RS5444, a novel PPARγ agonist, regulates aspects of the differentiated phenotype in nontransformed intestinal epithelial cells

Molecular and Cellular Endocrinology 251 (2006) 17–32

RS5444, a novel PPAR␥ agonist, regulates aspects of the differentiated phenotype in nontransformed intestinal epithelial cells Lu Chen a , Craig R. Bush b , Brian M. Necela a , Weidong Su a , Masahiro Yanagisawa a , Panos Z. Anastasiadis a , Alan P. Fields a , E. Aubrey Thompson a,∗ a

Department of Cancer Biology, Mayo Clinic Comprehensive Cancer Center, 4500 San Pablo Road, Griffin Cancer Research Bldg., Rm 310, Jacksonville, FL 32224, United States b University of Texas Medical Branch, Galveston, TX, United States Received 2 September 2005; received in revised form 16 January 2006; accepted 14 February 2006

Abstract Peroxisome proliferator-activated receptor-gamma (PPAR␥) is expressed in the intestinal epithelium, yet little is known about the physiological role of PPAR␥ in the small bowel or the effects of PPAR␥ on small intestinal epithelial cells. The present studies investigate cellular and genomic effects of PPAR␥ in nontransformed rat intestinal epithelial cells (RIE). These cells were engineered to express mouse PPAR␥1, and thereby to model the molecular phenotype that obtains upon induction of PPAR␥ at the crypt/villus junction in the small intestine. In these studies, we have used a novel third generation thiazolidinedione derivative, RS5444, which activates PPAR␥ with an EC50 about 1/50th that of rosiglitazone and has no effect on RIE cells that do not express PPAR␥. We used Affymetrix oligonucleotide microarrays to identify potential PPAR␥-regulated processes in RIE cells, including lipid metabolism, cell proliferation and differentiation, remodeling of the extracellular matrix, cell morphology, cell–cell adhesion, and motility. The genomic profile reflects cellular events that occur following PPAR␥ activation: RS5444 inhibited culture growth and caused irreversible G1 arrest, but did not induce apoptosis. In addition, RS5444 caused dramatic changes in cellular morphology which were associated with increased motility and diminished cellular adherence, but no increase in the ability of such cells to digest and invade Matrigel. Inhibition of proliferation, cell cycle arrest, increased motility, and altered adherence are aspects of the differentiated phenotype of villus epithelial cells, which withdraw from the cell cycle at the crypt/villus interface, migrate to the villus tips, and are subsequently shed by loss of contact with the epithelium and the underlying extracellular matrix. Our results are consistent with the hypothesis that PPAR␥ regulates critical aspects of differentiation in the small intestinal epithelium. Many nuclear receptors regulate differentiation. However, our results point to novel effects of PPAR␥ on cell–cell and cell–matrix interactions, which are not typical of other nuclear receptors. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: PPAR-gamma; Differentiation; Small intestine

1. Introduction PPAR␥, a member of the nuclear receptor superfamily, forms heterodimers with retinoid X receptors (principally RXR␣), and regulates gene expression in a ligand-dependent manner (Miyata et al., 1994; Tontonoz et al., 1994). The PPAR␥ gene encodes three known transcripts which arise though alternative promoter utilization and alternative splicing and encode three PPAR␥ isoforms, called PPAR␥1, ␥2, and ␥3 (Zhu et al., 1993, 1995; Greene et al., 1995; Mukherjee et al., 1997; Fajas et al., 1997, 1998). The two major isoforms, PPAR␥1 and ␥2, are identical

∗

Corresponding author. Tel.: +1 904 953 6226; fax: +1 904 9530277. E-mail address: [email protected] (E.A. Thompson).

0303-7207/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mce.2006.02.006

in sequence except for an additional 30 amino-terminal amino acids in PPAR␥2. The expression of PPAR␥2 is largely restricted to adipocytes, where it functions as a master switch in differentiation (Mukherjee et al., 1997; Fajas et al., 1997). One of the best studied aspects of PPAR␥ action is regulation of genes that are involved in fatty acid transport and oxidation, and in energy metabolism. The role of PPAR␥ in regulating these metabolic processes has been studied in detail, largely because of the observation that thiazolidinedione agonists of PPAR␥ increase insulin sensitivity in individuals with type II diabetes (Lehmann et al., 1995; Saltiel and Olefsky, 1996; Willson et al., 1996). Unlike PPAR␥2, PPAR␥1 is expressed in many epithelial and immune cells. The gut is a major site of PPAR␥ expression, and the abundance of PPAR␥1 in the colon is similar to that observed in adipocytes (Fajas et al., 1997). PPAR␥1 mRNA is

18

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

expressed in the small intestine of rodents and humans (Lefebvre et al., 1999; Escher et al., 2001; Huin et al., 2000; Fajas et al., 1997). PPAR␥2 mRNA is expressed at very much lower abundance in these tissues (Fajas et al., 1997). Rosiglitazone, a thiazolindiedione PPAR␥ agonist, inhibits ischemia reperfusion injury in the small intestine (Nakajima et al., 2001), indicating that PPAR␥ is functional in that tissue. Furthermore, two studies have shown that PPAR␥ agonists suppress the number and size of polyps that form in the small intestine of mice with APC mutations (Niho et al., 2003; Yang et al., 2005), although two other reports indicated that PPAR␥ agonists had no effect on polyp formation in the small intestine of APC mice (Lefebvre et al., 1998; Saez et al., 1998). Although there is some controversy, the preponderance of data indicate that PPAR␥ is expressed and is functional in epithelial cells of the small intestine, where this receptor may play an important role in transformation and/or inflammation. However, the physiological functions of PPAR␥ in the small bowel are unclear. One possible clue concerning the role of PPAR␥ in the intestine comes from studies of human colon cancer cell lines. PPAR␥1 is highly expressed in some human colon cancer cell lines (DuBois et al., 1998), and PPAR␥ agonists have been shown to induce aspects of differentiation of some colon cancer cells (Sarraf et al., 1998; Kitamura et al., 1999; Chang and Szabo, 2000; Gupta et al., 2003; Kato et al., 2004; Yoshizumi et al., 2004; Brockman et al., 1998). PPAR␥ has also been implicated in embryonic development of the small intestine (Drori et al., 2005). These data suggest that PPAR␥ may regulate differentiation of intestinal epithelial cells. We have therefore carried out a series of studies to determine the cellular and genomic consequences of PPAR␥ activation in nontransformed rat intestinal epithelial cells (RIE). RIE cells exhibit contact inhibition (Ko et al., 1998), do not form colonies in soft agar (Zhang et al., 2004), and are not tumorigenic in nude mice (Sheng et al., 1999). RIE cells are believed to be derived from the proliferative transit amplifying cells within the crypts of Lieberk¨uhn, and these cells have been widely studied as a model of transforming growth factor-beta (TGF␤) regulation of enterocyte proliferation (Ko et al., 1998; Smith, 1994; Sheng et al., 1997; Winesett et al., 1996). Immunohistochemical analysis of PPAR␥ in the adult mouse indicates that this receptor is expressed at very low levels in the transit amplifying cells within the crypts of Lieberk¨uhn (Drori et al., 2005). RIE cells, which are derived from this compartment, also express very low levels of PPAR␥. However PPAR␥ is induced at the crypt/villus junction, where intestinal epithelial cells cease to proliferate and begin to differentiate. Therefore, we engineered RIE cells to express mouse PPAR␥1 so as to model the transition that occurs at the crypt/villus junction and to determine how induction of this receptor affects the properties of cells in the villus epithelium of the small intestine. Since induction of PPAR␥ coincides with differentiation of intestinal epithelial cells during embryogenesis (Drori et al., 2005), our initial goal was to test the hypothesis that PPAR␥ regulates differentiation of intestinal epithelial cells in culture. Our functional definition of differentiation includes global changes in gene expression, inhibition of proliferation, and cytodifferentiation. We report here that PPAR␥ regulates a large cohort

of genes that regulate metabolism, signal transduction, proliferation, adhesion, migration, and morphology. The cellular response to activation of PPAR␥ recapitulates the genomic profile, in that PPAR␥ agonists cause irreversible G1 arrest of RIE cells, with profound changes in cellular morphology, motility, and adhesion. These responses reflect aspects of the differentiated phenotype of normal intestinal epithelial cells and indicate that PPAR␥ regulates processes that control renewal of the intestinal epithelium. These studies represent, to our knowledge, the first systematic attempt to define the physiological role of PPAR␥ in nontransformed epithelial cells from the small intestine, and our observations reveal important potential PPAR␥ functions in the small bowel. Furthermore, our observations reveal a novel aspect of PPAR␥ action: regulation of migration, motility, and adhesion of intestinal epithelial cells. These processes are generally not associated with activation of nuclear receptors. 2. Materials and methods 2.1. Reagents, cell lines, and immunological assays RS5444 was provided by Sankyo Ltd., Tokyo, Japan. Troglitazone and rosiglitazone were from obtained from our colleague Dr. Al Copland (Mayo Clinic). Carbaprostacyclin (cPGI), 15-deoxy-
12,14 -prostglandin J2 (15-d-
PGJ2) and linoleic acid (LA) were purchased from Cayman. Dimethyl sulphoxide (DMSO) and fenofibrate were purchased from Sigma. Transforming growth factor-beta1 (TGF-␤1) was from purchased from R&D Systems and dissolved in 4 mM HCl, 0.1% BSA. Mouse PPAR␥1 expression vector pSG5-mPPAR␥1 (Lehmann et al., 1995) and the PPAR␥ reporter PPRE3 -TK-LUC (Kliewer et al., 1992) were provided by Dr. Steven A. Kliewer; pSG5-PPAR␣ and pSG5-PPAR␦ were obtained from Dr. Al Copland. RIE/Smad3 (RIE/S3) and RIE/PKC␤II (RIE/␤II) cells were generated by infecting RIE-1 cells with retrovirus pBabepuro3-Flag-Smad3 or pBabe-puro3-PKC␤II as previously described (Conery et al., 2004; Yu et al., 2003). Stable RIE/S3 and RIE/␤II cells were selected and maintained in 5 ␮g/ml of puromycin. RIE/␥1, RIE/S3␥1, and RIE/␤II␥1 cells were generated by transfecting RIE-1, RIE/S3, and RIE/␤II cells, respectively, with pSG5-mPPAR␥1 and pREP4 (Invitrogen) using calcium phosphate coprecipitation. RIE/␥1, RIE/S3␥1, and RIE/␤II␥1 cells that were hygromycin B resistant were screened for stable PPAR␥ expression. For immunoblotting, cells were washed once with ice-cold phosphate-buffered saline (PBS) and lysed by sonication in protein lysis buffer consisting of 50 mM HEPES pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 10% glycerol, 100 mM NaF, and protease inhibitor cocktail (Sigma–Aldrich). Protein concentration was determined by the method of Bradford (Bio-Rad). An aliquot of total protein (5 ␮g) was resolved by electrophoresis in 10% Tris–glycine gels (Invitrogen) and transferred to a PVDF membrane (Millipore). The membrane was blocked with 5% nonfat milk in 10 mM Tris–HCl pH 7.4, 150 mM NaCl, and 0.1% Tween 20 (TBST) overnight at 4 ◦ C. Thereafter, the membrane was incubated with antibodies to PPAR␥ (Santa Cruz), or actin (Santa Cruz) diluted in 1% milk/TBST at room temperature for 2 h, after which the membrane was washed in TBST three times for 10 min each. The membrane was then incubated with horseradish peroxidase-conjugated anti-mouse (Santa Cruz), or anti-goat (Santa Cruz) secondary antibodies in 1% milk/TBST at room temperature for 1 h. The membrane was washed three times in TBST for 10 min each. Antigen–antibody complexes were detected using the ECL Plus chemiluminescent system (Amersham Bioscience). For immunofluorescence analysis, cells cultured on glass cover slips were fixed in 3% paraformaldehyde for 30 min then permeabilized in 0.2% Triton X-100 for 5 min. Cells were blocked with 3% nonfat milk/PBS for 10 min at room temperature, and then incubated with anti PPAR␥ (Santa Cruz) and anti ␤-catenin (whole antiserum developed in rabbits, Sigma) diluted in 3% nonfat milk/PBS for 30 min at room temperature. PBS-washed cover slips were incubated with biotin–XX goat anti-mouse IgG (Molecular Probes) for 15 min, and

