Negative regulation of the oncogenic transcription factor FoxM1 by thiazolidinediones and mithramycin

Abstract

The Forkhead Box transcription factor FoxM1 regulates expression of genes that promote cell cycle progression, and it plays essential roles in the development of liver, lung, prostate and colorectal tumors. Thiazolidinediones (TZDs) activate the peroxisome proliferator-activated receptor gamma (PPARγ), a ligand-activated nuclear receptor transcription factor. We found that treatment of the human hepatoma cell lines HepG2 and PLC/PRF/5 cells with TZDs leads to inhibition of FoxM1 gene expression. No PPARγ/retinoid X receptor (RXR) consensus DNA binding sites were detected in the FoxM1 promoter extending to −10 kb upstream, and knockdown of PPARγ had no impact on TZD mediated downregulation of FoxM1 expression. Previously, others showed that PPARγ agonists inhibit the expression and DNA-binding activity of the Sp1 transcription factor. Here we show that Sp1 binds to the FoxM1 promoter region and positively regulates FoxM1 transcription, while mithramycin, a chemotherapy drug that specifically binds GC rich sequences in the DNA and inhibits activities of Sp1, inhibits expression of FoxM1. Our data suggest that TZD mediated suppression of Sp1 is responsible for downregulation of FoxM1 gene expression. Inhibition of FoxM1 expression by TZDs provides a new mechanism for TZD mediated negative regulation of cancer cell growth. FoxM1 expression and activity in cancer cells can be targeted using PPARγ agonists or the anti-neoplastic antibiotic mithramycin.

Introduction

Forkhead Box M1 (FoxM1) is a member of the Forkhead Box family of transcription factors that share a conserved winged helix DNA binding domain.¹ Its expression is limited to proliferating normal cells and cell lines derived from tumors.^2,3 Embryonic cells and actively dividing adult cells express FoxM1 in mice, while terminally differentiated and quiescent cells that exit the cell cycle show no detectable levels of FoxM1 expression. Expression of FoxM1 is regulated by cell cycle progression, as it is phase and continually expressed during S induced during the G₁ phase and mitosis.^2–4 The human FoxM1 locus contains 10 exons and FoxM1 transcripts are alternatively spliced and encode three different isoforms. Two of these isoforms, FoxM1B and FoxM1C are transcriptionally active, while FoxM1A is transcriptionally inactive.^5–7

FoxM1 promotes cell growth and stimulates the expression of genes critical for the G₁/S transition, S-phase progression, the G₂/M transition and M-phase progression. FoxM1 regulates expression of genes involved in DNA repair, mitosis, chromatin assembly and protein metabolism (reviewed in refs. ^{5, 8} and ⁹). Target genes include Skp2 and Cks1 that are a part of the SCF ubiquitin ligase complex stimulating the degradation of CDK inhibitors p21^Cip, p27^Kip and p130,¹⁰ as well as KIS kinase that is essential for the phosphorylation of the Ser10 residue of p27^Kip protein causing it to translocate to the cytoplasm where it is degraded.¹¹ Other transcriptional targets of FoxM1 include CDC25B phosphatase, which is important for the activation of Cdk1, and a number of the genes encoding proteins essential for the correct execution of mitosis such as Plk1 kinase, AurkB, CenpB and Survivin.¹⁰ FoxM1 also regulates expression of the matrix metalloproteinases 2 and 9, which contribute to tumor cell invasion.¹² Activity of FoxM1 is regulated by the Ras-mitogen-activated protein kinase (MAPK) pathway and CDK-dependent phosphorylation during the cell cycle.^13,14 Recently FoxM1 was reported to regulate oncogene induced oxidative stress that contributes to malignant transformation and cell survival.¹⁵

