3,3′-Diindolylmethane Is a Novel Mitochondrial H⁺-ATP Synthase Inhibitor that Can Induce p21^Cip1/Waf1 Expression by Induction of Oxidative Stress in Human Breast Cancer Cells

Epidemiologic evidence suggests that high dietary intake of Brassica vegetables, such as broccoli, cabbage, and Brussels sprouts, protects against tumorigenesis in multiple organs. 3,3′-Diindolylmethane, one of the active products derived from Brassica vegetables, is a promising antitumor agent. Previous studies in our laboratory showed that 3,3′-diindolylmethane induced a G₁ cell cycle arrest in human breast cancer MCF-7 cells by a mechanism that included increased expression of p21. In the present study, the upstream events leading to p21 overexpression were further investigated. We show for the first time that 3,3′-diindolylmethane is a strong mitochondrial H⁺-ATPase inhibitor (IC₅₀∼20 μmol/L). 3,3′-Diindolylmethane treatment induced hyperpolarization of mitochondrial inner membrane, decreased cellular ATP level, and significantly stimulated mitochondrial reactive oxygen species (ROS) production. ROS production, in turn, led to the activation of stress-activated pathways involving p38 and c-Jun NH₂-terminal kinase. Using specific kinase inhibitors (SB203580 and SP600125), we showed the central role of p38 and c-Jun NH₂-terminal kinase (JNK) pathways in 3,3′-diindolylmethane-induced p21 mRNA transcription. In addition, antioxidants significantly attenuated 3,3′-diindolylmethane-induced activation of p38 and JNK and induction of p21, indicating that oxidative stress is the major trigger of these events. To further support the role of ROS in 3,3′-diindolylmethane-induced p21 overexpression, we showed that 3,3′-diindolylmethane failed to induce p21 overexpression in mitochondrial respiratory chain deficient ρ⁰ MCF-7 cells, in which 3,3′-diindolylmethane did not stimulate ROS production. Thus, we have established the critical role of enhanced mitochondrial ROS release in 3,3′-diindolylmethane-induced p21 up-regulation in human breast cancer cells. (Cancer Res 2006; 66(9): 4880-7)

3,3′-Diindolylmethane is a promising antitumor agent derived from Brassica food plants, including cabbage, broccoli, and Brussels, and other common dietary vegetables (1, 2). Several studies have shown that 3,3′-diindolylmethane can inhibit the development of both carcinogen-induced and implanted tumors in rodents (3, 4). In addition, 3,3′-diindolylmethane and its precursor indole-3-carbinol (I3C) are considered to be the preferred treatment for recurrent respiratory papillomatosis in humans (5).

Evidence is accumulating that 3,3′-diindolylmethane inhibits proliferation of cancer cells by inducing cell cycle arrest and apoptosis (4, 6–9). In previous studies of the cytostatic mode of action of 3,3′-diindolylmethane, we showed that this indole could cause a G₁ cell cycle arrest regardless of estrogen receptor or p53 tumor suppressor status of breast tumor cells (8). Analysis of G₁-acting cell cycle components indicated that the enzymatic activity of cyclin-dependent kinase 2 (CDK2) was strongly reduced by 3,3′-diindolylmethane by a mechanism that involved a large increase in expression of the CDK inhibitor, p21^cip1/waf1 (p21). p21^Cip1/Waf1 is a well-characterized Cip/Kip family CDK inhibitor, which can induce G₁ arrest and block entry into S phase by inactivating CDKs or by inhibiting activity of proliferating cell nuclear antigen (10). Further promoter deletion analysis of the p21 promoter showed that 3,3′-diindolylmethane responsiveness was partially dependent on the binding of the Sp1 and Sp3 transcription factors to the GC-rich region of the proximal promoter.