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32 subsequently, with streptavidin Alexa Fluor 594 conjugate (Molecular Probes) along with goat anti-rabbit IgG Alexa 488 (Molecular Probes) for 30 min at room temperature in the dark. Cover slips were mounted on glass slides using Aqua Poly.Mount (Polysciences). For immunohistochemistry, cells grown on cover slips were fixed and permeabilized as discussed above. They were then blocked with Antibody Diluent (DAKO) for 30 min, and incubated with a nonspecific rabbit polyclonal antibody (Biomol) diluted in Antibody Diluent for 1 h, after which the cover glasses were washed three times with PBS/0.05% Tween 20 (PBST). The cover slips were then incubated with Biotinylated Link Universal and Streptavidin-HRP from the DakoCytomation LSAB+ System-HRP kit (DAKO), according to the manufacture’s instruction. Following three PBST washes, cells were stained with DAKO Liquid DAB+ Substrate–Chromogen System (DAKO). Cover glasses were mounted on glass slides using Poly.Mount Xylene (Polysciences). Cells were observed with an Olympus BX51/52 system microscope using a 20× objective. Images were captured using an Olympus DP70 microscope digital camera and digital image capture software and compiled into Adobe Photoshop.

2.2. Cell proliferation, culture growth, and apoptosis Cells were maintained in mid log phase growth in DMEM supplemented with 5% charcoal/dextran treated fetal bovine serum (C-FBS, Hyclone). Cell proliferation was measured after plating 2 × 105 cells/well in 6-well tissue culture plates. Duplicate wells were harvested by trypsinization and the cells were counted using a hemocytometer. The Hoechst 33258 (bisbenzimide, Sigma) fluorometric assay was used for DNA quantification (Houston et al., 2003). Briefly, cells were plated 1 × 104 cells/well in 12-well tissue culture plates. The media was replaced with ddH2 O to lyse cells and release DNA. The plates were frozen at −20 ◦ C overnight. After allowing the plates to warm to room temperature, 100 ␮l of lysate was transferred to a 96-well tissue culture plate and 2 ␮g/ml Hoechst dye solution was added to each well in the dark. Calf-thymus DNA (Sigma) was used as a standard to determine DNA concentration. Fluorescence was measured at 458 nm following excitation at 365 nm using a Spectra-Max Gemini fluorometer (Molecular Devices). All experiments were done in triplicate. For analysis of S phase, cells were incubated with 10 ␮M bromodeoxyuridine (BrdU) for 30 min to pulse label S-phase cells, or continuously to measure entry of serum-stimulated cells into S phase. Cells were permeablized and stained with FITC-conjugated anti-BrdU using the BrdU Flowkit (BD PharMingen) according to the manufacturer’s recommendations. Total DNA was stained with 7-amino-actinomycin D (7-AAD) provided by the manufacturer. 1 × 104 cells from each culture were analyzed using a Becton Dickinson flow cytometer. Apoptosis was measured using Cell Death Detection ELISAPLUS (Roche), according to the manufacturer’s instruction. DNA fragmentation was analyzed by light absorbance measured at 405–490 nm on the EMAX precision microplate reader (Molecular Devices). Colony formation was assayed by plating 200 cells per 60 mm tissue culture dishes. The cultures were grown for 10 days, medium was removed, the cells were washed once with PBS and then fixed with icecold 100% methanol for 10 min at −20 ◦ C. Fixed cells were washed once with PBS and stained with Geimsa Staining solution (J.T. Baker) for 20 min at room temperature. Stained colonies were washed once with PBS and counted.

2.3. Transient transfection pSG5-mPPAR␥1, pSG5-hPPAR␣, or pSG5-hPPAR␦ plus PPRE3 -TK-LUC were co-transfected with pRL-SV40 (Promega) into cells at 70–80% confluence in 6-well plates using Fugene 6 (Roche) at a DNA:liposome ratio 1:3. Transfection was carried out in serum-free DMEM overnight. Fresh medium was added and cells were harvested after 24 h. Total cell extracts were prepared for dual-luciferase assay according to the manufacturer’s instructions (Promega) using a Monolight 2010 luminometer (Analytical Luminescence Laboratory). The activity of Renilla luciferase was used as an internal control. Results are expressed as the mean of triplicate determinations ± standard deviation.

2.4. Migration and invasion assays Cells were treated with 10 nM RS5444 for 24 h, harvested using trypsin, washed once with PBS, and suspended in serum-free DMEM containing

19

250 ␮g/ml heat-inactivated BSA (DMEM/BSA). For migration assays, Transwell chambers (6.5 mm diameter, 8 ␮m pore size, Costar) membranes were coated with 15 ␮g/ml collagen I (Vitrogen, Collagen Biomaterials) for 30 min at 37 ◦ C. 1 × 105 cells suspended in DMEM/BSA were added to the upper chamber and DMEM containing 5% C-FBS was added to the lower chamber. After incubating for 16 h at 37 ◦ C, nonmigrating cells were removed from the upper chamber with a cotton swab, and cells that had migrated to the lower surface of the membrane were washed once with PBS and stained with 0.2% (w/v) crystal violet in 2% ethanol. Migration was quantified by counting cells using brightfield optics. For invasion assays, Matrigel invasion chambers (8 ␮m pore size, BD BioCoat) were rehydrated according to the manufacturer’s instructions, and 1 × 105 cells were added to the chamber and allowed to invade for 6 h. Cells that had invaded were stained and counted as described above. All experiments were done in triplicate.

2.5. Aggregation assay Cells were tested for their ability to aggregate in hanging drop suspension cultures, as previously described (Thoreson et al., 2000). Briefly, cells were suspended using Cellstripper (Mediatech, Cellgro), washed twice in PBS, and suspended at 5 × 106 cells/ml in DMEM + 5% C-FBS. 1.5 × 105 cells in 30 ␮l of media were suspended as hanging drops from the lid of a 24-well culture dish and allowed to aggregate overnight in a humid 5% CO2 incubator at 37 ◦ C. Corresponding wells were filled with PBS to prevent drying of the drops. Aggregation was assessed 18 h after plating. Cells were subjected to shear force by passing them 10 times through a standard 200 ␮l Finnpipet tip. Cells were photographed within 10 min through a Leica AS-MDW using a 10× phase-contrast objective.

2.6. Microarray analysis Gene profiling analysis was performed using RG-U230 2.0 Microarrays (Affymetrix). Total RNA was extracted using RNAqueous (Ambion), according to the manufacturer’s protocols. First-strand cDNA synthesis was performed using total RNA (10–25 ␮g), a T1-(dT) 24 oligomer (5 GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG-dT24-3 ) and SuperScript II reverse transcriptase (Invitrogen). The cDNA was converted to double-stranded DNA and cRNAs were synthesized using bacteriophage T7 RNA polymerase in the presence of biotinylated nucleotides. Biotin-labeled target RNAs were fragmented to a mean size of 200 bases. Hybridization was performed at 45 ◦ C for 6 h in 0.1 M morpholenoethane sulfonic acid (MES), pH 6.6, 1 M sodium chloride (NaCl), 0.02 M ethylenediaminetetraaetic acid (EDTA), and 0.01% Tween 20. Arrays were washed using both nonstringent (1 M NaCl, 25 ◦ C) and stringent (1 M NaCl, 50 ◦ C) conditions prior to staining with phycoerythin-labeled streptavidin (10 ␮g/ml). Data were collected using a Gene Array Scanner (Hewlett Packard). Insightful Splus 7 with ArrayAnalyzer 2.0 (O’Connell, 2003) was used to perform quality control, and preprocessing of gene expression. Three different statistical models were employed to determine differential expression: (1) one-way ANOVA using Benjamini–Hochberg (Benjamini and Hochberg, 1995) false-discovery rate (BH.FDR) and p-value adjustment from the error model; (2) local pooled error (LPE) (Jain et al., 2003) methods using BH.FDR adjustment; and (3) significance analysis of microarrays (SAM) v.2.0 (Tusher et al., 2001). Quality control of microarray chips was performed by first visually inspecting images using Affymetrix CEL files. If chip artifacts such as scratches, ghosting, or uneven liquid pooling were found, the replicate would be excluded from further analysis. MVA (Bland-Altman) plots and box-plots analyzed chip bias and variance by pair-wise comparison (see Supplemental data). Finally, principle component analysis (PCA) was used to determine how well the replicate arrays clustered relative to each other. We found very favorable clustering and low variance/bias across replicates prior to any data manipulation. Although one of the control replicates showed a very small artifact region, both MVA plots as well as PCA suggested this error was negligible and was therefore kept for preprocessing. Irizarry’s GCRMA preprocessing method was chosen over robust multichip average (RMA) because we found much lower variance in gene expression by accounting for increased hybridization within GC-rich regions of labeled messenger (Irizarry et al., 2003; Wu et al., 2003). GCRMA preprocessing resulted

20

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

in a normalized dataset with exceptionally low variance across replicates and nearly uniform expression profiles across both control and treatment experimental conditions. Differential gene expression by LPE and ANOVA was scored as the intersection of genes displaying an adjusted p-value equal to or lower than 0.01 and a fold change equal to or greater than 1.5. Likewise, differential gene expression by SAM was scored as the intersection of genes having a delta of 5.0 and a fold change equal to or greater than 1.5. In SAM, a delta of five for our dataset correlated to a false detection rate of 0.

