The current options for treating breast cancer are limited to excision surgery, general chemotherapy, radiation therapy, and, in a minority of breast cancers that rely on estrogen for their growth, antiestrogen therapy. The naturally occurring chemical indole-3-carbinol (I3C), found in vegetables of the Brassica genus, is a promising anticancer agent that we have shown previously to induce a G1 cell cycle arrest of human breast cancer cell lines, independent of estrogen receptor signaling. Combinations of I3C and the antiestrogen tamoxifen cooperate to inhibit the growth of the estrogen-dependent human MCF-7 breast cancer cell line more effectively than either agent alone. This more stringent growth arrest was demonstrated by a decrease in adherent and anchorage-independent growth, reduced DNA synthesis, and a shift into the G1 phase of the cell cycle. A combination of I3C and tamoxifen also caused a more pronounced decrease in cyclin-dependent kinase (CDK) 2-specific enzymatic activity than either compound alone but had no effect on CDK2 protein expression. Importantly, treatment with I3C and tamoxifen ablated expression of the phosphorylated retinoblastoma protein (Rb), an endogenous substrate for the G1 CDKs, whereas either agent alone only partially inhibited endogenous Rb phosphorylation. Several lines of evidence suggest that I3C works through a mechanism distinct from tamoxifen. I3C failed to compete with estrogen for estrogen receptor binding, and it specifically down-regulated the expression of CDK6. These results demonstrate that I3C and tamoxifen work through different signal transduction pathways to suppress the growth of human breast cancer cells and may, therefore, represent a potential combinatorial therapy for estrogen-responsive breast cancer.
I3C 3 is a naturally occurring compound found in vegetables of the Brassica genus, such as broccoli and Brussels sprouts (1 , 2) , and it has been shown to reduce the incidence of spontaneous and carcinogen-induced mammary tumors in rodents (3 , 4) . When ingested, the stomach acid catalyzes the conversion of I3C into a number of derivatives. Two of the most prevalent derivatives that have been identified are 3,3′-diindolymethane and indolo[3,2-b]carbazole (5) . These compounds are thought to be responsible for the long-term antiestrogenic biological activities of ingested I3C that may contribute to protection against mammary tumor formation. We have documented that the treatment of MCF-7 human breast cancer cells with I3C but not with its acid-catalyzed derivatives rapidly suppresses cell growth. This reversible G1 cell cycle arrest is preceded by a decrease in the expression of CDK6 and is independent of estrogen receptor signaling (6) . However, many of the molecular details of this antiproliferative pathway are unknown and an understanding of the mechanism of I3C action in human breast cancer cells could potentially lead to a novel approach to control breast cancer. Currently, the only specific therapy for breast cancer is antiestrogen treatment, and this treatment is only effective on tumors that rely on estrogen for their growth. Because I3C has been shown to suppress the growth of both estrogen-dependent and estrogen-independent human breast cancer cell lines (6) , a combination of these two growth suppressors may potentially provide beneficial effects for breast cancer patients.
Tamoxifen has been a clinically useful antiestrogen for breast cancer patients for >20 years (7, 8, 9) . Tamoxifen, a nonsteroidal triphenyl ethylene, can adopt a structural conformation that resembles steroids in the nucleus and act as a competitive inhibitor of E2 binding to the estrogen receptor (10 , 11) . Although tamoxifen generally acts as an estrogen receptor antagonist in breast cancer cells, in certain other cell types, tamoxifen can act as an estrogen receptor agonist (reviewed in Ref. 12 ). Several mechanisms are proposed for the modulation of breast cancer cell proliferation by tamoxifen, including down-regulation of oncogenes, modulation of growth factor signaling, and regulation of the cell cycle machinery (13, 14, 15) . Regulated changes in the expression and/or activity of cell cycle components that act within G1 have been closely associated with alterations in the proliferation rate of normal and transformed mammary epithelial cells (16) . Regulation of these events by both antiproliferative and proliferative extracellular signals are well documented. For example, progesterone inhibition of breast cancer cell growth has been linked to the inactivation of G1 CDKs by modulating the components of the CDK complex (17) . Also, the estrogen-induced activation of CDK4 and CDK2 during progression of human breast cancer cells between the G1 and S phases is accompanied by the increased expression of cyclin D1 and decreased association of the CDK inhibitors with the cyclin E-CDK2 complex (18) .
