A new class of peroxisome proliferator-activated receptor γ (PPARγ) agonists that inhibit growth of breast cancer cells: 1,1-Bis(3′-indolyl)-1-(p-substituted phenyl)methanes

Peroxisome proliferator-activated receptor γ (PPARγ) is a ligand-activated transcription factor and a member of the nuclear hormone receptor (NHR) superfamily (1–7). PPARγ is highly expressed in fatty tissues and up-regulated in tumors. PPARγ participates in several metabolic pathways including adipocyte differentiation, insulin resistance, atherosclerosis, and hypertension, and this receptor is a major target for development of new drugs to treat metabolic diseases (1–7). PPARγ, like other members of the NHR superfamily, contains a modular structure that includes a NH₂-terminal activation function 1 and a COOH-terminal activation function 2, which also binds ligands. The DNA binding domain of PPARγ and other NHRs contain two zinc fingers, and PPARγ partners with the retinoid X receptor (RXR) to bind PPAR response elements (PPRE) containing AGGNCA half-sites separated by 1 bp (1–7). Ligand-receptor interactions induce conformational changes that facilitate recruitment of coactivator and coregulatory proteins required for transactivation.

PPARγ is widely expressed in cancer cell lines from multiple tumors (8–27), and a recent clinical study reported PPARγ mRNA expression in tumors from breast, lung, colon, prostate, osteosarcomas, acute myelogenous leukemia, adult T-cell leukemia, glioblastomas, B-cell acute lymphoblastic leukemia, non-Hodgkin's lymphoma, and myelodisplastic syndrome (28). PPARγ ligands such as 15-deoxy-Δ12,14-prostaglandin J2 (PGJ2) and synthetic thiazolidinediones inhibit growth of cancer cell lines, and this has been associated with modulation of multiple cell context-dependent pathways including inhibition of G₀/G₁-S phase cell cycle progression, down-regulation of cyclin D1, enhanced expression of p21 or p27 cyclin-dependent kinase inhibitors, and induction of apoptosis (8–27).

Research in this laboratory has investigated inhibitory cross-talk between the aryl hydrocarbon receptor (AhR) and estrogen receptor (ER) α signaling pathways in breast cancer cells (29, 30). Ligands for the AhR inhibit 17β-estradiol (E2)-induced growth and gene expression in ER-positive MCF-7 cells and also block growth of hormone-dependent mammary tumors in rodent models (29–34). Diindolylmethane (DIM) and several ring-substituted DIM derivatives have been characterized as selective AhR modulators that inhibit growth of carcinogen-induced mammary tumors in rats (31–33). We have also investigated the antiestrogenic activity of methylene (C)-substituted DIMs containing bulky aromatic substituents; these C-substituted DIMs did not activate the AhR but inhibited growth of breast cancer cells (in vitro) and tumors (in vivo; 35 and data not shown). This article reports the structure-dependent PPARγ agonist activities of DIMs containing substituted phenyl groups at the methylene carbon in MCF-7 cells, and 1,1-bis(3′-indolyl)-1-(p-trifluoromethylphenyl)methane (DIM-C-pPhCF₃) is used as a prototype for mechanistic studies. DIM-C-pPhCF₃ and compounds containing p-t-butyl, cyano, dimethylamino, and phenyl substituents were most active as PPARγ agonists in transactivation assays, whereas analogues containing methyl, hydrogen, methoxy, hydroxyl, or halogen groups were relatively inactive. DIM-C-pPhCF₃ and related compounds resemble PGJ2 in MCF-7 cells, and both PPARγ agonists induce selective recruitment of coactivators in a mammalian two-hybrid assay, inhibit G₀/G₁-S phase cell cycle progression, induce apoptosis, and activate proteasome-dependent degradation of cyclin D1 and ERα proteins.

MCF-7 cells and 3T3-L1 preadipocytes were obtained from American Type Culture Collection (Manassas, VA). MCF-7 cells were maintained in MEM supplemented with 0.22% sodium bicarbonate, 10% fetal bovine serum (FBS), 0.011% sodium pyruvate, 0.1% glucose, 0.24% HEPES, 0.00001% insulin, and 2 ml/l antibiotic solution (Sigma Chemical Co., St. Louis, MO). Cells were grown in 150 cm² culture plates in an air/CO₂ (95:5) atmosphere at 37°C and passaged every 5 days. For most experiments, cells were seeded in DMEM:F12 with 5% FBS. The caspase inhibitors IETD-CHO and LEHD-CHO were purchased from Alexis Biochemicals (Carlsbad, CA). PGJ2 (PG-050) was purchased from Biomol Research Laboratories Inc. (Plymouth Meeting, PA). Rosiglitazone and troglitazone were purchased from IKT Laboratories (St. Paul, MN). MG132, calpain inhibitor II (CII), and actinomycin D were purchased from Calbiochem-Novabiochem Co. (San Diego, CA). The PPARγ antagonist N-(4′-aminopyridyl)-2-chloro-5-nitrobenzamide (T007) was synthesized in this laboratory and confirmed by gas chromatography-mass spectrometry. Antibodies for cyclin D1 (sc-718 and sc-246), poly(ADP-ribose) polymerase (PARP), ERα (sc-544), PPARγ (sc-7196), and Sp1 (sc-59 and sc-420) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Horseradish peroxidase substrate for Western blot analysis was purchased from NEN Life Science Products (Boston, MA). RNase ONE, pGL2 luciferase reporter vector, cell lysis buffer, and luciferase reagent were purchased from Promega (Madison, WI), and β-galactosidase (β-gal) reagent was from Tropix (Bedford, MA). Plasmid preparation kits were purchased from Qiagen, Inc. (Santa Clarita, CA), and LipofectAMINE2000 transfection kits were obtained from Invitrogen (Carlsbad, CA). Small inhibitory RNA (siRNA) duplexes for lamin A/C (iLMN), luciferase (iGL2), and PPARγ (iPPARγ) were prepared by Dharmacon Research (Lafayette, CO) and contain the sequences listed in Table 1.