Thiazolidinediones (TZDs) are synthetic ligands of Peroxisome Proliferator Activated Receptor gamma (PPARγ). Commonly used thiazolidinediones (TZDs) include troglitazone, pioglitazone and rosiglitazone (reviewed in ref. 16). PPARγ forms a heterodimer with retinoid X receptor, and in the presence of ligand, binds to specific DNA sequences and activates transcription of target genes.^17,18 TZDs have been shown to have antiproliferative effects in many epithelial-derived human cancer cell lines.^19,20 phase cell cycle arrest This effect is caused by G₁ through the accumulation of the CDK Inhibitor proteins p21^Cip and p27^Kip.^21,22

A number of in vivo studies have shown that troglitazone has anti-tumor effects in colon, prostate and liver cancers.^19,23,24 Interestingly, an increasing amount of evidence suggests that the TZD-induced inhibition of tumor growth is independent of PPARγ activation. So far, there are a number of PPARγ-independent targets of troglitazone that have been characterized as anti-tumor mediators, such as the c-Jun N-terminal protein kinase and p38,²⁵ Early growth response-1 (EGR1)²⁶ and the tumor suppressor protein p53 and the p53-responsive stress protein Gadd45.²⁷ It has also been shown that troglitazone stimulates ubiquitin-mediated degradation of transcriptional factor Sp1 through a PPARγ-independent mechanism.²⁸

FoxM1 plays a major role in promoting the proliferation of cancer cells.^5,29,30 Here we examined the impact that TZDs have on FoxM1 regulated transcription. Although the FoxM1 gene has no detectable binding sites for ligand activated PPARγ receptors, we detected significant downregulation of FoxM1 transcription upon TZD treatment. This effect appears to be mediated by downregulation of the Sp1 transcription factor through a previously described, PPARγ-independent mechanism.^28,31

Results

Troglitazone downregulates expression of FoxM1 and its target genes in a human hepatocellular carcinoma cell line

Troglitazone inhibits expression of Skp2, an F-box protein that is a component of a SCF ubiquitin ligase complex in hepatic cell lines,³³ while FoxM1 activates Skp2 gene expression.¹⁰ To determine whether troglitazone treatment affects the expression of endogenous FoxM1 in hepatic cells, we incubated HepG2 cells in media supplemented with either 50 μM troglitazone or DMSO alone. After the 12-hour incubation period, total RNA was isolated from these cells and analyzed by real time RT-PCR to determine the relative mRNA expression levels of FoxM1 and its target genes Skp2 and Centromere protein B (CenpB)¹⁰ (Fig. 1A–C). As a control, we also examined expression of the NFκB1/p50 subunit (Fig. 1D). This analysis revealed that expression of FoxM1 and FoxM1 regulated genes was diminished by troglitazone, but expression of p50 was not significantly altered. These data suggest that the previously described troglitazone-induced downregulation of Skp2 could be due to FoxM1 deficiency in these cells. Accumulation of nuclear p27^Kip protein was also observed in cells with reduced FoxM1 expression following troglitazone treatment (data not shown). Troglitazone mediated inhibition of FoxM1 expression could lead to downregulation of FoxM1 target genes and ultimately nuclear accumulation of p27^Kip and cell cycle arrest.

Figure 1.

Troglitazone treatment induces downregulation of FoxM1 expression and its target genes. HepG2 cells were treated with either 50 μM troglitazone (TGZ) or solvent (DMSO) alone. Total RNA was isolated from the cells 12 hours following the treatment and subjected to real time RT-PCR analysis with primers specific for human FoxM1 (A) and its target genes Skp2 (B) and CenpB (C). p50 expression was examined as a control and was not significantly altered following treatment with troglitazone, pioglitazone (PIO) or rosiglitazone (ROSI). (D) The single asterisks in (A–D) indicate statistically significant changes, with the p values <0.05 calculated by the Student’s t test.