The sequential activation of CDKs and subsequent phosphorylation of specific substrates govern progress through the cell cycle on multiple levels. CDKs are inactive in the absence of cyclin binding; therefore, cyclin abundance is a major determinant of cyclin-CDK activity (19, 20, 21, 22) . Each cyclin is typically present for only a restricted portion of the cell cycle, and alterations in cyclin abundance are sufficient to alter the rate of cell cycle progression (19 , 21 , 22) . CDK activity is also regulated by a network of kinases and phosphatases (reviewed in Ref. 20 ), which can either activate or inactivate the complex. A further level of control results from the actions of two families of specific CDK-inhibitory proteins. Members of the p16INK4 family specifically target the kinases that associate with the D-type cyclins, CDK4 and CDK6 (21 , 23 , 24) . Members of the p21 (WAF1, Cip1) family interact with a broader range of CDKs, including CDK2, CDK4, and CDK6 (25 , 26) . One of the key endogenous substrates of the G1 CDKs is the retinoblastoma protein (Rb). Its phosphorylation is an important step in the transition between the G1 and S phases of the cell cycle because, when sufficiently phosphorylated, Rb releases a transcription factor of the E2F family that drives cells into S phase (27) .
Tamoxifen has been shown to decrease the activity of the estrogen receptor, and it does not have an antiproliferative effect on estrogen receptor-negative cell lines. In contrast, I3C can suppress the growth of cells regardless of estrogen receptor status (6) . Taking into account the known features of the growth suppression cascades induced by I3C or tamoxifen, we tested the combinatorial effects of these two breast cancer cell growth suppressors in estrogen-responsive MCF-7 cells. Most significantly, a combination of tamoxifen and I3C displayed a more effective growth suppression response, a more stringent inhibition of CDK2 specific activity, and more endogenous Rb phosphorylation than either compound alone. Our results suggest the possibility of developing I3C and tamoxifen as a potential combinatorial therapy to control estrogen-responsive breast cancers.
MATERIALS AND METHODS
DMEM, FBS, calcium- and magnesium-free PBS, l-glutamine, and trypsin-EDTA were supplied by BioWhittaker (Walkersville, MD). Insulin (bovine) and tamoxifen ([Z]-1-[p-dimethylaminoethoxyphenyl]-1,2-diphenyl-1-butene) citrate salt were obtained from Sigma Chemical Co. (St. Louis, MO). [3H]Thymidine (84 Ci/mmol) and [γ-33P]ATP (3,000 Ci/mmol) were obtained from NEN Life Science Products (Boston, MA). I3C was purchased from Aldrich (Milwaukee, WI). I3C was recrystallized in hot toluene prior to use. The sources of other reagents used in the study are either listed below or were of the highest purity available.
The MCF-7 human breast adenocarcinoma cell line was obtained from the American Type Culture Collection (Manassas, VA). MCF-7 cells were grown in DMEM supplemented with 10% FBS, 10 μg/ml insulin, 50 units/ml penicillin, 50 units/ml streptomycin, and 2 mml-glutamine and maintained at subconfluency at 37°C in humidified air containing 5% CO2. I3C and tamoxifen were dissolved in DMSO (99.9% high-performance liquid chromatography grade; Aldrich) at concentrations that were 1000-fold higher than the final medium concentration. In all experiments, 1 μl of the concentrated agent was added per 1 ml of medium, and for the vehicle control, 1 μl of DMSO was added per 1 ml of medium.
Estrogen Receptor Binding.
Rat uterine cytosol was prepared as described (28) . Briefly, 2.5 g of uterine tissue from five 12-week-old Sprague Dawley rats were excised and homogenized for 1 min with 30 ml of ice cold TEDG buffer [10 mm Tris, 1.5 mm EDTA, 1 mm DTT, and 10% glycerol (pH 7.4)], using a Polytron homogenizer at medium speed. The homogenate was centrifuged at 2800 rpm for 10 min at 4°C, transferred to a new tube, and centrifuged at 39,000 rpm for 90 min at 4°C. The supernatant was quickly frozen in a dry ice/ethanol bath and stored at −80°C.