The series of C-substituted DIMs used in this study were prepared using the same protocol as described previously (36) by condensation of 0.02 mol of indole with 0.01 mol of a substituted benzaldehyde at pH 2.5 in dilute aqueous acetic acid. Indole and the substituted benzaldehydes were purchased from Aldrich Chemical Co. (Milwaukee, WI). Indole and the appropriate substituted benzaldehyde were added to 50 ml of water and 0.6 ml of acetic acid at room temperature. The reaction mixture was stirred and progress of the condensation reaction was monitored by thin-layer chromatography; when 80–90% of the reaction were complete, the resulting solid was filtered and crystallized from benzene or benzene/hexane. High yields of the C-substituted DIMs were obtained (70–90%) and structures were confirmed by gas chromatography-mass spectrometry and/or nuclear magnetic resonance spectroscopy. A more extensive study on spectral properties of the C-substituted DIMs will be published separately. The compounds are light sensitive and are stored in the solid form or in Me₂SO solution in dark vials.

PPRE3Luc luciferase reporter was constructed using pGL2 with a minimal TATA sequence and triple consensus PPREs. The Gal4 reporter containing 5× Gal4DBD (Gal4Luc) was kindly provided by Dr. Marty Mayo (University of North Carolina, Chapel Hill, NC). Gal4DBD-PPARγ construct (gPPARγ) was a gift of Dr. Jennifer L. Oberfield (Glaxo Wellcome Research and Development, Research Triangle Park, NC), and PPARγ expression plasmid and pM-PPARγ coactivator-1 (PGC1) were gifts of Dr. Bruce M. Spiegelman (Harvard University, Boston, MA). The PPARγ₂-VP16 fusion plasmid (VP-PPARγ) contained the DEF region of PPARγ (amino acids 183–505) fused to the pVP16 expression vector and was kindly provided by Dr. Shigeaki Kato (University of Tokyo, Tokyo, Japan). The Gal4-coactivator fusion plasmids pM-AIB1, pM-steroid receptor coactivator (SRC) 1, pM-TIFII, and pM-thyroid hormone receptor-associated protein 220 (TRAP220) plasmids were also provided by Dr. Shigeaki Kato.

MCF-7 cells were seeded in 12-well plates with DMEM:F12 with 5% FBS 24 h before transfection. Various amounts of DNA [i.e., Gal4Luc (0.2 μg), PPRE3Luc (0.2 μg), β-gal DNA (0.04 μg), PPARγ (0.15 μg), gPPARγ (0.03 μg), pVP-PPARγ (0.03 μg), pM-AIB1 (0.03 μg), pM-SRC1 (0.03 μg), pM-TIFII (0.03 μg), and pM-TRAP220 (0.03 μg)] were mixed with LipofectAMINE2000 reagent at a ratio of 1:3 or 1:5 (μg of DNA/μl of reagent) and added to cells following the manufacturer's instructions. After incubation for 6 h, cells were treated for 24 h in fresh media as indicated. Cells were lysed with 200 μl of 1× Reporter Lysis Buffer, and 25 μl of cell extract were subjected to luciferase and β-gal assays. Luciferase and β-gal activities were quantitated by Lumicount, and luciferase activities were normalized to β-gal activity. All experiments were repeated at least thrice for each treatment group.

For transfection experiments, MCF-7 cells were cultured in six-well plates in 2 ml DMEM:F12 supplemented with 5% FBS. When cells were ∼50–60% confluent, siRNA duplexes and/or reporter gene constructs were transfected using OligofectAMINE reagent (Invitrogen). Based on results of previous studies (37–39), siRNAs were transfected in each well to give a final concentration of 70 nm. Cells were harvested in 48–56 h after transfection by manual scraping in 1× lysis buffer (Promega). Whole cell extracts were frozen in liquid nitrogen for 30 s, vortexed by 30 s, and centrifuged at 12,000 × g for 1 min to give lysates that were assayed for luciferase activity as described above. In a parallel experiment, whole cell lysates from this experiment were analyzed for PPARγ protein by Western blot analysis essentially as described (39). Relative intensities of the PPARγ protein were compared with a nonspecific band and results of three replicate studies were determined in the various treatment groups.