Troglitazone and its analogs downregulate FoxM1 expression in different liver cell lines in a PPARγ-independent manner

In addition to troglitazone, there are a number of chemical compounds with similar structure and function classified as thiazolidinediones or “glitazones,” including rosiglitazone and pioglitazone.^34–37 To test whether compounds of the same class regulate FoxM1 expression in different cell lines, we treated HepG2 cells as well as the PLC/PRF/5 hepatoma cell line (ATCC CRL-8024) with either DMSO alone, or 50 μM troglitazone, pioglitazone or rosiglitazone, and we determined the relative expression of FoxM1 mRNA as described above (Fig. 2). Expression of FoxM1 was reduced by troglitazone and its analogs at similar levels in HepG2 cells (Fig. 2A) and in PLC/PRF/5 cells (Fig. 2B).

Figure 2.

Troglitazone and its analogs have similar effects on FoxM1 expression in different liver cell lines, independent of PPARγ expression. PLC/PRF/5 (A) or HepG2 (B) cells were treated with either DMSO, or 50 μM troglitazone (TGZ), pioglitazone (PIO) or rosiglitazone (ROSI) for 12 hours. After this period of incubation, total RNA was extracted from all the samples and subjected to real time RT-PCR analysis with primers specific for human FoxM1 mRNA. (C) TZD mediated downregulation of FoxM1 is not affected by PPARγ expression. HepG2 cells were transfected with siRNA targeting PPARγ (siPPARγ) or control siRNA (siCTRL). There was no significant change in FoxM1 gene expression as measured by real time RT-PCR following knockdown of PPARγ in DMSO treated cells. Following addition of TGZ in DMSO, FoxM1 expression was repressed in the absence of significant PPARγ expression.

The downregulation of FoxM1 expression by TZDs led us to ask if the FoxM1 gene contains putative PPARγ binding sites. The human FoxM1 promoter sequence extending to −10 kb was analyzed using the JASPAR transcription factor binding profile database (http://jaspar.cgb.ki.se/).³⁸ The analysis was performed with the relative profile score threshold set to 85%, and did not reveal any relevant regions that shared even a moderate level of homology with predicted PPARγ binding sites. No conserved PPARγ binding sites were identified in the FoxM1 gene. To further explore possible contributions of PPARγ to TZD induced downregulation of FoxM1 expression, we used siRNA to knockdown PPARγ expression. Knockdown of PPARγ did not impair TGZ mediated downregulation of FoxM1 mRNA expression (Fig. 2C), providing further evidence that TZD mediated downregulation of FoxM1 gene expression occurs in a PPARγ-independent manner.

Troglitazone-mediated downregulation of FoxM1 mRNA expression temporally precedes cell cycle arrest

Troglitazone causes cells to accumulate in the G₂/M phase of the cell cycle.³³ FoxM1 deficiency results in a G₂/M block and cell cycle arrest.¹⁰ However FoxM1 itself is regulated throughout the cell cycle, and is expressed exclusively in actively proliferating cells.¹³ To determine whether FoxM1 downregulation is a consequence of a general cell cycle arrest following troglitazone treatment, we isolated cells at indicated time points after addition of troglitazone and either measured their relative DNA content by flow cytometry or FoxM1 mRNA expression by real time RT-PCR (Fig. 3A and B). This analysis indicated that the FoxM1 mRNA levels are diminished by 6 hours following the addition of troglitazone (Fig. 3A). Significant changes in cell cycle distribution, however, were observed only 24 hours after troglitazone was added (Fig. 3B). These data show that inhibition of FoxM1 expression by troglitazone is not caused by troglitazone-mediated cell cycle arrest. Based on previous studies showing that FoxM1 stimulates cell cycle progression,^10,13,30 FoxM1 deficiency probably contributes to troglitazone induced cell cycle arrest.

Figure 3.

Downregulation of FoxM1 mRNA precedes the accumulation of cells in G₂/M phase of the cell cycle upon troglitazone treatment. (A) HepG2 cells were grown with 50 μM troglitazone in the media for the indicated time periods after which total RNA extracts were isolated and subjected to real time RT-PCR analysis with primers specific for human FoxM1 as described previously. (B) HepG2 cells were grown with 50 μM troglitazone in the media for the indicated time periods, after which they were fixed with cold 75% ethanol, stained with propidium-iodide and analyzed by flow cytometry to determine the cell cycle distribution of the total cell population.

