Abstract
Peroxisome proliferator-activated receptor-γ (PPARγ) is a member of the nuclear hormone superfamily and has multiple endogenous and pharmacological ligands, including 15-deoxy-Δ 12,14-prostaglandin J2 and two thiazolidinediones (TZD), rosiglitazone and pioglitazone, which are used clinically to treat type-2 diabetes mellitus. PPARγ agonists regulate development, cellular growth and metabolism in various tissues and have been documented to decrease cellular proliferation and to induce apoptosis of various tumour phenotypes, including breast cancer. However, the full spectrum of anti-tumour effects occurs only at suprapharmacological doses. In this study, we investigated the mechanism of rosiglitazone-induced anti-tumour effects of MDA-MB-231 human breast cancer cells, and used that information to predict rosiglitazone-induced sensitization of breast cancer cells to the effects of other compounds. We first confirmed that 100 μM rosiglitazone, but not lower doses, decreases MDA-MB-231 cell viability in vitro. We then used microarray gene expression analysis to determine early rosiglitazone-induced gene expression changes after 4-h exposure, which included 1298 genes that we grouped into functional categories. We selectively confirmed rosiglitazone-mediated effects on expression of key regulators of breast cancer proliferation and apoptosis, including p53, p21 and Bax. Finally, we used this information to predict that rosiglitazone would sensitize MDA-MB-231 cells to the anti-tumour effects of CH11, which trimerizes Fas, as well as tumour necrosis factor-α. Moreover, we used the confirmed array data to predict cooperative activity of rosiglitazone and R-roscovitine (CYC202), an inhibitor of multiple cyclin-dependent kinases. We conclude that microarray analysis can determine early TZD-modulated changes in gene expression that help to predict effective in vitro drug combinations.
Introduction
Peroxisome proliferator-activated receptor-γ (PPARγ), a member of the nuclear hormone superfamily, is expressed in many tissues. PPARγ has endogenous and pharmacological ligands including 15-deoxy-Δ12,14-prostaglandin J2 and two thiazolidinediones (TZD), rosiglitazone and pioglitazone, which enhance insulin sensitivity and are used clinically to treat type-2 diabetes mellitus ( Theocharis et al. 2004). Ligand binding induces heterodimerization with retinoid X receptor, and the resultant complex transcriptionally regulates many genes ( Theocharis et al. 2004). However, recent evidence suggests that non-receptor effects of PPARγ agonists may be biologically important ( Clay et al. 2002, Shiau et al. 2005). PPARγ agonists regulate development and cellular growth and metabolism in various tissues. In particular, they mediate adipocyte differentiation, which has led to an examination of their potential anti-tumour effects. PPARγ agonists have been documented to decrease cellular proliferation and to induce apoptosis of various tumour phenotypes, including breast cancer ( Nahle 2004, Theocharis et al. 2004). However, anti-tumour effects occur only at suprapharmacological doses; this has been a source of criticism, even though doses up to 1500 times those used clinically in humans were well tolerated in mouse tumour models ( Heaney et al. 2002, 2003). Regardless, investigators are reluctant to use very high doses in clinical trials. Therefore, it was our aim to examine, in vitro, a short-course of high-dose rosiglitazone treatment, and to develop a rational approach to predict how rosiglitazone might enhance the activity of other agents, and/or how rosiglitazone might be used for short periods of time in combination therapy strategies.
We investigated the mechanism of rosiglitazone-induced anti-tumour effects on ER-negative MDA-MB-231 human breast cancer cells, and used that information to determine rational combination therapies that include roziglitazone and other agents. We first confirmed a rosiglitazone dose that inhibited viability of MDA-MB-231 cells in vitro. We then used microarray gene expression analysis to determine very early rosiglitazone-induced primary gene expression changes. We confirmed expression changes of key regulators of breast cancer proliferation and apoptosis. Finally, we used this information to predict rosiglitazone-mediated sensitization of MDA-MB-231 cells to apoptosis by CH11, which trimerizes Fas, and tumour necrosis factor-α (TNFα). Moreover, we used the confirmed array data to predict synergistic activity of rosiglitazone and R-roscovitine (CYC202), an inhibitor of multiple cyclin-dependent kinases (CDK).
Materials and methods
Reagents
A stock solution of 14.3 mM rosiglitazone (Cayman Chemicals, Ann Arbor, MI, USA) in DMSO was stored at −20 °C and a working solution was prepared in the culture medium just before use. CH11 was obtained from Cell Signaling (Beverly, MA, USA). TNFα was obtained from Sigma–Aldrich, CYC202 was obtained from Cyclacel Ltd (Dundee, Scotland, UK) and GW9662 was obtained from EMD Chemicals, Inc. (Gibbstown, NJ, USA). PPRE-x3-TK-LUC reporter vector that contains three copies of a consensus PPAR-response element (PPRE) upstream of the thymidine kinase promoter-luciferase fusion gene, was kindly provided by Dr Bruce Spiegelman (Harvard University, Cambridge, MA, USA).
Cell culture
MDA-MB-231 human breast cancer cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C in a humidified incubator containing 5% CO2. Cell numbers were determined using a Z2 Coulter Counter (Beckman Coulter, Fullerton, CA, USA).