MCF-7 cells were seeded in DMEM:F12 with 5% FBS and treated with 1, 5, or 10 μm of either DIM-C-pPhCF₃, DIM-C-pPhtBu, DIM-C-pPhCN, 1,1-bis(3′-indolyl)-1-(p-biphenyl)methane (DIM-C-pPhC₆H₅), or solvent (Me₂SO) as indicated. Cell numbers were determined using a Coulter Z1 counter and results for each treatment group were determined for at least three replicate studies.

Female Sprague-Dawley rats (Harlan, Indianapolis, IN) were received at 39 days of age. At 50 days of age, each animal received 7,12-dimethylbenz(a)anthracene (20 mg) suspended in corn oil by gavage. Tumors developed in these animals within 30–75 days after treatment. After palpable tumors were detected (250–400 mm³), rats were treated with corn oil vehicle control (4 ml/kg) or DIM-pPhC₆H₅ (1 mg/kg) in corn oil (4 ml/kg) every second day as described (31–33). Tumor volumes were measured throughout the experiment and were calculated as π × L/2 × W/2 × D/2. On day 21 (2 days after the final dose), animals were sacrificed by CO₂ asphyxiation followed rapidly by cervical dislocation, and organ and body weights were determined and shown to be unchanged by the treatment (data not shown).

After synchronization of cells for 24 h in serum-free medium, MCF-7 cells were treated for 20 h with different concentrations of DIM-C-pPhCF₃. Cells were then trypsinized, syringed, and collected by centrifugation. Single cell suspensions were prepared by resuspending cell pellets in 1× staining solution [50 μg/ml propidium iodide, 30 units/ml RNase ONE, 4 mm sodium citrate, and Triton X-100 (pH 7.8)]. Cells were incubated at 37°C for 10 min, and sodium chloride solution was added to the final concentration of 0.15 m. Stained cells were analyzed on a FACSCalibur Flow Cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA) using Cell Quest (Becton Dickinson Immunocytometry Systems) acquisition software as described (35).

3T3-L1 preadipocytes were cultured on poly-lysine-coated coverslips with DMEM and 10% FBS at 5% CO₂ in 24-well plates. At 2 days postconfluence, cells were incubated with fresh media supplemented with 3-isobutyl-1-methylxanthine (0.5 mm), dexamethasone (1 μm), insulin (1.7 μm), and PGJ2 or DIM-C-pPhCF₃ (1 μm). After 48 h, cells were changed to fresh media and treated with Me₂SO, PGJ2 (1 μm), or DIM-C-pPhCF₃ for 5 days. Cells without any treatment for the entire 7 days were used as control. The cells were then fixed with 10% formalin. Fixed cells were washed with 60% isopropanol and stained with filtered 60% Oil Red O in deionized water. After staining, cells were washed with water and photographed.

Whole cell lysates were extracted using 1× Western sampling buffer. Protein samples were heated at 100°C for 5 min, separated on 10% SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membrane (Amersham Pharmacia Biotech, Piscataway, NJ). The PVDF membrane was blocked for 1 h and incubated with 1:1000 (cyclin D1), 1:500 (PPARγ), or 1:2000 (Sp1) primary antibody (Santa Cruz Biotechnology) for 1 h or with 1:500 (ERα) for 3 h. After vigorous washing for 30 min, 1:2000 secondary antibody was incubated with shaking for 1 h. The membrane was washed for 30 min, incubated with enhanced chemiluminescence substrate (NEN Life Science Products) for 1.0 min, and exposed to Kodak X-Omat AR autoradiography film (Rochester, NY). The membrane was reused and probed with other antibodies as indicated for individual experiments.

MCF-7 cells (10⁶) were plated in 145 mm dishes and treated with DIM-C-pPhC₆H₅ or Me₂SO (0.2%) for the indicated times. Floating and adherent cells were collected and resuspended in 100 μl of TE buffer (pH 8.0) and transferred to prechilled 2 ml epitubes and incubated on ice for 5 min. The cell suspension was then mixed with 500 μl of sonication buffer [1% SDS, 10 mm Tris-HCl, 1 mm EDTA (pH 8.0)] and incubated on ice for an additional 20 min. After several extractions with phenol/chloroform/isoamyl alcohol (25:24:1), 0.5 volume of sodium chloride (5 m) and 2.5 volumes of 95% ethanol were added to the aqueous layer and nucleic acids were precipitated at −20°C overnight and collected by centrifugation at 12,000 × g for 30 min. The pellet was washed with 80% ethanol, air dried, and resuspended in TE buffer with DNase-free RNase A (40 μg/ml) and incubated at 37°C for 2 h. DNA was run in 2% agarose gels and visualized by ethidium bromide staining. For studies on PARP cleavage, whole cell lysates from treated cells were electrophoresed in 10% SDS-PAGE gels, and proteins were transferred to PVDF membranes (Immobilon-P; Millipore, Bedford, MA). Anti-PARP mouse monoclonal antibody (F-2; Santa Cruz Biotechnology) was incubated at 1:400 dilution at room temperature for 2 h in 1× PBS with 0.04% Tween 20. Bound antibodies were detected with horseradish peroxidase-conjugated donkey anti-mouse IgG and luminol regent (Santa Cruz Biotechnology). z-VAD-fmk, a broad-spectrum, irreversible caspase inhibitor (Santa Cruz Biotechnology), was used at a final concentration of 100 μm for determining effects of DIM-C-pPhCF₃ (7.5 μm) on inducing apoptotic detached MCF-7 cells after treatment for 16 h.