Our previous studies also showed that 3,3′-diindolylmethane strongly induced the stress-activated protein kinase pathways, c-Jun NH₂-terminal kinase (JNK) and p38, in human breast cancer MCF-7 cells, which contribute to 3,3′-diindolylmethane-induced IFN-γ expression (11). JNK and p38 kinases are important in controlling cell growth and apoptosis in response to chemical stress, radiation, and growth factors (12). The involvement of JNK and p38 activation in 3,3′-diindolylmethane-induced p21 expression in MCF-7 cells has not been reported.

Results of our most recent studies have shown that high concentrations of 3,3′-diindolylmethane (≥75 μmol/L) can induce a G₂ cell cycle arrest in human hepatoma cells by a mechanism that involves induction of cellular oxidative stress and activation of DNA damage checkpoint signaling pathways.³

Because oxidative stress has been shown to activate p21 expression in other systems (13, 14), we sought to determine whether reactive oxygen species (ROS) might contribute to the mechanism of increased p21 expression by 3,3′-diindolylmethane in breast tumor cells. Here, we show that 3,3′-diindolylmethane could strongly enhance mitochondrial ROS release by inhibition of mitochondrial H⁺-ATPase, leading to activation of the p38 and JNK stress pathways and up-regulation of cell cycle inhibitor p21 expression.

Chemicals. All the cell culture reagents except fetal bovine serum (Omega Scientific, Inc., Tarzana, CA) were from Life Technologies/Invitrogen (Carlsbad, CA). 3,3′-Diindolylmethane was purchased from LKT Laboratory, Inc. (St. Paul, MN). RNase, propidium iodide, α-tocopherol, ascorbic acid, antimycin A, oligomycin, digitonin, ATP, NADH, phosphoenolpyruvate (PEP), pyruvate kinase (PK), lactate dehydrogenase (LDH), succinate, phenylmethylsulfonyl fluoride (PMSF), DTT, leupeptin, aprotinin, pepstatin, β-glycerophosphate, sodium fluoride, sodium orthovanadate, and glutathione assay kit were from Sigma (St. Louis, MO). Rhodamine 123, CM-H₂DCFDA, and tetramethylrhodamine, methyl ester, perchlorate (TMRM) were purchased from Molecular Probes (Eugene, OR). SB203580 and SP600125 were from Calbiochem (San Diego, CA). ENLITEN rLuciferase/Luciferin Reagent was a product of Promega (Madison, WI). All other reagents were of the highest grade available.

The antibody against p21 was from Oncogene Research Products (San Diego, CA). The antibody against β-tubulin was from Santa Cruz Biotechnology (Santa Cruz, CA). The antibodies against p-p38, p-INK, p-retinoblastoma (pRb), p38, JNK, and Rb were all purchased from Cell Signal Technology (Beverly, MA).

Cell culture and generation of mitochondrial DNA-deficient cells. The human mammary carcinoma MCF-7 and MDA-MB-231 cell lines were obtained from the American Type Culture Collection (Manassas, VA). Cells were cultured in 10-cm Petri dishes in DMEM supplemented with 10% fetal bovine serum, 100 units/mL penicillin, and 100 units/mL streptomycin in a humidified incubator with 5% CO₂ at 37°C. MCF-7 cells lacking mitochondrial DNA (mtDNA; ρ⁰) were generated as described previously by growing MCF-7 cells in DMEM supplemented with 10% fetal bovine serum, 1 mmol/L pyruvate, 50 μg/mL uridine, and 50 ng/mL ethidium bromide for 8 weeks (15). The absence of mtDNA was confirmed by PCR analysis using the following specific mtDNA primers (16): mtDNA-1, 5′-CCTAGGGATAACAGCGCAAT-3′; mtDNA-2, 5′-TAGAAGAGCGATGGTGAGAG-3′.