2.7. Quantitative reverse transcriptase-polymerase chain reaction The abundance of individual mRNAs was determined using two-step quantitative reverse transcriptase-mediated real-time PCR (QPCR). Reverse transcription of total RNA was performed using the High-Capacity cDNA Archive kit (Applied Biosystems). QPCR reactions were performed using 10 ng of input cDNA for both target genes and endogenous controls using the TaqMan Universal PCR master mix (Applied Biosystems). Amplification data were collected using an Applied Biosystems Prism 7900 sequence detector and analyzed using the Sequence Detection System software (Applied Biosystems). The primers and probes detecting rat ACAA2, ADAMTS-1, ALCAM, AQP1, ANGPL-2, ANGPL-4, CD36, COX2, CROT, CYCLIN D1, DAB2, ECH1, FST, ID2, GPDH, MGP, MHMGS, MTE-1, NRP, PAI-1, PDK4, and GAPDH were purchased from Applied Biosystems. Data were normalized to GAPDH and mRNA abundance was calculated using the

CT method (Livak and Schmittgen, 2001).

2.8. Statistical analysis Statistical analysis of microarray data was carried out as described above. Statistical analysis of other quantitative data was carried out using the t-test function of SigmaStat 32.

3. Results 3.1. PPARγ expression in rat intestinal epithelial cell lines Proliferative transit amplifying cells within the crypts of Lieberk¨uhn express very low levels of PPAR␥, as shown in Fig. 1A. Similar results have been reported in the intestinal epithelium from 2-week-old mice (Drori et al., 2005). RIE cells are derived from the transit amplifying cells and also express very low levels of PPAR␥, as shown in Fig. 1B. PPAR␥ is induced at the crypt/villus interface (Fig. 1A) where intestinal epithelial cells cease to proliferate and differentiate into mature, functional intestinal epithelial cells. This histological coincidence between induction of PPAR␥ and differentiation suggests that PPAR␥ may play a role in differentiation intestinal epithelial cells. To test this hypothesis, we generated a number of RIE cell lines that stably express mouse PPAR␥1, as shown in Fig. 1B. RIE/␥1 was generated from RIE-1 cells, RIE/S3␥1 was generated from RIE/S3 cells, and RIE/␤II␥I was generated from RIE/PKC␤II cells, as described in Section 2. PPAR␥ abundance in these cells was comparable to that observed in Jurkat cells (Fig. 1B). The abundance of PPAR␥ was slightly higher in RIE/S3␥1 cells and most of our analyses were carried out initially with this cell line, but all the three cell lines exhibit identical responses to activation of PPAR␥. Immunocytochemical analysis indicated that PPAR␥ was confined to the nucleus of RIE/S3␥1 cells, as shown in Fig. 1C. These RIE/PPAR␥ cell lines therefore model the transition that occurs at the crypt/villus interface and our objective was to determine the cellular phenotype that results from activation of PPAR␥ in these cells. To this

Fig. 1. PPAR␥ expression in mouse small intestine and rat intestinal epithelial cells. In Panel A, a section from mouse small intestine was stained with antibody against PPAR␥. Panel B shows expression of PPAR␥1 in the stable transgenic cell lines. Construction of cell lines was described in Section 2. Five micrograms of total protein lysate from each cell line was subjected to Western blotting analysis using antibodies specific for PPAR␥ and actin. Jurkat cells were used as a positive control for PPAR␥ expression. In Panel C, RIE/S3 cells and RIE/S3␥1 cells were subjected to immunofluorescent analysis using antibodies against PPAR␥ and ␤-catenin. Images were merged to show expression and nuclear localization of PPAR␥ (orange) in RIE/S3␥1 cells.

end, we have used a novel third generation thiazolidinedione, RS5444. Since this drug has not previously been characterized, we carried out a series of experiments to confirm that RS5444 is a specific, high affinity PPAR␥ agonist. 3.2. RS5444 is a high affinity PPARγ agonist RS5444 was used to activate PPAR␥ in RIE/S3␥1 transiently transfected with the PPAR␥ reporter PPRE3 -TK-LUC, as shown in Fig. 2A. RS5444 activated this reporter in RIE/S3␥1 cells, but not in RIE/S3 cells. Likewise, the PPAR␥ reporter was activated by 15-d-
12,14 -PGJ2, rosiglitazone, and troglitazone in RIE/S3␥1, but none of these PPAR␥ agonists activated the reporter in RIE/S3 cells. The EC50 for RS5444 activation of the PPAR␥ reporter (PPRE-tk/luc) in RIE/S3␥1 cells was about 0.5 nM, as shown in Fig. 2B. Rosiglitizone exhibits an EC50 of about 20 nM in these cells. We also measured the EC50s for induction of an endogenous PPAR␥ target gene, pyruvate dehydrogenase kinase 4 (PDK4), which was originally identified as a PPAR␥ target in white adipose tissue (Way et al., 2001) and is induced by PPAR␥ agonists in many human colon cancer cell lines (unpublished data, Thompson Laboratory). As shown in Fig. 2C, the EC50s for RS5444 and rosiglitazone for induction of the endogenous gene corresponded to those observed for activation of the PPRE-tk/luc reporter gene, indicating that RS5444 activates PPAR␥ at significantly lower concentrations than those

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

21

required for rosiglitazone. RS5444 activated PPAR␥ but not PPAR␣ or PPAR␦ when transiently transfected into COS-1 cells, as shown in Fig. 2D, although PPAR␣ and PPAR␦ were effectively activated by specific ligands for these receptors. These data indicate that RS5444 is a specific, very high affinity PPAR␥ agonist, and this drug was subsequently used for genomic profiling of PPAR␥ targets in RIE/S3␥1 cells. 3.3. PPARγ regulates gene expression in RIE cells RIE/S3␥1 cells were treated for 24 h with 10 nM RS5444. RNA was extracted from quadruplicate cultures and applied to Affymetrix 230V2 chips. Background correction and normalization of these chips was carried out using Bioconductor GCRMA, and the results were analyzed by three statistical methods (LPE, ANOVA, and SAM2) setting the false detection ratio at 0.01 with a fold change cutoff of 1.5. Details of these analyses are given in Section 2. Using this approach, we identified 415 probesets that were scored as targets by all three analyses. Among these, 291 probesets were induced and 124 were repressed. A complete list of these targets is given as a supplement. We had previously analyzed PPAR␥ targets in colonic epithelial cells treated with RS5444 in vivo (Su et al., manuscript in preparation), and we determined that almost all of the targets could be classified in one of three functional cohorts: genes that are involved in metabolism; in signal transduction, including genes that are involved in DNA synthesis and/or chromatin structure and function; and in migration, motility, cell structure, and cell–cell interaction (Su et al., manuscript in preparation). Many of the PPAR␥-regulated genes identified in RIE/S3␥1 cells segregated into similar cohorts. The largest cohort of PPAR␥ targets included 47 probesets that correspond to genes that are known to be involved in cellular motility, adhesion, and structure. These probesets, shown in Table 1, correspond to 39 known transcripts. Among these probesets 29 were induced and 18 repressed. We selected a subset of these putative targets for confirmation by QPCR, as shown in Fig. 3. These genes included a disintegrin and metalloproteinase with thrombospondin repeats1 (ADAMTS-1), angiopoietin related protein 2 (ANGPTL-2), angiopoietin related protein 4 (ANGPTL-4), the fatty acid transporter/thrombospondin receptor CD36, neuropilin 1 (NRP1), plasminogen activator inhibitor 1 (PAI-1), matrix Gla protein (MGP), activated leukocyte cell adhesion molecule (ALCAM), and disabled homology 2 (DAB2). In every case, our gene chip data were confirmed by QPCR, and no false positives have

Fig. 2. Characterization of PPAR␥ transcriptional activity in RIE/S3␥1 cells. In Panel A, RIE/S3 and RIE/S3␥1 cells were transiently transfected with PPRE3 -TK-Luc and pRL-SV40. Following transfection, cells were treated with 0.1% DMSO, 5 ␮M 15-d-
-PGJ2, 10 nM RS5444, 1 ␮M rosiglitazone, or 10 ␮M troglitazone for 24 h. Normalized luciferase activity was determined and plotted as fold activation relative to cells treated with DMSO. Data represent the mean ± S.D., n = 3. In Panel B RIE/S3␥1 cells were transiently transfected with PPRE3 -TK-Luc and pRL-SV40. Following transfection, cells were treated with the indicated concentration of RS5444 (circles) or rosiglitazone (squares) for

24 h. Normalized luciferase activity was determined and plotted as percentage of maximum induction. Data represent the mean ± S.D., n = 3. EC50 was calculated using Prism Graph Pad. QPCR was used to measure the abundance of PDK4 mRNA in Panel C. RIE/S3␥1 cells were treated 24 h with various concentrations of RS5444 (circles) or rosiglitazone (squares). RNA was extracted and PDK4 mRNA expression was assayed, relative to GAPDH mRNA, as described in Section 2. Cos-1 cells were transiently co-transfected with empty vector pSG5, vectors expressing PPAR␥, PPAR␣, or PPAR␦, and PPRE3 -TK-Luc, as shown in Panel D. Cells were treated with 0.1% DMSO, 10 nM RS5444, 50 ␮M fenofibrate, or 20 ␮M cPGI for 24 h. PPRE3 -TK-Luc activities were normalized to pRL-SV40 and plotted as fold activation relative to cells treated with DMSO. Data represent the mean ± S.D., n = 3.