The 2.3 Å resolution crystal structure of the ligand binding domain (residues 206–477) from human PPARγ derived from the liganded PPARγ LBD/RXRα LBD heterodimer complexed with GW409544, 9-cis-retinoic acid, and two LxxLL peptides from SRC1 has been published (40). The atomic coordinates and structure factors were downloaded from the Protein Data Bank (PDB; http://www.rcsb.org; PDB identification code: 1K74). RXRα, GW409544, 9-cis-retinoic acid, and the two peptides from SRC1 were deleted leaving the coordinates for the subunit PPARγ. DIM-C-pPhCF₃ and PGJ2 were created in ISIS/Draw and were energy minimized using semi-empirical quantum mechanical AM1 methods (HyperChem, Ottawa, ON, Canada). This file was converted to a PDB file by HyperChem and inserted into the Y-shaped ligand binding pocket of the crystal structure subunit of PPARγ using Swiss PDB Deep Viewer v3.7b2 from Glaxo Wellcome Experimental Research (Geneva, Switzerland). Quantitative structure-activity relationships were computed using HyperChem for DIM-C-pPhCF₃ and PGJ2.

Statistical differences between different groups were determined by ANOVA and Scheffe's test for significance. Data are presented as means ± SD for at least three separate determinations for each treatment.

Ring-substituted DIMs activate the AhR and inhibit E2-induced growth of breast cancer in vitro and mammary tumor growth in carcinogen-induced female Sprague-Dawley rats (31–33). In contrast, a series of C-substituted DIMs, which did not activate the AhR, also inhibited breast cancer cell growth. These lipophilic compounds, typified by DIM-C-pPhCF₃, were systematically investigated in transactivation assays as ligands for other receptors, including PPARγ, PPARα, retinoic acid receptor (RAR), and RXR. PGJ2, DIM-C-pPhCF₃, and isomers with o- and m-trifluoromethyl groups induced luciferase activity in MCF-7 cells transfected with PPRE3Luc/PPARγ expression plasmid (Fig. 1A). PGJ2 and DIM-C-pPhCF₃ (10 and 20 μm) induced luciferase activity in both assays, whereas o-CF₃ and m-CF₃ analogues were less active at the 20 μm concentration. Comparable results were obtained in cells transfected with Gal4Luc/PPARγ. DIM-C-pPhCF₃ did not induce transactivation in MCF-7 cells using Gal4Luc and Gal4-receptor chimeras containing PPARα, RXR, or RAR, and activation of AhR- and ERα-mediated responses were also not observed (data not shown).

Thiazolidinediones such as rosiglitazone and troglitazone are synthetic PPARγ agonists that have been extensively investigated. Results in Fig. 1B compare concentration-dependent activation of PPARγ by DIM-C-pPhCF₃ (-CF₃) and DIM-C-pPhtBu (-tBut) and the thiazolidinediones in MCF-7 cells transfected with Gal4-luc/pM-PPARγ. DIM-C-pPhCF₃ and DIM-C-pPhtBu (5–10 μm) induced a maximal 5.5–6.5-fold response, and this decreased at higher concentrations due to cytotoxicity. Rosiglitazone and troglitazone induced transactivation in MCF-7 cells at lower concentrations; however, the maximum fold induction (4–5-fold) was lower than observed for the C-substituted DIMs. This was consistently observed in replicate experiments.

The effects of substituent structure on the activity of a series of 1,1-bis(3′-indolyl)-1-(p-substituted phenyl)methane analogues was also investigated in PPARγ-mediated transactivation assays (Fig. 1, C and D), which gave complementary results. At concentrations of 10 or 20 μm, the most active compounds contained the p-CF₃, p-t-butyl, p-cyano, p-phenyl (C₆H₅), or p-dimethylamino substituents. Higher concentrations of these compounds were cytotoxic. In contrast, minimal to nondetectable responses were observed after treatment of MCF-7 cells with 10 or 20 μm concentrations of compounds containing p-hydrogen, hydroxyl, fluoro, bromo, methoxy, or methyl substituents.