Assays of cellular viability and DNA synthesis
Cellular viability was measured by 3-(4,5-dimethylthia-zol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay. MDA-MB-231 cells were seeded in multiplicate (n = 6) at 8 × 103 cells per well of a 96-well plate in DMEM. After overnight incubation, cells were replenished with fresh media containing 0.1, 1, 10, and 100 μM rosiglitazone and incubated for 4 or 6 days. Media containing drugs or vehicle were changed on alternate days. The effect of rosiglitazone on DNA synthesis was carried out using a BrdU Cell Proliferation ELISA kit purchased from Roche following the manufacturer’s instruction.
The effect of rosiglitazone on sensitizing MDA-MB-231 cells to CH11, TNFα or CYC202 was assessed using the MTT assay. MDA-MB-231 cells were seeded in triplicate at 1 × 104 cells per well of a 24-well plate in DMEM. After overnight incubation, cells were replenished with fresh media containing 1.4 or 100 μM rosiglitazone and incubated for the indicated time periods. A dose of 1.4 μM is analogous to the peak plasma concentration of an 8 mg dose of rosiglitazone given to treat diabetes; i.e. Cmax = 598 ± 117 ng/ml. The molecular weight of rosiglitazone equals 357.44; so the molar concentration in serum equals 1.4 μM.
Cells were pretreated with 100 μM rosiglitazone for 24 h followed by exposure to 50 ng/ml CH11 or 20 ng/ml TNFα for an additional 48 h. Control cells received vehicle (DMSO) at a concentration equal to that of the drug-treated cells. Media containing drug or vehicle was changed on alternate days. For readings at each time point, cells were incubated with fresh media containing 1 mg/ml MTT (Sigma–Aldrich). Following 2-h incubation at 37 °C, media containing MTT was removed and 200 μl of vehicle was added per well. Cells were incubated at room temperature for 20 min with rotation and then transferred to a 96-well plate. The absorbance was determined at 570 nm with a microtiter plate reader.
MDA-MB-231 cells were also subjected to simultaneous treatment with 100 μM rosiglitazone plus vehicle (DMSO), or vehicle (DMSO) plus 20 nM CYC202, or 100 μM rosiglitazone plus 20 nM CYC202 for 96 h. Media containing drug or vehicle was changed on alternate days. At 96 h, MTT assays were performed as previously described.
Western blot analysis
MDA-MB-231 cells were seeded at 1 × 106 cells per well in a six-well plate in DMEM. After treatment with 100 μM rosiglitazone for the indicated time periods, cells were washed twice in cold 1 × PBS, then scraped and collected in chilled lysis buffer (50 mM Tris–HCl (pH 7.5), 120 mM NaCl, 5 mM EDTA, 0.5% NP40, 0.2 mM Na3VO4, 1 mM dithiothreitol and protease inhibitor cocktail). Cells were lysed by gently vortexing for 5 s every 5 min for a total of 20 min; cells were kept cold on ice in between vortexing. After centrifugation at 10 000 g for 5 min at 4 °C, the supernatant (lysate) was transferred to a fresh tube. Protein concentration was then determined using the Bradford Assay (Bio-Rad). Total extract (50 μg) was mixed with 2 × Laemmli sample buffer (Bio-Rad) and boiled for 5 min. Samples were run on 4–20% Tris–HCl precast gel (Bio-Rad) along with Precision Plus Kaleidoscope Protein Standards (Bio-Rad). After electrophoresis at 30 mA, proteins were transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA) at 100 V for 1 h at 4 °C. The membrane was blocked with PBS/0.1% Tween-20 (PBST) containing 5% non-fat dry milk for 1 h at room temperature and washed three times for 5 min each in PBST. The membrane was then incubated with the appropriate primary antibody overnight at 4 °C. Bax rabbit polyclonal antibody (Cell Signaling Technology, Danvers, MA, USA) was diluted in the ratio of 1:1000 in PBST containing 5% non-fat dry milk; and p53 mouse monoclonal antibody (Cell Signaling) was diluted in the ratio of 1:2000 in PBST containing 5% BSA. The membrane was washed three times for 5 min each in PBST and then incubated with HRP-conjugated goat anti-mouse IgG (Pierce, Rockford, IL, USA) diluted in the ratio of 1:2000 or HRP-conjugated goat anti-rabbit (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted in the ratio of 1:10 000 in PBST containing 5% non-fat dry milk for 1 h at room temperature. The immunoblot was visualized by chemiluminescence using SuperSignal West Dura Extended Duration Substrate (Pierce) on a Kodak 2000MM Multimodal Imaging System.
Quantitative RT-PCR
Total RNA was isolated using Trizol reagent (Invitrogen) following the manufacturer’s instructions. Total RNA (1 μg) was digested using DNase I (Invitrogen). The DNase-treated RNA was then reverse transcribed using iScript cDNA Synthesis kit (Bio-Rad) which contains a blend of oligo (dT) and random hexamer primers under the following conditions: 25 °C for 5 min; 42 °C for 30 min; and 85 °C for 5 min. Triplicate samples of cDNA were amplified by PCR using the ABI Prism 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The amplification reaction mixture (25 μl total volume) contained 1 μl cDNA, 20 pmol forward primer (p21 F primer CCT AAT CCG CCC ACA GGA A, Bax F primer ATG GAC GGG TCC GGG GAG CAG CCC, and 18S F primer GCC GCT AGA GGT GAA ATT CTT G), 20 pmol reverse primer (p21 R primer AAG ATG TAG AGC GGG CCT TTG, Bax R primer GGT GAG CAC TCC CGC CAC AAA GAT, and 18S R primer CAT TCT TGG CAA ATG CTT TCG), 12.5 μl iTaq SYBR Green Supermix (Bio-Rad), and 9.5 μl DEPC-H2O. RRNA (18S) was used as an endogenous control to normalize the quantification of the target transcripts in each sample.