Western blot. Cells were scraped from 10-cm Petri dishes, washed twice with PBS, and then suspended in 1 mL of radioimmunoprecipitation assay buffer containing 50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 1% NP40, 0.1% SDS, 1 mmol/L NaF, 100 μg/mL PMSF, 1 mmol/L DTT, 2 μg/mL leupeptin, 2 μg/mL aprotinin, 1 μg/mL pepstatin, 10 mmol/L β-glycerophosphate, and 1 mmol/L sodium orthovanadate on ice for 30 minutes. After centrifugation at 13,200 rpm for 15 minutes at 4°C, the supernatants were collected, and the proteins were separated on either 12% or 15% SDS-PAGE. After electrophoresis, proteins were transferred to an Immobilon-P membrane (Millipore, Bedford, MA). The membrane was blocked with 5% nonfat milk or bovine serum albumin in TBST and incubated overnight with the corresponding primary antibodies at 4°C. After washing thrice with TBST, the membrane was incubated at room temperature for 1 hour with horseradish peroxidase–conjugated secondary antibody diluted with TBST (1:1,000). The detected protein signals were visualized by an enhanced chemiluminescence reaction system (Amersham, Arlington Heights, IL).

Reverse transcription-PCR. Total RNA was extracted using TRIzol reagent as recommended by the manufacturer (Invitrogen). cDNA was prepared from 5 μg of total RNA by reverse transcription in a volume of 50 μL with superscript reverse transcriptase (Invitrogen). The resultant cDNA was subjected to PCR (94°C for 45 seconds, 57°C for 1 minute, and 72°C for 1.5 minutes) using specific primers synthesized by Genosys (Woodlands, TX) in a Robocycler Gradient 96 (Stratagene, La Jolla, CA). PCR fragments were analyzed on 1% agarose/TAE gels. Primer sequences for PCR amplification are as follows: p21-1, 5′-CAGGTCCACATGGTCTTCCT-3′; p21-2, 5′-CTGCCCAAGCTCTACCTTCC-3′; glyceraldehyde-3-phosphate dehydrogenase (GAPDH)-1, 5′-ACCACAGTCCATGCCATCAC-3′; GAPDH-2, 5′-TCCACCACCCTGTTGCTGTA-3′.

Measurement of mitochondrial membrane potential. Cells were seeded into six-well culture plates at a density of 5 × 10⁵/mL, and 24 hours later, they were treated with different concentrations of 3,3′-diindolylmethane or 5 μmol/L carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; proton gradient uncoupler) for 1 hour. During the last 30 minutes of treatment, either rhodamine 123 (100 nmol/L) or TMRM (40 nmol/L) was added to the medium to measure the mitochondrial membrane potential (MMP). Cells were collected by gentle trypsinization and analyzed on a Beckman-Coulter EPICS XL flow cytometer, and data were processed by WinMDI 2.8 software.

Measurement of ROS production in whole cells and in isolated mitochondria. Intracellular ROS were detected with the cell permeable fluorescent probe CM-H₂DCFDA. MCF-7 cells were seeded into six-well culture plates at a density of 5 × 10⁵/mL, and 24 hours later, they were treated with different concentrations of 3,3′-diindolylmethane for 1 hour in the presence or the absence of antioxidants. During the last 30 minutes of treatment, cells were loaded with 1 μg/mL CM-H₂DCFDA for 30 minutes in the dark at 37°C. Cells were then trypsinized and analyzed on a Beckman-Coulter EPICS XL flow cytometer. Data were processed by WinMDI 2.8 software.

Mitochondria were isolated from cultured MCF-7 cells by differential centrifugation according to the Pierce mitochondria isolation kit manual (Rockford, IL). To analyze the mitochondrial ROS production, isolated mitochondria were suspended in 500 μL analysis buffer [250 mmol/L sucrose, 20 mmol/L 3-(N-morpholino) butane sulfonic acid, 10 mmol/L Tris-base, 100 μmol/L KH₂PO₄, 0.5 mmol/L Mg²⁺, 1 μmol/L cyclosporin A (pH 7.0)] containing 5 mmol/L succinate and 1 μg/mL CM-H₂DCFDA (17). In some samples, antimycin A (an inhibitor of complex III; 2.5 μg/mL) was added to enhance ROS production. Mitochondria suspension were incubated at room temperature, protected from light for 30 minutes and then analyzed by Beckman-Coulter EPICS XL flow cytometer. Data were processed with WinMDI 2.8 software.

