LXRα pathway: Findings from in vitro and ex vivo studies

Atherosclerosis 219 (2011) 141–150

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Pioglitazone enhances cholesterol efflux from macrophages by increasing ABCA1/ABCG1 expressions via PPAR␥/LXR␣ pathway: Findings from in vitro and ex vivo studies Hideki Ozasa a,b , Makoto Ayaori a,∗ , Maki Iizuka a , Yoshio Terao a,b , Harumi Uto-Kondo a , Emi Yakushiji a,c , Shunichi Takiguchi a,c , Kazuhiro Nakaya a,c , Tetsuya Hisada d , Yoshinari Uehara e , Masatsune Ogura a,c , Makoto Sasaki a,c , Tomohiro Komatsu a , Shunpei Horii a , Seibu Mochizuki b , Michihiro Yoshimura b , Katsunori Ikewaki a a

Division of Anti-aging, Department of Internal Medicine, National Defense Medical College, 3-2 Namiki, Tokorozawa 359-8513, Japan Division of Cardiology, Department of Internal Medicine, Jikei University School of Medicine, Japan c Division of Cardiology, Department of Internal Medicine, National Defense Medical College, Japan d Department of Public Health and Preventive Medicine, National Defense Medical College, Japan e Department of Cardiology, Fukuoka University, School of Medicine, Japan b

a r t i c l e

i n f o

Article history: Received 1 April 2011 Received in revised form 17 July 2011 Accepted 25 July 2011 Available online 4 August 2011 Keywords: PPAR␥ Macrophage Cholesterol efflux ABCA1 ABCG1

a b s t r a c t Objective: Pioglitazone, a peroxisome proliferator-activated receptor ␥ (PPAR␥) agonist, reportedly reduces cardiovascular events in diabetic patients. ATP cassette binding transporters (ABC) A1 and G1 are pivotal molecules for cholesterol efflux (ChE) from macrophages and high density-lipoprotein biogenesis, and the A1 transporter is regulated by a PPAR␥-liver receptor X (LXR) pathway. Also, pioglitazone induces ABCG1 expression, though the exact mechanism remains unclear. We therefore investigated the effects of pioglitazone on ABCA1/G1 expression in vitro and ex vivo. Methods: The effects of pioglitazone on ChE and ABCA1/G1 expressions in macrophages were assessed. Then, mRNA was quantified in macrophages when PPAR␥/LXR inhibition by siRNA or overexpression of oxysterol sulfotransferase was performed. ABCA1/G1 promoter activity with mutated LXR-responsive elements was also measured. As an ex vivo study, 15 type 2 diabetic patients were administered pioglitazone or placebo, and ChE assays and protein expressions were determined using macrophages cultured with the corresponding sera. Results: Pioglitazone increased LXR␣/ABCA1/G1 expressions, which enhanced ChE from macrophages. Inhibition of PPAR␥/LXR pathways revealed that LXR was primarily involved in pioglitazone’s transactivation of ABCA1 but only partially involved for ABCG1. Promoter assays showed that ABCG1 was regulated more by the promoter in intron 4 than that upstream of exon 1 but both promoters were responsive to LXR activation. Sera obtained after pioglitazone treatment promoted ChE and ABCA1/G1 expressions in macrophages. Conclusion: Pioglitazone enhanced ChE from macrophages by increasing ABCA1/G1 in LXR-dependent and -independent manners. Our comparable in vitro and ex vivo results shed new light on pioglitazone’s novel anti-atherogenic property. © 2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Type 2 diabetes is associated with increased risk of cardiovascular diseases, leading to significantly increased morbidity and mortality. Management of blood glucose levels is reportedly important in avoiding microvascular complications of diabetes. However, there is limited evidence for an effect on macrovascular outcomes

∗ Corresponding author. Tel.: +81 4 2995 1597; fax: +81 4 2996 5200. E-mail address: [email protected] (M. Ayaori). 0021-9150/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2011.07.113

or for any particular therapy having advantages in this respect [1]. Pioglitazone, a peroxisome proliferator-activated receptor ␥ (PPAR␥) agonist, is an oral anti-diabetic agent which positively affects blood glucose and lipids by improving insulin sensitivity. Recently, it has been reported to have an anti-atherogenic effect in type 2 diabetic patients, in comparison with a sulfonylurea agent [2,3]. High density-lipoprotein (HDL) removes cholesterol pathologically accumulated in atherosclerotic lesion macrophages and transports it back to the liver for subsequent conversion to bile in a process called reverse cholesterol transport [4]. ATP-binding cas-

142

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

sette transporters A1 and G1 (ABCA1, ABCG1) play essential roles in cholesterol efflux from macrophages and HDL formation by acting in a sequential manner: ABCA1 generates nascent HDL particles from lipid-poor apolipoprotein A-I (apoA-I) [5] which then facilitate cholesterol efflux via ABCG1, followed by formation of mature HDL particles [6]. Deletion of both ABCA1 and ABCG1 in macrophages reportedly accelerated atherosclerotic legion development as compared to deletion of either [7], indicating that they have a synergetic role in anti-atherogenesis. ABCA1/G1 in macrophages are transcriptionally regulated by ligand-dependent nuclear receptors: PPAR␥-liver receptor X (LXR) pathway [8,9]. Induction of ABCA1 by PPAR␥ agonists is completely cancelled in the absence of LXR␣/␤ expression in mice [9], indicating that the effects of PPAR␥ are dependent on LXRs. Further, ABCG1 levels are reportedly increased by PPAR␥ agonists, though the mechanisms for PPAR␥-mediated ABCG1 expression have not been fully characterized. In addition, it remains unclear whether the in vitro finding that PPAR␥ agonists enhance cholesterol efflux from macrophages translates into human circumstances. In this study, we demonstrated that pioglitazone increased ABCA1/G1 expressions, resulting in enhanced cholesterol efflux from macrophages, and that their individual expressions were completely and partially regulated in an LXR-dependent fashion, respectively. Our ex vivo experiment, where macrophages were cultured with sera from type 2 diabetic patients after oral administration of pioglitazone or placebo, showed that the former significantly promoted cholesterol efflux to apoA-I and HDL and enhanced ABCA1/G1 expressions compared to the latter.