Luciferase assay
To test for transcriptional activity of PPARγ, cells were seeded at a density of 2.5 × 105 cells per well of a six-well plate in triplicate 1 day prior to transfection. Cells were transiently co-transfected with PPRE-x3-TK-LUC reporter (2 μg) and β-galactosidase vector (1 μg) using FuGENE6 Transfection Reagent (Roche Diagnostics) according to the manufacturer’s instructions. After 12 h, the FuGENE6-DNA mix-containing media was replaced with fresh media and incubated for another 24 h. This was followed by treatment with rosiglitazone or vehicle for the indicated time periods and concentrations. Cells were then harvested and lysed with 1 × Reporter Lysis Buffer (Promega). Luciferase and -β-galactosidase activity were measured using the Luciferase Assay System and β-Galactosidase Enzyme Assay System respectively (Promega). Luciferase activities were normalized against β-galactosidase levels to adjust for transfection efficiency.
Microarray analysis
Sample preparation and processing
MDA-MB-231 cells were seeded at 4.0 × 105 cells per 25 cm2 flask in three replicates each for vehicle and rosiglitazone treatment. The following day, cells were treated with 100 μM rosiglitazone or vehicle for 4 h. Total RNA was isolated using Trizol reagent (Invitrogen) and subsequently purified with RNeasy Mini Kit (Qiagen). Sample quality was assessed on RNA 6000 Nano LabChip using the Bioanalyzer 2100 system (Agilent Technologies, Palo Alto, CA, USA). Only samples that yielded high quality RNA with minimal degradation and sharp, clear 18S/28S ribosomal peaks were chosen for microarray analysis, which was performed in triplicate for vehicle and rosiglitazone treated samples. In vitro transcription (IVT), microarray hybridization and scanning were performed according to Affymetrix protocols. Briefly, total RNA from each sample was synthesized into double-stranded cDNA by reverse transcription. IVT reactions were performed on cDNA to generate biotinylated cRNA targets. These were then chemically fragmented and hybridized overnight at 45 °C to Affymetrix Human Genome Focus Array chips, which represent over 8500 well-described human genes. Hybridized chips were washed and stained with streptavidin–phycoerythrin using the Affymetrix Fluidics Station 400 and scanned with an Affymetrix GeneChip 3000 confocal scanner. Fluorescent images from scanned chips were processed using Affymetrix GeneChip Operating System.
Microarray data analysis
Comparative analysis between expression profiles for vehicle- and rosiglitazone-treated samples was performed using GeneSpring software v7.0 (Agilent Technologies). Gene expression data were normalized, by selecting ‘per chip normalization’. In this case, all expression data on a chip were normalized to the 50th percentile of all values on that chip. The ‘cross gene error model for deviation from 1.0′ was inactive.
There were three replicates per group (vehicle versus rosiglitazone). The genes were filtered on the basis of flag value ‘present’ in all six samples. This criterion allowed for selection of genes based on the data quality flags in the original files. The filtered gene list was subsequently used to identify genes that were differentially expressed between vehicle and rosiglitazone-treated cells using a one-way ANOVA parametric test with a false discovery rate of 0.05. The expression of each gene was reported as the ratio of the value obtained after treatment relative to the control after normalization of the data.
Results
Rosiglitazone-mediated Inhibition of cell viability and DNA synthesis
In order to determine the rosiglitazone dose that inhibited MDA-MB-231 cell viability, cells were treated with 0.1, 1, 10 or 100 μM rosiglitazone and MTT assay was carried out after 4 and 6 days. A significant decrease in viability was observed only when cells were treated with a 100 μM rosiglitazone (Fig. 1a ).
Other groups noted decreases in cell proliferation as measured by thymidine H3 incorporation with 10 μM rosiglitazone; we, therefore, evaluated the effects of 0.1, 1, 10 and 100 μM rosiglitazone on DNA synthesis as measured by BrdU incorporation. We showed that the lower doses decreased DNA synthesis, but the effect was not dose-dependent and was much smaller than the effect of 100 μM rosiglitazone (Fig. 1b ). Moreover, these effects were seen only after 4 and 6 days of treatment and not at earlier time points.
Transcriptional activation of a PPARγ response element by rosiglitazone
PPARγ is expressed in MDA-MB-231 cells. To assess the functional activity of PPARγ, MDA-MB-231 cells were transfected with a reporter construct (PPRE-x3-TK-LUC) that contains three copies of PPRE upstream of the firefly luciferase gene ( Hortobagyi 1998, Lu et al. 2005). In Fig. 2a cells were exposed to vehicle, 100 μM rosiglitazone, 100 μM GW9662, or 100μM rosiglitazone plus 100 μM GW9662. Vehicle-treated cells exhibited only basal luminescence; however, PPARγ promoter activity was significantly induced after 4-h treatment with 100 μM rosiglitazone but not with GW9662. Furthermore, GW9662 abrogated the rosiglitazone-mediated transactivation indicating the effect was specific. Moreover, 1.4 μM rosiglitazone was unable to significantly enhance PPRE activity (Fig. 2b ).
