PPARdelta is an APC-regulated target of nonsteroidal anti-inflammatory drugs

Abstract

PPARB was identified as a target of APC through the analysis of global gene expression profiles in human colorectal cancer (CRC) cells. PPARdelta expression was elevated in CRCs and repressed by APC in CRC cells. This repression was mediated by beta-catenin/Tcf-4-responsive elements in the PPARdelta promotor. The ability of PPARs to bind eicosanoids suggested that PPARdelta might be a target of chemopreventive non-steroidal anti-inflammatory drugs (NSAIDs). Reporters containing PPARdelta-responsive elements were repressed by the NSAID sulindac. Furthermore, sulindac was able to disrupt the ability of PPARdelta to bind its recognition sequences. These findings suggest that NSAIDs inhibit tumorigenesis through inhibition of PPARdelta, the gene for which is normally regulated by APC.

Figure 1. Expression of PPARδ in Human CRC Cells

(A) Decreased expression of PPARδ following induction of APC in CRC cells. Expression of APC (HT29-APC) or β-galactosidase (HT29-GAL) was induced with (110 μM) ZnCl₂ for the indicated times in HT29 CRC cells containing the respective genes under the control of a modified metallothionein promotor. Total RNA (10 μg) was isolated and analyzed by Northern blot analysis with probes specific for PPARδ and PPARγ. (B) Increased expression of PPARδ in primary CRCs. Northern blot analyses with probes specific to PPARδ and PPARγ were performed on total RNA (10 μg) isolated from matched primary CRCs (C) and normal colon epithelium (N) removed from four different patients. (C) Expression of PPARδ in CRCs is dependent on Tcf-4-mediated transcription. CRC cells with increased β-catenin/Tcf-4-mediated transcription due to either APC (SW480, DLD1) or β-catenin (HCT116) mutations were either mock infected (Con) or infected with adenovirus expressing GFP (GFP) or a dominant-negative mutant of Tcf-4 (dnTcf). Total RNA (10 μg) was isolated and analyzed by Northern blot analysis with probes specific for PPARδ and PPARγ.

Figure 2. APC Regulates PPARδ Expression through CRT

(A) PPARδ promotor. A restriction map of the 3.1 kb region upstream of the first exon of PPARδ is shown. Restriction fragments BE, NE, HE, DE, BN, NH, HD, and NP were used to construct reporters for measuring APC and β-catenin responsiveness. Filled boxes represent potential Tcf-4-binding sites, and open boxes represent the same sites engineered to contain mutations that abolish Tcf-4 binding. mNP represents fragment NP with both potential Tcf-4-binding sites mutated. TRE1 and TRE2 contain four repeats of the two Tcf-4-binding sites, respectively. mTRE1 and mTRE2 are mutant forms of TRE1 and TRE2. (B) The PPARδ promotor is repressed by APC and dnTcf. SW480 CRC cells were transfected with the indicated PPARδ promotor luciferase reporters (0.4 μg), with a β-galactosidase expression vector (0.2 μg pCMVβ), and with 1.0 μg of either a control (Vector), APC, or dnTcf expression vector. Luciferase activity is reported relative to the vector control after normalizing for transfection efficiency through β-galactosidase activity. Bars represent the means of three independent replicates with error bars representing the unbiased standard deviations. (C) APC and dnTcf responsiveness is mediated by two putative Tcf-4-binding sites. PPARδ promotor fragments with intact and mutated Tcf-4-binding sites were tested for APC and dnTcf-4 responsiveness as described in (B). Bars represent the means of three independent replicates, with error bars representing the unbiased standard deviations. (D) β-catenin transactivation maps to the same promotor regions mediating APC and dnTcf responsiveness. 293 cells were transfected with the indicated PPARδ promotor luciferase reporters (0.4 μg), with a β-galactosidase expression vector (0.2 μg pCMVβ), and with 1.0 μg of either a no insert control (Vector) or an oncogenic β-catenin (β-catenin) expression vector. Luciferase activity was reported as described for (B). Bars represent the means of three independent replicates, with error bars representing the unbiased standard deviations. (E) Putative Tcf-4-binding sites in the PPARδ promotor bind Tcf-4. GEMSA was performed using ³²P-labeled probes containing either putative Tcf-4-binding sites TRE1 or TRE2. GEMSA was performed in the presence of a GST fusion protein containing the Tcf-4 DNA-binding domain as indicated. Wild-type (wt) or mutant (mut) competitors corresponding to the Tcf-4-binding sites were used as indicated.