Sp1 stimulates the expression of FoxM1

It has been reported that troglitazone treatment facilitates the ubiquitin-dependent proteasomal degradation of the transcription factor Sp1 through a PPARγ-independent pathway in the LNCaP prostate cancer cell line.²⁸ To test if troglitazone affects Sp1 protein levels in HepG2 cells, we treated the cells with either DMSO alone, or troglitazone and isolated total protein extracts, which were used for immunoblotting with FoxM1 and Sp1 specific antibodies (Fig. 4A). We observed a reduction in FoxM1 and Sp1 protein expression due to troglitazone treatment. To determine whether Sp1 is essential for FoxM1 expression, we treated HepG2 cells with mithramycin A, an anticancer drug that binds GC rich sequences in the DNA and acts as an Sp1 inhibitor.³⁹ Upon isolating total RNA extracts from these cells, we performed real time RT-PCR analysis to measure relative expression of FoxM1 mRNA (Fig. 4B). We saw significant downregulation of FoxM1 mRNA expression following addition of mithramycin, suggesting that Sp1 is necessary for the transcription of FoxM1. We also examined expression of other Sp1 regulated genes in cells that were treated with either troglitazone or mithramycin as described above. Real time RT-PCR analysis of these samples with primers specific for human BRCA1, SOX9 and TOPO 2α revealed that the expression of each of these genes is suppressed in the presence of troglitazone, confirming that troglitazone-mediated downregulation of Sp1 in HepG2 cells affects all of the Sp1 target genes (Fig. 4C).

Figure 4.

Downregulation of Sp1 leads to reduced FoxM1 expression. (A) Troglitazone treatment reduces expression of Sp1 and FoxM1. HepG2 cells were treated with either DMSO or troglitazone and total cell protein extracts were subjected to immunoblotting with antibodies specific for FoxM1 and Sp1 transcription factors. (B) The Sp1 inhibitor mithramycin inhibits FoxM1 gene expression. HepG2 cells were treated with either DMSO or two different concentrations of mithramycin, after which the total RNA extracts were processed for real time RT-PCR analysis as described previously. (C) Troglitazone and mithramycin negatively regulate Sp1 target genes. HepG2 cells were treated with DMSO, 50 μM mithramycin (MM) or 50 μM troglitazone (TGZ) and analyzed by real time RT-PCR with primers specific for Sp1 regulated genes encoding human BRCA1, SOX9 and TOPO 2α. The asterisks in (B and C) indicate statistically significant changes, with the following p values calculated by the Student’s t test: *p < 0.05; **p < 0.01.

Troglitazone inhibits binding of Sp1 to regulatory elements in the FoxM1 promoter region and represses its transcription

Analysis of the human FoxM1 promoter sequence⁶ using JASPAR transcription factor binding profile database revealed several potential Sp1 binding sites in close proximity of the FoxM1 transcription initiation region. The matrix model analysis was done with the relative profile score threshold set to 90%.³⁸ The −891 bp putative binding site contains the Sp1 consensus sequence ACGCCC.⁴⁰ Potential sites are represented with dark boxes in the schematic diagram in Figure 5A, with numbering according to Korver and colleagues.⁶

Figure 5.

S p1 binds the FoxM1 promoter and stimulates FoxM1 transcription. (A) Schematic diagram showing putative Sp1 binding sites within the proximal FoxM1 promoter region.⁴⁰ (B) Troglitazone or DMSO treated HepG2 cells were processed by a quantitative ChIP assay. The cross-linked and sonicated human chromatin was immunoprecipitated with an antibody specific for Sp1 or purified rabbit IgG (control) and the amount of promoter DNA associated with the IP was quantified by real time PCR with primers specific for the indicated FoxM1 promoter regions predicted to contain Sp1 binding sites using computer analysis (JASPAR/CONSITE database;³⁸). The human proximal DHFR promoter region was used as a positive control in the Sp1 binding assay. (C) Luciferase reporter constructs containing −1.1 kb proximal human FoxM1 promoter or empty luciferase construct used for cloning were used to cotransfect HepG2 cells along with a plasmid expressing Sp1 protein. Protein extracts were prepared and analyzed for dual luciferase activity. The single asterisks in (B) indicate statistically significant changes, with the p values <0.05 calculated by the Student’s t test.