For the competitive binding assay, 1 μl of competitor (100× solution in DMSO) was added to 5 μl of [3H]E2 mixture [20 nm [3H]E2 (NEN) in 50% ethanol, 5 mm Tris (pH 7.5), 5% glycerol, 0.5 mg/ml BSA, and 0.5 mm DTT] and incubated at room temperature with 95 μl of uterine cytosol (3 mg/ml) for 2 h. The range of concentrations of competitive ligand were 0.1 nm–1 μm for E2, 10 nm–10 μm for tamoxifen, and 10 μm–1 mm for I3C. The reaction was then incubated on ice for 15 min with 100 μl of 50% HAP slurry (hydoxylapatite; Bio-Rad BioGel HTP) rinsed and swelled overnight in TE buffer [50 mm Tris and 1 mm EDTA (pH 7.4)] and then adjusted to 50%. The slurry was vortexed to resuspend HAP every 5 min. Next, 1.0 ml of ice-cold wash buffer [40 mm Tris (pH 7.4), 100 mm KCl, 1 mm EDTA, and 1 mm EGTA] was added, and the mixture was vortexed and centrifuged for 5 min at 10,000 rpm at 4°C. The pellet was washed twice more with ice-cold wash buffer and then resuspended in 400 μl of ethanol and transferred to a scintillation vial for measurement. Relative binding affinities were calculated using the concentration of competitor needed to reduce [3H]E2 binding by 50%, as compared to the concentration of unlabeled estrogen needed to achieve the same result.
MCF-7 cells were plated onto 24-well Corning tissue culture dishes. Triplicate samples of asynchronously growing mammary cells were treated for the indicated times with either vehicle control (1 μl of DMSO per 1 ml of medium) or varying concentrations of I3C and/or tamoxifen. The cells were pulsed for 3 h with 3 μCi of [3H]thymidine (84 Ci/mmol), washed three times with ice-cold 10% trichloroacetic acid, and lysed with 300 μl of 0.3 n NaOH. Lysates (150 μl) were transferred into vials containing liquid scintillation cocktail, and radioactivity was quantitated by scintillation counting. Triplicates were averaged and expressed as cpm per well.
Crystal Violet Staining of Low Confluency Cultures.
MCF-7 cells were plated onto 100-mm Corning tissue culture dishes (10,000 cells per plate). The cells were treated with the indicated concentrations of I3C and tamoxifen for 8 days. At the end of the treatment, the cells were washed with PBS and incubated in a solution of 0.5% crystal violet and 10% formalin for 10 min and then rinsed with water. The integrated density of the colonies on each plate was determined using NIH Image software.
Soft Agar Colony Formation.
Two layers of Difco agar with different concentrations were set in individual wells of a 24-well plate. The lower layer contained 0.5 ml of 0.6% soft agar in medium with the indicated combinations of I3C and tamoxifen. The upper layer was composed of 0.5 ml of 0.3% soft agar in medium with MCF-7 cells (500 cells/well) and the corresponding combinations of I3C and tamoxifen in triplicate. After 4.5 weeks, all of the colonies that were <50 μm in diameter were counted.
Flow Cytometric Analyses of DNA Content.
MCF-7 cells (4 × 104) were plated onto Corning six-well tissue culture dishes. Triplicate samples were treated with the indicated concentrations of I3C and tamoxifen. The medium was changed every 24 h. Cells were incubated for 96 h and hypotonically lysed in 1 ml of DNA staining solution (0.5 mg/ml propidium iodide, 0.1% sodium citrate, and 0.05% Triton X-100). Nuclear emitted fluorescence with wavelength of >585 nm was measured with a Coulter Elite instrument with laser output adjusted to deliver 15 mW at 488 nm. Nuclei (10,000) were analyzed from each sample at a rate of 300–500 nuclei/s. The percentages of cells within the G1, S, and G2-M phases of the cell cycle were determined by analysis with the Multicycle computer program provided by Phoenix Flow Systems in the Cancer Research Laboratory Microchemical Facility of the University of California, Berkeley.
Western Blot Analysis.