Figure 3. Development of a PPARδ-Specific Reporter

(A) RXRα consensus binding site. PCR products of a randomized oligonucleotide template that bound a GST fusion protein containing the DNA-binding domain of RXRα were selected, cloned, and sequenced. The sequences of 28 clones are shown, manually aligned to derive the consensus binding sequence indicated at the bottom. (B) PPARδ consensus binding site. PCR products of a randomized oligonucleotide template that bound a GST fusion protein containing the DNA-binding domain of PPARδ were selected, cloned, and sequenced. The sequences of 20 clones are shown, manually aligned to derive the consensus binding sequence indicated at the bottom. (C) Binding specificity of PPARα, PPARδ, and PPARγ. Oligonucleotides containing the indicated binding elements (DRE or ACO) were ³²P-labeled and incubated with GST fusion proteins containing either the PPARα, PPARδ, PPARγ, RXRα, or no DNA binding domain (−). DNA binding was assessed by GEMSA, where “Probe” indicates the unbound probe and “Shifted” indicates bound probe. (D) Binding specificity of PPARδ/RXRα and PPARγ/RXRα heterodimers. Oligonucleotides containing the indicated binding elements (DRE or ACO) were ³²P-labeled and incubated with in vitro translated PPARδ, PPARγ, and RXRα as indicated. Bindings were supplemented with PPARδ ligand cPGI (10 μM) or PPARγ ligand BRL 49653 (10 μM) as indicated. DNA binding was assessed by GEMSA, where “Probe” indicates the unbound probe and “Shifted” indicates bound probe. (E) DRE confers PPARδ but not PPARγ responsiveness. 293 cells were transfected with DRE luciferase reporter (0.3 μg), a β-galactosidase expression vector (0.2 μg pCMVβ), and with 1.0 μg of either empty vector (Control), PPARδ, or PPARγ expression vectors. Where indicated, cells were treated with the PPARδ ligand cPGI (20 μM) or the PPARγ ligand BRL 49653 (20 μM). Luciferase activity was reported as relative luciferase activity after correction for transfection efficiency using β-galactosidase activity. Bars represent the means of three independent replicates with the error bars representing the unbiased standard deviations. (F) ACO confers PPARγ but not PPARδ responsiveness. The same as in (E) except an ACO luciferase reporter was used.

Figure 4. PPARδ Activity Is Regulated by APC, β-Catenin, and Sulindac

(A) APC and dnTcf specifically repress PPARδ activity. PPARδ and PPARγ activity was assessed with the DRE and ACO luciferase reporters, respectively. SW480 CRC cells were transfected with the indicated luciferase reporters (0.4 μg of DRE or ACO), with a β-galactosidase expression vector (0.2 μg pCMVβ), and with 1.0 μg of either a control (Vector), APC, or dnTcf expression vector. Luciferase activity was calculated as described in Figure 2. Bars represent the means of three independent replicates with the error bars being the unbiased standard deviations. (B) β-catenin expression increases PPARδ activity. The 293 human cell line was transfected with the indicated luciferase reporters (0.4 μg of DRE or ACO), with a β-galactosidase expression vector (0.2 μg pCMVβ), and with 0.8 μg of either a no insert control (Vector) or an oncogenic β-catenin expression vector. (C) Sulindac and indomethacin specifically repress PPARδ activity. PPARδ and PPARγ activity was assessed as transcriptional activity of the DRE and ACO luciferase reporters, respectively. HCT116 and SW480 CRC cells were transfected with the indicated luciferase reporters (1.0 μg of DRE or ACO) and with a β-galactosidase expression vector (0.2 μg of pCMVβ). Cells were allowed to recover for 20 hr after transfection and then treated for 10 hr with the indicated concentrations (μM) of sulindac sulfide or indomethacin. Luciferase activity was reported relative to the control (0) after normalizing for transfection efficiency.

Figure 5. PPARδ Can Protect Colon Cancer Cells from Sulindac-Induced Apoptosis

(A–D) PPARδ inhibits sulindac sulfide–induced apoptosis. HCT116 and SW480 cells were either mock infected (Uninfected) or infected with adenovirus expressing GFP (AdGFP) or PPARδ (AdPPARδ). Twenty hours after infection, cells were treated for 42 hr with sulindac sulfide. Apoptosis was assessed by the presence of apoptotic nuclei (condensation and fragmentation) after Hoechst 33258 staining. (A–C) Fluorescence microscopy of uninfected (A), AdGFP (B), or AdPPARδ (C) infected HCT116 cells treated with 125 μM of sulindac sulfide. (D) Bars represent the fraction of apoptotic nuclei after treatment with the indicated adenoviruses and concentration of sulindac sulfide (μM). (E) PPARδ rescues sulindac sulfide inhibition of clonal growth. Cells were infected with the indicated adenovirus, treated with the indicated concentrations of sulindac sulfide, and plated. Clonal growth was scored as colony formation after 6 days. Colonies were visualized by staining with Crystal Violet (upper panel) and enumerated (lower panel).

Figure 6. Mechanism of Suppression of PPARδ by NSAIDs

(A) NSAIDs do not affect PPARδ, PPARγ, or RXRα expression. HCT116 and SW480 cells were treated with the indicated concentration (μM) of sulindac sulfide for 36 hr, and RNA was isolated. Northern blot analysis was performed on 10 μg of total RNA with a probe specific for PPARδ, PPARγ, or RXRα. (B) NSAIDs suppress PPARδ DNA binding. The DRE-binding element was ³²P-labeled and incubated with no lysate (Probe only) or in vitro translated PPARδ (δ), RXRα (RXRα), or both (δ + RXRα). PPARδ + RXRα + cPGI (10 μM) was included in all lysates treated with the indicated NSAIDs. DNA binding was assessed by GEMSA, where “Probe” indicates the unbound probe and “Shifted” indicates bound probe. (C) NSAIDs do not suppress PPARγ DNA binding. The ACO-binding element was ³²P-labeled and incubated with no lysate (Probe only) or in vitro translated PPARγ (γ), RXRα (RXRα), or both (γ + RXRα). PPARγ + RXRα + BRL 49653 (10 μM) was included in all lysates treated with the indicated NSAIDs. DNA binding was assessed by GEMSA, where “Probe” indicates the unbound probe and “Shifted” indicates bound probe.

Figure 7. Unified Model for APC- and NSAID-Mediated Suppression of CRC

Elements indicated in blue have been shown to have a tumor-suppressive effect, whereas elements in red have been shown to promote tumor formation. The effects of items in boxes have been demonstrated by genetic alterations. LOX, 5-lipoxygenase; sPLA2, secretory phospholipase 2; and COX, cyclooxygenase.