Recent studies identified T007 as a PPARγ antagonist that inhibited thiazolidinedione-induced transactivation in HEK 293 cells and adipocyte differentiation in 3T3-L1 cells (41). In this study, T007 also inhibited activation of PPARγ by PGJ2 and DIM-C-pPhCF₃ in MCF-7 cells transfected with Gal4Luc/pM-PPARγ or PPARE3Luc/PPARγ expression plasmid (Fig. 2, A and B). In contrast, T007 (20 μm) alone did not significantly inhibit activity in cells transfected with Gal4Luc/pM-PPARγ, whereas inhibition (40%) was observed in cells transfected with PPRE3Luc/PPARγ expression plasmid (data not shown). T007 (20 μm) inhibited ∼50% of PGJ2 (20 μm)-induced transactivation in both assays, whereas a higher percentage of inhibition of DIM-C-pPhCF₃-induced activity was observed. This was particularly evident for the Gal4Luc/pM-PPARγ assay where T007 (0.2 μm) inhibited 80% of DIM-C-pPhCF₃-induced transactivation. In parallel studies, we also used iPPARγ in an RNA interference assay, and results in Fig. 2C show that in MCF-7 cells transfected with iPPARγ, there was a significant decrease in PPARγ protein as determined by Western blot analysis of whole cell lysates. In two separate experiments, iPPARγ inhibited PGJ2 (Fig. 2D)- and DIM-C-pPhCF₃ (Fig. 2E)-induced luciferase activity in MCF-7 cells transfected with PPARE3Luc compared with cells treated with iLMN (a siRNA targeted to lamin A/C). iGL2, which targets the bacterial luciferase mRNA, also decreased activity in cells treated with Me₂SO or PPARγ agonist. In contrast, iPPARγ did not affect induction of luciferase activity in MCF-7 cells treated with E2 (10 nm) and transfected with an E2-responsive pERE₃ construct (data not shown). Thus, both iPPARγ and T007, a PPARγ antagonist, inhibited PGJ2- and DIM-C-pPhCF₃-mediated transactivation.

PPARγ agonists characteristically induce cellular differentiation in 3T3-L1 adipocytes, and this is characterized by altered gene expression and accumulation of fat droplets, which can be detected by Oil Red O staining. Results in Fig. 3 show that in control 3T3-L1 cells and after treatment with Me₂SO, PGJ2 (1 μm), and DIM-C-pPhCF₃ (1 μm), a significant increase in Oil Red O staining of fat droplets was only observed in the latter two treatment groups. Thus, DIM-C-pPhCF₃, like PGJ2 and other synthetic PPARγ agonists, induced PPARγ-dependent differentiation in 3T3-L1 cells.

Ligand-induced transactivation mediated by PPARγ and other NHRs is dependent on recruitment of coactivator proteins. These include SRCs (42), mediator complex (43–46) proteins such as PBP/TRAP220/DRIP 205 (45), and PGC1 (47–53). Results in Fig. 4 summarize ligand-dependent transactivation in MCF-7 cells transfected with pM-coactivator and VP-PPARγ chimeric expression plasmids and a Gal4Luc reporter construct. This mammalian two-hybrid assay determines the effects of PPARγ agonists to induce conformational changes in the ligand binding domain, which are required for subsequent binding to coactivators. The pattern of activation by PGJ2 and DIM-C-pPhCF₃ exhibited both similarities and differences. Both PPARγ ligands activated gene expression in MCF-7 cells transfected with pM-SRC1 and pM-TIFII (SRC2) but not pM-AIB1 (SRC3). DIM-C-pPhCF₃ (1–10 μm) induced transactivation in cells transfected with pM-TRAP220, whereas only a submaximal response was observed for PGJ2 (2.5 μm). In contrast, previous studies showed that PGJ2 induced activity in COS-1 cells transfected with VP-PPARγ and chimeric SRC1, TIFII, AIB1, and TRAP220-Gal4 proteins (54). The major difference between PGJ2 and C-substituted DIMs in the mammalian two-hybrid assay in MCF-7 cells was observed using PGC1. Both DIM-C-pPhCF₃ and DIM-C-pPhC₆H₅ induced interactions between PPARγ and PGC1, whereas treatment with PGJ2 decreased activity. These results further demonstrate that C-substituted DIMs induce interactions between PPARγ and coactivators, which is characteristic of PPARγ agonists.

PPARγ agonists typically inhibit growth of breast cancer cells (22, 24, 25, 35), and results in Fig. 5 illustrate the effects of 1,1-bis(3′-indolyl)-1-(p-substituted phenyl) methane analogues (1, 5, and 10 μm) as inhibitors of MCF-7 cell growth. These compounds contain p-trifluoromethyl, p-phenyl (C₆H₅), p-t-butyl, and p-cyano substituents on the phenyl ring and were among the most active PPAR agonists in the screening assay (Fig. 1). Significant inhibition of MCF-7 cell growth was observed at concentrations of ≥ 1 μm and the higher concentrations (5 and/or 10 μm) were cytotoxic; cell numbers after treatment with cytotoxic doses were lower than the number of attached cells on day 1, and floating dead cells were observed. The effects of DIM-C-pPhCF₃ on MCF-7 cell cycle progression was investigated by fluorescence-activated cell sorting (FACS) analysis after treatment with DIM (10 or 30 μm) for 20 h. The percentage of cells in G₀/G₁ after treatment with Me₂SO, DIM-C-pPhCF₃ (10 μm), and DIM-C-pPhCF₃ (30 μm) was 56.5%, 76.6%, and 90.8%, respectively, and the percentage in S phase was 43.2%, 23.1%, and 9.0%, respectively (Fig. 5E). Thus, DIM-C-pPhCF₃ inhibited G₀/G₁-S phase progression of MCF-7 cells, and this is consistent with previous studies on the effects of PGJ2 in MCF-7 cells (35). One of these compounds (DIM-C-pPhC₆H₅) was also investigated as a mammary tumor growth inhibitor in 7,12-dimethylbenz(a)anthracene-induced female Sprague-Dawley rats, and significant (P < 0.05) inhibition was observed at a dose of 1.0 mg/kg every second day (×10; Fig. 5F). Ongoing studies in rodents with C-substituted DIMs and ring-substituted DIMs (31–33) indicated that total doses as high as 300 mg/kg did not cause changes in body or organ weights or organ histopathology. This compound was selected for the initial in vivo studies due to its activity as a PPARγ agonist (Fig. 1) and chemical stability (resistance to photodegradation). Inhibition of breast cancer cell growth, decreased G₀/G₁-S phase progression, and tumor growth inhibition by C-substituted DIMs are consistent with their activity as PPARγ agonists in mammary cancer cells.