Rosiglitazone-induced differential gene expression
The Affymetrix Human Genome Focus Array, which represents over 8500 verified human sequences, was used to identify genes induced or repressed in MDA-MB-231 cells treated with 100 μM rosiglitazone, the dose that attenuated cell viability. A short treatment time (4 h) was used with the intent of identifying direct PPARγ target genes and to avoid detecting the later, secondary effects of protein changes induced by rosiglitazone. By filtering the data based on flag value ‘present’ in all of the replicates analysed, 4002 out of 8793 genes passed the filter in MDA-MB-231 cells. After performing one-way ANOVA with a false discovery rate of 0.05 on the filtered data set, a total of 1298 genes were identified as being differentially expressed between vehicle and rosiglitazone treatment in MDA-MB-231 cells.
To assess the biological process or function of the genes that showed a significant alteration of expression level, the expression analysis systematic explorer (EASE) classification system database (http://david.niaid.nih.gov/david/ease.htm) was queried using the gene list derived from our microarray analysis. We focused our attention on categories that were relevant to this study (i.e. cell cycle, cell growth/maintenance, cell communication, regulation of transcription, apoptosis and cell proliferation). We reasoned that this approach would allow us to determine an association between the altered gene set and the functional pathways that might be involved in rosiglitazone-mediated apoptosis/growth inhibition. Based on this analysis, Table 1 lists genes that are overexpressed (≥1.25-fold, P≤0.05) and underexpressed (≤0.75-fold, P≤0.05) after rosigli-tazone treatment relative to the vehicle. A comprehensive list of statistically significant gene changes can be found in the supplemental data.
Validation of selected targets by western blot and quantitative RT-PCR
Microarray analysis of MDA-MB-231 cells established p53, p21 and Bax as potential target genes that were upregulated by rosiglitazone that are involved in tumour suppression and/or apoptosis. Therefore, we focused on these three genes for validation studies. Protein levels were assessed by western blot for later time points than were used for the microarray screening in order to allow enough time for translation to occur. MDA-MB-231 cells treated with 100 μM rosiglitazone for 24 and 48 h demonstrated an increase in p53 and p21 protein levels when compared with vehicle-treated cells (Fig. 3a ). However, no change in the Bax protein level was detected. Bax protein levels were then analysed 4- and 8-h post-treatment, which also failed to show a change in expression (Fig. 3b ). Therefore, Bax mRNA expression was analysed by real-time RT-PCR at 4- and 24-h post-treatment. Results indicate an increase in expression by 1.43-fold at 4 h, closely correlating with the array data; however, no change in mRNA levels was detected at 24 h (data not shown).
Rosiglitazone sensitizes MDA-MB-231 cells to CH11 and TNFα
Since rosiglitazone-treated cells demonstrated higher levels of p53 and p21 protein, we hypothesized that rosiglitazone could sensitize MDA-MB-231 cells to agents that also act on these apoptosis/cell cycle pathways. MDA-MB-231 cells were pretreated with 100 μM rosiglitazone for 24 h followed by the treatment with either CH11 or TNFα for an additional 48 h at doses that are known to induce apoptosis ( Lejeune et al. 2006, Wajant 2006). This resulted in enhanced growth inhibition when compared with treatment with any of the agents alone. Rosiglitazone followed by CH11 treatment resulted in 42% cell death, whereas vehicle followed by CH11 treatment inhibited cell growth by only 1% (Fig. 4 ). Rosiglitazone followed by TNFα treatment resulted in a 45% cell death, while vehicle followed by TNFα inhibited cell growth by only 5%. Rosiglitazone followed by vehicle treatment inhibited growth by only 17% (Fig. 5a ). Since the ultimate goal of these studies is to determine whether a short exposure and/or lower doses of TZD can ultimately be used clinically, we evaluated whether 1.4 μM rosiglitazone was able to enhance the effects of TNFα on DNA synthesis. As seen in Fig. 5b , 1.4 μM rosiglitazone was unable to augment the effect of TNFα.
Rosiglitazone enhances CYC202 effects
MDA-MB-231 cells co-treated with 100 μM rosiglitazone and 20 nM CYC202 demonstrated reduced viability when compared with either agent alone. Rosiglitazone plus CYC202 demonstrated a 91% inhibition of growth by 96 h when compared with 45 and 21% for CYC202 plus vehicle and rosiglitazone plus vehicle respectively (Fig. 6a ). Moreover, 1.4 μM rosiglitazone was not able to enhance the effects of CYC202 on DNA synthesis (Fig. 6b ).
Discussion
PPARγ agonists facilitate adipocyte differentiation when combined with other agents. Several lines of evidence indicate that fibroblasts differentiate into adipocytes when exposed to glucocorticoid plus either insulin or insulin-like growth factor I plus a PPARγ agonist ( Kletzien et al. 1992, Tontonoz et al. 1995). Hu et al.(1995) were able to transdifferentiate myoblasts into adipocytes using dexamethasone plus 5,8,11,14-eicosatraynoic acid (an endogenous PPARγ ligand) plus insulin. Several pathways regulate adipogenesis, including PPARγ-induced degradation of β-catenin during preadipocyte differentiation, and CCAAT/enhancer binding proteins (C/EBPβ), which induces C/EBPα, enhances PPARγ production, and is required for PPARγ-associated adiponectin expression during the terminal stages of adipogenesis ( Farmer 2005). In agreement, Wnt-1, an upstream activator of β-catenin, blocks adipogenesis in vitro ( Ross et al. 2000). These data indicate a complex, tightly regulated pathway where multiple agents are needed to fully induce differentiation. Therefore, we surmised that TZDs might be more effective anti-tumour agents in combination with other compounds.