2. Materials and methods 2.1. Materials Pioglitazone was donated by Takeda Pharmaceutical Co. (Osaka, Japan) Ltd., rosiglitazone was purchased from Santa Cruz (Santa Cruz, CA, USA), 8-bromo cyclic adenosine monophosphate (8-BrcAMP), T-0901317 (TO1317) and 22-hydroxycholesterol (22HC) from Sigma (St. Louis, MO, USA), phorbol 12-myristate 13-acetate (PMA) from Wako Pure Chemical (Tokyo, Japan), and human apoAI from Sigma. HDL was isolated by sequential ultracentrifugation and acetylated LDL (AcLDL) was prepared according to the methods previously reported [10]. 2.1.1. Cell cultures THP-1 cells (Riken Cell Bank, Tsukuba, Japan) were maintained in RPMI 1640 (Sigma) containing 10% fetal bovine serum (FBS). The differentiation of THP-1 monocytes into macrophages was induced in the presence of 320 nmol/L of PMA for 72 h. RAW264.7 (RAW) and HEK293 cells (Riken Cell Bank) were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% FBS. 2.1.2. Determination of cholesterol efflux Cholesterol efflux experiments were performed as previously described [10–12]. In brief, THP-1 macrophages or RAW cells were radiolabeled with [3 H]-cholesterol (1.0 ␮Ci/mL) pre-equilibrated with AcLDL (50 ␮g/mL) in media containing 0.2% bovine serum albumin (BSA) for 24 h, and in the case of RAW cells, in the presence of 0.3 mmol/L 8-bromoadenocine 3 ,5 -cyclic monophosphate (8Br-cAMP). The cells were washed twice with PBS and replaced with DMEM containing 0.2% BSA plus the indicated doses of pioglitazone or the vehicle in the presence or absence of apoA-I (10 ␮g/ml) or HDL (50 ␮g/ml), and incubated for 24 h. The percentage of cholesterol efflux was calculated by dividing the media-derived radioactivity by the sum of the radioactivity in the media and the

cells. The within-assay coefficient of variation of the cholesterol efflux assay was 5.5%. 2.1.3. Western blot analyses Cells were harvested and protein extracts prepared as previously described [10–12]. They were then subjected to Western blot analyses (10% SDS-PAGE; 25 ␮g protein per lane) using rat antiABCA1 antiserum (kindly donated by Dr. S. Yokoyama of Nagoya City University) [13], and rabbit anti-ABCG1 (Novus Biologicals, Littleton, CO, USA)-, LXR␣ (PP-PPZ0412-00, Perseus Proteomics, Tokyo, Japan)-, LXR␤ (PP-K8917-00, Perseus Proteomics)-, PPAR␥ (sc-7196, Santa Cruz Santa Cruz, CA)- and ␤-actin (OOOO, Santa Cruz)-specific antibodies. The proteins were visualized and quantified using a chemiluminescence method (ECL Plus Western Blotting Detection System; GE Healthcare UK Ltd) and the NIH image analysis software program. 2.1.4. Real-time quantitative RT-PCR At the indicated hours after treatment with PPAR␥ agonists or other compounds, total RNA was extracted from the cells, and first-strand cDNA was synthesized from the total RNA (500 ng) by placing in a Reverse Transcription Reagent (Applied Biosystems, Foster City, CA, USA). Quantitative PCR was performed with a Perkin–Elmer 7900 PCR machine, TaqMan PCR master mix and FAM-labeled TaqMan probes (Assays-on-Demand, Applied Biosystems) for human or mouse ABCA1 (Hs00194045), ABCG1 (total transcripts (Hs00245154) and transcripts containing exons 1–2 (Hm207627)), LXR␣ (Hs01105193), sterol responsive element binding protein-1 (SREBP-1, Hs00231674), CD36 (Hs01120071) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, Hs99999905). Expression data were normalized for GAPDH levels. To determine the relative abundance of total and exons 1–2-containing ABCG1, we used a standard curve based on serial dilution of an ABCG1 cDNA containing 2 target sequences for the TaqMan primer probe. The 994-bp standard cDNA was prepared by PCR amplification using the RT transcripts of THP-1 macrophages as a template. The primer sequences were designed based on exon 1 (forward: 5 -CCTGCAGTGGTGAAAACTGA-3 ) and exon 15 (reverse: 5 -reverse, AGACCACCTGGAAGCAGGA-3 ). 2.1.5. siRNA-mediated macrophage RNA interference Human PPAR␥-specific small interfering RNA (siRNA) and scrambled control RNA oligonucleotides were purchased from Ambion Inc (Austin, TX, USA). The transfection of siRNA was performed using TransIT-TKO Transfection Reagent according to the manufacturer’s instructions (Mirus Bio Corporation, Madison, WI, USA). Briefly, 40 nmol/L of scrambled control RNA oligonucleotide or PPAR␥-siRNA were added to THP-1 cells 48 h after treatment with PMA for differentiation. The cells were incubated for a further 24 h, washed and the indicated doses of pioglitazone added. They were harvested 24 h later and mRNA levels determined as described above. The oligonucleotide sequences used to construct siRNA used in this study were: 5 -GGAUGCAAGGGUUUCUUCCtt-3 and 5 -GGAAGAAACCCUUGCAUCCtt-3 for PPAR␥ (PPAR␥-siRNA). 2.1.6. Cloning and generation of recombinant adenoviruses encoding for mouse sulfotransferase family cytosolic 2B member 1 A recombinant adenovirus expressing mouse cholesterol sulfotransferase family cytosolic 2B member 1 (Ad-mSult2b1) was produced using the ViraPower Adenoviral Expression System (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Briefly, to generate an entry clone of the Gateway system (Invitrogen), cloning of the open reading frame into a pENTR/D-TOPO vector (Invitrogen) was carried out using first strand complementary DNA derived from mouse skin as a template and the specific primers as follows: forward: 5 -CAC CAT GGA

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

143

Fig. 1. Genomic organization of human ABCG1 and LXREs. (A) The 23 exons (indicated by shaded boxes, numbers indicate size of each exon) and the introns (numbers indicate size of each intron) of the human ABCG1 gene are illustrated (window from the Ensembl Genome Browser (http://uswest.ensembl.org/index.html)). Putative promoters (bold lines), LXRE (closed ellipses), translation initiation sites (horizontal arrows), transcription start sites driven by promoter A and B (vertical arrow heads) and locations of Taqman primer and probe sets for real-time RT-PCR (arrow heads) are also indicated. Reported transcripts are indicated by closed boxes connected by dashed lines with the corresponding reference numbers in parentheses. (B) DNA sequences of LXREs and direct repeat 4 (DR4) hexanucleotide repeat bases are shown in uppercase. Mutated nucleotides for the electrophoretic mobility shift assay and the promoter assays are indicated in lowercase. Numbers indicate first and last nucleotide of DR4 in human chromosome 22. To facilitate comparison, sequences are written in the direction in which the DR4 motifs follow the script RGGTYActnnMGKTCA. However, the numbers at the top of each alignment ascend in the centromeric-to-telomeric direction for the human ABCG1 gene.

