Promoter-specific roles for liver X receptor/corepressor complexes in the regulation of ABCA1 and SREBP1 gene expression

Abstract

Liver X receptors (LXRs) regulate the expression of genes involved in cholesterol and fatty acid homeostasis, including the genes for ATP-binding cassette transporter A1 (ABCA1) and sterol response element binding protein 1 (SREBP1). Loss of LXR leads to derepression of the ABCA1 gene in macrophages and the intestine, while the SREBP1c gene remains transcriptionally silent. Here we report that high-density-lipoprotein (HDL) cholesterol levels are increased in LXR-deficient mice, suggesting that derepression of ABCA1 and possibly other LXR target genes in selected tissues is sufficient to result in enhanced HDL biogenesis at the whole-body level. We provide several independent lines of evidence indicating that the repressive actions of LXRs are dependent on interactions with the nuclear receptor corepressor (NCoR) and the silencing mediator of retinoic acid and thyroid hormone receptors (SMRT). While dissociation of NCoR and SMRT results in derepression of the ABCA1 gene in macrophages, it is not sufficient for derepression of the SREBP1c gene. These findings reveal differential requirements for corepressors in the regulation of genes involved in cholesterol and fatty acid homeostasis and raise the possibility that these interactions may be exploited to develop synthetic ligands that selectively modulate LXR actions in vivo.

FIG. 1.

LXRs function as activators and inhibitors of ABCA1 expression and cholesterol efflux. (a) Effect of LXR expression on plasma lipids. Plasma samples from LXR^+/+ and LXR^−/− mice were analyzed for HDL and triglyceride content. An asterisk indicates that the LXR^−/− value is significantly different (P < 0.001) from that for LXR^+/+. Plasma Conc., concentration in plasma; TG, triglycerides. (b and c) LXR functions as an activator and inhibitor of ABCA1 mRNA and protein expression in peritoneal macrophages. Peritoneal macrophages were isolated from LXR^+/+ and LXR^−/− mice as indicated in the text. Cells were treated in culture with 1 μM T1317 or vehicle for 18 h before analysis of ABCA1 mRNA by RT PCR (b) and ABCA1 protein by Western blotting (c). ABCA1 mRNA levels determined by RT PCR in panel b are normalized to cyclophilin levels. (d) Unliganded LXRs function as inhibitors of ABCA1 and ABCG1 expression in bone marrow-derived macrophages. The cells were treated as described for panel b prior to isolation of total RNA and analysis by Northern blotting with the indicated probes. (e) Unliganded LXRs repress cholesterol efflux in peritoneal macrophages. The cells were treated as described for panel b prior to measurement of ApoAI-dependent cholesterol efflux.

FIG. 2.

LXR-mediated repression is gene and tissue specific. (a) LXR target gene expression in LXR^+/+ and LXR^−/− peritoneal macrophages treated with vehicle or 1 μM T091317. SREBP1c, SCD-1, LPL, and ApoE levels were analyzed by RT PCR analysis and normalized to that of cyclophilin. (b) LXR represses ABCA1 in a tissue-specific manner. LXR^+/+ and LXR^−/− mice were dosed daily for 7 days by oral gavage with vehicle or 10 mg of T1317/kg of body weight. Four mice per group were used in this study. MEFs were isolated and treated in culture with vehicle or 1 μM T1317. RT PCR analysis was used to determine the levels of ABCA1 mRNA relative to those of cyclophilin in the intestinal mucosa, liver, quadriceps, and MEFs.

FIG. 3.