After the indicated treatments, cells were harvested in radioimmunoprecipitation assay buffer (150 mm NaCl, 0.5% deoxycholate, 0.1% NP40, 0.1% SDS, and 50 mm Tris) containing protease and phosphatase inhibitors (50 μg/ml phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 5 μg/ml leupeptin, 0.1 μg/ml NaF, and 10 μg/ml β-glycerophosphate). Equal amounts of total cellular protein were mixed with loading buffer [25% glycerol, 0.075% SDS, 1.25 ml of 14.4 m β-mercaptoethanol, 10% bromphenol blue, 3.13% 0.5 m Tris-HCl, and 0.4% SDS (pH 6.8)] and fractionated on 10% (7.5% for Rb) polyacrylamide/0.1% SDS resolving gels by electrophoresis. Rainbow marker (Amersham Life Sciences, Arlington Heights, IL) was used as the molecular weight standard. Proteins were electrically transferred to nitrocellulose membranes (Micron Separations, Inc., Westboro, MA) and blocked overnight at 4°C with Western wash buffer-5% NFDM [10 mm Tris-HCl (pH 8.0), 150 mm NaCl, and 0.05% Tween 20–5% nonfat dry milk). Blots were subsequently incubated for 1 h at room temperature for rabbit anti-CDK2, CDK4, CDK6, p16, p21, and cyclin D1 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and overnight at 4°C for mouse anti-Rb and cyclin E antibodies (PharMingen, San Diego, CA). The working concentration for all antibodies was 1 μg/ml in Western wash buffer. Immunoreactive proteins were detected after incubation with horseradish peroxidase-conjugated secondary antibody diluted to 3 × 10−4 in Western wash buffer-1% NFDM [goat antirabbit IgG (Bio-Rad, Hercules, CA); rabbit antimouse IgG (Zymed, San Francisco, CA)]. Blots were treated with enhanced chemiluminescence reagents (NEN Life Science Products), and all proteins were detected by autoradiography. Equal protein loading was confirmed by Ponceau S staining of blotted membranes.
IP and CDK Kinase Assay.
MCF-7 cells were cultured in growth medium with combinations of tamoxifen and I3C for the indicated times and then rinsed twice with PBS, harvested, and stored as dry pellets at −70°C. For the IP, cells were lysed for 15 min in IP buffer [50 mm Tris-HCl (pH 7.4), 200 mm NaCl, and 0.1% Triton X-100] containing protease and phosphatase inhibitors (50 μg/ml phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 5 μg/ml leupeptin, 0.1 μg/ml NaF, 10 μg/ml β-glycerophosphate, and 0.1 mm sodium orthovanadate). Samples were diluted to 500 μg of protein in 1 ml of IP buffer. Samples were precleared for 30 min at 4°C with 20 μl of a 1:1 slurry of protein A-Sepharose beads (Pharmacia Biotech, Sweden) in IP buffer and 1 μg of rabbit IgG. After a brief centrifugation to remove precleared beads, 0.5 μg of anti-CDK2 or anti-CDK6 antibody was added to each sample and incubated on a rocking platform at 4°C for 2 h. Then, 20 μl of protein A-Sepharose beads were added to each sample, and the slurries were incubated on the rocking platform at 4°C for 30 min. The beads were then washed five times with IP buffer and twice with kinase buffer (50 mm HEPES, 10 mm MgCl2, 5 mm MnCl2, 0.1 μg/ml NaF, 10 μg/ml β-glycerophosphate, and 0.1 mm sodium orthovanadate). Half of the immunoprecipitated sample was checked by Western blot analysis to confirm the IP.
For the kinase assay, the other half of the sample was resuspended in 25 μl of kinase buffer containing 20 mm ATP, 5 mm DTT, 0.21 μg of Rb COOH-terminal domain protein substrate (Santa Cruz Biotechnology), and 10 μCi of [γ-33P]ATP (3000 Ci/mmol). Reactions were incubated for 15 min at 30°C and stopped by the addition of an equal volume of 2× loading buffer [10% glycerol, 5% β-mercaptoethanol, 3% SDS, 6.25 mm Tris-HCl (pH 6.8), and bromphenol blue]. Reaction products were boiled for 10 min and then electrophoretically fractionated in SDS-10% polyacrylamide gels. Gels were stained with Coomassie blue to monitor loading and destained overnight with 3% glycerol in 10% acetic acid. Subsequently, gels were dried and quantitated on a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) and visualized by autoradiography.
Quantitation of Autoradiography.
Autoradiographic exposures were scanned with a UMAX UC630 scanner, and band intensities were quantified using the NIH Image program. Autoradiographs from a minimum of three independent experiments were scanned per time point.