Measurement of cellular ATP and GSH levels. For ATP content analysis, cells were seeded in six-well plates at a density of 5 × 10⁵/mL and treated with different concentrations of drugs for the indicated times. Cells were washed with cold PBS twice and lysed in a solution containing 0.5% trichloroacetic acid and 4 mmol/L EDTA. The supernatants obtained following a 5-minute spin at 13,000 rpm were diluted 50× with distilled water for bioluminescent ATP content measurement using Promega ENLITEN rLuciferase/Luciferin reagent as instructed by the manufacturer.

For GSH content analysis, cells were seeded in a 10-cm Petri dish at a density of 5 × 10⁶/mL and treated with different concentrations of drugs for the indicated times. The intracellular GSH levels were measured by GSH kit from Sigma.

PK/LDH–coupled ATPase assay. Fresh isolated mitochondria were solubilized in 1% digitonin solubilizing buffer containing 50 mmol/L Tris-HCl (pH 7.4), 120 mmol/L NaCl, 5 mmol/L KCl, 1 mmol/L MgSO₄, 1 mmol/L CaCl₂, and 10% glycerol for 30 minutes at 4°C. The ATPase activity of solubilized mitochondria was measured spectrophotometrically at 340 nm by coupling the production of ADP to the oxidation of NADH via pyruvate kinase and lactate dehydrogenase as described previously (18). The reaction mixture contained a final volume of 400 μL at 25°C and included the following: Tris-HCl (pH 8), 2 mmol/L ATP, 2 mmol/L MgCl₂, 50 mmol/L KCl, 0.2 mmol/L EDTA, 0.2 mmol/L NADH, 1 mmol/L PEP, 5 units pyruvate kinase, 8 units lactate dehydrogenase, and 25 to 50 μg mitochondrial proteins. These assay conditions minimized the contribution of other transport ATPases, such as Na⁺,K⁺-ATPase. Test compounds were added before the reaction was started by addition of 25 to 50 μg of solubilized mitochondria at a constant temperature of 25°C.

Statistical analysis. The statistical differences between groups were determined using Student's t test. The levels of significance are noted for P < 0.01 and P < 0.05. The results are expressed as means ± SD for at least three replicate determinations for each assay.

3,3′-Diindolylmethane increased MMP in both MCF-7 and MDA-MB-231 cells. We reported previously that 3,3′-diindolylmethane induced apoptosis in human breast cancer cells (8). Because depolarization of the mitochondrial membrane is established to be part of the mitochondria-mediated apoptotic pathway (19), we examined MMP in 3,3′-diindolylmethane-treated MCF-7 cells using rhodamine 123 and TMRM dyes. These cell-permeable fluorescent probes are actively sequestered by mitochondria according to the membrane potential without causing cytotoxic effects. Surprisingly, the flow cytometry results for both probes presented in Fig. 1A show that 3,3′-diindolylmethane concentration-dependently increased MMP 1 hour after treatment in MCF-7 cells. 3,3′-Diindolylmethane (50 μmol/L) almost doubled the fluorescence intensity, indicating a hyperpolarization of the mitochondrial membrane. FCCP, a mitochondrial proton gradient dissipater, showed the opposite effect by decreasing MMP. Similar effects were observed in MDA-MB-231 breast cancer cells after 3,3′-diindolylmethane or FCCP treatment (data not shown). Results presented in Fig. 1B show that after 24 hours of 3,3′-diindolylmethane treatment, most cells displayed depolarized MMP. Thus, 3,3′-diindolylmethane induced a biphasic change in MMP that involved a concentration-dependent increase within the first hour of treatment followed by a reduction after 24 hours of treatment.

