS-induced oxidation of H₂DCFDA in DU145 cells was reduced below control level in the presence of 25 μmol/L desferrioxamine. The DU145 cells treated with desferrioxamine alone also exhibited statistically significant decrease in H₂DCFDA oxidation compared with DMSO-treated control, which is consistent with involvement of iron in ROS production (38, 39). We used iron-saturated desferrioxamine (DFO-Fe) as a control to confirm iron dependence of DATS-induced ROS production. Unlike desferrioxamine, iron-saturated desferrioxamine did not offer protection against DATS-induced ROS generation (Fig. 2A). The protective effect of desferrioxamine, but not iron-saturated desferrioxamine, against DATS-mediated ROS production was also observed in PC-3 cells (Fig. 2B). These results suggested that the DATS-induced ROS generation in DU145 and PC-3 cells might be caused by an increase in labile iron.

Desferrioxamine protected against DATS-induced G₂-M arrest. Next, we determined the effect of desferrioxamine on DATS-induced cell cycle arrest, and the results are summarized in Fig. 3. Exposure of DU145 (Fig. 3A) and PC-3 (Fig. 3B) cells to 40 μmol/L DATS for 8 hours resulted in about 2.6- and 1.9-fold increase, respectively, in percentage of cells with G₂-M phase DNA content compared with DMSO-treated control. The DATS-induced G₂-M phase cell cycle arrest was statistically significantly attenuated (P < 0.05, DATS alone group versus DATS + desferrioxamine group by one-way ANOVA followed by Bonferroni's multiple comparison test) in the presence of desferrioxamine in both cell lines (Fig. 3). Similar to ROS generation, however, the DATS-induced G₂-M block was not affected in the presence of iron-saturated desferrioxamine (Fig. 3). These results indicated that the DATS-induced cell cycle arrest in both cell lines was dependent on availability of labile iron.

DATS treatment increased labile iron by promoting ferritin degradation. Ferritin is an iron storage protein whose degradation leads to liberation of chelatable iron (40, 41). As can be seen in Fig. 4A, treatment of DU145 cells with 40 μmol/L DATS resulted in a time-dependent decrease in the protein levels of light (ferritin L) and heavy chains of ferritin (ferritin H), although the decrease in ferritin L protein level was relatively more pronounced compared with ferritin H. The DATS-mediated decrease in ferritin L protein level in DU145 cells was evident as early as 1 hour after treatment and increased progressively with increasing exposure time (Fig. 4A).

Next, we directly tested the possibility whether DATS-mediated degradation of ferritin caused an increase in labile iron. The level of labile iron was measured by two independent but complementary methods: (a) analysis of labile iron by an HPLC procedure and (b) fluorescence microscopy after staining with iron-sensitive probe calcein, the fluorescence of which is quenched on binding to non–ferritin-bound iron (35). The HPLC assay revealed a modest yet statistically significant increase in the level of labile iron on treatment of DU145 and PC-3 cells with 40 μmol/L DATS (Fig. 4B). Fluorescence microscopy using calcein confirmed an increase in labile iron in DATS-treated cells (Fig. 4C). In DMSO-treated control DU145 cells, the calcein fluorescence was intense and evenly distributed across the cell (Fig. 4C,, a). By contrast, cells treated with DATS (40 μmol/L, 4 hours) exhibited greatly reduced intracellular calcein fluorescence, indicating an increase in non–ferritin-bound labile iron (Fig. 4C,, b). We argued that if quenching of calcein fluorescence in DATS-treated cells (Fig. 4B,, b) was due to release of iron from intracellular stores, then the calcein fluorescence should be dequenched by trapping the released iron with desferrioxamine to prevent the interaction between free iron and the fluorescent probe. The DATS-mediated quenching of calcein fluorescence was indeed abolished upon 2-hour pretreatment with 25 μmol/L desferrioxamine (Fig. 4C,, c), providing further evidence for an increase in labile iron. The desferrioxamine treatment alone did not have any appreciable effect on calcein fluorescence (Fig. 4C , d), which is understandable because an effect on calcein fluorescence is not expected in the absence of labile iron.

Next, we designed experiments to systematically explore the possibility whether DATS-mediated increase in labile iron was associated with proteasomal degradation of ferritin. We addressed this question by using proteasomal inhibitor MG132. As can be seen in Fig. 4C (e), the DATS-mediated quenching of calcein fluorescence was reduced in the presence of MG132, whereas MG132 alone did not have any appreciable effect on calcein fluorescence (Fig. 4C,, f). Consistent with these results, MG132 offered a marked protection against DATS-induced degradation of ferritin (Fig. 4D,, inset) and cell cycle arrest in DU145 cells (Fig. 4D). These results clearly indicated that DATS treatment promoted proteasome-mediated degradation of ferritin to cause release of labile iron in DU145 cells.