I3C and Tamoxifen Cooperate to Arrest the Growth of MCF-7 Cells.
To determine the potential combinatorial effects of I3C and tamoxifen on the growth of an estrogen-dependent breast cancer cell line, MCF-7 cells were grown in medium supplemented with 10% FBS, which contains enough estrogen to support the proliferation of these cells. The cells were treated with 100 μm I3C, 1 μm tamoxifen, or a combination of I3C and tamoxifen over a 96-h time course (Fig. 1) ⇓ . These concentrations of I3C and tamoxifen were previously shown to decrease cell growth without affecting viability (6 , 29) . The cells were then pulse-labeled with [3H]thymidine for 3 h at each time point to provide a measure of their proliferative state. Analysis of [3H]thymidine incorporation revealed that tamoxifen caused a steady decrease in DNA synthesis over the time course with a 60% inhibition after 96 h of treatment. I3C treatment also resulted in a time-dependent decrease in DNA synthesis with a 90% inhibition after 96 h. The combination of I3C and tamoxifen yielded statistically similar results as I3C alone for the 24- and 48-h time points. However, by the 72- and 96-h time points, the combination of I3C and tamoxifen resulted in a more effective growth suppression than either agent alone, resulting in a 95% inhibition after 96 h of treatment.
Because both I3C and the antiestrogen tamoxifen inhibited the growth of estrogen-responsive breast cancer cells, we wanted to determine whether I3C has any effect on estrogen receptor ligand binding. An in vitro competition binding assay for receptor-ligand interactions was used to examine the relative affinities of I3C and tamoxifen for the estrogen receptor. As a control and point of reference for the relative ligand affinity, unlabeled E2 was shown to effectively compete with [3H]E2 binding to the estrogen receptor, with half-maximal competition occurring at ∼3 nm(Fig. 2) ⇓ . As expected, tamoxifen has a relatively high affinity for the estrogen receptor with a half maximal [3H]E2 displacement of ∼200 nm. In contrast, I3C caused no significant displacement of [3H]E2 binding to the estrogen receptor, even at 1 mm. For the remainder of this study, the highest concentration of tamoxifen used was 1 μm, which is within the range of E2 competition, and the highest concentration of I3C used was 100 μm, which does not compete with E2 for receptor binding.
Effects of I3C and Tamoxifen on Adherent Cell Growth and Anchorage-independent Cell Growth.
To characterize the inhibitory effects of I3C and tamoxifen on adherent cell growth, MCF-7 cells were plated at low confluency (10,000 single cells per 100-mm plate) and grown for 8 days in medium containing the vehicle control or various doses of each agent alone or in combination. To visualize the cell colonies, the cells were stained and fixed in crystal violet/formalin. Representative plates of vehicle control, high doses of I3C (100 μm) or tamoxifen (1 μm) and the combination of I3C and tamoxifen are shown in Fig. 3A ⇓ . The average integrated density of replicate areas on each plate was determined by NIH Image and normalized by dividing that value by the area that was measured on each plate. This measurement takes into account both the number and size of the colonies and is representative of the number of cells on each plate. Treatment with the high doses of I3C or tamoxifen alone inhibited cell colony formation by 80 and 65%, respectively, whereas a combination of both agents inhibited adherent cell growth by >95%. An expanded analysis of the dose-dependent suppression of adherent cell growth by I3C and tamoxifen is shown in Fig. 3B ⇓ .
The effects of combinations of I3C and tamoxifen on the anchorage-independent growth of MCF-7 cells in soft agar was examined. Cells were cultured for 4.5 weeks in 0.3% agar and complete medium containing the vehicle control, I3C, tamoxifen, or a combination of both agents. Cell colonies that were >50 μm in diameter were counted. As shown in Fig. 3C ⇓ , increasing doses of I3C or tamoxifen caused a decrease in cell colony formation, and consistent with the adherent cell growth properties, a combination of both agents caused a more pronounced decrease in cell colony formation. For both growth conditions, the combined inhibitory effect of suboptimal concentrations of I3C (50 μm) and tamoxifen (0.5 μm) approximated that observed with the highest doses of either I3C or tamoxifen alone.
Cell Cycle Effects of I3C and Tamoxifen.