3,3′-Diindolylmethane inhibited mitochondrial H⁺-ATPase. MMP results from a proton gradient established during the electron transport, which is used by mitochondrial H⁺-ATP synthase to produce ATP. Inhibition of the mitochondrial H⁺-ATP synthase by drugs, such as oligomycin, often results in an increase in MMP (20, 21) Thus, we examined whether 3,3′-diindolylmethane could inhibit mitochondrial ATP synthase.

Mitochondrial H⁺-ATP synthase (H⁺-ATPase) can catalyze both ATP synthesis and ATP hydrolysis. When there is a proton gradient, the enzyme catalyzes the forward reaction (ATP synthesis), and when there is no gradient, it displays H⁺-ATPase activity (22). The effect of 3,3′-diindolylmethane on mitochondrial H⁺-ATPase activity was determined using 1% digitonin-solubilized, freshly isolated mitochondria from both MCF-7 cells and mouse liver. Data presented in Fig. 2A show that 3,3′-diindolylmethane concentration-dependently inhibited H⁺-ATPase activity in mitochondria extracted from MCF-7 cells, with IC₅₀ of ∼20 μmol/L. Similar levels of inhibition were seen in mitochondria extracted from mouse liver (data not shown). I3C, the precursor of 3,3′-diindolylmethane in plants, showed no effect on enzyme activity. The extent of inhibition by 3,3′-diindolylmethane (50 μmol/L) was comparable with the effect of the well-known H⁺-ATPase inhibitor, oligomycin (10 μmol/L). Ouabain, a selective Na⁺,K⁺-ATPase inhibitor, produced no significant effect on the total H⁺-ATPase activity, indicating Na⁺,K⁺-ATPase activity did not significantly interfere with our reaction.

The inhibitory effects of 3,3′-diindolylmethane on mouse liver mitochondrial H⁺-ATPase activity are summarized in the double reciprocal plot shown in Fig. 2B and indicate that 3,3′-diindolylmethane is a noncompetitive inhibitor of mitochondrial H⁺-ATPase.

3,3′-Diindolylmethane decreased ATP levels in MCF-7 cells. 3,3′-Diindolylmethane-mediated mitochondrial H⁺-ATPase inhibition was further confirmed by determination of a change in cellular ATP level. The results presented in Fig. 2C show that 3,3′-diindolylmethane (50 μmol/L) time-dependently decreased cellular ATP level by 15 minutes after treatment. By 2 hours, the level of intracellular ATP had dropped by 36%. A greater and more rapid decrease to a maximum of 55% inhibition was observed in the positive control oligomycin-treated cells. 2-Deoxy-d-glucose (5 mmol/L), a glycolytic pathway competitive inhibitor (23), decreased ATP concentration by 30% after a 2-hour treatment. Because the glycolytic pathway and mitochondrial oxidative phosphorylation are the two major cellular ATP-generating pathways, we also measured the effects of cotreatment with 3,3′-diindolylmethane and 2-deoxy-d-glucose on cellular ATP levels. The results indicated that the combination produced a roughly additive decrease in cellular ATP level, especially at the early time points.