CGG GCC GCA GCC-3 – reverse: 5 -TTA TTG TGA GGA TCC TGG GTT GGG-3 . An expression clone for adenoviral vector was then generated by performing an LR recombination reaction between the entry clone and a pAd/CMV/V5-DEST (Invitrogen) according to the manufacturer’s protocol. The recombinant adenoviral plasmid was purified, and then transfected into 293A cells. After a sufficient cytopathic effect had been observed in the 293A cells, the adenovirus was purified using the Adeno-X Virus Purification Kit (Clontech, Pao Alto, CA, USA). Adenoviral vector expressing luciferase (Ad-Luc) was kindly donated by Dr. Santamarina-Fojo S of National Institute of Health (NIH), and used as a control. The adenovirus titer in plaque-forming units was determined by a plaque formation assay following infection of HEK293 cells. The multiplicity of infection (MOI) was defined as the ratio of the total number of plaque-forming units to the total number of cells that were infected. 2.1.7. Construction of luciferase reporter plasmids The human ABCA1 promoter region spanning −940 to +110 bp (hABCA1-940 + 110) was PCR-amplified using ABCA1-specific primers with HindIII sites (forward: 5 -CCCAAGCTTAAGTTGGAGGTCTGGAGTGG-3 – reverse: 5 -CCCAAGCTTACCGGCTCTGTTGGTGCGCGG-3 ) and a plasmid containing the nucleotides–2245 to +110 of the human ABCA1 promoter region derived from

the human BAC clone as a template (kindly provided by Dr. S. Santamarina-Fojo of the NIH). The PCR-amplified product was ligated into a pGL3 Basic vector (Promega, Madison, WI, USA) and confirmed by DNA sequencing. To obtain a reporter construct with mutations in direct repeat 4 (DR4) element (designated as DR4mut) within the LXR-responsive element (LXRE) of the ABCA1 promoter, we performed site-directed mutagenesis using a Quick Change II Site-Directed Mutagenesis Kit (Stratagene La Jolla, CA, USA) and the respective primers – forward: 5 -CGAGCGCAGAGGTTACTATCTGCAGAAGCCTGTGCTCTCCC-3 – reverse: 5 -GGGAGAGCACAGGCTTCTGCAGATAGTAACCTCTGCGCTCG-3 . Fig. 1A shows the genomic organization and reported transcripts of human ABCG1. Lorkowski et al. [14] identified several transcripts containing variable combinations of exons 1–6 by performing 5 RACE with a primer corresponding to exon 7 and cDNA derived from THP-1 macrophages. These transcripts are assumed to result from alternative splicing and/or the use of four putative promoters postulated to be upstream of exons 1, 4, 5, or 6. Four additional transcripts were identified by Kennedy et al. [15] when they performed 5 -RACE with a primer corresponding to exon 15 and RNA from THP-1 macrophages treated with an LXR ligand. These new transcripts resulted in the identification of exons 8–10, a putative promoter, and two LXREs upstream of exon 8 (Fig. 1A). However, they observed that the activity of this promoter in luciferase

144

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

constructs was only 2-fold of a promoter-less luciferase vector, implying that it did not greatly contribute to overall ABCG1 expression. Recently, Uehara et al. [16] reported an LXRE in the promoter upstream of exon 1, which is designated as promoter A in the present study (Fig. 1A and B), and Sabol et al. [17] reported that the promoter upstream of exon 5, designated as promoter B, was LXR ligand-responsive in the presence of three LXREs located in intron 5 or 7 (Fig. 1). To date, the promoter activity of the 5 -flanking region upstream of exons 4 and 6 has not been examined. Therefore, to investigate the mechanisms for transcriptional regulation of human ABCG1 by PPAR␥ ligands, we focused on the expression of transcripts driven by promoters A and B. Luciferase reporter plasmids that contain the human ABCG1 promoter A spanning −1104 to +37 bp (upstream of exon 1) with or without mutations in DR4 (hABCG1A-1104 + 37 and hABCG1A1104 + 37DR4mut, respectively), and those that contain promoter B spanning −1180 to +144 bp (upstream of exon 5, hABCG1B1180 + 144) were generated as previously reported [16]. To obtain a reporter construct containing human ABCG1 promoter spanning −5544 to +144 bp, DNA fragments were PCR-amplified from human genomic DNA using primers with KpnI/MluI sites – forward: 5 -GGGGTACCAGAGCTGCAGTATGGGCTGT-3 – reverse: 5 GGACGCGTGCGGTGCCGACCGAGAA-3 . The PCR-amplified product was ligated into a pGL3 Basic vector (hABCG1B-5544 + 144). The intronic DNA sequence LXRE-A, which was reported to be LXRE with DR4 (Fig. 1 A and B) [17], was PCR amplified from human genomic DNA with primers – forward: 5 -GCTGTAAGCCCACAGTGCAT-3 – reverse: 5 -TGTCTGCGGCCCCATTCTAT. The amplified products were cloned into the SalI site downstream of the firefly luciferase gene in the hABCG1B-1180 + 144 plasmid (hABCG1B-LXRE-A). Mutagenesis in DR4 of LXREA (hABCG1B-LXRE-A DR4mut) was performed using a KOD mutagenesis kit (TOYOBO, Tokyo, Japan) with primers – forward: ctaGtCTGTCAATGGAGAACGGAACTCGAG-3 – reverse: 5 -TTAGGACGACTCGGGACGACTGAAGT-3 (lowercase indicates mutated bases). We prepared the LXRE-B region (47 bp), which is located in intron 7 (Fig. 1), with SalI ends and with or without mutations in DR4 as synthetic complementary oligodeoxynucleotides, and then cloned these regions (hABCG1B-LXRE-B/DR4mut). The oligodeoxynucleotide sequence for LXRE-B was 5 -CCGCCCGCGCCGGGGTTACTACCGGTCAACGCTCGCTAGTAACCTCC-3 , and the sequence for LXRE-B mut-DR4 was 5 -CCGCCCGCGCCGGcGaTtCTACCtGagAACGCTCGCTAGTAACCTCC-3 (lowercase indicates mutated bases). 2.1.8. DNA transfection and luciferase assays RAW and HEK293 cells, which had been cultured in 24-well plates, were transfected with 500 ng of luciferase reporter plasmids and 12.5 ng of phRL-TK (Promega) per well using Lipofectamine LTX reagent (Invitrogen) according to the manufacturer’s instructions. Ten h after transfection, the media were replaced with DMEM containing pioglitazone, TO1317, or the vehicle and incubated for an additional 24 h. Luciferase assays were performed as previously described [10,11]. 2.1.9. Human study Fifteen type 2 diabetic patients, who had been treated with pioglitazone at least for 6 months, were recruited to participate in a crossover study where pioglitazone was compared with placebo. The subjects were randomly assigned to either of 2 groups, which were first orally treated with pioglitazone (30 mg) or placebo after an overnight (12 h) fast. Venous blood samples were taken at 0 and 2 h after the respective treatments. At the next visit to physicians (4–8 weeks later), the treatments were switched and the same protocol was applied. Timing of post-administration blood