To examine if Sp1 directly binds to one of the putative binding sites within the human FoxM1 promoter, DMSO or troglitazone treated cells were subjected to quantitative ChIP analysis. The cross-linked and sonicated human chromatin was immunoprecipitated with antibodies specific for Sp1 or purified rabbit IgG (control), and the amount of human FoxM1 promoter DNA associated with the immunoprecipitated protein was quantified by real-time PCR. The primers for PCR amplification were designed to amplify the segments considered as putative Sp1 binding sites. A fragment of the human DHFR promoter containing a confirmed Sp1 binding site was used as a positive control.⁴¹ These ChIP assays demonstrated that Sp1 protein binds to the putative element at −891 bp of the endogenous FoxM1 promoter region, while the another putative Sp1 binding site at +79 bp showed no association (Fig. 5B). We also did not detect Sp1 association with the putative Sp1 sites at −214 bp and +20 bp.

To further examine the ability of Sp1 to regulate the FoxM1 promoter, −1.1 kb of the human FoxM1 promoter region was amplified by high-fidelity PCR from human genomic DNA and cloned in front of a luciferase reporter gene in the pGL3 luciferase reporter construct. We performed cotransfection assays with the CMV Sp1 expression vector or control plasmid, and the FoxM1 promoter luciferase reporter or original pGL3 plasmid as a control. These were used to measure dual luciferase enzyme activity at 24 hours post transfection in HepG2 cells. Cotransfection of the Sp1 expression construct caused over a threefold increase in transcriptional activity of the FoxM1 promoter containing plasmid, as compared with the vector pGL3, demonstrating that Sp1 can transcriptionally activate the FoxM1 gene (Fig. 5C).

Discussion

A variety of studies indicate that the FoxM1 transcription factor and PPARγ agonists play opposing roles in cancer. FoxM1 promotes expression of genes that positively regulate proliferation.⁸ It is expressed at high levels in a number of different cancers, including liver, lung, breast and colon cancer, and its expression correlates with increased proliferation (reviewed in refs. ^{5, 42} and ⁴³). In contrast, PPARγ agonists negatively regulate proliferation of liver, lung, breast, colon (reviewed in ref. 44), and prostate tumor cell lines.^19,45 FoxM1 has been shown to play an essential role in the development of hepatocellular carcinomas,³⁰ lung tumors⁴⁶ and colon tumors⁴⁷ in animal models in vivo. It promotes tumor development in vivo,⁴⁸ while PPARγ agonists inhibit the development of some of the same types of tumors in mice.⁴⁹

Mechanisms by which TZDs negatively regulate cell growth are not well understood. Accumulating evidence suggests that TZD mediated inhibition of tumor cell growth may be independent of PPARγ activation.³¹ TZD analogs that do not bind to PPARγ display equal anti-tumor effects in tumor-derived cell lines with different levels of PPARγ expression.⁴⁹ The effect of troglitazone on FoxM1 expression in HepG2 cells is unchanged when the cells are depleted of PPARγ using siRNA (Fig. 2C), suggesting that troglitazone mediated downregulation of FoxM1 expression occurs independent of PPARγ activation.