Heaney et al.(2002, 2003) investigated rosiglita-zone’s anti-tumour effects using a pituitary tumour model. They demonstrated that PPARγ expression is restricted to normal pituitary corticotroph cells in rodents and humans, but is overexpressed in all pituitary tumour phenotypes. Furthermore, suprapharmacological doses of rosiglitazone significantly inhibited growth of incipient and established gonadotroph, somatotroph and corticotroph tumours in mice. In each case, rosiglitazone decreased the tumour-mediated hormone oversecretion ( Heaney et al. 2002, 2003). Because suprapharmacological doses of rosiglitazone were required, this suggests that 1) anti-tumour effects might occur only after saturation of endogenous PPARγ receptors, 2) co-treatment with other agents might potentiate TZD effects and 3) rosiglitazone doses up to 1500 times those used in humans are well tolerated by mice.
In this study, we confirmed that a high dose of rosiglitazone was required to inhibit MDA-MB-231 cell viability, but that lower doses could inhibit DNA synthesis over the course of 4 or 6 days. However, only 100 μM rosiglitazone was able to transactivate a PPRE reporter construct at 1 and 4 h, which are consistent with findings of other in vitro and in vivo studies ( Heaney et al. 2002, 2003, Sheu et al. 2006). One possibility for the requirement of very high doses to inhibit cell viability is that the endogenous PPARγ receptors, which are often overexpressed in cancer, may need to be saturated prior to intracellular levels of TZD rising sufficiently to mediate anti-tumour effects. Indeed, Seargent et al.(2004) demonstrated that the PPARγ antagonist GW9662 selectively occupied PPARγ receptors, yet enhanced rosiglitazone’s anti-tumour effects in breast cancer cells, supporting the concept that by binding to PPARγ, GW9662 permitted a higher intracellular level of rosiglitazone ( Seargent et al. 2004). Alternatively, the amount of TZD that can enter a tumour cell may be limited. Another possibility is that the molecular machinery required for growth inhibitory/apoptotic effects may be incomplete or less functional in breast cancer cells. Finally, the potency of TZDs as activators of PPARγ and their PPARγ-independent anti-tumour effects may not correlate. For example, rosiglitazone and troglitazone are both potent PPARγ activators; however, rosiglitazone’s anti-tumour effects are seen only at suprapharmacological doses, whereas troglitazone demonstrates some anti-tumour activity at pharmacologic doses ( Sharma et al. 2004). We attempted to simulate a clinically relevant dose (an 8 mg dose of rosiglitazone leads to a serum concentration of ~1.4 μM) and to increase the dose to 70 times of that level (100 μM). In contrast, Heaney et al. used 1500 times the maximal clinical dose in in vivo murine studies. Of course, direct exposure of cells in culture to rosiglitazone is not necessarily analogous to in vivo exposure, regardless of the dose.
In an effort to evaluate only primary direct PPARγ target gene changes, we performed a microarray analysis only after 4-h rosiglitazone treatment in order to avoid detecting secondary effects of protein changes induced by rosiglitazone, which then might further alter gene expression. A similar approach was taken by Gerhold et al.(2002) who investigated the gene expression profile of adipocyte differentiation by performing microarray analysis only after 6-h treatment with a PPARγ agonist. They used the Affymetrix platform to identify 579 out of 6347 genes, which were either up- or downregulated. Their findings support the notion that early genomic changes are diverse and robust. Our array analysis demonstrated that rosiglitazone altered expression of 1298 genes only after 4 h. These changes affected key pathways involved in growth and apoptosis. Based on EASE analysis, genes were grouped into functional categories. We focused on categories most relevant to anti-cancer effects, including cell cycle, growth/maintenance, communication, transcriptional regulation, apoptosis and proliferation.
We hypothesized that rosiglitazone would alter expression of genes that regulate cell death, thereby sensitizing cancer cells to known apoptosis-inducing drugs. Since several pro-apoptotic genes were upregulated in our microarray screen, we predicted cooperative interactions with agents that also act in the apoptotic pathway. We chose to focus on members of the TNF death-receptor family including TNFα and Fas, which induce apoptosis through pathways involving ligand-mediated receptor activation. TNFα and CH11 (which trimerizes the Fas receptor) have been shown to promote apoptosis in tumour cells ( Toillon et al. 2002, Butt et al. 2005). As seen in Fig. 4a and b , co-treatment with rosiglitazone plus CH11 or rosiglitazone plus TNFα resulted in greater cell death than treatment with either of the agents alone, but only at suprapharmacological doses.
Since rosiglitazone increased p53 and p21 protein levels (Fig. 3a ) and has previously been shown to induce G1/S phase arrest in a variety of cell lines, we aimed to determine whether co-treatment with a drug that arrested cells at the G2/M phase would potentiate the anti-tumour effects of rosiglitazone in breast cancer cells ( Artwohl et al. 2005, Gruszka et al. 2005, Sheu et al. 2006). As can be seen in Fig. 5 , co-treatment of MDA-MB-231 cells with rosiglitazone and CYC202 reduced proliferation greater than treatment with either drug alone, but again only at suprapharmacological doses. CYC202 is an ATP analogue that inhibits several cellular CDKs at various levels of sensitivities but has greatest effect against CDK2/cyclin E ( Gray et al. 1999, McClue et al. 2002). CYC202 has been shown to inhibit the growth of a variety of tumour cell lines, including B cell lymphomas, and breast, lung and colon cancer cell lines ( Mgbonyebi et al. 1999, McClue et al. 2002, Wojciechowski et al. 2003, Alvi et al. 2005). Treatment of tumour cell lines with CYC202 induced p53 expression resulting in growth arrest and apoptosis ( Mgbonyebi et al. 1999, Wojciechowski et al. 2003, Alvi et al. 2005). These results suggest that targeting two different cell cycle checkpoints simultaneously might be an effective therapeutic strategy in breast cancer. Moreover, CYC202 may be advantageous in co-treatment as it has proven effective against p53-null and p53-positive tumour cell lines ( Alvi et al. 2005). Furthermore, CYC202 had little toxicity in phase I clinical trials ( Fischer & Gianella-Borradori 2003).