sampling was determined from the pharmacokinetic information for pioglitazone which states that blood levels peak at 2 h after oral administration in humans. For cholesterol efflux assays, RAW cells were cultured in DMEM containing 20% sera obtained before and after pioglitazone or placebo for 24 h in the presence of [3 H]cholesterol pre-equilibrated with AcLDL as described above. After washing, the cells were incubated in the presence of 10 ␮g/mL of human apoA-I or 50 ␮g/mL of HDL for 24 h. Determination of cholesterol efflux from the cells was performed as described above. All subjects provided written informed consent. The study was approved by the Ethical Committee of the National Defense Medical College. 2.1.10. Statistical analysis Statistical analyses were performed using the Stat View Version 5.0 software package (SAS Institute Inc., NC, USA). The paired t-test was used for paired samples (human study) and one-factor factorial ANOVA followed by post hoc analysis using Bonferroni/Dunn test for cell culture studies, with a value of P < 0.05 considered to be significant. All results were expressed as the mean ± SD. 3. Results 3.1. Pioglitazone promotes apoA-I- and HDL-mediated cholesterol efflux by increasing ABCA1 and ABCG1 expression in macrophages As shown in Fig. 2A and B, pioglitazone dose-dependently promoted both apoA-I- and HDL-mediated cholesterol efflux from human THP-1 macrophages and mouse RAW cells, observations consistent with previous studies. To examine underlying molecular mechanisms, we determined ABCA1, ABCG1, PPAR␥ and LXR expressions. Fig. 2C shows that all molecules, except PPAR␥ and LXR␤, were increased by treatment with pioglitazone in a dose-dependent fashion. Rosiglitazone, another thiazolidinedione, exerted a similar effect, indicating that the expression of these molecules was enhanced via PPAR␥ activation. 3.2. ABCA1 and ABCG1 are transcriptionally regulated by PPAR with complete and partial dependence on LXR˛, respectively In parallel with protein levels, pioglitazone dose-dependently increased ABCA1 and ABCG1 mRNA levels in both THP-1 macrophages (Fig. 3A) and RAW cells (Fig. 3B). We confirmed that rosiglitazone also stimulated ABCA1/G1 expressions. Concomitant treatment with actinomycin D abolished ABCA1/G1 mRNA induction due to pioglitazone (Fig. 3C), indicating that pioglitazone transcriptionally regulated ABCA1/G1 expressions. To investigate whether PPAR␥ was involved in the pioglitazonemediated mRNA induction, we determined mRNA levels in THP-1 macrophages under knockdown of the expression of PPAR␥ using siRNA. As expected, PPAR␥-siRNA brought about an 80% and 85% reduction in mRNA and protein expression, respectively (Fig. 4A). Under PPAR␥ knockdown, pioglitazone-induced ABCA1/G1 and LXR␣ expressions were almost completely cancelled (Fig. 4B), indicating that pioglitazone induces ABCA1/G1 in a PPAR␥-dependent fashion. Next, to further assess the role of LXR in pioglitazone-mediated effects on ABCA1/G1, we used an adenovirus vector encoding a mouse oxysterol catabolic enzyme, cholesterol sulfotransferase (Ad-mSult2b1), to inactivate LXR signaling by depleting oxysterols, which are LXR ligands [18]. Fig. 4C shows that Sult2b1 overexpression completely cancelled the increased expressions of ABCA1 and ABCG1 induced by oxysterol 22HC. Sult2b1 overexpression also resulted in complete abolishment of pioglitazone-induced ABCA1 expression. In contrast, it partially attenuated the effect of pioglitazone on ABCG1 expression, suggesting that the effects of

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

Fig. 2. Pioglitazone enhances cholesterol efflux and increases ABCA1, ABCG1 and LXR␣/␤ expressions in THP-1 macrophages and RAW264.7 cells. After THP-1 macrophages and RAW264.7 cells (RAW) were labeled with 3 H-cholesterol, the indicated doses of PPAR␥ agonists or the vehicle were added to the cultures, which were then incubated in the presence of 10 ␮g/mL of human apoA-I (A) or 50 ␮g/mL of HDL (B) for 24 h. Cholesterol efflux was determined as described in Section 2. The results from 4 samples are representative of 3 or more experiments and are presented as mean ± SD. *P < 0.05 vs. control. C, Western blot analysis was performed as described in Section 2 to determine ABCA1, ABCG1, PPAR␥ and LXR␣/␤ protein levels in the absence (for THP-1 macrophages and ABCG1/SR-BI/LXRs in RAW) or presence (for ABCA1 in RAW) of 0.3 mmol/L of 8-Br-cAMP. Incubation conditions were the same as those used for the cholesterol efflux experiments.