A major emerging target of TZDs is the transcription factor Sp1. Although Sp1 sites are often found in housekeeping genes, they are also found in many genes that control growth.⁵⁰ Higher expression of Sp1 has been correlated with more advanced gastric^51–53 and pancreatic⁵⁴ cancers. While ubiquitously expressed, activities of Sp1 can be differentially controlled by posttranslational modifications including glycosylation and phosphorylation (reviewed in ref. 55). TZDs have been shown to regulate Sp1 activities by both PPARγ dependent and independent mechanisms.^28,56,57

Following addition of TGZ, we detected a reduction in Sp1 levels (Fig. 4A). Others have reported that TGZ may promote the proteosomal degradation of Sp1.²⁸ We identified Sp1 binding sites in the FoxM1 gene and show that Sp1 binds to a site at position −891 bp within the FoxM1 promoter. We also demonstrated that ectopic expression of Sp1 increases levels of FoxM1 reporter gene expression. In addition, treatment with TGZ or the Sp1 inhibitor mithramycin leads to reduced FoxM1 gene expression. Recently FoxM1 was shown to stimulate its own expression,⁵⁸ and it may synergize with Sp1 to positively regulate its transcription.

PPARγ agonists (reviewed in ref. 59), the anti-neoplastic antibiotic mithramycin,^39,60–65 and specific FoxM1 inhibitors (reviewed in refs. ⁴³ and ⁶⁶) have potential therapeutic benefits in the treatment of certain cancers. Here we showed that TZDs and mithramycin inhibit FoxM1 expression, and expression of FoxM1 target genes. Recently the FoxM1 gene was shown to be repressed by the tumor suppressor protein p53,^67,68 underscoring its potential role in the etiology of different cancers. Targeting FoxM1 in cancers with mutant p53 may have particular advantages, and could lead to enhanced growth suppression and/or apoptosis of cancer cells. PPARγ agonists³¹ and FoxM1 inhibitors⁶⁹ have both also been shown to promote apoptosis of cancer cells, and it will be interesting to determine if different combination therapies will have greater efficacy in targeting some cancers. The finding that TZDs inhibit FoxM1 expression through downregulation of Sp1 provides a new mechanism by which TZDs may negatively regulate tumor cell growth and promote apoptosis.

Materials and Methods

Cell culture and chemical treatment

HepG2 or PLC/PRF/5 cells were maintained as a monolayer in Dulbecco’s modified Ham’s F-12 medium supplemented with 10% fetal calf serum (FCS), 100 IU/ml penicillin, 100 μg/ml streptomycin, 1 nM MEM amino acids and 2 mM L-glutamine. Cells were grown at 37°C in an atmosphere of 95% air and 5% CO₂. Troglitazone (Sigma Aldrich T2573-5MG), Rosiglitazone (Cayman Chemical 71740-50), Pioglitazone (Actos, Takeda Pharmaceutical Company Limited NDC 64674-151-05), Mithramycin (Sigma Aldrich M6891-1MG) and Actinomycin D (Sigma Aldrich A4262) were dissolved in DMSO and then diluted to appropriate concentrations with culture medium.