22

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

Table 1 PPAR␥ targets involved in migration, motility, adhesion, and ECM Probeset 1372031 1367568 1395753 1368202 1367570 1398354 1367859 1370959 1376850

at a at at a at at at at at a at

1393129 1389533 1368824 1370043 1386857 1370158 1373401 1370155 1398294 1369958 1367973 1372473 1370244 1374525

at at at at at at at at at at at at at at

1381533 at 1379252 at 1393558 at 1370245 at 1369716 s at 1367858 at 1384132 at 1376657 at 1398270 at 1377000 at 1382913 at 1368223 at

1381748 at 1387655 1368384 1369633 1369443 1373577 1370570 1392264

at at at at at at s at

1368519 at 1386901 at 1367689 a at 1388924 at

Gene title

Gene symbol

LPE.fold change

Raw p-value

Disabled homolog 2 (Drosophila) Matrix Gla protein Elastin Disabled homolog 2 (Drosophila) Transgelin Catenin alpha-like 1 (predicted) Transforming growth factor, beta 3 Collagen, type III, alpha 1 Chemokine (C–C motif) ligand 27 (predicted) Hypothetical LOC293146 Fibulin 2 Caldesmon 1 Activated leukocyte cell adhesion molecule Stathmin 1 Myosin heavy chain 10, nonmuscle Tenascin C Procollagen, type I, alpha 2 Actinin, alpha 1 rhoB gene Chemokine (C–C motif) ligand 2 Tight junction protein 1 (predicted) Cathepsin L Ras association (RalGDS/AF-6) and pleckstrin homology domains 1 (predicted) Rho family GTPase 1 (predicted) Immunoglobulin superfamily, member 4A (predicted) Integrin, alpha 6 Cathepsin L Lectin, galactose binding, soluble 5///lectin, galactose binding, soluble 9 Matrix metalloproteinase 11 Immunoglobulin superfamily, member 4A (predicted) Immunoglobulin superfamily, member 4A (predicted) Bone morphogenetic protein 2 Similar to junction-mediating and regulatory protein Cortactin binding protein 2 A disintegrin-like and metalloprotease (reprolysin type) with thrombospondin type 1 motif, 1 Ras association (RalGDS/AF-6) and pleckstrin homology domains 1 (predicted) Chemokine (C-X-C motif) ligand 12 Kallikrein 6 Chemokine (C-X-C motif) ligand 12 Angiopoietin-like 2 Neuropilin 1 Neuropilin 1 Serine (or cysteine) proteinase inhibitor, clade E, member 1 Serine (or cysteine) proteinase inhibitor, clade E, member 1 cd36 antigen cd36 antigen Angiopoietin-like protein 4

Dab2 Mgp Eln Dab2 Tagln Catnal1 predicted Tgfb3 Col3a1 Ccl27 predicted

−1.64 −1.63 −1.56 −1.43 −1.19 −0.99 −0.98 −0.96 −0.95

5.32E−63 2.63E−51 5.85E−61 1.33E−58 1.09E−33 2.77E−21 6.03E−15 1.46E−25 4.02E−22

P4ha3 Fbln2 Cald1 Alcam Stmn1 Myh10 Tnc Col1a2 Actn1 Rhob Ccl2 Tjp1 Ctsl Raph1 predicted

−0.90 −0.74 −0.65 −0.63 −0.63 −0.62 −0.61 −0.61 −0.60 0.63 0.65 0.68 0.72 0.84

4.57E−29 8.62E−18 1.64E−15 5.09E−14 1.08E−18 8.86E−14 5.30E−7 2.27E−12 2.44E−15 0 4.44E−16 0 0 0

Rnd1 predicted –

0.89 0.90

4.22E−12 0

Itga6 Ctsl Lgals 5///Lgals9

0.92 0.96 1.00

2.22E−16 0 0

Mmp11 Igsf4a predicted

1.03 1.03

0 0

Igsf4a predicted

1.12

0

Bmp2 LOC293057

1.41 1.48

0 0

Cttnbp2 Adamts1

1.58 1.91

0 0

Raph1 predicted

2.07

0

Cxcl12 Klk6 Cxcl12 Angptl2 Nrp Nrp1 Serpine1

2.24 2.26 2.29 2.75 3.12 3.22 3.27

0 0 0 0 0 0 0

Serpine1

3.58

0

Cd36 Cd36 Angptl4

5.55 5.98 6.69

0 0 0

Fold change (log 2) and p-values were calculated using the local pooled error model, as described in Section 2. Genes indicated in bold were confirmed by QPCR.

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

Fig. 3. Real-time PCR analysis of selected RS5444 regulated genes in RIE/S3␥1 cells. Mid log phase RIE/S3␥1 cells were treated with 10 nM RS5444 for 4 or 24 h. Total RNA was extracted and QPCR was used to measure mRNA abundance of selected genes.

CT values were calculated relative to GAPDH, as described in Section 2. Data represent mean ± S.D., n = 3.

been identified in this analysis, reflecting the stringency of our genomic analysis protocol. The second largest cohort of PPAR␥ targets included 40 probesets, corresponding to 31 known transcripts, which were classified as metabolic genes. Only two of these genes were repressed, as shown in Table 2. The majority of the target genes are involved in lipid metabolism, consistent with the known role of PPAR␥ in adipocytes. However, regulation of lipid metabolism by PPAR␥ has not previously been demonstrated in nontransformed epithelial cells from the small intestine. Induction of a subset of these metabolic genes was confirmed by QPCR, as shown in Fig. 3. These genes included acyl-Coenzyme A acyltransferase 2 (ACAA2), carnitine-O-octanyltransferase (CROT), peroxisomal enoylCoA hydratase 1 (ECH1), glycerol phosphate dehydrogenase (GPDH), mitochondrial hydroxymethyl glutarylCoA synthetase (MHMGS), and mitochondria acylCoA thioesterase 1 (MTE-1). PKD4 was included in this cohort as a positive control, although the abundance of PDK4 mRNA was below the levels of detection in our gene chip analysis, and this gene is not listed in Table 2. We also included cyclooxygenase 2 (Cox2) in this cohort because of its role in arachidonic acid metabolism. However, we note that both Cox2 and ALCAM are potentially involved in inflammation; and PPAR␥ is known to have anti-inflammatory properties in the intestinal epithelium (Su et al., 1999; Delerive et al., 2001). We also included aquaporin 1 (AQP1) in this cohort of metabolic genes because of observations that indicate that PPAR␥ may regulate water transport (Zhang et al., 2005). Thirty-three probesets, corresponding to 26 known or inferred genes were classified as signal transduction components, as shown in Table 3. About half of these probesets were induced, and half repressed. The cohort listed in Table 3 includes a number of growth factor receptors, suggesting that PPAR␥ may regulate proliferation of RIE cells, and a smaller cohort of 11 probesets was identified by virtue of their potential role in cell proliferation. As shown in Table 4, these probesets correspond

23

to 10 known regulators of cell proliferation. The majority of these were repressed. Two of these targets (follistatin FST and inhibitor of DNA binding 2 ID2) were confirmed by QPCR, as shown in Fig. 3. Cyclin D1 was also repressed by RS5444, as shown in Fig. 3, although the level of inhibition was modest. Most of the target genes that we selected for confirmation by QPCR were near maximally induced or repressed within 4 h after addition of RS5444 implying that they are primary targets. Cox2, for example, was repressed at 4 h, and only minimally repressed at 24 h. In general, the genomic data suggest that PPAR␥ regulates a number of important cellular properties of intestinal epithelial cells. Specifically, the data indicate that PPAR␥ probably inhibits proliferation and is likely to have some effect on cellular motility and/or adhesion. To confirm these suggestions, we examined the cellular responses that prevail upon activation of PPAR␥ by RS5444. 3.4. PPARγ inhibits RIE culture growth The genomic profile of RS5444-treated cells suggested that activation of PPAR␥ inhibited RIE culture growth. We noticed that RIE/S3␥1 cells exhibited a population doubling time of 37 h in 5% fetal bovine serum, whereas RIE/S3 cells, which lack PPAR␥, exhibited a population doubling time of 30 h in the same medium. However, both RIE/S3 and RIE/S3␥1 cultures exhibited similar population doubling times (32 and 33 h, respectively) when grown in charcoal-stripped 5% fetal bovine serum. This observation suggested that activation of PPAR␥ by substances in fetal bovine serum inhibited growth of RIE cells. This hypothesis was confirmed by measuring growth of RIE-derived cell lines treated for 3 days with RS5444 in charcoal-stripped serum. As shown in Fig. 4A, RS5444 significantly inhibited proliferation of RIE/␥1, RIE/S3␥1, and RIE/␤II␥1 cells (p < 0.01 for all three cell lines). RS5444 had no effect on proliferation of RIE-1, RIE/S3, or RIE/␤II cultures, which do not express PPAR␥. Inhibition of proliferation was confirmed by measuring culture growth on a daily basis over 5 days, as shown in Fig. 4B. RS5444 inhibited proliferation of RIE/S3␥1 cells, but had no effect on RIE/S3. About 50% inhibition of proliferation was observed at 3 days, consistent with the data shown in Fig. 4A, and about 85% inhibition was observed at 5 days. The EC50 for inhibition of growth of RIE/S3␥1 cells was about 1 nM, as shown in Fig. 4C, consistent with the EC50 for activation of PPAR␥ (Fig. 2). RS5444 did not induce apoptosis in RIE/S3␥1 cells (Fig. 4D), under conditions in which these cells die when exposed to TGF␤ (Conery et al., 2004). The experiment shown in Fig. 4D was carried out in serum-free medium, since the apopotic effects of troglitazone are known to be suppressed by serum (Okura et al., 2000). Prolonged treatment with RS5444 (up to 72 h) did not induce apoptosis of RIE/S3␥1 cells under these conditions (data not shown). These observations indicate that PPAR␥ inhibits proliferation of nontransformed small intestinal epithelial cells with no effect on apoptosis. Inhibition of proliferation was confirmed by flow cytometric analysis of DNA labeling and content in mid log phase RIE/S3␥1 cultures treated with RS5444 for 24 h and then pulse-labeled