To assess the effect of combinations of I3C and tamoxifen on the cell cycle, MCF-7 cells were treated with the indicated concentrations of each compound for 96 h and then hypotonically lysed in the presence of propidium iodide to stain the nuclear DNA. Flow cytometry profiles revealed that treatment with increasing doses of I3C or tamoxifen lead to a dose-dependent shift in percentage of cells with a G1-like DNA content. As shown in Fig. 4A ⇓ , I3C or tamoxifen altered the DNA content of the MCF-7 cell population from an asynchronous population of growing cells in all phases of the cell cycle (61.4% in G1, 29.0% in S, and 9.7% in G2-M) to one in which most of the treated breast cancer cells were in the G1 phase of the cell cycle (Fig. 4A ⇓ , top right for I3C and bottom left for tamoxifen). Consistent with the cell growth studies, a combination of both agents caused a more striking shift to G1 (Fig. 4A ⇓ , bottom right). Incubation with suboptimal concentrations of both I3C and tamoxifen (Fig. 4A ⇓ , middle) induced the G1 shift to approximately the same extent as observed with the highest concentrations of either compound alone. As shown graphically in Fig. 4B ⇓ , the I3C- and tamoxifen-mediated shift in number of G1 cells (top) appeared to result from a decrease in S-phase cells (middle), whereas the G2-M phase values did not significantly change (bottom).
The expression levels of specific cell cycle proteins that are responsible for progression through G1 and/or transition into S phase were examined in cells treated for 96 h with combinations of 100 μm I3C and/or 1 μm tamoxifen. Western blot analysis revealed that I3C selectively decreased the level of CDK6 protein and increased the level of the p21 CDK inhibitor at this time point (Fig. 5) ⇓ . In contrast, tamoxifen had no effect on each of these cell cycle proteins under our cell culture conditions. Our previous studies show that I3C also slightly increases the level of the p27 CDK inhibitor protein (6) , whereas tamoxifen has no effect on p27 protein levels (data not shown). The expression of the other cell cycle proteins tested, such as CDK2, CDK4, cyclin D1, and cyclin E, were not affected by either I3C or tamoxifen. The p16 CDK inhibitor was not detectable in the MCF-7 cells used for our study.
I3C and Tamoxifen Cooperate to Decrease the in Vitro Activity of CDK2 and the Phosphorylation of Endogenous Rb.
The control of G1 CDK enzymatic activity is critical for regulating cell cycle progression (22) . The activity of specific CDKs is regulated, in part, by the composition of the holoenzyme, which includes the appropriate cyclin and/or CDK inhibitory proteins. Therefore, although the levels of CDK2 and CDK4 remain unaltered after I3C and/or tamoxifen treatment, we examined the potential effects of I3C and tamoxifen on G1 CDK specific activities. Because one of the key endogenous substrates for the G1 CDKs is the Rb protein, we determined the ability of the individual G1 CDKs to phosphorylate Rb in vitro. MCF-7 cells were treated for 48 h with I3C and/or tamoxifen and then CDK2, CDK4, or CDK6 were immunoprecipitated from total cell extracts. For the kinase assays, half of each immunoprecipitated sample was incubated with the COOH-terminal domain of Rb fused to GST and [γ-33P]ATP. The electrophoretically fractionated reaction products were quantitated on a PhosphorImager and then visualized by autoradiography. The other half of the immunoprecipitated samples were analyzed by Western blot and densitometry to confirm the efficiency and specificity of each IP. The CDK-specific activity was calculated by dividing the in vitro kinase activity by the corresponding protein expression. As shown in the CDK2 specific activity graph, increasing doses of I3C alone or tamoxifen alone resulted in dose-dependent decreases in CDK2-specific activity (Fig. 6) ⇓ . Treatment with combinations of these two growth inhibitors resulted in a more stringent decrease in CDK2-specific activity than either agent alone. In contrast, I3C inhibited CDK6 protein expression, but the specific activity of the residual CDK6 protein remained unaltered. Tamoxifen had no effect on CDK6 protein expression or specific activity (Fig. 6) ⇓ . Thus, the decrease in CDK6 activity appears to result from the I3C-mediated decrease in CDK6 protein levels and not due to an effect on CDK6 enyzmatic activity. CDK4-specific activity was not affected by either tamoxifen or I3C (data not shown).