3,3′-Diindolylmethane induced mitochondrial ROS production. Release of ROS in the form of superoxide anion occurs normally during oxidative phosphorylation to the extent of about 3% to 5% of total oxygen consumed (24). This rate of ROS production can be increased greatly, however, under certain conditions when oxidative phosphorylation is inhibited by drugs (20, 25–27). To determine whether 3,3′-diindolylmethane functions in a similar manner, we examined the effect of 3,3′-diindolylmethane on cellular ROS levels using the ROS-sensitive dye CM-H₂DCFDA. Our results indicate that 3,3′-diindolylmethane (50 μmol/L) clearly increased the level of cellular ROS, as did hydrogen peroxide and the other oxidative phosphorylation inhibitors oligomycin and antimycin A (data not shown). The ROS production occurred within 5 minutes of 3,3′-diindolylmethane treatment, and their levels were reduced on coincubation with antioxidants (shown in Fig. 3A). The results presented in Fig. 3B indicate that the induction of ROS occurred in a clear 3,3′-diindolylmethane concentration-dependent manner with 30 μmol/L, the apparent threshold for the activity in the cells. To confirm that the 3,3′-diindolylmethane-mediated ROS release produces an oxidative stress on the cells, Fig. 3C illustrates the effects of 3,3′-diindolylmethane on intracellular glutathione levels. The results indicate that the positive control, hydrogen peroxide, and 3,3′-diindolylmethane treatment caused similar 40% and 50% decreases, respectively, in total glutathione level.

To further verify the role of mitochondria in the 3,3′-diindolylmethane-induced oxidative burst, we directly treated mitochondria freshly isolated from MCF-7 cells with different concentrations of 3,3′-diindolylmethane and examined the ROS production over time. The results presented in Fig. 3D indicate that antimycin A (10 μmol/L), a mitochondrial complex III inhibitor, significantly stimulated mitochondrial ROS production, as did 3,3′-diindolylmethane in a concentration-dependent manner. Taken together, these results show that 3,3′-diindolylmethane treatment caused oxidative stress in cells through stimulating ROS release from mitochondria.

3,3′-Diindolylmethane induced p21 upregulation through oxidative stress. We reported previously that 3,3′-diindolylmethane induced a G₁ cell cycle arrest in MCF-7 cells that was accompanied by a strong increase in the expression of the cell cycle inhibitor p21 (8). Here, we examined the role of oxidative stress in 3,3′-diindolylmethane-induced p21 expression. Control experiments confirmed that hydrogen peroxide treatment increased p21 expression in MCF-7 cells (data not shown). Results of Western blot analysis showed that 3,3′-diindolylmethane concentration-dependently increased p21 expression in MCF-7 cells, which could be abolished by antioxidants (Fig. 4A and B).

To further test our hypothesis, we examined the effects of 3,3′-diindolylmethane on ROS releases and p21 expression in ρ⁰ MCF-7 cells, which are mtDNA-deficient cells and are not capable of carrying out oxidative phosphorylation (15). Reverse transcription-PCR (RT-PCR) analysis showed that mtDNA content was decreased by >95% in the ρ⁰ MCF-7 cells (data not shown). Results presented in Fig. 4C and D show that whereas 3,3′-diindolylmethane (50 μmol/L) treatment produced ∼4-fold increase in ROS-associated fluorescence in wild-type MCF-7 cells, this indole did not significantly elevate ROS levels or increase p21 expression in ρ⁰ MCF-7 cells. Taken together, these results support the hypothesis that 3,3′-diindolylmethane can induce p21 expression in MCF-7 cells by enhancing ROS release from mitochondria.