24

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

Table 2 PPAR␥ targets involved in metabolic processes Probeset

Gene title

Gene symbol

LPE.fold change

Raw p-value

1371421 at 1368565 at

3-Oxoacid CoA transferase 1 (predicted) Solute carrier family 1 (glial high affinity glutamate transporter), member 3 Monoglyceride lipase Brain acyl-CoA hydrolase Farnesyltransferase, CAAX box, beta Aldo-keto reductase family 1, member B8 Fatty acid desaturase 3 Diacylglycerol O-acyltransferase homolog 2 (mouse) Acetyl-Coenzyme A dehydrogenase, long-chain Acetyl-Coenzyme A dehydrogenase, medium-chain Hydroxyacyl-Coenzyme A dehydrogenase/3-ketoacyl-Coenzyme A thiolase/enoyl-Coenzyme A hydratase (trifunctional protein), alpha subunit Diacylglycerol O-acyltransferase homolog 2 (mouse) Stearoyl-Coenzyme A desaturase 1 Acyl-Coenzyme A oxidase 1, palmitoyl Hydroxysteroid (17-beta) dehydrogenase 4 Acyl-CoA synthetase long-chain family member 3 Carnitine palmitoyltransferase 1, liver Peroxisomal biogenesis factor 11A Malic enzyme 1 Hydroxysteroid (17-beta) dehydrogenase 7 2,4-Dienoyl CoA reductase 1, mitochondrial Fatty acid binding protein 5, epidermal Carnitine palmitoyltransferase 1, liver Cytosolic acyl-CoA thioesterase 1///mitochondrial acyl-CoA thioesterase 1 Malic enzyme 1 Carnitine O-octanoyltransferase Oxidized low density lipoprotein (lectin-like) receptor 1 Hydroxysteroid 11-beta dehydrogenase 1 Glycerol-3-phosphate dehydrogenase 1 (soluble) Lipoprotein lipase Mitochondrial acyl-CoA thioesterase 1 Carnitine O-octanoyltransferase Mitochondrial acyl-CoA thioesterase 1 Mitochondrial acyl-CoA thioesterase 1 Acetyl-Coenzyme A acyltransferase 2 (mitochondrial 3-oxoacyl-Coenzyme A thiolase) Enoyl Coenzyme A hydratase 1, peroxisomal 3-Hydroxy-3-methylglutaryl-Coenzyme A synthase 2 Glycerol-3-phosphate dehydrogenase 1 (soluble) Glycerol-3-phosphate dehydrogenase 1 (soluble) Fatty acid binding protein 4, adipocyte

– Slc1a3

−1.13 −0.67

1.06E−29 1.05E−12

Mgll Bach Fntb Akr1b8 Fads3 Dgat2

0.61 0.63 0.68 0.74 0.74 0.76

1.99E−07 1.33E−15 5.35E−11 0 0 2.22E−16

Acadl

0.78

0

Acadm

0.80

1.73E−14

Hadha

0.80

0

Dgat2

0.86

4.44E−16

Scd1 Acox1 Hsd17b4 Acsl3

0.86 0.87 0.89 0.95

0 0 0 0

Cpt1a Pex11a Me1 Hsd17b7 Decr1 Fabp5 Cpt1a Cte1///Mte1

1.02 1.12 1.14 1.16 1.17 1.19 1.22 1.22

0 0 0 0 0 0 0 0

Me1 Crot Oldlr1

1.22 1.32 1.42

0 0 0

Hsd11b1 Gpd1

1.64 1.83

0 0

Lpl Mte1 Crot Mte1 Mte1 Acaa2

1.87 1.90 1.91 1.92 1.93 1.94

0 0 0 0 0 0

Ech1 Hmgcs2

2.05 2.99

0 0

Gpd1

5.10

0

Gpd1

6.18

0

Fabp4

10.10

0

1388644 1370313 1370829 1370902 1372476 1371615

at at at at at at

1367735 at 1367702 at 1370164 at

1391045 at 1370355 1367680 1367672 1368177

at at at at

1386946 1379361 1370870 1387233 1367777 1370281 1367836 1388211

at at at at at at at s at

1370067 at 1368426 at 1368683 at 1386953 at 1378960 at 1386965 1388210 1387183 1391433 1384115 1386880

at at at at at at

1386885 at 1370310 at 1369560 at 1371363 at 1368271 a at

Gene indicated in bold were confirmed by QPCR.

with bromodeoxyuridine (BrdU). A significant increase in cells with G1 DNA content was observed in RS5444-treated cells, as shown in Fig. 5A (open bars), with a corresponding decrease in S phase (BrdU-labeled) cells. No increase in cells with less

than 2N DNA content was observed in the RS5444-treated cells (data not shown), consistent with the observation that RS5444 does not induce apoptosis of small intestinal epithelial cells. Although the data shown in Fig. 5A indicate an increase in cells

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

25

Table 3 PPAR␥ targets involved in signal transduction Probeset 1370331 1367859 1375661 1368918 1368114 1390426 1387275 1371887

at at at at at at at at

1383210 1374601 1368919 1368896 1379651 1388856 1371450 1373804 1388773

at at at at at at at at at

1372592 1371101 1369156 1387502 1389349 1373928 1379662 1368073 1367771 1370209 1370256 1370699 1375898

at at at at s at at a at at at at at a at at

1388109 at 1369405 a at 1398326 at

Gene title

Gene symbol

LPE.fold change

Raw p-value

Interleukin 11 receptor, alpha chain 1 Transforming growth factor, beta 3 SRY-box containing gene 11 Placental growth factor Fibroblast growth factor 13 Notch gene homolog 1, (Drosophila) SRY-box containing gene 11 Similar to high mobility group protein homolog HMG4 SRY-box containing gene 11 Interferon gamma receptor 2 (predicted) Placental growth factor MAD homolog 7 (Drosophila) Forkhead box P1 (predicted) Kit ligand SRY-box containing gene 11 Forkhead box P1 (predicted) Tumor necrosis factor, alpha-induced protein 2 (predicted) Histone deacetylase 6 Receptor-like tyrosine kinase fyn-related kinase Serine/threonine kinase 17b (apoptosis-inducing) Similar to interleukin 17 receptor E isoform 1 Similar to interleukin 17 receptor E isoform 1 SNF-related kinase Interferon regulatory factor 1 Delta sleep inducing peptide, immunoreactor Basic transcription element binding protein 1 Frizzled homolog 1 (Drosophila) Epidermal growth factor receptor///peptidase D Similar to RNA-binding protein with multiple splicing (RBP-MS) G protein-coupled receptor 116 Cholinergic receptor, nicotinic, beta polypeptide 4 Similar to Nur77 downstream protein 2

Il11ra1 Tgfb3 Sox11 Pgf Fgf13 Notch1 Sox11 LOC305373

−1.27 −0.98 −0.91 −0.91 −0.90 −0.89 −0.87 −0.82

9.53E−43 6.03E−15 1.41E−20 1.02E−11 9.11E−24 1.69E−39 1.83E−22 4.00E−21

Sox11 Ifngr2 predicted Pgf Madh7 Foxp1 predicted Kitl Sox11 Foxp1 predicted –

−0.80 −0.75 −0.74 −0.73 −0.71 −0.68 −0.67 −0.63 0.63

1.96E−23 3.91E−13 1.10E−13 6.62E−10 1.88E−10 9.95E−15 3.29E−11 1.98E−12 2.22E−15

with G1 DNA content, one cannot, by pulse labeling, determine if this increase is due to cells dividing more slowly or to accumulation of cells that have altogether ceased to divide. A continuous labeling experiment was carried out to discriminate between these alternative responses. Initially, RIE/S3␥1 cells were synchronized in G1 by serum starvation. Such cultures were then stimulated to proliferate

Hdac6 Ryk Frk Stk17b Il17re Il17re Snrk Irf1 Dsipi Bteb1 Fzd1 Egfr///Pepd –

0.66 0.671 0.68 0.71 0.80 0.86 0.90 0.95 1.13 1.14 1.15 1.40 1.40

1.77E−15 1.33E−15 1.10E−9 7.51E−10 0 0 0 0 0 0 0 0 0

Gpr116 Chrnb4 MGC105647

1.48 1.67 1.93

0 0 0

by addition of fresh serum containing BrdU. Under these circumstances BrdU was present continuously. Consequently, after serum stimulation, BrdU-labeled cells accumulate continuously until all of the cells are labeled. At that point, culture growth ceases due to inability to replicate BrdU-labeled DNA. Representative data are shown in Fig. 5B, in which we observed that about 60% of the control RIE/S3␥1 cells incorporated

Table 4 PPAR␥ targets involved in control of cell proliferation Probeset 1368870 1376648 1375532 1372016 1387004 1387843 1387769 1368650 1379440 1368947 1370699

at at at at at at a at at at at a at

Gene title

Gene symbol

LPE.fold change

Raw p-value

Inhibitor of DNA binding 2, dominant negative helix–loop–helix protein V-myc myelocytomatosis viral related oncogene, neuroblastoma derived (avian) Inhibitor of DNA binding 2, dominant negative helix–loop–helix protein Growth arrest and DNA-damage-inducible 45 beta (predicted) Neuroblastoma, suppression of tumorigenicity 1 Follistatin Inhibitor of DNA binding 3, dominant negative helix–loop–helix protein TGFB inducible early growth response Follistatin-like 3 Growth arrest and DNA-damage-inducible 45 alpha Epidermal growth factor receptor///peptidase D

Id2 Mycn Id2 Gadd45b predicted Nbl1 Fst Id3 Tieg Fstl3 Gadd45a Egfr///Pepd

−1.96 −1.65 −1.538 −1.11 −0.75 −0.69 −0.65 0.69 0.81 1.22 1.40

2.64E−83 1.02E−57 3.87E−85 1.87E−27 1.49E−7 7.11E−8 1.06E−16 5.10E−15 1.81E−10 0 0

Gene indicated in bold were confirmed by QPCR.

26

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

Fig. 4. Effects of PPAR␥ on culture growth. The growth of RIE-1, RIE/␥1, RIE/S3, RIE/S3␥1, RIE/␤II, and RIE/␤II␥1 cultures is shown in Panel A. Cells were plated in 6-well plates at 2 × 104 cells/well in 5% C-FBS. The cultures were treated with DMSO or RS5444 for 72 h. Cell number was measured by hemocytometer count. Data represent mean ± S.D., n = 3. In Panel B, RIE/S3 and RIE/S3␥1 cells were plated in 6-well plates at 2 × 104 cells/well. Culture growth in the presence of 0.1% DMSO or 10 nM RS5444 was measured by hemocytometer count on a daily basis. Data represent mean ± S.D., n = 3. In Panel C, RIE/S3␥1 cells were plated in 12-well plates at 1 × 104 cells/well. Cultures were treated with increasing concentration of RS5444 for 96 h. DNA content was measured after Hoechst staining, as described in Section 2. In this experiment maximum inhibition by RS5444 was 0.15% of control. Data represent the mean ± S.D., n = 3. In Panel D, RIE/S3␥1 cells were plated in 12-well plates at 2 × 104 cells/well in DMEM with 0.1% C-FBS. Cultures were treated with DMSO, 10 nM RS5444, or 1 ng/ml TGF-␤1 for 24 h. Cell death was determined by measuring cytoplasmic histone-associated DNA fragments. Data represent the mean ± S.D., n = 3.