PPAR␥ activation on ABCA1 mRNA levels were completely LXR␣ pathway-dependent; as opposed to partial dependence for ABCG1. We confirmed that Sult2b1 also abolished the stimulatory effect of 22HC on another LXR target gene, SREBP-1, which was not significantly affected by pioglitazone. The canonical PPAR␥ target gene CD36 was indeed increased by pioglitazone, though this was not affected by Sult2b1 overexpression. 3.3. Pioglitazone activates ABCA1 promoter in LXR-dependent manner To further elucidate the molecular mechanisms for pioglitazone-induced ABCA1 expression, we performed luciferase assays with reporter constructs containing human ABCA1 promoters. Since DR4 in LXRE is a pivotal sequence for LXR binding among human ABCA1 promoters, we introduced mutations in DR4 (Fig. 5A). For the wild type DR4, pioglitazone activated ABCA1 promoter 3 fold, which was comparable to TO1317, a synthetic LXR agonist. In contrast, the stimulatory effects of TO1317 and pioglitazone were completely abolished when mutant DR4 was introduced, supporting the above observation that pioglitazone enhanced ABCA1 expression in an LXR-dependent manner.

145

Fig. 3. Pioglitazone transcriptionally upregulates ABCA1 and ABCG1 expressions in THP-1 macrophages and RAW cells. THP-1 macrophages (A) and RAW cells (B) were treated with the indicated concentrations of pioglitazone (Pio) or 10 ␮mol/L of rosiglitazone (Ros), or the vehicle for 24 h. For ABCA1 in RAW cells, 8-Br-cAMP (0.3 mmol/L) was added together with the other compounds. C, THP-1 macrophages were incubated with 10 ␮mol/L of pioglitazone or the vehicle in the absence or presence of actinomycin D (ActD, 10 ␮g/mL) for 12 h. RNA extraction and real-time quantitative RT-PCR were performed as described in Section 2. The mRNA levels of each gene were standardized for GAPDH levels. The results from 3 samples are representative of 3 or more experiments, expressed relative to the controls and presented as mean ± SD. *P < 0.05 vs. control.

3.4. ABCG1 transcripts are driven by two promoters, A and B, via PPAR-LXR pathway In consideration of the genomic organization, reported transcripts, promoters (A and B), and LXREs of human ABCG1 (Fig. 1A), we next explored mechanisms for transcriptional regulation of human ABCG1 by PPAR␥ with particular focus on the expression of transcripts driven by promoters A and B. We performed quantitative real-time RT-PCR with exons 1–2 and 13–14 as specific primers and probe sets (Fig. 1A) in order to determine levels of ABCG1 transcripts containing exons 1–2 driven by promoter A, and total ABCG1 transcripts. As shown in Fig. 6A, the levels of the transcripts containing exons 1–2 were substantially lower than those of total transcripts in THP-1 macrophages. Luciferase assays further revealed that promoter A activity was 20-fold lower than that of promoter B (Fig. 6B), suggesting the reason for amounts of exons 1–2-transcripts being relatively small compared to total transcripts. The levels of both exons 1–2-transcripts and total transcripts were increased by pioglitazone in a dose-dependent manner (Fig. 6C and D). Fig. 6E–H summarizes a series of time course experiments where LXR␣, CD36, ABCA1, and ABCG1 expressions in THP-1 macrophages

146

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

Fig. 4. Involvement of PPAR␥ and LXR in pioglitazone-induced ABCA1 and ABCG1 expressions in THP-1 macrophages. (A and B) Incubation was conducted with scrambled siRNA or PPAR␥-siRNA during THP-1 differentiation into macrophages as described in Section 2. Twenty-four hours after treatment with the siRNAs, RNA and protein extractions were performed, followed by real-time quantitative RT-PCR and Western blot analysis for PPAR␥ (A) as described in Section 2. Further incubation for 24 h was performed with 10 ␮mol/L of pioglitazone (Pio) or the vehicle, and then cells were harvested for mRNA measurement (B). (C) THP-1 macrophages were treated with 100 MOI of adenoviral vectors expressing mouse Sult2b1/luciferase (Luc) for 24 h as described in Section 2. After washing, the cells were treated with 22-hydroxycholesterol (22HC, 10 ␮mol/l) or Pio (10 ␮mol/l) for 24 h. Real-time quantitative RT-PCR was performed as described in Section 2. The mRNA levels of each gene were standardized for GAPDH levels. The results from 3 samples are representative of 3 or more experiments, expressed relative to the controls and presented as mean ± SD. *P < 0.05 vs. control.

were monitored in the presence of 10 ␮mol/L of pioglitazone. LXR␣ and CD36, genes directly targeted by PPAR␥, were already increased at 2 h after treatment, whereas ABCA1 and ABCG1 (total transcripts) levels were more modestly elevated at 6–12 h relative to LXR␣ and CD36. These observations imply that ABCA1/G1 are not directly transactivated by PPAR␥, except for exons 1–2-transcripts of ABCG1, which were increased 2–6 h after treatment. To explore the LXR-independent mechanisms for PPAR␥mediated ABCG1 expression, we determined the activities of human ABCG1 promoter A and B in the absence or presence of reported LXREs. The activity of the less dominant promoter A was stimulated by pioglitazone as well as by TO1317 (Fig. 7A) and these LXR␣/PPAR␥ agonist–induced activations were completely abolished by the introduction of the mutated DR4 (Figs. 1B and 7A). Fig. 7B shows the results for human ABCG1 promoter B. In the presence of the sequences containing LXRE-A or -B downstream of the luciferase gene, LXR␣/PPAR␥ agonists exerted stimulatory effects on promoter B, and we confirmed that mutagenesis in these LXREs (Fig. 1B) cancelled LXR␣/PPAR␥-mediated transactivation. In view of the above findings, we further extended upstream sequences of human ABCG1 promoter B up to 5544 bp from the transcription start site in order to further explore the LXRindependent pathways of PPAR␥, but no PPAR␥-responsiveness was evident (Fig. 7B). Rosiglitazone had similar effects on ABCA1/G1 promoters to pioglitazone (data not shown).

