Sterol-independent repression of low density lipoprotein receptor promoter by peroxisome proliferator activated receptor gamma coactivator-1alpha (PGC-1alpha)

Abstract

Peroxisome proliferator activated receptor (PPAR) gamma coactivator-1alpha (PGC-1alpha) may be implicated in cholesterol metabolism since PGC-1alpha co-activates estrogen receptor alpha (ERalpha) transactivity and estrogen/ERalpha induces the transcription of LDL receptor (LDLR). Here, we show that overexpression of PGC-1alpha in HepG2 cells represses the gene expression of LDLR and does not affect the ERalpha-induced LDLR expression. PGC-1alpha suppressed the LDLR promoter-luciferase (pLR1563- luc) activity regardless of cholesterol or functional sterol-regulatory element-1. Serial deletions of the LDLR promoter revealed that the inhibition by PGC-1alpha required the LDLR promoter regions between -650 bp and -974 bp. Phosphorylation of PGC-1alpha may not affect the suppression of LDLR expression because treatment of SB202190, a p38 MAP kinase inhibitor, did not reverse the LDLR down-regulation by PGC-1alpha. This may be the first report showing the repressive function of PGC-1alpha on gene expression. PGC-1alpha might be a novel modulator of LDLR gene expression in a sterol-independent manner, and implicated in atherogenesis.

Figure 1

Repression of endogenous LDLR mRNA by overexpression of PGC-1α. HepG2 cells were infected with AdLacZ or AdPGC-1α for 48 h at a multiplicity of infection of 50. Total RNA or cell lysate was prepared from the cells. RT-PCR (RT), Northern blot (Northern) and Western blot (WB) analyses of PGC-1α or LDLR in HepG2 cells were executed as described in "Methods". A. The PCR fragments or hybridized signals of PGC-1α, LDLR and β-actin are shown. Equivalent loading of RNA was verified by the β-actin PCR product or 28S and 18S rRNAs on the agarose gel stained with ethidium bromide. B. The antibody-directed PGC-1α bands (upper band, the ectopically expressed PGC-1α with extra 35 amino acids; lower band, endogenous PGC-1α) are as shown (WB). Equivalent loading of protein was verified by β-actin protein.

Figure 2

Repression of LDLR promoter activity by PGC-1α that co-activates the ERα/ERE-dependent transcription. HepG2 cells were transfected with LacZ expression vector (400 ng), expression vectors (400 ng) for PGC-1α and/or ERα, and a reporter gene (800 ng), pERE-luc (panel A) or pLR1563-luc (panel B), as indicated. The cells were treated for 24 h with either vehicle only (EtOH, ethanol) or 17β-ethinyl estradiol (E₂, 10^-8 M) in phenol red-free MEM with 10% charcoal-treated FBS and harvested for luciferase assay. Normalized luciferase expressions from triplicate samples were calculated relative to the LacZ expressions, and the results were expressed as n-fold activation or % control over the value obtained with the reporter alone in EtOH. Values are mean ± SD of three independent duplicate experiments. ^*: P < 0.05 vs. the vehicle-treated.

Figure 3

PGC-1α repressed the LDLR promoter activity regardless of cholesterol. HepG2 cells were transfected with LacZ expression vector (400 ng), expression vectors (400 ng) for PGC-1α or mock vector (pcDNA3.1/HisC), and pLR1563-luc, as indicated. The transfected cells were incubated for 24 h in media containing 10% FBS (FBS), 10% lipoprotein deficient serum (LPDS) or serum free media (SFM), and harvested for luciferase assay. Normalized luciferase expressions from triplicate samples were calculated relative to the LacZ expressions, and the results were expressed as n-fold activation over the value obtained with the reporter alone. Values are mean ± SD of three independent duplicate experiments. ^*: P < 0.05 vs. mock control.