In summary, we conclude that rosiglitazone induces diverse early genomic changes that include key regulators of apoptosis and cell cycle progression. In addition, these early genomic changes can be used to guide investigators to rationally combine rosiglitazone with other agents. Finally, early rosiglitazone-induced genomic changes have functional relevance and can predict some early protein expression changes.
List of upregulated and downregulated genes in rosiglitazone-treated MDA-MB-231 cells
| P value | Fold change | Common | Genbank | Description | |
|---|---|---|---|---|---|
| a. Upregulated genes Induction of apoptosis/apoptotic program | |||||
| 0.015 | 2.39 | BAD | U66879 | BCL2-antagonist of cell death | |
| 0.020 | 1.57 | DAP | NM_004394 | Death-associated protein | |
| 0.013 | 2.89 | BAK1 | NM_001188 | BCL2-antagonist/killer 1 | |
| 0.046 | 1.47 | BAX | NM_004324 | BCL2-associated X protein | |
| 0.038 | 3.91 | BCL2L1 | AL117381 | BCL2-like 1 (bcl-xL) | |
| 0.0002 | 1.55 | DEDD | AK022531 | Death effector domain containing | |
| 0.006 | 2.26 | DDX41 | NM_016222 | DEAD (Asp–Glu–Ala– Asp) box polypeptide 41 | |
| 0.0187 | 2.26 | TP53 | NM_000546 | Tumour protein p53 | |
| Programmed cell death | 0.015 | 3.01 | LTBR | NM_002342 | Lymphotoxin β receptor (TNFR superfamily, member 3) |
| Cell cycle regulation | 0.014 | 1.93 | NBL1 | D28124 | Neuroblastoma, suppression of tumourigenicity 1 |
| 0.034 | 1.54 | G0S2 | NM_015714 | Putative lymphocyte G0/G1 switch | |
| 0.014 | 1.93 | NBL1 | D28124 | Neuroblastoma, suppression of tumourigenicity 1 | |
| 8.39 × 10−6 | 4.16 | PPP2R1A | NM_014225 | Protein phosphatase 2 (formerly 2A), regulatory subunit Aα | |
| 0.032 | 1.82 | PPM1G | NM_002707 | Protein phosphatase 1G (formerly 2C), magnesium-dependent, γ | |
| 0.019 | 1.82 | CDK4 | NM_000075 | Cyclin-dependent kinase 4 | |
| 0.048 | 1.61 | CDKN1A | NM_000389 | Cyclin-dependent kinase inhibitor 1A (p21, Cip1) | |
| 0.00948 | 6.64 | TGFB1 | BC000125 | Transforming growth factor, β 1 | |
| 0.013 | 7.79 | RAB1B | NM_030981 | RAB1B, member RAS oncogene family | |
| 0.018 | 2.32 | PPP1R15A | U83981 | Protein phosphatase 1, regulatory (inhibitor) subunit 15A | |
| b. Downregulated genes Induction of apoptosis/apoptotic program | |||||
| 0.00263 | 0.66 | CASP3 | NM_004346 | Caspase 3 | |
| 0.00778 | 0.67 | CASP4 | U25804 | Caspase 4 | |
| 0.018 | 0.68 | DAP3 | U18321 | Death-associated protein 3 | |
| 0.00141 | 0.59 | APAF1 | NM_013229 | Apoptotic protease activating factor | |
| Cell cycle regulation | 0.014 | 0.69 | RAN | AF054183 | RAN, member RAS oncogene family |
| 0.033 | 0.61 | CCNL1 | NM_020307 | Cyclin L1 | |
| 0.018 | 0.65 | CCNC | AL137784 | Cyclin C | |
| 0.019 | 0.65 | ANAPC10 | NM_014885 | Anaphase promoting complex, subunit 10 | |
| 0.044 | 0.61 | CETN3 | BC005383 | Centrin | |
| 0.00285 | 0.65 | CDK7 | L20320 | Cyclin-dependent kinase 7 | |
| Cell proliferation | 0.00563 | 0.68 | STAMBP | NM_006463 | STAM-binding protein |
| 0.047 | 0.73 | FGFR1OP | NM_007045 | FGFR1 oncogene partner | |
| 0.00228 | 0.50 | NRP1 | BE620457 | Neuropilin 1 | |
Effect of rosiglitazone on cellular viability and DNA synthesis. (a) MDA-MB-231 cells were treated with vehicle, 0.1, 1, 10 or 100 mM rosiglitazone and cell viability was measured by MTT assay at 4 and 6 days. Data are expressed as the mean of three independent experiments and compared using Student’s t-test. *P<0.05, 100 μM Rosiglitazone vs DMSO control. **P<0.01, 100 μM Rosiglitazone vs DMSO control. (b) MDA-MB-231 cells were treated with vehicle, 0.1, 1, 10 or 100 μM rosiglitazone and DNA synthesis was measured by BrdU ELISA assay at 4 and 6 days. Data are expressed as the mean of three independent experiments and compared using Student’s t-test. *P<0.01, 100 μM Rosiglitazone vs DMSO control.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