3,3′-Diindolylmethane-induced ROS release resulted in pRb inactivation by activation of JNK and p38 stress-activated pathways. We next examined the effect of 3,3′-diindolylmethane-induced oxidative stress on upstream and downstream signaling events associated with p21 expression and cell cycle arrest. We found in a recent study that high concentrations of 3,3′-diindolylmethane (≥75 μmol/L) could induce oxidative DNA damage and activate the DNA checkpoint 2 pathway in human hepatoma cells, thus leading to G₂ cell cycle arrest. In this study, lower concentrations of 3,3′-diindolylmethane (≤50 μmol/L) were used to avoid DNA damage–induced signaling pathways. JNK and p38 signaling pathways are well established as mediators of stress-induced cell cycle arrest and apoptosis (28), and both pathways were activated by 3,3′-diindolylmethane in our previous study (11, 29–31). Here, we investigated the involvement of these two pathways in 3,3′-diindolylmethane-induced p21 expression. Results presented in Fig. 5 show that these stress pathways were activated by 3,3′-diindolylmethane by a mechanism that was blocked by antioxidants and which resulted in p21 expression and Rb hypophosphorylation. Thus, results of Western blot analysis presented in Fig. 5A show that 3,3′-diindolylmethane treatment produced a rapid and persistent increase in phosphorylation of both JNK (p54 and p46 isoforms) and p38 kinases. The results presented in Fig. 5B indicate that the 3,3′-diindolylmethane-induced phosphorylation of JNK and p38 were strongly reduced by cotreatment with antioxidants. Furthermore, inhibitors of JNK kinase (SP600125) and of p38 kinase (SB203580), which, respectively, prevented JNK and p38 phosphorylation (see Fig. 5C), effectively inhibited the 3,3′-diindolylmethane-induced increase in p21 mRNA level, as shown by the results of RT-PCR analysis presented in Fig. 5D. Results of Western blot analysis presented in Fig. 5E show that 3,3′-diindolylmethane strongly induced Rb dephosphorylation and thus activated this tumor suppressor. The specific JNK and p38 kinase inhibitors blocked both the inducing effect of this indole on p21 expression and partially reversed the inhibitory effect of 3,3′-diindolylmethane on Rb phosphorylation. Taken together, these results confirm that 3,3′-diindolylmethane can increase p21 expression by a mechanism that involves ROS-mediated activation of JNK and p38 kinases and results in hypophosphorylation and activation of Rb protein.

We have previously published that 3,3′-diindolylmethane could induce a G₁ cell cycle arrest in human breast cancer MCF-7 cells that was associated with strong induction of cell cycle inhibitor, p21 (8). As summarized in Fig. 6, we now further report that 3,3′-diindolylmethane is a potent mitochondrial H⁺-ATPase inhibitor (IC₅₀∼20 μmol/L). 3,3′-Diindolylmethane not only induced mitochondrial membrane hyperpolarizaton and reduced intracellular [ATP], but also enhanced mitochondrial ROS release, which led to the activation of stress-sensitive JNK and p38 pathways and increased expression of p21. These effects were not produced by 3,3′-diindolylmethane in mitochondrial respiratory chain deficient ρ⁰ MCF-7 cells, thus indicating the essential role of mitochondria in the induction of p21 expression by 3,3′-diindolylmethane in human breast tumor cells. We have shown previously that the activation of p38 and JNK pathways was also responsible for 3,3′-diindolylmethane-induced IFN-γ secretion in MCF-7 cells (11). Now by showing that antioxidants prevented the activation of these two pathways, we further conclude that 3,3′-diindolylmethane-enhanced mitochondrial ROS release is the upstream signaling event leading to the induced IFN-γ expression.

Results of our double-reciprocal analysis establish that 3,3′-diindolylmethane is a direct, noncompetitive inhibitor of mitochondrial H⁺-ATPase. This activity is particularly promising for the development of 3,3′-diindolylmethane as a therapeutic agent because a family of H⁺-ATPase inhibitors, including oligomycin, was shown recently to be among the most cell line selective cytostatic agents of 37,000 substances tested against the 60 human cancer cell lines of the National Cancer Institute (32). Our previous studies revealed that both cell cycle arrest and apoptosis contributed to 3,3′-diindolylmethane-induced cytostatic effects in MCF-7 cells (8, 9). The identification of 3,3′-diindolylmethane as a mitochondrial H⁺-ATPase inhibitor may explain the cytostatic effects of 3,3′-diindolylmethane.

We have shown that 2 hours of 3,3′-diindolylmethane treatment depleted the cellular ATP levels by one third. ATP is the major readily useable form of energy involved in many crucial cellular processes. It is known that if cells fail to coordinate energy-supply and energy-consumption, they either arrest or initiate apoptosis to avoid possible replication errors (33, 34). Inhibitors of oxidative phosphorylation, which decrease the intracellular ATP concentration, induce cell cycle arrest and apoptosis in various types of tumor cells (34–36). Thus, by lowering intracellular ATP level, 3,3′-diindolylmethane imposes cytostatic restraints in fast-growing tumor cells.