It was important to determine whether the in vitro kinase assay results reflected the phosphorylation status of endogenous Rb after treatment with the antiproliferative agents. Therefore, the levels of phosphorylated and hypophosphorylated Rb were examined in MCF-7 cells treated for 48 h with I3C, tamoxifen, or a combination of both agents. The extent of Rb phosphorylation was determined by probing Western blots with a Rb-specific antibody and analyzing the characteristic mobility shift of the hyperphosphorylated Rb protein. As shown in Fig. 7 ⇓ , I3C treatment, and to a lesser extent tamoxifen treatment, caused a decrease in total Rb protein levels and an increase in the relative levels of hypophosphorylated Rb (pRb). Most significantly, a combination of I3C and tamoxifen virtually ablated the expression of the hyperphosphorylated form of Rb (ppRb), which likely explains the more potent growth arrest observed in the presence of these two growth suppressors.
A wide range of extracellular signaling molecules either inhibit or stimulate the proliferation of mammalian cells through pathways that ultimately target specific components within G1 (19 , 22 , 30) . It is well established that tamoxifen can exert its growth-inhibitory effects in human breast cancer cells by antagonizing the estrogen receptor stimulation of cell cycle progression (18) . We have recently documented that the dietary indole I3C can induce a G1 block in cell cycle progression of breast cancer cell lines in the absence of estrogen receptor signaling (6) . Depending on the antiproliferative pathways, cell cycle progression can be more effectively blocked through the coordinate actions of specific networks of cell cycle-regulated gene products (31) . Our results have established that the antiproliferative cascades initiated by I3C and tamoxifen can cooperate to induce a more stringent growth suppression in breast cancer cells treated with a combination of both agents compared to treatment with either I3C or tamoxifen alone. We propose that the I3C- and tamoxifen-activated antiproliferative responses are mediated by distinct signal transduction pathways (see Fig. 8 ⇓ ) that converge on the inhibition of CDK2-specific activity as a common target with a subsequent decrease in endogenous Rb phosphorylation.
The distinguishing feature of the I3C antiproliferative pathway is the rapid down-regulation of CDK6 expression (6) , which accounts for the inhibition of total CDK6 activity in breast cancer cells. Tamoxifen had no apparent effect on either CDK6 protein levels or enzymatic activity. In contrast to tamoxifen, I3C did not compete with E2 for binding to the estrogen receptor and has been shown to exert its growth-suppressive effects in breast cancer cells irrespective of estrogen receptor status, such as in estrogen receptor-deficient breast cancer cells (6) . The immediate cellular target for I3C or a specific metabolite is unknown. We propose (Fig. 8) ⇓ that this extracellular signal binds to a target molecule that mediates the growth-inhibitory cascade. The sustained growth suppression of colonies in soft agar in the absence of a medium change for 4.5 weeks implies that I3C could be working through a stable compound that binds to a specific cellular component. I3C regulates several G1-acting cell cycle components, including the down-regulation of CDK6 expression and the inhibition of CDK2-specific activity. Tamoxifen inhibited MCF-7 cell growth through its antiestrogenic effects because the cells were cultured in medium supplemented with 10% FCS, which contains enough endogenous estrogen to maintain the cells in a proliferative state. Other studies have shown that tamoxifen interferes with the estrogen stimulation of the cell cycle, which includes a disruption in the stimulation of CDK2 activity and Rb phosphorylation (32) . Our results are consistent with these observations and we propose that the inhibition of CDK2-specific activity and Rb phosphorylation are common targets for both the I3C and tamoxifen pathways (Fig. 8) ⇓ and, thereby, provides one mechanistic explanation for the more stringent growth suppression observed with both agents.
Several potential mechanisms could account the inhibition of CDK2-specific activity by the I3C and/or the tamoxifen pathways. For example, treatment with either agent could alter the composition of CDK2 holoenzyme complex resulting in inactive complexes that do not support the phosphorylation of Rb. Alternatively, because full activation of CDK2 requires phosphorylation on specific residues and removal of phosphates from other residues (17 , 20 , 33) , the CDK2-specific kinases and phosphatases may be targets for either the I3C or tamoxifen pathways. The unique features of each pathway likely play a role in the more stringent inhibition of growth observed in cells treated with both I3C and tamoxifen. The I3C down-regulation of CDK6 expression could release other cell cycle factors contained within the CDK6 protein complexes and, thereby, shift the cellular equilibrium of these components in manner that alters the composition of the other G1 CDK protein complexes. For example, enough CDK-inhibitory molecules could conceivably be released from the CDK6 protein complexes to negatively modulate the specific activity of CDK2. Furthermore, longer I3C treatments increase the level of the p21 CDK inhibitor, which could also influence CDK2-specific activity. However, within the 24-h time point when CDK2 activity is affected, neither p21 protein levels nor p21’s ability to coimmunoprecipitate with CDK2 changes (data not shown). Thus, p21 does not appear to be responsible for I3C’s ability to decrease CDK2 activity.