DATS-mediated increase in labile iron was regulated by JNK kinase. Recent studies have indicated that activation of JNK, which is best known for its role in regulation of cell survival and apoptosis in response to different stimuli (25–28), can cause ROS generation in some systems (42, 43). For example, JNK activation has been linked to ROS production by tumor necrosis factor (TNF; ref. 42), although the mechanism of this effect remains elusive. Because our previous work showed rapid activation of JNK in DATS-treated PC-3 and DU145 cells (20), we designed experiments to test a hypothesis that DATS-mediated ROS production in our model was regulated by JNK signaling axis. We tested this hypothesis using DU145 cells transiently transfected with a catalytically inactive mutant of JNKK2, which is a JNK-specific upstream kinase (44). As can be seen in Fig. 5A, DATS treatment (40 μmol/L, 30 minutes) caused an ∼2.3-fold increase in the kinase activity of JNK in vector-transfected control DU145 cells as determined by immunoprecipitation kinase assay by monitoring phosphorylation of c-Jun, a down-stream target of JNK. The DATS-mediated activation of JNK was reduced markedly in cells transfected with catalytically inactive JNKK2 (Fig. 5A). Interestingly, DATS-mediated ROS generation was observed in vector-transfected control DU145 cells but not in cells overexpressing catalytically inactive JNKK2 (Fig. 5B). Next, we explored possible involvement of JNKK2 in DATS-mediated degradation of ferritin and increase in labile iron pool. Similar to untransfected DU145 cells (Fig. 4A), DATS treatment caused a marked reduction in protein level of ferritin L (Fig. 5C) in vector-transfected control cells. In addition, DATS treatment caused quenching of calcein fluorescence in vector-transfected DU145 cells (Fig. 5D). DU145 cells overexpressing catalytically inactive JNKK2, where DATS-mediated JNK activation was impaired, were resistant to both DATS-induced degradation of ferritin L (Fig. 5C) and quenching of calcein fluorescence (Fig. 5D). Collectively, these results indicated that the DATS-induced ferritin degradation and increase in labile iron was regulated by JNK signaling axis.

We confirmed involvement of JNK signaling axis in regulation of DATS-induced ROS production by RNA interference of SEK1, which is an upstream kinase in JNK signal transduction pathway (25). As shown in Fig. 6A (inset), the protein level of SEK1 was reduced by >85% in DU145 cells stably transfected with SEK1-specific siRNA (hereafter abbreviated as siSEK1#2 cells) compared with cells transfected with empty vector pSilencer 2.1-U6 hygro (hereafter abbreviated as pSilencer cells). Similar to untransfected DU145 cells (Fig. 1A), DATS treatment caused a time-dependent and statistically significant increase in ROS generation in pSilencer cells compared with DMSO-treated control, whereas ROS generation was not observed in siSEK1#2 cells (Fig. 6A). Consistent with these results, DATS and H₂O₂ (positive control, 25 μmol/L for 30 minutes) treatments resulted in ferritin L degradation in pSilencer cells but not in siSEK1#2 cells (Fig. 6B). We tested the possibility whether DATS treatment caused ubiquitination of ferritin by immunoprecipitation of ferritin followed by immunoblotting using anti-ubiquitin antibody. Polyubiquitin conjugates upon treatment with 40 μmol/L DATS were observed in pSilencer cells but not in siSEK1#2 cells (data not shown). DATS as well as H₂O₂ treatment caused quenching of calcein fluorescence in vector-transfected control pSilencer cells (Fig. 6C). On the other hand, DATS or H₂O₂ treatment failed to significantly quench calcein fluorescence in siSEK1#2 cells (Fig. 6C). Moreover, the siSEK1#2 cells were resistant to DATS-induced G₂-M phase cell cycle arrest (Fig. 6D). On the other hand, DATS treatment caused a significant increase in G₂-M fraction in vector-transfected pSilencer cells compared with DMSO-treated control (Fig. 6D). Collectively, these results indicated that JNK signaling axis regulated DATS-mediated ROS production and cell cycle arrest by controlling ferritin degradation and release of free iron.