3.5. Sera obtained from diabetic patients after acute administration of pioglitazone promote cholesterol efflux from macrophages by increasing ABCA1/G1 Next, we investigated whether the above in vitro observations translate to human physiology. Ex vivo experiments were performed using 15 type 2 diabetic patients (clinical characteristics are shown in Supplemental Table) to assess the effects of acute administration of pioglitazone on cholesterol efflux. Pioglitazone and placebo were administered in a cross-over design and RAW cells were cultured for 24 h in the presence of 20% sera obtained before administration and 2 h afterwards, and then cholesterol efflux assays were performed on the cells. Fig. 8A and B summarize individual results for apoA-I- and HDL-mediated cholesterol efflux at baseline and after administration. Surprisingly, a single administration of pioglitazone significantly promoted both apoA-I- and HDL-mediated cholesterol efflux from the macrophages, by 32% and 7.5%, respectively, compared to baseline. As expected, placebo did not affect cholesterol efflux (Fig. 8C and D). Finally, we performed immunoblots for ABCA1/G1 to further elucidate the molecular mechanisms for the observed acute effects of pioglitazone. Supporting the increase in apoA-I- and HDLmediated cholesterol efflux, ABCA1/G1 protein levels were also increased by pioglitazone, but not placebo, in 3 representative subjects (Fig. 8E and F). Quantifying the resulting ABCA1 and ABCG1

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

147

Fig. 5. Pioglitazone transactivates ABCA1 via LXRE. (A) Shows wild type and mutated direct repeat 4 (DR4) in LXRE of human ABCA1 promoter. Upper case characters indicate half-sites of the DR4 element. Mutated nucleotides for the promoter assays are indicated in lowercase and by vertical arrows. (B) After transfection with hABCA1940 + 110 or hABCA1-940 + 110 DR4mut, RAW cells were treated with 10 ␮mol/L of pioglitazone (Pio), 1 ␮mol/L of TO1317, or the vehicle (Cont). Twenty-four h after the treatments, the cells were lysed and a luciferase assay performed. The results from 3 samples are representative of 3 or more experiments, expressed relative to the controls and presented as mean ± SD. *P < 0.05 vs. control.

protein levels using a chemiluminescence method revealed significant increases of 1.8 ± 0.4 fold and 1.4 ± 0.3 fold, respectively, in RAW cells cultured with post-pioglitazone sera as compared to prepioglitazone sera (Fig. 8G and H). Again, there was no difference between pre- and post-placebo sera (data not shown). 4. Discussion The present study demonstrated that pioglitazone enhanced cholesterol efflux from macrophages by increasing ABCA1 and ABCG1 expressions, which were completely and partially mediated by an LXR-dependent pathway, respectively. Promoter assays revealed that pioglitazone activated ABCA1 and ABCG1 promoters in an LXR-dependent manner. We also demonstrated, for the first time, that human ABCG1 promoter B was the primary contributor to total transcripts of ABCG1, rather than promoter A. Furthermore, in a cross-over human study using diabetic subjects, macrophages treated with sera following a single oral dose of pioglitazone significantly enhanced apoA-I- and HDL-mediated cholesterol efflux by increasing ABCA1/G1. Our comparable in vitro and ex vivo results therefore indicate that potential anti-atherogenic properties of pioglitazone could, at least in part, be due to up-regulation of cholesterol efflux from macrophages by ABCA1 and ABCG1 via LXR-dependent or–independent pathways. Previous studies in this respect had been limited to mice [19,20] but we observed for the first time that PPAR␥ agonists promoted LXR␣ expression in a PPAR␥-dependent manner in human cells, resulting in increased ABCA1 mRNA levels in them. Ogata et al. [21] reported that LXR␣-knockdown resulted in abolishment of pioglitazone-induced LXRE-containing thymidine kinase promoter in human embryonic fibroblast, WI38 cells. Although their findings are likely to apply to the ABCA1 promoter, we demonstrated in the present study that ligands for LXR␣/PPAR␥ activated human ABCA1 promoter in a LXRE-dependent manner. Furthermore, inhibition of LXR activation by overexpression of Sult2b1 totally cancelled pioglitazone-induced ABCA1 expression in THP1 macrophages (Fig. 4C). Also, previously, deletion of both LXR␣/␤ was seen to completely cancel ABCA1 induction due to treatment

Fig. 6. Relative expression levels of the transcriptional variants and effects of pioglitazone, and relative activities of dual promoters A and B. (A) After RNA extraction and cDNA synthesis from differentiated THP-1 macrophages, real-time quantitative RT-PCR was performed to determine the levels of ABCG1 transcripts with specific primers for exons 1–2 or 13–14, as described in Fig. 1A and Section 2. To determine the relative abundance of total transcripts and exons 1–2-containing ABCG1, we used a standard curve based on serial dilution of an ABCG1 cDNA containing two target sequences for the TaqMan primer probe as described in Section 2. The results from 3 samples are expressed relative to the expression levels of transcripts containing exons 1–2. (B) Twenty hours after transfection with hABCG1A-1104 + 37, hABCG1B-1180 + 144, or pGL3 Basic (empty) vector, HEK293 cells were lysed and a luciferase assay was performed. C, D, THP-1 macrophages were treated with the indicated dose of pioglitazone or the vehicle for 24 h, and RNA extraction, cDNA synthesis and determination of total (C) or exons 1–2-containing (D) transcripts of ABCG1 were performed using primers and probes as described in Fig. 1A. (E–H) THP-1 macrophages were treated with 10 ␮mol/L of pioglitazone for the indicated times. (F and H) Closed triangles and circles respectively indicate total transcripts of ABCG1 and those containing exons (ex) 1–2. RNA extraction and real-time quantitative RT-PCR were performed as described in Section 2. The mRNA levels of each gene were standardized for GAPDH levels. The results from 3 samples are expressed relative to the controls (or the empty vector for luciferase assays) and presented as mean ± SD. *P < 0.05 vs. control.