BrdU within 24 h after release from G1 arrest (closed circles), and >90% of control cells incorporated BrdU within 48 h after serum stimulation, indicating that essentially all of the serumstimulated RIE/S3␥1 cells had initiated DNA synthesis within this time period. In contrast, no more than 60% of the RS5444treated cells entered S phase within 3 days after serum stimulation (Fig. 5B, triangles). These results suggest that about one-third of the RIE/S3␥1 cells are incapable of exiting G1 and entering S phase in the presence of RS5444. A clonogenic (colony formation) assay was carried out to determine whether or not this apparent G1 arrest was reversible. RIE/S3␥1 cells were treated with RS5444 for 4 days, after which RS5444 was removed. Control (DMSO-treated) and RS5444-treated cells were washed to remove the agonist, harvested, and seeded into fresh medium without RS5444. The ability of RS5444-treated cells to form colonies was approximately 30% less than that of control cells, as shown in Fig. 5C. The decrease in plating efficiency of RS5444-treated cells corresponded to the percent inhibition of BrdU incorporation that we observed upon serum stimulation of G1-arrested cells in the presence of RS5444 (Fig. 5B). These data suggest that activation of PPAR␥ caused RIE/S3␥1 cells to withdraw irreversibly from the cell cycle. This conclusion is consistent with the observation that PPAR␥ expression and proliferation are inversely proportional in the intestinal epithelium (Fig. 1A).

3.5. Activation of PPARγ promotes motility of RIE cells Our genomic profile of RS5444-treated cells revealed a large cohort of PPAR␥-regulated genes that are involved in cell–cell contact, aggregation, association with the extracellular matrix, and motility. This observation was unanticipated, and prompted us to examine the morphology of RS5444-treated RIE/S3␥1 cells. As shown in the representative micrographs in Fig. 6A, RIE/S3␥1 cells grow as loosely associated colonies which are probably clonal in origin, reflecting the fact that these cells exhibit very little intrinsic motility. RIE/S3 cells exhibit the same colony phenotype (Fig. 6A), as do RIE-1 cells (data not shown). Addition of RS5444 to RIE/S3␥1 cells resulted in a dramatic change in colony morphology, as shown in Fig. 6A. This effect was dependent upon expression of PPAR␥, as evidenced by the lack of response exhibited by RIE/S3 cells. The morphological response demonstrated in Fig. 6A is characteristic of the response that epithelial cells undergo when exposed to the polypeptide ‘scatter factor’ HGF, and this response is response is called ‘scattering’. Both rosiglitizone and troglitizone induced colony scattering, indicating that this effect is not specific for RS5444. The morphological changes observed upon activation of PPAR␥ are consistent with increased motility. To test this hypothesis, we measured migration of RIE/S3␥1 cells through

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

27

collagen-coated Transwell chambers. As shown in Fig. 6B, activation of PPAR␥ caused a profound increase in cellular motility of RIE/S3␥1 cells, but had no effect on motility of RIE/S3 cells. However, RIE/S3␥1 cells treated with RS5444 showed no significant increase in invasiveness, as measured by the ability to penetrate Matrigel-coated Transwell membranes (Fig. 6B). These results suggest that activation of PPAR␥ by RS5444 induced dramatic morphological changes in RIE/S3␥1 cells, accompanied by an increase in motility, but with no increased invasive phenotype. We interpret these observations to be evidence of cytodifferentiation, particularly in view of the role that enterocyte migration plays in maintenance of cellular homeostasis in the gut. 3.6. PPARγ regulates cellular adhesion The ultimate step in differentiation of enterocytes is dissociation from the epithelium and exfoliation into the intestinal lumen. Our genomic data indicate that PPAR␥ regulates a number of genes that might affect cell–cell interaction. We employed a hanging drop adhesion assay to test this hypothesis. Cells were suspended in an inverted 30 ␮l drop for 24 h, after which the cells were disaggregated by trituration, as described in Section 2. Representative data are shown in Fig. 7. RIE/S3 and RIE-1 cells form dense aggregates which do not dissociate when treated with RS5444. Conversely, RIE/S3␥1 and RIE/␥1 form smaller aggregates, and these completely dissociate upon addition of RS5444. These data indicate that activation of PPAR␥ decreased cell–cell adhesion between nontransformed small intestinal epithelial cells in culture and are consistent with the hypothesis PPAR␥ regulates processes that may impinge upon the ability of such cells to dissociate from the epithelium during exfoliation in vivo. 4. Discussion

Fig. 5. Effects of PPAR␥ on cell cycle progression. (Panel A) mid log phase RIE/S3␥1 cells were treated with 0.1% DMSO or 10 nM RS5444 for 24 h. Cultures were then pulse-labeled with BrdU, and subsequently stained with FITC-anti-BrdU and 7-AAD. Cell cycle distribution was determined by flow cytometry. Data represent the mean ± S.D., n = 3. In Panel B, G1-arrested RIE/S3␥1 cells were stimulated with 5% C-FBS containing 0.1% DMSO or 10 nM RS5444 in the presence of BrdU. Cells were harvested on a daily basis, and BrdU incorporation was determined by flow cytometric analysis. Data represent the mean ± S.D., n = 3. Colony formation is shown in Panel C. RIE/S3␥1 cells were treated with 0.1% DMSO or 10 nM RS5444 for 4 days. The cultures were washed three times with warm medium to remove RS5444. Cells were harvested and 200 cells were then plated in 60 mm dishes and cultured in complete medium (containing 5% C-FBS) without RS5444 for 10 days. Colonies were stained with Giemsa and counted. Data represent the mean ± S.D., n = 10.

The intestinal epithelium is one of the largest and most dynamic tissues in the adult, and the entire tissue is renewed every 3 days or so. Renewal of the intestinal epithelium initiates with stem cells within the crypts of Lieberk¨uhn, which give rise to rapidly proliferating transit amplifying cells within the crypts. As these cells emerge from the crypts they cease to proliferate at the crypt/villus interface and begin to acquire differentiated functions. One of the differentiated functions is migration, since intestinal epithelial cells must migrate from the crypt/villus junction to the villus tips where they are ultimately shed into the intestinal lumen by exfoliation. Exfoliation involves loss of adhesion to the epithelium and dissociation from the underlying extracellular matrix. Broadly speaking, renewal of the intestinal epithelium may therefore be viewed as the sum of four processes: proliferation, differentiation, migration, and exfoliation. Homeostasis in the intestinal epithelium is exquisitely controlled by a number of signals that determine the rate at which these four processes occur. Our objective was to determine to what extent PPAR␥ regulates any or all of these processes in the epithelium of the small intestine. To this end, we engineered RIE cells to express PPAR␥, thereby providing a cellular model of the transition that occurs at the crypt/villus interface in mice. We

28

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

Fig. 6. Effects of PPAR␥ on cell morphology, motility and invasiveness. In Panel A, RIE/S3 or RIE/S3␥1 cells were treated with 0.1% DMSO, 10 nM RS5444, 1 ␮M rosiglitazone, or 10 ␮M troglitazone for 24 h, as indicated. Cultures were fixed and stained with NS-antibody, as described in Section 2. Micrographs were taken at 400×. In Panel B, RIE/S3 or RIE/S3␥1 cells were treated with DMSO or RS5444 for 24 h and migration was measured through collagen-coated Transwell chambers, as described in Section 2. Invasiveness was analyzed by plating DMSO- or RS5444-treated cells on Matrigel-coated Transwell filters. Data represent the mean ± S.D., n = 3.

were unable to identify any published report of PPAR␥ localization in the human small intestine. Consequently, we can only assume that a similar transition in PPAR␥ expression occurs at the crypt/villus junction in humans. Nevertheless, we have used the rat intestinal epithelial cell model to determine the effects of expression and activation of PPAR␥ in nontransformed small intestinal epithelial cells. In all of our studies we have used a novel third generation thiazolidinedione derivative, RS5444. This compound activates PPAR␥ at significantly lower concentrations than the second generation thiazolidinedione, rosiglitazone. Some thiazodidinediones, particularly the first generation drug troglitazone, have nonspecific effects that are not mediated by PPAR␥ (Okura et al., 2000; Hattori et al., 1999; Wang et al., 1999; Baek et al., 2003). We have observed no effect of RS5444 on RIE cells, in which PPAR␥ expression is below the limits of detection. We suspect that failure to observe nonspecific effects is

largely due to the very high affinity of PPAR␥ for RS5444, which enables us to carry out our experiments at low nanomolar concentrations of agonists, as opposed to micromolar concentrations that are routinely used with first and second generation thiazolidinediones. Initially, we obtained a genomic profile of PPAR␥-regulated genes in RIE cells. Our analysis indicates that a large number of genes are regulated by PPAR␥ in nontransformed intestinal epithelial cells, consistent with the hypothesis that induction and consequent activation of PPAR␥ at the crypt/villus interface is likely to have profound physiological consequences in the villus epithelium of the small intestine. Many of these PPAR␥regulated genes fall into one or more functional cohorts, and these provide insight into the physiological function of PPAR␥ in intestinal epithelial cells. PPAR␥ regulates lipid metabolism in intestinal epithelial cells. This response is clear from the large number of PPAR␥

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

29

Fig. 7. Effects of PPAR␥ on cell adhesion. RIE, RIE/S3, RIE/␥1, and RIE/S3␥1 cells were treated with DMSO or RS5444 for 24 h. Aggregation assays were performed using a hanging drop assay, as described in Section 2. Photographs were taken at 100×.

target genes that are involved in mitochondrial and peroxisomal biosynthesis, fatty acid metabolism, and lipid transport. Many of the genes that are induced by PPAR␥ in RIE cells have been described as markers of adipocyte differentiation, but it seems more likely that regulation of lipid metabolism is a general property of PPAR␥ in cells of diverse origin and function. PPAR␥ is known to be activated by dietary lipids (Forman et al., 1997), and it is plausible that one function of PPAR␥ in the intestine is to activate lipid metabolism in response to high concentrations of fatty acids in the intestine. For example, we routinely measure induction of PKD4 mRNA as a marker for PPAR␥ activity. PDK4 regulates pyruvate metabolism, and induction of PDK4 would be predicted

to shut down oxidation of carbohydrates in favor of lipid oxidation. Our genomic analysis indicates that PPAR␥ has a profound affect on the manner in which RIE cells interact with each other and respond to their environment. One of the largest cohorts of PPAR␥ targets in RIE cells is made up of genes that regulate signal transduction. A number of important growth factor and cytokine receptors are regulated. These data suggest that PPAR␥ may have significant and potentially important effects on cross talk between signal transduction pathways. Three of the most important signaling pathways in the intestinal epithelium are TGF␤, Notch, and Wnt; and our data suggest that PPAR␥ regulates expression of TGF-␤3, Smad7, Notch 1, Dab2, and Frizzled