LXR represses basal transcription and interacts with corepressors. (a) CV-1 cells were transiently cotransfected with a luciferase reporter under the control of four copies of a GAL4 response element (4 × UAS) and either empty pCMX vector, a plasmid containing the GAL4 DNA binding domain (DBD) alone, or a construct with the GAL4 DBD fused to full-length LXRα or LXRβ. The cells were then treated with vehicle or 1 μM T1317 overnight. TK, thymidine kinase minimal promoter; Luc, luciferase. (b) LXRα and LXRβ interact with the nuclear receptor IDs of NCoR and SMRT. A mammalian two-hybrid system was established by transiently cotransfecting CV-1 cells with VP16 fusions of LXRα or LXRβ ligand binding domains together with GAL4 fusions of the receptor ID1 and ID2 of NCoR and SMRT and a luciferase reporter under the control of four copies of a GAL4 response element. The cells were treated with agonist overnight prior to being assayed for reporter activity. L.U., luciferase units. (c) Repression by LXR requires NCoR expression. MEFs isolated from NCoR^+/+ and NCoR^−/− embryos were transiently cotransfected with the GAL4 response element-luciferase reporter and the full-length LXRα- or LXRβ-GAL4 fusion constructs described for panel a. After an overnight incubation in the absence of ligand, the cells were lysed and analyzed for promoter activity. (d) The effect of a ligand was analyzed with MEFs transfected with the GAL4 one-hybrid system. NCoR^+/+ and NCoR^−/− MEFs were transfected with the same constructs as described for panel c, followed by addition of vehicle or T1317 (1 μM). (e) Overexpression of exogenous NCoR rescues repression of basal transcription in NCoR^−/− MEFs. The cells were cotransfected with the constructs described for panel c together with empty vector (pCMX) or full-length NCoR (pCMX NCoR). For panels a through e, all cells were cotransfected with a cytomegalovirus-β-Gal expression vector and luciferase values were normalized to those of β-Gal activity. In panel e, the luciferase activity for each control sample (cells transfected with GAL4-DBD) was assigned a value of 100%. The luciferase activities of cells transfected with GAL4-LXRs are represented as percentages of the respective control activity.

FIG. 4.

RXR is recruited to the ABCA1 and SREBP1c promoters in an LXR-dependent manner. Bone marrow-derived macrophages were obtained from LXR^+/+ and LXR^−/− mice. Control macrophages from each genotype were treated with vehicle. The effect of an LXR ligand was assessed in LXR^+/+ cells by stimulating them with T1317 (1 μM) for 90 min. ChIP analysis was performed with antibodies specific to RXRα and control rabbit preimmune serum. Primers specific to the LXRE-containing regions of the ABCA1 and SREBP1c promoters and the control RAR-response element-containing region of the RARβ2 promoter were used for PCR analysis. Quantitation of the bands was performed by densitometry. Intensity values obtained for the immunoprecipitated products were first normalized to their respective input controls. The indicated ratios represent the quotient derived from each normalized value and the value obtained for unstimulated wild-type cells.

FIG. 5.

ChIP analysis of the ABCA1 and SREBP1c promoters. Bone marrow-derived macrophages from LXR^+/+ and LXR^−/− mice were treated with vehicle or 1 μM T1317 for various lengths of time (0 to 240 min) prior to ChIP analysis. Rabbit preimmune serum or antibodies specific to NCoR (a and c) or acH3 and acH4 (b and d) were used to immunoprecipitate the ABCA1, SREBP1c, hsf2, and RARβ2 promoters as indicated in the text. Quantitation of the bands was performed by densitometry. Intensity values obtained for the immunoprecipitated products were first normalized to those of their respective input controls. The indicated ratios represent the quotient derived from each normalized value and the value obtained for unstimulated wild-type cells.

FIG. 6.

Model for differential derepression of the ABCA1 and SREBP1c genes in LXR^−/− macrophages. (a) EMSA detecting a DNA binding activity in nuclear extracts prepared from bone marrow-derived macrophages that recognizes an E-box in the ABCA1 promoter. This activity is supershifted by antibodies directed against USF1 or USF2 but not by an antibody directed against LXRα. (b) ChIP analysis of the ABCA1 and SREBP1c promoters by using antibodies directed against USF1 or USF2. Both antibodies immunoprecipitate the ABCA1 promoter but not the SREBP1c promoter. (c) Derepression of the ABCA1 gene in LXR^−/− macrophages is proposed to be due to binding of additional sequence-specific activators, such as USF1 and USF2, that are capable of recruiting HAT-containing complexes. These factors are derepressed in the absence of LXR/NCoR complexes and are sufficient to drive ABCA1 expression. In contrast, HAT-containing coactivators also appear to be recruited to the SREBP1c gene, but these factors are unable to stimulate transcription in the absence of agonist-bound LXRs. CoA, coactivator.