We have shown further that enhanced mitochondrial ROS release by 3,3′-diindolylmethane leads to increased expression of the important cell cycle inhibitor p21, contributing to 3,3′-diindolylmethane-induced G₁-S arrest. ROS are well-known regulators of cell cycle progression and apoptosis (24, 37–39). We observed that depending on the concentration, 3,3′-diindolylmethane could rapidly induce ROS production in all the cell lines that we examined, including human breast cancer cells, prostate cancer cells, hepatoma cells, leukemia cells, and primary endothelial cells (data not shown). Induction of p21 through a p53-independent pathway is quite common in ROS-induced cell cycle arrest (14, 40–42). These p53-independent pathways activate the binding of the Sp1 family of transcription factors to the six Sp1-binding sites located in the proximal promoter of p21. Mitogen-activated protein kinase pathway components, especially c-jun, which are subject to redox control (29–31), can transactivate the promoter of human p21 by acting as a superactivator of Sp1 (43). Thus, our observations that 3,3′-diindolylmethane could produce ROS-dependent and p38 and JNK-mediated induction of p21 are consistent with our previous findings that 3,3′-diindolylmethane induced p21 expression through a p53-independent and Sp1-dependent pathway (8).

Both of the above two mechanisms (ATP depletion and ROS production) seem involved in 3,3′-diindolylmethane-induced cytostatic effects. Antioxidants could block ROS generation and induction of p21, but they could not completely reverse 3,3′-diindolylmethane-induced G₁ cell cycle arrest or the 3,3′-diindolylmethane-induced inhibition of DNA synthesis (data not shown). This partial effect of antioxidants on reversing the antiproliferative activities of 3,3′-diindolylmethane suggests that both of the above mechanisms are important for 3,3′-diindolylmethane-induced cytostatic effects.

3,3′-Diindolylmethane induced a biphasic change in MMP that involved an increase within the first hour of treatment followed by a reduction after 24 hours of treatment (see Fig. 1). Because the alterations in mitochondrial structure and function play key roles in the regulation of apoptosis, both mitochondrial depolarization and hyperpolarization have been recognized as early events in apoptosis (44, 45). Thus, 3,3′-diindolylmethane-induced alterations in mitochondrial membrane integrity could contribute to 3,3′-diindolylmethane-induced apoptosis. However, the direct involvement of alteration of mitochondrial membrane integrity in p21 expression and cell cycle arrest has not yet been defined.

The bulk of our current studies were done at a 50 μmol/L concentration of 3,3′-diindolylmethane, at which p21 was highly induced; however, only minimal (up to 15%) apoptosis was seen (9). It is noteworthy that the roles of p21 induction in cell cycle arrest and apoptosis are different. Whereas p21 plays a crucial rule in inducing cell cycle arrest, p21 may act as a major inhibitor of both p53-dependent or p53-independent apoptosis (46). It is possible that the inhibitory effect of p21 on apoptosis might account for the low amount of apoptosis observed at this concentration of 3,3′-diindolylmethane.

Our data indicate that 3,3′-diindolylmethane is a potent mitochondrial H⁺-ATPase inhibitor that can decrease cellular ATP levels, induce mitochondrial ROS release, activate stress-responsive p38 and JNK pathways, and up-regulate p21 expression in human breast cancer cells. Our study identified a further important mode of action for this intriguing dietary component by showing the distinct link between mitochondrial activity regulation and cell cycle regulation by 3,3′-diindolylmethane that might be exploited for the therapeutic development.

Grant support: Department of Defense, Army Breast Cancer Research Program grant DAMDI7-96-1-6149 and NIH grants CA69056 and CA102360.

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.