Real-time RT-PCR and siRNA gene expression knockdown

HepG2 cells or PLC/PRF/5 were harvested for preparation of total RNA using RNA-STAT-60 (Tel-Test B Inc., Friendswood, TX). Following DNase I (RNase free; New England BioLabs) digestion of total RNA to remove contaminating genomic DNA, we used the Bio-Rad cDNA synthesis kit containing both oligo(dT) and random hexamer primers to synthesize cDNA from 10 μg of total RNA. The following reaction mixture was used for all PCR samples: 1x IQ SybrGreen supermix (Bio-Rad, Carlsbad, CA), 100 to 200 nM of each primer, and 2.5 μl of cDNA in a 25-μl total volume. Reactions were amplified and analyzed in triplicate using a MyiQ single-color real-time PCR detection system (Bio-Rad, Carlsbad, CA). The following sense (S) and anti-sense (AS) primer sequences and annealing temperatures (Ta) were used to amplify and measure the amount of human mRNA by real-time RT-PCR: FoxM1-S, 5′-GGA GGA AAT GCC ACA CTT AGC G-3′, and FoxM1-AS, 5′-TAG GAC TTC TTG GGT CTT GGG GTG-3′ (Ta, 55.7°C); CENPB-S, 5′-ATT CAG ACA GTG AGG AAG AGG ACG-3′, and CENPB-AS, 5′-CAT CAA TGG GGA AGG AGG TCA G-3′ (Ta, 58°C); Skp2-S, 5′-GGT GTT TGT AAG AGG TGG TAT CGC-3′, and Skp2-AS, 5′-CAC GAA AAG GGC TGA AAT GTT C-3′ (Ta, 62°C); PPARγ-S, 5′-GCA GTG GGG ATG TCT CAT AAT GC-3′, and PPARγ-AS, 5′-TTG CTT TGG TCA GCG GGA AG-3′ (Ta, 62°C); SOX9-S, 5′-CAG TAC CCG CAC TTG CAC AA-3′, and SOX9-AS, 5′-CTC GTT CAG AAG TCT CCA GAG CTT-3′ (Ta, 62°C); BRCA1-S, 5′-ACA GCT GTG TGG TGC TTC TGT G-3′, and BRCA1-AS, 5′-CAT TGT CCT CTG TCC AGG CAT C-3′ (Ta, 62°C); TOPO 2α-S, 5′-GGG TTT ACG ATG AAG ATG TTG GC G-3′, and TOPO 2α-AS, 5′-CCT TTG TTT GTT GTC CGC AGC-3′ (Ta, 62°C). These real-time RT-PCR RNA levels were normalized to human cyclophilin mRNA levels, and these primers are as follows: cyclophilin-S, 5′-GCA GAC AAG GTC CCA AAG ACA G-3′, and cyclophilin-AS, 5′-CAC CCT GAC ACA TAA ACC CTG G-3′ (Ta, 55.7°C).

siRNA Transfection-siRNA duplexes specific for human PPARγ (ON-TARGETplus SMARTpool, cat. no. L-003436-00) and siGENOME RISC-Free Control siRNA (cat. no. D-001220-01-05) were synthesized by Dharmacon Research (Lafayette, CO). These siRNA duplexes were transfected into HepG2 cells using Lipofectamine 2000 reagent (Invitrogen) in serum-free tissue culture medium following the manufacturer’s protocol. The cells were harvested 48 h after siRNA transfection for total RNA.

Flow cytometry assays to determine cell cycle profiles

HepG2 cells were treated with either DMSO or troglitazone and subjected to flow cytometry to analyze their cell cycle profile. For flow cytometry, cells were trypsinized and fixed in 70% ethanol for 2 h at 4°C. Cells were incubated with 40 μg/ml propidium iodide and 100 μg/ml RNase A (Sigma Aldrich) in phosphate-buffered saline (PBS) for 1 h at 37°C. After washing, cells were resuspended in PBS for further analysis. Data were acquired using a Beckman Coulter EPICS Elite ESP apparatus (Hialeah, FL) and then analyzed using Multicycle AV (Phoenix Flow Systems, San Diego, CA). The flow cytometry and analysis were performed in the Research Resource Center at the University of Illinois at Chicago.

Immunoblotting and antibodies

To prepare protein extracts, HepG2 cells were harvested in ice-cold PBS, pelleted by centrifugation, and used to make whole-cell protein extracts using the NP-40 lysis buffer as described previously.¹³ Protein concentrations were determined by the Bradford method with the Bio-Rad protein assay reagent. Equal amounts of proteins from each set of experiments were fractionated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membrane (Bio-Rad). The membrane was subjected to immunoblotting with antibodies against proteins of interest as described previously.³⁰ The signals from the primary antibodies were amplified by horseradish peroxidase-conjugated anti-mouse immunoglobulin G (IgG; Bio-Rad, Hercules, CA) and detected with enhanced chemiluminescence (ECL Plus; Amersham Pharmacia Biotech, Piscataway, NJ). The following commercially available antibodies and dilutions were used for western blotting: mouse anti-β-actin (AC-15; 1:5,000) (Sigma); rabbit anti Sp1 (Upstate CS 200631), rabbit antisera specific to the human C-terminal FoxM1b protein region (1:5,000) was obtained as described previously.¹⁰