148

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

Fig. 7. Effects of pioglitazone on transcriptional activity of human ABCG1 promoters. (A) After transfection with hABCG1A-1104 + 37/DR4mut (Fig. 1B), RAW cells were treated with 1 ␮mol/L of TO1317, 10 ␮mol/L of pioglitazone (Pio), or the vehicle (Cont). (B) After transfection with hABCG1B-1180 (-5544) + 144, hABCG1B-LXREA/DR4mut and hABCG1B-LXRE-B/DR4mut (Fig. 1B), RAW cells were treated with TO1317, Pio, or the vehicle (Cont) as described above. Twenty-four h after the treatments, the cells were lysed and a luciferase (Luc) assay performed. The results from 3 samples are representative of 3 or more experiments, expressed relative to the controls and presented as mean ± SD. *P < 0.05 vs. control.

with PPAR␥ ligands in mice [9]. LXR␤ reportedly compensated for LXR-mediated transactivation of the respective target genes, including ABCA1/G1, in LXR␣ deficient mice [22,23]. In our study, however, LXR␤ was not affected by PPAR␥ agonists (Fig. 2C) consistent with previous studies [9,20]. Taken together with the similar findings of our previous study [11], these observations indicate that, similar to mice, PPAR␥-mediated ABCA1 induction in human macrophages is dependent on LXR␣. Some investigators have reported that ABCG1 transcription is regulated by several promoters but it is still not clear which promoters contribute to overall ABCG1 mRNA expression. Though Kennedy et al. [24] found that there were transcripts containing exon 8 in THP-1 macrophages (Fig. 1A), they observed that the sequences upstream of exon 8 had scarcely any promoter activity as described above. Lorkowski et al. [14] identified transcripts in THP-1 monocytes that were driven by promoter A, located 19.3 kb upstream from the originally described promoter B [25]. In the present study, we investigated the relative contribution of these promoters in macrophages and found, for the first time, that the activity of promoter B was substantially greater than that of promoter A. This was supported by the observation that the levels of ABCG1 transcripts containing exons 1–2, which are driven by promoter A, only account for negligible amounts of total ABCG1 transcripts in THP-1 macrophages (Fig. 6A). Regarding LXR-mediated regulation of ABCG1, Uehara et al. [16] and Sabol et al. [17] identified an LXRE in promoter A and

Fig. 8. Acute administration of pioglitazone in diabetic patients enhances ABCA1/G1 expressions and cholesterol efflux from macrophages. (A–D) Sera were isolated before and 2 h after oral administration of pioglitazone (30 mg) or placebo from 15 type 2 diabetic patients. RAW cells were cultured in DMEM containing 20% sera obtained before (pre) and after (post) pioglitazone or placebo for 24 h in the presence of [3 H]-cholesterol pre-equilibrated with AcLDL, and subjected to cholesterol efflux and Western blot analysis. After washing, the cells were incubated in the presence of 10 ␮g/mL of human apoA-I or 50 ␮g/mL of HDL for 24 h. Determination of cholesterol efflux from the cells was performed as described in Section 2. Individual changes in mean cholesterol efflux from 4 replicate wells are shown, and the results are also expressed as the mean ± SD. E, F, Representative blots for ABCA1, ABCG1 and ␤-actin from 3 subjects before and after administering pioglitazone (E) or placebo (F). (G and H) Protein levels were quantified as described in Section 2. Data from 3 independent experiments are expressed as mean ± SD. *P < 0.05 vs. pre.

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

LXRE-A/B in introns 5 and 7, respectively, as described above. Nonetheless, it remains unclear whether such LXREs confer the LXR ligand responsiveness of human ABCG1 transcription in vivo. However, these sequences might be pivotal in the LXR pathway regulating ABCG1 expression because they are evolutionarily highly conserved among various species, and other researchers [26] determined that LXRs and their cofactors were recruited for these intronic LXREs in vivo based on ChIP analysis. It was also observed that luciferase constructs containing the LXREs (especially LXREB) downstream from the reporter gene in either orientation were highly responsive to LXR agonist treatment [17], indicating that LXREs might act as an enhancer. Furthermore, their greater distance from the ABCG1 promoter B would not be prohibitively long, as previously reported [27]. It is noteworthy that the ligands for PPAR␥/LXR␣ activated both promoter A and B LXRE-dependently (Fig. 7). In conjunction with promoter B, the 2 intronic LXREs, LXRE-A and -B, produced PPAR␥/LXR␣ ligand-responsiveness and this was cancelled when mutations were introduced into DR4. Similar to the case of ABCA1, a rise in ABCG1 mRNA levels followed pioglitazone stimulated LXR␣ expression (Fig. 6E and H) supporting the idea that regulation of ABCG1 levels by pioglitazone was also LXR␣-dependent. However, LXR pathway-inhibition by Sult2b1-mediated depletion of endogenous oxysterols partially attenuated pioglitazone-induced ABCG1 levels, in contrast to ABCA1, whose levels were completely cancelled by Sult2b1 overexpression. The observation for ABCG1, however, is supported by the finding of Li et al. [9] that a PPAR␥ agonist did stimulate ABCG1 expression in LXR␣/␤ double knockout mice. Also, a region further upstream, up to 5544 bp from the transcription starting site of promoter B, did not induce pioglitazone-responsiveness, indicating the need for future studies to determine the exact LXR-independent pathway involved in the pioglitazone-induced effect on ABCG1 expression. In the clinical setting, 2 trials demonstrated that pioglitazone treatment achieved significant regression of atherosclerotic plaques as detected by carotid ultrasonography [3] and coronary intravascular ultrasonography, as compared with glimepiride, though both drugs provided comparable glycemic control [2]. Overall, these favorable effects of pioglitazone on atherosclerosis can be interpreted as non-metabolic, pleiotropic effects on the vasculature. In the present study, the vitro findings were borne out by an ex vivo experiment which showed that ABCA1/G1 expressions and cholesterol efflux from macrophages were enhanced when cultured with sera obtained from diabetic patients after administration of a clinical dose of pioglitazone (Fig. 8). Previously, plasma concentrations were reported to be 3.9 ± 1.2 ␮mol/L at 2 h after oral administration of pioglitazone (30 mg) in adult males, concentrations where we observed increased cholesterol efflux from macrophages in vitro (Fig. 3). Based on our comparable in vitro and ex vivo observations, it is conceivable that pioglitazone would provide macrophage-associated anti-atherogenesis in humans, in vivo. However, there were limitations on the ex vivo study. There was the possibility of bias as it was not blinded and had a small sample size, but this was overcome by having it performed by investigators who were unaware of the randomization status. Also, single oral administration of pioglitazone may only represent an acute effect, differing from the chronic effect. In addition, experiments using THP-1 and human serum before and after administration of pioglitazone might have provided further evidence for our conclusions but we did not have enough samples to perform the assays. In conclusion, pioglitazone enhanced cholesterol efflux from macrophages in vitro and ex vivo through increases in ABCA1 and ABCG1, which were regulated in LXR-dependent and -independent manners, respectively. Anti-atherosclerotic properties of