Treatment of MCF-7 cells with Me₂SO or different concentrations of DIM-C-pPhC₆H₅ shows that the growth inhibitory effects of this compound (Fig. 6) are accompanied by apoptosis. PARP is a substrate for activated caspases (55), and incubation of MCF-7 cells for 36 h with DIM-C-pPhC₆H₅ (5 or 10 μm) induced cleavage of PARP to give the characteristic 85-kDa degradation product (Fig. 6A). Results in Fig. 6B show that DIM-C-pPhC₆H₅ (10 μm)-induced cleavage of PARP was not inhibited by IETD-CHO (20 μm) or LEHD-CHO (20 μm), which specifically inhibit the apical caspase-8 and caspase-9, respectively. The same concentrations of caspase inhibitors block cycloheximide (CHX)- and MG132-induced cleavage of PARP in MCF-7 cells. Both of these compounds induce apoptosis and activation of caspase-8 and caspase-9 (56, 57). Results indicate that DIM-C-pPhC₆H₅-induced cleavage of PARP was not dependent on activation of caspase-8 or caspase-9. Induction of PARP cleavage by DIM-C-pPhC₆H₅ was also accompanied by increased DNA laddering (Fig. 6C), and the pan-caspase inhibitor z-VAD-fmk significantly protected from DIM-C-pPhC₆H₅-induced PARP cleavage (data not shown) and cell detachment (Fig. 6D; 58, 59). Inhibition of DIM-C-pPhC₆H₅-induced apoptosis by T007, and structurally related PPARγ antagonists gave inconclusive results because the compounds alone induced apoptosis (data not shown). Results demonstrate that C-substituted DIMs induce apoptosis in MCF-7 cells and this response is typically observed for PPARγ agonists in other cancer cell lines (reviewed in 4–6).

Treatment of MCF-7 cells with PGJ2 and ciglitazone inhibited G₀/G₁-S phase progression, and this was accompanied by decreased expression of cyclin D1 and ERα proteins through activation of proteasomes (35). Results in Fig. 7A show that both PGJ2 and DIM-C-pPhCF₃ induce a concentration-dependent decrease in cyclin D1 and ERα protein expression in whole cell extracts after treatment with both compounds for 8 h. Decreased levels of cyclin D1 and ERα proteins were observed in cells treated with DIM-C-pPhCF₃ (concentrations between 10 and 20 μm), whereas levels of immunoreactive PPARγ and Sp1 proteins were unchanged (Fig. 7A). In contrast, levels of other cell cycle regulatory proteins including cyclin A, cyclin E, p21, and p27 were unaffected by this treatment. A time course study using relatively high concentrations of PGJ2 (30 μm) and DIM-C-pPhCF₃ (30 μm; Fig. 7B) showed that cyclin D1 and ERα protein levels are decreased as early as 1–3 and 3–15 h after treatment with the PPARγ agonists, respectively. Treatment of MCF-7 cells for a longer time period (20 h) showed that DIM-C-pPhCF₃-induced degradation of cyclin D1 and ERα was observed at concentrations of 10–12.5 μm (Fig. 7C). Results in Fig. 7D show that PGJ2-induced down-regulation of cyclin D1 and ERα was blocked by the proteasome inhibitor MG132 but not by the protease inhibitor CII (Fig. 7D). Similar results were observed in MCF-7 cells cotreated with DIM-C-pPhCF₃ plus these inhibitors, whereas in solvent (Me₂SO)-treated cells, MG132 and CII had minimal effects on ERα and cyclin D1 levels. These studies used the short time period (8 h) and higher concentrations of PGJ2 (30 μm) and DIM-C-pPhCF₃ (30 μm) because MG132 is cytotoxic after incubation for longer time periods. Thus, both PGJ2 and DIM-C-pPhCF₃ induced proteasome-dependent degradation of ERα and cyclin D1 in MCF-7 cells, and this was also observed in a recent study with ciglitazone and PGJ2 (35).