ChIP assays

Troglitazone or DMSO treated HepG2 cells were processed for ChIP assay using published methods with additional modifications.³² For the immunoprecipitation step, specific amounts of antibody as indicated were added to the pre-cleared and clarified sample, which was incubated at 4°C with rotation for 12 to 16 h and washed according to the Upstate ChIP assay protocol. Five μg of Sp1 antibody (Upstate CS 200631) was used for immunoprecipitation, as well as equal amount of purified rabbit IgG (control). Cross-links were reversed on all samples, including 20% input, by addition of 100 μl TE (1 mM EDTA, 10 mM Tris-HCl, pH 7.4) containing 10 μg of RNase A and then incubated for 15 min at 25°C. Proteinase K (10 μg) and NaCl (4 μl of 5 M solution) were then added, and samples were digested for 16 h at 65°C. DNA was extracted from the digested samples using PCR purification columns following the manufacturer’s instructions (QIAGEN, Maryland). We then used 2.5 μl of this ChIP DNA sample in the subsequent 25-μl real-time PCR mixture. The total input sample was diluted 1:10, and 2.5 μl was used for real-time PCR (10% total input).

The primers used to amplify the following human gene fragments are annotated with the binding position relevant to the transcription start site, annealing temperature, and whether in the sense or antisense orientation: FoxM1 +59 S, 5′-CGA TTG GCG ACG TTC CGT CA-3′, and FoxM1 +59 AS, 5′-TGG CAC CGG AGC TTT CAG TTT GTT-3′ (Ta, 62°C), and FoxM1 −891 S, 5′-AAG TGC TGG GAT TAC AG GCG TG-3′, and FoxM1 −891 AS, 5′-CAA TGG CAG ACA AGG TTC TTT CC-3′ (Ta, 62°C), and DHFR −48 S, 5′-TCG CCT GCA CAA ATA GGG AC-3′, and DHFR −48 AS, 5′-AGA ACG CGC GGT CAA GTT T-3 (Ta, 62°C). The following reaction mixture was used for all PCR samples: 1x IQ SybrGreen Supermix (Bio-Rad, Carlsbad, CA), 100 nM of each primer, and 2.5 μl of each purified ChIP extract in a 25-μl total volume. Reactions were amplified and analyzed in triplicate using a MyiQ single-color real-time PCR detection system (Bio-Rad, Carlsbad, CA). Normalization was carried out using the Ct method. Briefly, IP samples and total input threshold cycles (Ct) for each treatment were subtracted from the Ct of the corresponding serum control IP (rabbit serum). The resulting corrected value for the total input was then subtracted from the corrected experimental IP value (Ct), and these values were raised to the power of 2 (2Ct). These values were then expressed as a relative promoter binding ± SD.

Transfections and luciferase assays

We used PCR of human U2OS genomic DNA to amplify −1.1 kb of the human FoxM1 promoter region. This PCR-amplified promoter region was cloned in the correct orientation in the pGL3-Basic Luciferase reporter plasmid (Promega). The following PCR primers were used to amplify the human FoxM1 promoter region: forward, 5′-GCG GTA CCA GCC TCC ACC TCT CGG CC-3′; reverse, 5′-GCA AGC TTT CCA ACC TGG GGG CC GAG-3′. The FoxM1 promoter region was confirmed by DNA sequencing (University of Illinois at Chicago Sequencing Facility).

Transfection of HepG2 cells and dual luciferase assays were performed as described previously.¹³ Promoter transactivation was expressed as the fold induction of transcriptional activity by the Sp1 expression vector ± the SD, where promoter activity resulting from transfection with CMV empty vector was set at 1.

Statistical analysis

All the experiments were performed in triplicate, and the Microsoft Excel tools were used to calculate the standard deviation and perform the Student’s t test to determine statistically significant differences between samples. The t test was set to two-sample unequal variance, two-tailed distribution type and p values of <0.05 were considered significant with asterisks in each graph marking statistically significant changes.