In addition to the differential cell cycle effects of I3C and tamoxifen, treatment with both compounds virtually eliminated colony formation in soft agar. This suggests the possibility that I3C could enhance the effectiveness of tamoxifen to control the growth of breast tumors in vivo. Tamoxifen has been shown to act as an estrogen agonist or antagonist, depending on cell type (reviewed in Ref. 12 ). The use of tamoxifen to treat estrogen-responsive breast cancers has extended the lives of many women (7, 8, 9) . In addition, the preliminary results of a clinical trial using tamoxifen as a preventative treatment for women that are at a high risk for breast cancer are very encouraging (34) . However, approximately two-thirds of breast cancer patients have estrogen receptor-positive tumors, only half of which respond to tamoxifen therapy (35) . Moreover, after 12–18 months of treatment, resistance to tamoxifen develops in all patients (36) , and tamoxifen has been shown to actually stimulate the growth of breast cancer cells after prolonged treatment (37 , 38) . The molecular mechanism underlying the development of acquired resistance to antiestrogens is unclear, but continued exposure of cells to tamoxifen may select for cells able to grow without estrogen stimulation (39) .
Our results suggest that I3C, in combination with tamoxifen, could overcome some of the drawbacks of tamoxifen therapy while capitalizing on the positive effects of this proven therapy. Lower doses and/or pulses of tamoxifen are two of the proposed methods of circumventing tamoxifen resistance (40) . In this regard, our results showed that lower doses of tamoxifen and I3C inhibited MCF-7 cell growth and CDK2-specific activity to the same extent as higher doses of either agent added individually. In principle, this response could be exploited to circumvent acquired drug resistance to sustained high doses of tamoxifen. Alternatively, patients could conceivably receive intermittent pulses of tamoxifen while undergoing I3C treatment. I3C has been shown to reduce the formation of both spontaneous and carcinogen-induced mammary tumors in rodents with no apparent side effects (4 , 41, 42, 43) . Human subjects who ingested I3C also had no side effects (44 , 45) . To extend our current studies, we plan to examine the in vivo effectiveness of combinations of I3C and tamoxifen on the growth of tumors derived from estrogen-responsive and estrogen-nonresponsive breast cancer cell lines and to determine the mechanism by which the I3C and tamoxifen pathways cooperate to block cell cycle progression.
We express our appreciation to members of both the Firestone and Bjeldanes laboratories for their helpful comments throughout the duration of this work. We also thank Jerry Kapler for his excellent photography and Peter Schow, Meredith L. Leong, Minnie Wu, and Tran Van for their technical assistance.
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.
↵1 Supported by Department of Defense Army Breast Cancer Research Program Grant DAMD17-96-1-6149 (to L. F. B.) and by University of California Breast Cancer Research Program Grant 31B-0110 (to G. L. F.). C. M. C. is a recipient of a predoctoral fellowship supported by NIH National Research Service Grant CA-09041. E. J. C. is a recipient of a predoctoral fellowship supported by the United States Army Medical Research and Materiel Command Breast Cancer Research Program Award BC971062.
↵2 To whom requests for reprints should be addressed, at Department of Molecular and Cell Biology, 591 LSA, University of California at Berkeley, Berkeley, CA 94720-3200. Phone: (510) 642-8319; Fax: (510) 643-6791; E-mail:
↵3 The abbreviations used are: I3C, indole-3-carbinol; CDK, cyclin-dependent kinase; E2, β-estradiol; FBS, fetal bovine serum; IP, immunopreciptation; GST, glutathione S-transferase.
- Received July 15, 1998.
- Accepted January 13, 1999.
- ©1999 American Association for Cancer Research.