We have shown previously that the DATS-induced G₂-M phase cell cycle arrest in PC-3 and DU145 human prostate cancer cells is associated with generation of ROS, which causes rapid degradation and Ser²¹⁶ phosphorylation of Cdc25C, leading to accumulation of inactive Cdk1 (22). In the present study, we confirmed ROS dependence of DATS-induced cell cycle arrest using an antioxidant EUK134, which is a combined mimetic of superoxide dismutase and catalase. We show that EUK134 confers significant protection not only against DATS-mediated ROS production (Fig. 1A) but also the cell cycle arrest (Fig. 1B) and degradation and hyperphosphorylation of Cdc25C (Fig. 1C). These results clearly indicate that ROS generation is a critical event in DATS-mediated cell cycle arrest in our model. A role for ROS in cellular responses to another garlic-derived OSC (DADS) has been suggested previously (19), but the mechanism by which DADS or DATS cause oxidative stress was unclear.

The main goal of the present study was to gain insights into the mechanism by which DATS causes ROS generation in DU145 and PC-3 cells. We show that the DATS-mediated ROS production in our model is not caused by depletion of GSH level (data not shown); rather, it is associated with an increase in labile iron. Iron is necessary for normal cellular proliferation because iron-containing proteins catalyze various key biochemical processes, such as energy metabolism, respiration, folate metabolism, and DNA synthesis (45). Intracellular iron is either labile (chelatable) or stored by ferritin as a nonchelatable pool (38–41). Although labile iron represents only a minor fraction of total cellular iron pool, it plays an important role in cellular metabolism due to its rapid accessibility (39). Even trace amounts of free iron can lead to generation of highly toxic hydroxyl radicals (38, 39). Repression of ferritin expression has been shown to increase the labile iron, leading to oxidative stress in human erythroleukemia cells (47). The present study reveals that the DATS-induced ROS generation and consequently the cell cycle arrest in DU145 and PC-3 cells is caused by an increase in labile iron due to ferritin degradation. This conclusion is based on the following observations: (a) the DATS-mediated ROS generation and G₂-M phase cell cycle arrest in DU145 and PC-3 cells are significantly attenuated in the presence of desferrioxamine (Figs. 2 and 3), which is a chelator of labile iron. (b) The protective effect of desferrioxamine is abolished if it is saturated with iron before addition to the cell culture. (c) The DATS treatment causes a rapid decrease in the protein level of ferritin (Fig. 4A), especially the light chain. (d) The DATS treatment increases the level of labile iron as revealed by HPLC-based quantitation of chelatable iron (Fig. 4B) and quenching of the fluorescence of iron-sensitive probe calcein (Fig. 4C). (e) The DATS-mediated quenching of calcein fluorescence, ferritin degradation, and cell cycle arrest is significantly attenuated in the presence of proteasomal inhibitor MG132. Thus, our data provide support for existence of a signaling mechanism involving ferritin degradation, release of iron, and ROS generation in DATS-induced cell cycle arrest.

The most interesting observation of the present study is that DATS-induced ROS generation in our model is regulated by JNK signaling axis, which is best known for its role in control of cell proliferation and death in response to diverse stimuli, including dietary cancer chemopreventive agents (25–28). Recently, it has been shown that JNK activation by TNF leads to an increase in ROS formation and this effect is significantly reduced in JNK^−/− fibroblasts (42). However, the mechanism of JNK-dependent ROS formation by TNF was not explained. We now show that the DATS-induced ROS generation and cell cycle arrest in our model is linked to JNK activation. This conclusion stems from the observations that DATS-induced ROS production and cell cycle arrest are significantly attenuated by ectopic expression of a catalytically inactive mutant of the upstream kinase (JNKK2) specific for activation of JNK (Fig. 5) as well as RNA interference of SEK1 (Fig. 6). We also found that DATS-mediated degradation of ferritin and release of labile iron are reduced markedly by interference of JNK activation (Figs. 5 and 6). Thus, it is reasonable to conclude that JNK signaling regulates stability of ferritin. A role for JNK in regulation of stability of several other proteins, including c-jun and c-myc, has been documented (48, 49).

In conclusion, the present study reveals that the DATS-induced ROS generation, which is a critical event in cell cycle arrest caused by this cancer chemopreventive agent (22), is mediated by an increase in labile iron due to proteasome-mediated degradation of ferritin. In addition, we provide experimental evidence to indicate involvement of JNK signaling axis in DATS-induced ROS production. To the best of our knowledge, the present study is the first published report to explain the mechanism by which JNK signaling may regulate cellular iron homeostasis and ROS production.

Grant support: National Cancer Institute/USPHS grants CA113363 and CA115498.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Michael Karin (University of California, San Diego, CA) for generous gift of plasmid expressing catalytically inactive mutant of JNKK2 and Dr. Yong J. Lee (University of Pittsburgh, Pittsburgh, PA) for the generous gift of pSilencer and siSEK1#2 cells.