Figure 4

SRE-1 is not required for PGC-1α-mediated inhibition of LDLR promoter activity. HepG2 cells were transfected with LacZ expression vector, expression vectors (400 ng) for PGC-1α or mock vector (pcDNA3.1/HisC), and one of three LDLR promoter reporter genes (800 ng), wild type pLR1563-luc (WT), pLR1563 without SRE-1 (dSRE-luc), pLR1563 with scrambled SRE-1 (mSRE-luc), as indicated. Normalized luciferase expressions from triplicate samples were calculated relative to the LacZ expressions, and the results were expressed as % control over the value obtained with the reporter alone. Values are mean ± SD of three independent duplicate experiments. ^*: P < 0.05 vs. mock control.

Figure 5

Identification of LDLR promoter region that is responsible for PGC-1α-mediated suppression. The original pLR1563-luc and the indicated series of 5'-deletion constructs linked to luciferase were cotransfected into HepG2 cells with LacZ expression vector, a PGC-1α expression vector or the pcDNA3.1/HisC mock vector alone. Normalized luciferase values (% Control) were averaged from three independent experiments. Arrows represent positions of two Sp1 and one SREBP binding sites of LDLR promoter. The results revealed that the PGC-1α-mediated suppression of LDLR promoter activity required the promoter region between -650 and -974. Values are mean ± SD of three independent duplicate experiments. ^*: P < 0.05 vs. mock control.

Figure 6

Effect of N-SREBP-2 dominant positive or PPARγ activation on the PGC-1α-repressed LDLR promoter activity. (A) Overexpression of N-SREBP2 dominates the repressive effect of PGC-1α. HepG2 cells were transfected with LacZ expression vector, expression vectors for PGC-1α (400 ng) and SREBP-2 dominant positive (N-SREBP-2), and pLR1563-luc (800 ng), as indicated. (B) Activation of PPARγ did not alter the inhibition by PGC-1α. HepG2 cells were transfected with LacZ expression vector, expression vectors for PGC-1α (400 ng) and PPARγ, and pLR1563-luc (800 ng), as indicated. The cells were treated for 24 h with either vehicle only (DMSO) or a PPARγ ligand troglitazone (TZD, 10 µM) in serum free media and harvested for luciferase assay. Normalized luciferase expressions from triplicate samples were calculated relative to the LacZ expressions, and the results were expressed as % control over the value obtained with the reporter alone. Values are mean ± SD of three independent duplicate experiments.

Figure 7

Inhibition of p38-MAPK had no effect on the PGC-1α-repressed LDLR mRNA expression. HepG2 cells were infected with AdLacZ alone (AdPGC -) or AdPGC-1 (AdPGC +) for 24 h at a multiplicity of infection of 50. The infected cells in serum free media were then incubated with p38-MAPK inhibitor SB202190 (20 µM) or vehicle DMSO for 24 h, and harvested for total RNA preparation. RT-PCR (A) and Northern blot (B) analyses of PGC-1α or LDLR in HepG2 cells. The PCR fragments or hybridized signals of PGC-1α, LDLR and β-actin are shown. Equivalent loading of RNA was verified by (A) the β-actin PCR product or (B) 28S and 18S rRNAs on the agarose gel stained with ethidium bromide.

Figure 8

PGC-1α may interact with the LDLR distal promoter through unknown factor. (A) The distal LDLR promoter (-974 to -633) contains E-box and AP1 binding motifs. (B) ChIP assay. HepG2 cells were infected with Ad-PGC-1 or Ad-LacZ control. After 36 h, chromatin-bound DNA was immunoprecipitated with polyclonal antibodies against PGC-1α, normal rabbit IgG (IgG, negative control) or acetyl-histoneH3 (Ac-H3, positive control). Immunoprecipitated DNA was analyzed by PCR using primer sets for the LDLR promoter regions between -974 and -650bp and β-actin coding region as a negative control. Ten percent of the soluble chromatin used in the reaction was used as inputs. C. A putative model of LDLR gene suppression by PGC-1α via dissociation of unknown factor (X) is suggested.