Rosiglitazone-mediated PPRE transactivation. (a) Cells were transiently transfected with a PPRE-luciferase reporter construct and treated with vehicle, 100 μM rosiglitazone, 100 μM GW9662, or both for the indicated time periods before assaying for promoter activity. Promoter activity is represented as relative luciferase activity after normalization with β-galactosidase. Data are expressed as the mean of three determinants from two independent experiments. Data were compared using Student’s t-test. (b) Cells were transiently transfected with a PPRE-luciferase reporter construct and treated with vehicle, 1.4 μM rosiglitazone, 1.4 μM GW9662, or both for the indicated time periods before assaying for promoter activity. Promoter activity is represented as relative luciferase activity after normalization with β-galactosidase. Data are expressed as the mean of three determinants from two independent experiments. Data were compared using Student’s t-test. *P<0.01 vs DMSO control.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
Western blot analysis to validate genes modulated by rosiglitazone (a) MDA-MB-231 cells were treated with 100 μM rosiglitazone or vehicle and harvested at indicated times. Cell extracts were separated on SDS-PAGE and protein levels were detected with antibodies against Bax, p21 and p53. (b) MDA-MB-231 cells were treated with 100 μM rosiglitazone or vehicle for the indicated time periods and then harvested. Cell extracts were separated on SDS-PAGE and protein levels were detected with antibody against Bax.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
Effect of rosiglitazone plus CH11 on cellular viability. Cells were pretreated with 100 μM rosiglitazone for 24 h, then treated with 50 ng/ml CH11 for an additional 48 h. Cell viability was measured by MTT assay. Data are expressed as the mean of three independent experiments. *P<0.01 vs DMSO; *P<0.01 vs Rosiglitazone; †P<0.01 vs CH11, Student’s t-test.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
Effect of rosiglitazone plus TNFα on cellular viability. (a) Cells were pretreated with 100 μM rosiglitazone for 24 h, then treated with 20 ng/ml TNFα for an additional 4 h. Cell viability was measured by MTT assay. Data are expressed as the mean of three independent experiments. (b) Cells were pretreated with 1.4 μM rosiglitazone for 24 h, then treated with 20 ng/ml TNFα for an additional 4 h. Cell viability was measured by MTT assay. Data are expressed as the mean of three independent experiments. *P<0.01 vs DMSO; **P<0.01 vs Rosiglitazone, £P<0.05 vs TNF, Student’s t-test.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
Effect of rosiglitazone plus CYC202 on cellular viability. (a) Cells were treated with vehicle, 100 μM rosiglitazone plus vehicle, CYC202 plus vehicle, or CYC202 plus 100 μM rosiglitazone for 96 h. Cell viability was measured by MTT assay. Results represent mean ± S.D. of three experiments and were compared by Student’s t-test. (b) Cells were treated with vehicle, 1.4 μM rosiglitazone plus vehicle, CYC202 plus vehicle, or CYC202 plus rosiglitazone for 96 h. Cell viability was measured by MTT assay. Results represent mean ± S.D. of three experiments and were compared by Student’s t-test. *P<0.05 vs DMSO. **P<0.01 vs Rosiglitazone. †P<0.01 vs CYC202.
Citation: Endocrine-Related Cancer Endocr Relat Cancer 14, 2; 10.1677/ERC-06-0003
This work was supported by NIH grant number CA101875 (JJP) AG12712 (TM, YZ). The authors declare that within the last twelve months JJP received honoraria from GlaxoSmithKline, Inc. and MDR received honoraria from Genzyme, Inc., Abbott, Inc., and Amgen, Inc.
References
Alvi AJ, Austen B, Weston VJ, Fegan C, MacCallum D, Gianella-Borradori A, Lane DP, Hubank M, Powell JE, Wei W et al.2005 A novel CDK inhibitor, CYC202 (R-roscovitine), overcomes the defect in p53-dependent apoptosis in B-CLL by down-regulation of genes involved in transcription regulation and survival. Blood 105 4484–4491.
Artwohl M, Furnsinn C, Waldhausl W, Holzenbein T, Rainer G, Freudenthaler A, Roden M & Baumgartner-Parzer SM 2005 Thiazolidinediones inhibit proliferation of microvascular and macrovascular cells by a PPARgamma-independent mechanism. Diabetologia 48 586–594.
Butt AJ, Dickson KA, Jambazov S & Baxter RC 2005 Enhancement of tumor necrosis factor-alpha-induced growth inhibition by insulin-like growth factor-binding protein-5 (IGFBP-5), but not IGFBP-3 in human breast cancer cells. Endocrinology 146 3113–3122.
Clay CE, Monjazeb A, Thorburn J, Chilton FH & High KP 2002 15-Deoxy-delta12,14-prostaglandin J2-induced apoptosis does not require PPARgamma in breast cancer cells. Journal of Lipid Research 43 1818–1828.
Farmer SR 2005 Regulation of PPARgamma activity during adipogenesis. International Journal of Obesity 29 S13–S16.