30

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

1. These genes may be potentially important in understanding how PPAR␥ regulates differentiation of intestinal epithelial cells. Perhaps the most dramatic, and certainly the most surprising, effect of PPAR␥ is the role of this receptor in regulating the manner in which RIE cells interact with each other and with the extracellular matrix. Our initial indication of this response came from analysis of the genomic profile, which suggests that this receptor regulates a large number of genes that are involved in organization and function of the cytoskeleton, interaction with the extracellular matrix, cell–cell adhesion, and motility. This outcome was unanticipated. Responses of this sort are generally associated with polypeptide hormones and their receptors, most notably HGF (also known as scatter factor) and its receptor c-Met. Nuclear receptors are not, in general, associated with such responses. However, nuclear receptors are known to be intimately involved in differentiation, and the changes that we have observed in RIE cells are consistent with the hypothesis that PPAR␥ regulates two very important differentiated functions of intestinal epithelial cells: migration and cell–cell adhesion. It has previously been reported that PPAR␥ agonists stimulate adhesion in colon cancer cells (Gupta et al., 2001) and inhibit adhesion in hepatocellular carcinoma cells (Schaefer et al., 2005). The contrast between these results may have to do with tissue-specific responses or with the effects of transformation on the manner in which gut epithelial cells respond to PPAR␥. Proliferation is the most obvious manifestation of differentiation of intestinal epithelial cells. PPAR␥ is known to inhibit proliferation of some colon cancer cell lines (Sarraf et al., 1998; Kitamura et al., 1999; Brockman et al., 1998), and it has been speculated that this response reflects one potential role of PPAR␥ in normal cells. We have determined that RS5444 inhibits proliferation of colonic epithelial cells in vivo (Su et al., manuscript in preparation), and we show here that PPAR␥ inhibits proliferation of nontransformed epithelial cells from the small intestine. Our data indicate that this response is irreversible, at least for a subset of RIE cells. Irreversible inhibition of proliferation is perhaps the gold standard for differentiation. It is unclear why only one-third of the cells seem to exhibit this response. An obvious possibility is cell to cell variation in PPAR␥ activity, which might be due to differences in PPAR␥ expression, or perhaps to genetic or epigenetic differences in expression of genes that modify PPAR␥ activity. A third alternative is that PPAR␥ collaborates with other growth inhibitory signals that are not operative under the conditions we have examined, so that one observes only a partial effect when PPAR␥ is activated in culture. This question warrants additional consideration, as does the question of the proximal cause of PPAR␥ inhibition of proliferation of intestinal epithelial cells. Given the number and diversity of growth factor receptors and signal transduction components that are regulated by PPAR␥, it is possible that inhibition of proliferation may be a secondary consequence of cross talk between PPAR␥ and serum growth factor-regulated processes. For example, inhibition of follistatin (FST) expression could inhibit cell proliferation through an activin-dependent mechanism. Alternatively, inhibition of proliferation could be a primary consequence

of regulation of genes like ID2, which might directly regulate proliferation and/or differentiation. These hypotheses remain to be explored. It should also be noted that all of our analyses have been carried out with pharmacological agonists. The identity of endogenous PPAR␥ agonists is unclear, and it remains to be demonstrated that such ligands, if they exist, elicit the same effects as thiazolidinediones. Like most genomic analyses, our data raise many more questions than they answer. Our initial objective was to define the major physiological functions of PPAR␥ in intestinal epithelial cells in culture, and from such observations, to infer the role of PPAR␥ in the villus epithelium. PPAR␥ regulates proliferation, motility, and cell–cell adhesion in RIE cells in culture. These are critical functions in renewal of the epithelium and are well-known aspects of the differentiated phenotype of intestinal epithelial cells. Our observations raise the question of whether or not PPAR␥ causes differentiation. We submit that induction of irreversible growth arrest and effects on migration and cellular adhesion are probably more accurately classified as aspects of the differentiated phenotype, rather than differentiation, per se. Stated another way, we speculate that induction of PPAR␥ at the crypt/villus interface is a manifestation, but not the cause, of differentiation of intestinal epithelial cells. This hypothesis is consistent with the observation that PPAR␥-mediated differentiation of intestinal epithelial cells appears to involved programmed induction of both the receptor and the Hic5 transactivator (Drori et al., 2005). However, our data indicate that PPAR␥ regulates very important physiological functions in renewal of the epithelium in the small intestine; and may play an important role in diseases of the small bowel, including cancer and inflammatory bowel disease. Acknowledgements This work was supported in part by grants from the National Cancer Institute (CA24347) and Sankyo Ltd. Craig R. Bush is supported by a training fellowship from the Keck Center for Computational and Structural Biology of the Gulf Coast Consortia (NLM Grant No. 5 T15-LM07093). We wish to express our thanks for excellent support from Ms. Pamela A. Kreinest of the Mayo Clinic Tumor Histology service. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.mce.2006.02.006. References Baek, S.J., Wilson, L.C., Hsi, L.C., Eling, T.E., 2003. Troglitazone, a peroxisome proliferator-activated receptor gamma (PPARgamma) ligand, selectively induces the early growth response-1 gene independently of PPARgamma. A novel mechanism for its anti-tumorigenic activity. J. Biol. Chem. 278, 5845–5853. Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57, 289–300.

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32 Brockman, J.A., Gupta, R.A., Dubois, R.N., 1998. Activation of PPARgamma leads to inhibition of anchorage-independent growth of human colorectal cancer cells. Gastroenterology 115, 1049–1055. Chang, T.H., Szabo, E., 2000. Induction of differentiation and apoptosis by ligands of peroxisome proliferator-activated receptor gamma in non-small cell lung cancer. Cancer Res. 60, 1129–1138. Conery, A.R., Cao, Y., Thompson, E.A., Townsend Jr., C.M., Ko, T.C., Luo, K., 2004. Akt interacts directly with Smad3 to regulate the sensitivity to TGF-beta induced apoptosis. Nat. Cell Biol. 6, 366–372. Delerive, P., Fruchart, J.C., Staels, B., 2001. Peroxisome proliferator-activated receptors in inflammation control. J. Endocrinol. 169, 453–459. Drori, S., Girnun, G.D., Tou, L., Szwaya, J.D., Mueller, E., Xia, K., Shivdasani, R.A., Spiegelman, B.M., 2005. Hic-5 regulates an epithelial program mediated by PPARgamma. Genes Dev. 19, 362–375. DuBois, R.N., Gupta, R., Brockman, J., Reddy, B.S., Krakow, S.L., Lazar, M.A., 1998. The nuclear eicosanoid receptor, PPARgamma, is aberrantly expressed in colonic cancers. Carcinogenesis 19, 49–53. Escher, P., Braissant, O., Basu-Modak, S., Michalik, L., Wahli, W., Desvergne, B., 2001. Rat PPARs: quantitative analysis in adult rat tissues and regulation in fasting and refeeding. Endocrinology 142, 4195–4202. Fajas, L., Auboeuf, D., Raspe, E., Schoonjans, K., Lefebvre, A.M., Saladin, R., Najib, J., Laville, M., Fruchart, J.C., Deeb, S., Vidal-Puig, A., Flier, J., Briggs, M.R., Staels, B., Vidal, H., Auwerx, J., 1997. The organization, promoter analysis, and expression of the human PPARgamma gene. J. Biol. Chem. 272, 18779–18789. Fajas, L., Fruchart, J.C., Auwerx, J., 1998. PPARgamma3 mRNA: a distinct PPARgamma mRNA subtype transcribed from an independent promoter. FEBS Lett. 438, 55–60. Forman, B.M., Chen, J., Evans, R.M., 1997. Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferatoractivated receptors alpha and delta. Proc. Natl. Acad. Sci. U.S.A. 94, 4312–4317. Greene, M.E., Blumberg, B., McBride, O.W., Yi, H.F., Kronquist, K., Kwan, K., Hsieh, L., Greene, G., Nimer, S.D., 1995. Isolation of the human peroxisome proliferator activated receptor gamma cDNA: expression in hematopoietic cells and chromosomal mapping. Gene Exp. 4, 281–299. Gupta, R.A., Brockman, J.A., Sarraf, P., Willson, T.M., DuBois, R.N., 2001. Target genes of peroxisome proliferator-activated receptor gamma in colorectal cancer cells. J. Biol. Chem. 276, 29681–29687. Gupta, R.A., Sarraf, P., Mueller, E., Brockman, J.A., Prusakiewicz, J.J., Eng, C., Willson, T.M., DuBois, R.N., 2003. Peroxisome proliferator-activated receptor gamma-mediated differentiation: a mutation in colon cancer cells reveals divergent and cell type-specific mechanisms. J. Biol. Chem. 278, 22669–22677. Hattori, Y., Hattori, S., Kasai, K., 1999. Troglitazone upregulates nitric oxide synthesis in vascular smooth muscle cells. Hypertension 33, 943–948. Houston, K.D., Copland, J.A., Broaddus, R.R., Gottardis, M.M., Fischer, S.M., Walker, C.L., 2003. Inhibition of proliferation and estrogen receptor signaling by peroxisome proliferator-activated receptor gamma ligands in uterine leiomyoma. Cancer Res. 63, 1221–1227. Huin, C., Corriveau, L., Bianchi, A., Keller, J.M., Collet, P., KremarikBouillaud, P., Domenjoud, L., Becuwe, P., Schohn, H., Menard, D., Dauca, M., 2000. Differential expression of peroxisome proliferatoractivated receptors (PPARs) in the developing human fetal digestive tract. J. Histochem. Cytochem. 48, 603–611. Irizarry, R.A., Hobbs, B., Collin, F., Beazer-Barclay, Y.D., Antonellis, K.J., Scherf, U., Speed, T.P., 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249–264. Jain, N., Thatte, J., Braciale, T., Ley, K., O’Connell, M., Lee, J.K., 2003. Local-pooled-error test for identifying differentially expressed genes with a small number of replicated microarrays. Bioinformatics 19, 1945–1951. Kato, M., Kusumi, T., Tsuchida, S., Tanaka, M., Sasaki, M., Kudo, H., 2004. Induction of differentiation and peroxisome proliferator-activated receptor gamma expression in colon cancer cell lines by troglitazone. J. Cancer Res. Clin. Oncol. 130, 73–79. Kitamura, S., Miyazaki, Y., Shinomura, Y., Kondo, S., Kanayama, S., Matsuzawa, Y., 1999. Peroxisome proliferator-activated receptor gamma