There are several possible mechanisms of PGJ2-induced down-regulation of cyclin D1, and these include PPARγ-independent pathways such as those observed for cyclopentenone, prostaglandin A2, and PGJ2 in breast cancer cells (60–62). Treatment of MCF-7 cells with PGJ2 (15 μm) alone or after cotreatment with actinomycin D (0.5 or 5.0 μm; Fig. 7E) resulted in decreased cyclin D1 and ERα protein levels, and similar results were observed for DIM-C-pPhCF₃ (data not shown). These data suggest a post-transcriptional mechanism of PGJ2-induced down-regulation of cyclin D1 and ERα and do not exclude the reported PPARγ-dependent competition for limiting levels of p300 as a mechanism for repression of cyclin D1 (25). However, results illustrated in Fig. 7F show that the PPARγ inhibitor T007 did not block down-regulation of cyclin D1 or ERα in cells treated with DIM-C-pPhCF₃ or PGJ2, suggesting that both compounds act through PPARγ-independent pathways.

The functional similarities between DIM-C-pPhCF₃ and PGJ2 were investigated by comparing their physical properties using HyperChem and by showing their interactions with the ligand binding domain of PPARγ (7, 40) using atomic coordinates downloaded from the PDB. Although the surface areas of DIM-C-pPhCF₃ (440.96 Ų) and PGJ2 (689.87 Ų) were different, their atomic volumes (1028.83 and 1096.81 ų, respectively) and octanol/water partition coefficients (log P = 5.30 and 5.30, respectively) showed minimal differences. Both DIM-C-pPhCF₃ and PGJ2 are readily accommodated by the Y-shaped ligand binding pocket of PPARγ. Future studies will investigate interactions of this new class of PPARγ agonists using X-ray crystallography to further define the overall structure of the ligand-bound complex and design new structurally related PPARγ agonists.

DIM is a metabolite of indole-3-carbinol, a phytochemical that contributes to the anticarcinogenic activity of cruciferous vegetables (63–66). DIM and related compounds bind the AhR (67), and research in this laboratory has characterized inhibitory AhR-ERα cross-talk in breast cancer cells and the anticarcinogenic activities of DIM and several ring-substituted DIMs in the rat mammary tumor model (31–34). Various acid-catalyzed condensation products derived from indole-3-carbinol bind the AhR and ERα (31–34, 67–69), and other structurally diverse indoles exhibit activities as cancer cell growth inhibitors, tubulin inhibitors, kinase inhibitors, and ligands for other receptors (70–75). For example, substituted 2-arylindoles bind the human neurokinin-1 receptor (74) and substituted indole-5-carboxylic acids activate PPARγ (75).

Introduction of a bulky substituted phenyl group on the methylene bridge of DIM eliminated the AhR agonist activity of these compounds in transactivation assays (data not shown). However, like DIM/ring-substituted DIMs, the C-substituted phenyl analogues inhibited growth of MCF-7 cells (Fig. 5); DIM-C-pPhC₆H₅ also inhibited MCF-7 cell proliferation and growth of carcinogen-induced mammary tumors in female Sprague-Dawley rats (Fig. 5). Results of transactivation assays using Gal4-receptor chimeras for receptors that bind lipophilic compounds showed that DIM-C-pPhCF₃ activated PPARγ and not other NHRs. Structure-activity studies among a series of 1,1-bis(3′-indolyl)-1-(p-substituted phenyl)methane analogues identified several compounds that were PPARγ agonists. Although the p-substituent on the C-phenyl markedly affected transactivation, the specific molecular determinants required for maximal activity were not apparent. For example, compounds containing both electron-donating (t-butyl) and electron-withdrawing (CN and CF₃) p-substituents were active in transactivation assays (Fig. 1), and molecular volumes of substituents were also not predictors of activity. These results are consistent with previous studies showing that PPARγ binds structurally diverse endogenous and synthetic compounds including fatty acids, prostaglandins, flavonoids, substituted terpenoids, thiazolidinediones, and substituted l-tyrosine and l-leucine derivatives (reviewed in 1–7). Molecular modeling studies show that the physicochemical properties of DIM-C-pPhCF₃ and PGJ2 are similar, and both compounds interact with the ligand binding pocket of PPARγ. Compared with the synthetic thiazolidinediones, DIM-C-pPhCF₃ was significantly more potent as an inducer of PPARγ-dependent transactivation in MCF-7 cells (Fig. 1B).