Fischer PM & Gianella-Borradori A 2003 CDK inhibitors in clinical development for the treatment of cancer. Expert Opinion on Investigational Drugs 12 955–970.
Gerhold DL, Liu F, Jiang G, Li Z, Xu J, Lu M, Sachs JR, Bagchi A, Fridman A, Holder DJ et al.2002 Gene expression profile of adipocyte differentiation and its regulation by peroxisome proliferator-activated receptor-gamma agonists. Endocrinology 143 2106–2118.
Gray N, Detivaud L, Doerig C & Meijer L 1999 ATP-site directed inhibitors of cyclin-dependent kinases. Current Medicinal Chemistry 6 859–875.
Gruszka A, Kunert-Radek J & Pawlikowski M 2005 Rosiglitazone, PPAR-gamma receptor ligand, decreases the viability of rat prolactin-secreting pituitary tumor cells in vitro. Neuro Endocrinology Letters 26 51–54.
Heaney AP, Fernando M, Yong WH & Melmed S 2002 Functional PPAR-gamma receptor is a novel therapeutic target for ACTH-secreting pituitary adenomas. Nature Medicine 8 1281–1287.
Heaney AP, Fernando M & Melmed S 2003 PPAR-gamma receptor ligands: novel therapy for pituitary adenomas. Journal of Clinical Investigation 111 1381–1388.
Hortobagyi GN 1998 Treatment of breast cancer. New England Journal of Medicine 339 974–984.
Hu E, Tontonoz P & Spiegelman BM 1995 Transdifferentiation of myoblasts by the adipogenic transcription factors PPAR gamma and C/EBP alpha. PNAS 92 9856–9860.
Kletzien RF, Clarke SD & Ulrich RG 1992 Enhancement of adipocyte differentiation by an insulin-sensitizing agent. Molecular Pharmacology 41 393–398.
Lejeune FJ, Lienard D, Matter M & Ruegg C 2006 Efficiency of recombinant human TNF in human cancer therapy. Cancer Immunity 6 6.
Lu M, Kwan T, Yu C, Chen F, Freedman B, Schafer JM, Lee EJ, Jameson JL, Jordan VC & Cryns VL 2005 Peroxisome proliferator-activated receptor gamma agonists promote TRAIL-induced apoptosis by reducing survivin levels via cyclin D3 repression and cell cycle arrest. Journal of Biological Chemistry 280 6742–6751.
McClue SJ, Blake D, Clarke R, Cowan A, Cummings L, Fischer PM, MacKenzie M, Melville J, Stewart K, Wang S et al.2002 In vitro and in vivo antitumor properties of the cyclin dependent kinase inhibitor CYC202 (R-roscovitine). International Journal of Cancer 102 463–468.
Mgbonyebi OP, Russo J & Russo IH 1999 Roscovitine induces cell death and morphological changes indicative of apoptosis in MDA-MB-231 breast cancer cells. Cancer Research 59 1903–1910.
Nahle Z 2004 PPAR trilogy from metabolism to cancer. Current Opinion in Clinical Nutrition and Metabolic Care 7 397–402.
Ross SE, Hemati N, Longo KA, Bennett CN, Lucas PC, Erickson RL & MacDougald OA 2000 Inhibition of adipogenesis by Wnt signaling. Science 289 950–953.
Seargent JM, Yates EA & Gill JH 2004 GW9662, a potent antagonist of PPARgamma, inhibits growth of breast tumour cells and promotes the anticancer effects of the PPARgamma agonist rosiglitazone, independently of PPARgamma activation. British Journal of Pharmacology 143 933–937.
Sharma C, Pradeep A, Wong L, Rana A & Rana B 2004 Peroxisome proliferator-activated receptor gamma activation can regulate beta-catenin levels via a proteasome-mediated and adenomatous polyposis coli-independent pathway. Journal of Biological Chemistry 279 35583–35594.
Sheu WH, Ou HC, Chou FP, Lin TM & Yang CH 2006 Rosiglitazone inhibits endothelial proliferation and angiogenesis. Life Sciences 78 1520–1528.
Shiau CW, Yang CC, Kulp SK, Chen KF, Chen CS & Huang JW 2005 Thiazolidenediones mediate apoptosis in prostate cancer cells in part through inhibition of Bcl-xL/Bcl-2 functions independently of PPARgamma. Cancer Research 65 1561–1569.
Theocharis S, Margeli A, Vielh P & Kouraklis G 2004 Peroxisome proliferator-activated receptor-gamma ligands as cell-cycle modulators. Cancer Treatment Reviews 30 545–554.
Toillon RA, Descamps S, Adriaenssens E, Ricort JM, Bernard D, Boilly B & Le Bourhis X 2002 Normal breast epithelial cells induce apoptosis of breast cancer cells via Fas signaling. Experimental Cell Research 275 31–43.
Tontonoz P, Hu E & Spiegelman BM 1995 Regulation of adipocyte gene expression and differentiation by peroxisome proliferator activated receptor gamma. Current Opinion in Genetics and Development 5 571–576.
Wajant H 2006 CD95L/FasL and TRAIL in tumour surveillance and cancer therapy. Cancer Treatment and Research 130 141–165.
Wojciechowski J, Horky M, Gueorguieva M & Wesierska-Gadek J 2003 Rapid onset of nucleolar disintegration preceding cell cycle arrest in roscovitine-induced apoptosis of human MCF-7 breast cancer cells. International Journal of Cancer 106 486–495.