31

induces growth arrest and differentiation markers of human colon cancer cells. Jpn. J. Cancer Res. 90, 75–80. Kliewer, S.A., Umesono, K., Noonan, D.J., Heyman, R.A., Evans, R.M., 1992. Convergence of 9-cis retinoic acid and peroxisome proliferator signalling pathways through heterodimer formation of their receptors. Nature 358, 771–774. Ko, T.C., Yu, W., Sakai, T., Sheng, H., Shao, J., Beauchamp, R.D., Thompson, E.A., 1998. TGF-beta1 effects on proliferation of rat intestinal epithelial cells are due to inhibition of cyclin D1 expression. Oncogene 16, 3445–3454. Lefebvre, A.M., Chen, I., Desreumaux, P., Najib, J., Fruchart, J.C., Geboes, K., Briggs, M., Heyman, R., Auwerx, J., 1998. Activation of the peroxisome proliferator-activated receptor gamma promotes the development of colon tumors in C57BL/6J-APCMin/+ mice. Nat. Med. 4, 1053–1057. Lefebvre, M., Paulweber, B., Fajas, L., Woods, J., McCrary, C., Colombel, J.F., Najib, J., Fruchart, J.C., Datz, C., Vidal, H., Desreumaux, P., Auwerx, J., 1999. Peroxisome proliferator-activated receptor gamma is induced during differentiation of colon epithelium cells. J. Endocrinol. 162, 331–340. Lehmann, J.M., Moore, L.B., Smith-Oliver, T.A., Wilkison, W.O., Willson, T.M., Kliewer, S.A., 1995. An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor gamma (PPAR gamma). J. Biol. Chem. 270, 12953–12956. Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods 25, 402–408. Miyata, K.S., McCaw, S.E., Marcus, S.L., Rachubinski, R.A., Capone, J.P., 1994. The peroxisome proliferator-activated receptor interacts with the retinoid X receptor in vivo. Gene 148, 327–330. Mukherjee, R., Jow, L., Croston, G.E., Paterniti Jr., J.R., 1997. Identification, characterization, and tissue distribution of human peroxisome proliferatoractivated receptor (PPAR) isoforms PPARgamma2 versus PPARgamma1 and activation with retinoid X receptor agonists and antagonists. J. Biol. Chem. 272, 8071–8076. Nakajima, A., Wada, K., Miki, H., Kubota, N., Nakajima, N., Terauchi, Y., Ohnishi, S., Saubermann, L.J., Kadowaki, T., Blumberg, R.S., Nagai, R., Matsuhashi, N., 2001. Endogenous PPAR gamma mediates antiinflammatory activity in murine ischemia-reperfusion injury. Gastroenterology 120, 460–469. Niho, N., Takahashi, M., Kitamura, T., Shoji, Y., Itoh, M., Noda, T., Sugimura, T., Wakabayashi, K., 2003. Concomitant suppression of hyperlipidemia and intestinal polyp formation in Apc-deficient mice by peroxisome proliferator-activated receptor ligands. Cancer Res. 63, 6090–6095. O’Connell, M., 2003. Differential expression, class discovery and class prediction using S-Plus and S+ array analyzer. SIGKDD Explor., 5. Okura, T., Nakamura, M., Takata, Y., Watanabe, S., Kitami, Y., Hiwada, K., 2000. Troglitazone induces apoptosis via the p53 and Gadd45 pathway in vascular smooth muscle cells. Eur. J. Pharmacol. 407, 227–235. Saez, E., Tontonoz, P., Nelson, M.C., Alvarez, J.G., Ming, U.T., Baird, S.M., Thomazy, V.A., Evans, R.M., 1998. Activators of the nuclear receptor PPARgamma enhance colon polyp formation. Nat. Med. 4, 1058–1061. Saltiel, A.R., Olefsky, J.M., 1996. Thiazolidinediones in the treatment of insulin resistance and type II diabetes. Diabetes 45, 1661–1669. Sarraf, P., Mueller, E., Jones, D., King, F.J., DeAngelo, D.J., Partridge, J.B., Holden, S.A., Chen, L.B., Singer, S., Fletcher, C., Spiegelman, B.M., 1998. Differentiation and reversal of malignant changes in colon cancer through PPARgamma. Nat. Med. 4, 1046–1052. Schaefer, K.L., Wada, K., Takahashi, H., Matsuhashi, N., Ohnishi, S., Wolfe, M.M., Turner, J.R., Nakajima, A., Borkan, S.C., Saubermann, L.J., 2005. Peroxisome proliferator-activated receptor gamma inhibition prevents adhesion to the extracellular matrix and induces anoikis in hepatocellular carcinoma cells. Cancer Res. 65, 2251–2259. Sheng, H., Shao, J., Hooton, E.B., Tsujii, M., DuBois, R.N., Beauchamp, R.D., 1997. Cyclooxygenase-2 induction and transforming growth factor beta growth inhibition in rat intestinal epithelial cells. Cell Growth Differ. 8, 463–470. Sheng, H., Shao, J., O’Mahony, C.A., Lamps, L., Albo, D., Isakson, P.C., Berger, D.H., DuBois, R.N., Beauchamp, R.D., 1999. Transformation

32

L. Chen et al. / Molecular and Cellular Endocrinology 251 (2006) 17–32

of intestinal epithelial cells by chronic TGF-beta1 treatment results in downregulation of the type II TGF-beta receptor and induction of cyclooxygenase-2. Oncogene 18, 855–867. Smith, R.D., 1994. The anti-proliferative action of transforming growth factor-beta 1 on a rat intestinal epithelial cell line (RIE-1) is dependent on cell population density and culture passage number. Int. J. Biochem. 26, 899–903. Su, C.G., Wen, X., Bailey, S.T., Jiang, W., Rangwala, S.M., Keilbaugh, S.A., Flanigan, A., Murthy, S., Lazar, M.A., Wu, G.D., 1999. A novel therapy for colitis utilizing PPAR-gamma ligands to inhibit the epithelial inflammatory response. J. Clin. Invest. 104, 383–389. Thoreson, M.A., Anastasiadis, P.Z., Daniel, J.M., Ireton, R.C., Wheelock, M.J., Johnson, K.R., Hummingbird, D.K., Reynolds, A.B., 2000. Selective uncoupling of p120(ctn) from E-cadherin disrupts strong adhesion. J. Cell Biol. 148, 189–202. Tontonoz, P., Graves, R.A., Budavari, A.I., Erdjument-Bromage, H., Lui, M., Hu, E., Tempst, P., Spiegelman, B.M., 1994. Adipocyte-specific transcription factor ARF6 is a heterodimeric complex of two nuclear hormone receptors, PPAR gamma and RXR alpha. Nucleic Acids Res. 22, 5628–5634. Tusher, V.G., Tibshirani, R., Chu, G., 2001. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U.S.A. 98, 5116–5121. Wang, M., Wise, S.C., Leff, T., Su, T.Z., 1999. Troglitazone, an antidiabetic agent, inhibits cholesterol biosynthesis through a mechanism independent of peroxisome proliferator-activated receptor-gamma. Diabetes 48, 254–260. Way, J.M., Harrington, W.W., Brown, K.K., Gottschalk, W.K., Sundseth, S.S., Mansfield, T.A., Ramachandran, R.K., Willson, T.M., Kliewer, S.A., 2001. Comprehensive messenger ribonucleic acid profiling reveals that peroxisome proliferator-activated receptor {{gamma}} activation has coordinate effects on gene expression in multiple insulin-sensitive tissues. Endocrinology 142, 1269–1277. Willson, T.M., Cobb, J.E., Cowan, D.J., Wiethe, R.W., Correa, I.D., Prakash, S.R., Beck, K.D., Moore, L.B., Kliewer, S.A., Lehmann, J.M., 1996. The structure–activity relationship between peroxisome proliferator-activated receptor gamma agonism and the antihyperglycemic activity of thiazolidinediones. J. Med. Chem. 39, 665–668.

Winesett, M.P., Ramsey, G.W., Barnard, J.A., 1996. Type II TGF(beta) receptor expression in intestinal cell lines and in the intestinal tract. Carcinogenesis 17, 989–995. Wu, Z., LeBlanc, R., Irizarry, RA., 2003. Stochastic models based on molecular hybridization theory for short oligonucleotide microarrays. Technical Report. Johns Hopkins University Department of Biostatistics Working Papers. Yang, K., Fan, K.H., Lamprecht, S.A., Edelmann, W., Kopelovich, L., Kucherlapati, R., Lipkin, M., 2005. Peroxisome proliferatoractivated receptor gamma agonist troglitazone induces colon tumors in normal C57BL/6J mice and enhances colonic carcinogenesis in Apc1638 N/+ Mlh1 +/− double mutant mice. Int. J. Cancer 116, 495– 499. Yoshizumi, T., Ohta, T., Ninomiya, I., Terada, I., Fushida, S., Fujimura, T., Nishimura, G., Shimizu, K., Yi, S., Miwa, K., 2004. Thiazolidinedione, a peroxisome proliferator-activated receptor-gamma ligand, inhibits growth and metastasis of HT-29 human colon cancer cells through differentiationpromoting effects. Int. J. Oncol. 25, 631–639. Yu, W., Murray, N.R., Weems, C., Chen, L., Guo, H., Ethridge, R., Ceci, J.D., Evers, B.M., Thompson, E.A., Fields, A.P., 2003. Role of cyclooxygenase 2 in protein kinase C beta II-mediated colon carcinogenesis. J. Biol. Chem. 278, 11167–11174. Zhang, J., Anastasiadis, P.Z., Liu, Y., Thompson, E.A., Fields, A.P., 2004. Protein kinase C (PKC) betaII induces cell invasion through a Ras/Mek, PKC iota/Rac 1-dependent signaling pathway. J. Biol. Chem. 279, 22118–22123. Zhang, H., Zhang, A., Kohan, D.E., Nelson, R.D., Gonzalez, F.J., Yang, T., 2005. Collecting duct-specific deletion of peroxisome proliferatoractivated receptor gamma blocks thiazolidinedione-induced fluid retention. Proc. Natl. Acad. Sci. U.S.A. 102, 9406–9411. Zhu, Y., Alvares, K., Huang, Q., Rao, M.S., Reddy, J.K., 1993. Cloning of a new member of the peroxisome proliferator-activated receptor gene family from mouse liver. J. Biol. Chem. 268, 26817–26820. Zhu, Y., Qi, C., Korenberg, J.R., Chen, X.N., Noya, D., Rao, M.S., Reddy, J.K., 1995. Structural organization of mouse peroxisome proliferatoractivated receptor gamma (mPPAR gamma) gene: alternative promoter use and different splicing yield two mPPAR gamma isoforms. Proc. Natl. Acad. Sci. U.S.A. 92, 7921–7925.