Recent studies have demonstrated that T007 is a specific PPARγ antagonist that inhibits PPARγ-induced transactivation (41). T007 covalently interacts with Cys³¹³ in helix 3 of human PPARγ and this affects the conformation of helix 12 and subsequent interactions with coactivators and coactivator-derived peptides. T007 also inhibits DIM-C-pPhCF₃- and PGJ2-induced transactivation in MCF-7 cells (Fig. 2), and this is consistent with the PPARγ agonist activity of PGJ2 and the new C-substituted DIMs. Moreover, using an RNA interference assay with iPPARγ, we could also show that transfected iPPARγ inhibits PGJ2- and DIM-C-pPhCF₃-dependent activation of PPARE3Luc via decreased expression of PPARγ protein in whole cell extracts (Fig. 2, C–E). It was reported previously that PGJ2, but not troglitazone or 9-hydroxyoctadienoic acid, induced transactivation in COS-1 cells transfected with chimeric VP-PPARγ and Gal4-SRC1, Gal4-TIFII, Gal4-AIB1, or Gal4-TRAP220 fusion proteins (54). In contrast, using the same two-hybrid assay and Gal4-coactivator constructs in MCF-7 cells, PGJ2 induced transactivation in cells transfected with Gal4-SRC1, Gal4-TIFII, and Gal4-TRAP220 (partial) but not Gal4-AIB1 (Fig. 4). Thus, PGJ2-induced coactivator-PPARγ interactions are cell context dependent, and it is noteworthy that AIB1, which is highly expressed in breast cancer cells (76–78), does not interact with VP-PPARγ in MCF-7 cells treated with PGJ2, ciglitazone (data not shown), or DIM-C-pPhCF₃. The major difference between PGJ2 and C-substituted DIMs was observed with PGC1 in which only the latter compounds induced PGC1-PPARγ interactions in MCF-7 cells.

PGJ2 and other PPARγ agonists typically inhibit cancer cell growth, and this is accompanied by several responses including inhibition of G₁-S phase progression, cyclin D1 down-regulation, and increased differentiation and/or apoptosis, and comparable responses were observed for C-substituted DIMs (8–27, 35). C-substituted DIMs inhibit MCF-7 cell proliferation, G₁-S phase progression, and mammary tumor growth in carcinogen-induced female Sprague-Dawley rats (Fig. 5). DIM-C-pPhC₆H₅ was among the most active compounds in the in vitro and in vivo cell proliferation assays, and results in Fig. 6 summarize the effects of DIM-C-pPhC₆H₅ on apoptosis in MCF-7 cells 36 h after treatment. Results show that caspase-dependent cleavage of PARP and DNA laddering were observed at the 5 and 10 μm concentrations of DIM-C-pPhC₆H₅, and the broad-spectrum caspase inhibitor z-VAD-fmk only slightly inhibited DIM-C-pPhC₆H₅-induced cell death. MCF-7 cells do not express caspase-3 (79, 80); however, several agents including hydrogen peroxide, tumor necrosis factor/CHX, and paclitaxel also induce PARP cleavage in MCF-7 cells (80–82). Using specific inhibitor (LEHD-CHO/IETD-CHO) and pan-caspase inhibitor (z-VAD-fmk), our results suggest that DIM-C-pPhC₆H₅-induced apoptosis is not dependent on activation of caspase-8 or caspase-9 through the death receptor or mitochondrial pathways, respectively. Although z-VAD-fmk partially blocks DIM-C-pPhC₆H₅-induced apoptosis (Fig. 6D), results suggest that caspase-independent pathways may be important, and these are currently being investigated.

PPARγ agonist-induced down-regulation of cyclin D1 is associated with inhibition of G₀/G₁-S phase progression in MCF-7 and other cell lines (25, 35, 60–62, 83). Results shown in Fig. 7 demonstrate that, like PGJ2, DIM-C-pPhCF₃ also decreased cyclin D1 and ERα proteins in MCF-7 cells within 1–3 h after treatment. Moreover, down-regulation of both proteins was reversed by proteasome (MG132) but not by protease (CII) inhibitors. Similar results were obtained for other C-substituted DIMs (data not shown) and ciglitazone (35) in MCF-7 cells. Actinomycin D did not block DIM-C-pPhC₆H₅- or PGJ2-induced down-regulation of cyclin D1 or ERα proteins, suggesting that these responses were associated with post-transcriptional pathways. The PPARγ inhibitor T007 also did not affect down-regulation of cyclin D1 or ERα by PGJ2 or DIM-C-pPhCF₃ (Fig. 7F), indicating that this response was PPARγ independent as reported previously for PGJ2 in MCF-7 cells (60–62). Current studies are investigating the PPARγ-independent mechanisms of cyclin D1 and ERα degradation by C-substituted DIMs, other synthetic PPARγ agonists, and PGJ2.

In summary, this study shows that C-substituted DIMs are a novel class of PPARγ agonists that induce PPARγ-dependent transactivation and PPARγ-coactivator interactions in MCF-7 cells. C-substituted DIMs also induce responses observed for PGJ2 and other PPARγ agonists, and these include apoptosis, inhibition of cancer cell growth, inhibition of G₀/G₁-S phase progression, and enhanced Oil Red O staining in 3T3-L1 preadipocytes. Both PGJ2 and C-substituted DIMs induce proteasome-dependent down-regulation of cyclin D1 and ERα in MCF-7 cells. This response was PPARγ independent and consistent with results of previous studies with PGJ2 in MCF-7 cells (60–62). Currently, we are adapting other RNA interference methods (84, 85) for high efficiency and stable suppression of PPARγ in breast and other cancer cell lines to further investigate PPARγ-dependent and -independent pathways induced by C-substituted DIMs, PGJ2, and thiazolidinediones.