149

pioglitazone might therefore be associated, at least partly, with enhancement of the anti-atherogenic function of HDL. Acknowledgement We thank Mr. Alexander Cox for his help in the preparation of this manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atherosclerosis.2011.07.113. References [1] Bolen S, Feldman L, Vassy J, Wilson L, Yeh HC, Marinopoulos S, et al. Systematic review: comparative effectiveness and safety of oral medications for type 2 diabetes mellitus. Ann Intern Med 2007;147:386–99. [2] Nissen SE, Nicholls SJ, Wolski K, Nesto R, Kupfer S, Perez A, et al. Comparison of pioglitazone vs glimepiride on progression of coronary atherosclerosis in patients with type 2 diabetes: the PERISCOPE randomized controlled trial. J Am Med Assoc 2008;299:1561–73. [3] Mazzone T, Meyer PM, Feinstein SB, Davidson MH, Kondos GT, D’Agostino Sr RB, et al. Effect of pioglitazone compared with glimepiride on carotid intimamedia thickness in type 2 diabetes: a randomized trial. J Am Med Assoc 2006;296:2572–81. [4] Lewis GF, Rader DJ. New insights into the regulation of HDL metabolism and reverse cholesterol transport. Circ Res 2005;96:1221–32. [5] Wang N, Silver DL, Thiele C, Tall AR. ATP-binding cassette transporter A1 (ABCA1) functions as a cholesterol efflux regulatory protein. J Biol Chem 2001;276:23742–7. [6] Wang N, Lan D, Chen W, Matsuura F, Tall AR. ATP-binding cassette transporters G1 and G4 mediate cellular cholesterol efflux to high-density lipoproteins. Proc Natl Acad Sci U S A 2004;101:9774–9. [7] Yvan-Charvet L, Ranalletta M, Wang N, Han S, Terasaka N, Li R, et al. Combined deficiency of ABCA1 and ABCG1 promotes foam cell accumulation and accelerates atherosclerosis in mice. J Clin Invest 2007;117:3900–8. [8] Chinetti G, Lestavel S, Bocher V, Remaley AT, Neve B, Torra IP, et al. PPARgamma activators induce cholesterol removal from human macrophage foam cells through stimulation of the ABCA1 pathway. Nat Med 2001;7:53–8. [9] Li AC, Binder CJ, Gutierrez A, Brown KK, Plotkin CR, Pattison JW, et al. Differential inhibition of macrophage foam-cell formation and atherosclerosis in mice by PPARalpha, beta/delta, and gamma. J Clin Invest 2004;114:1564–76. [10] Ayaori M, Sawada S, Yonemura A, Iwamoto N, Ogura M, Tanaka N, et al. Glucocorticoid receptor regulates ATP-binding cassette transporter-A1 expression and apolipoprotein-mediated cholesterol efflux from macrophages. Arterioscler Thromb Vasc Biol 2006;26:163–8. [11] Nakaya K, Ayaori M, Hisada T, Sawada S, Tanaka N, Iwamoto N, et al. Telmisartan enhances cholesterol efflux from THP-1 macrophages by activating PPARgamma. J Atheroscler Thromb 2007;14:133–41. [12] Uto-Kondo H, Ayaori M, Ogura M, Nakaya K, Ito M, Suzuki A, et al. Coffee consumption enhances high-density lipoprotein-mediated cholesterol efflux in macrophages. Circ Res 2010;106:779–87. [13] Iwamoto N, Abe-Dohmae S, Ayaori M, Tanaka N, Kusuhara M, Ohsuzu F, et al. ATP-binding cassette transporter A1 gene transcription is downregulated by activator protein 2alpha. Doxazosin inhibits activator protein 2alpha and increases high-density lipoprotein biogenesis independent of alpha1adrenoceptor blockade. Circ Res 2007;101:156–65. [14] Lorkowski S, Kratz M, Wenner C, Schmidt R, Weitkamp B, Fobker M, et al. Expression of the ATP-binding cassette transporter gene ABCG1 (ABC8) in Tangier disease. Biochem Biophys Res Commun 2001;283:821–30. [15] Kennedy MA, Venkateswaran A, Tarr PT, Xenarios I, Kudoh J, Shimizu N, et al. Characterization of the human ABCG1 gene: liver X receptor activates an internal promoter that produces a novel transcript encoding an alternative form of the protein. J Biol Chem 2001;276:39438–47. [16] Uehara Y, Miura S, von Eckardstein A, Abe S, Fujii A, Matsuo Y, et al. Unsaturated fatty acids suppress the expression of the ATP-binding cassette transporter G1 (ABCG1) and ABCA1 genes via an LXR/RXR responsive element. Atherosclerosis 2007;191:11–21. [17] Sabol SL, Brewer Jr HB, Santamarina-Fojo S. The human ABCG1 gene: identification of LXR response elements that modulate expression in macrophages and liver. J Lipid Res 2005;46:2151–67. [18] Chen W, Chen G, Head DL, Mangelsdorf DJ, Russell DW. Enzymatic reduction of oxysterols impairs LXR signaling in cultured cells and the livers of mice. Cell Metab 2007;5:73–9. [19] Akiyama TE, Sakai S, Lambert G, Nicol CJ, Matsusue K, Pimprale S, et al. Conditional disruption of the peroxisome proliferator-activated receptor gamma gene in mice results in lowered expression of ABCA1, ABCG1, and apoE in macrophages and reduced cholesterol efflux. Mol Cell Biol 2002;22:2607–19. [20] Chawla A, Boisvert WA, Lee CH, Laffitte BA, Barak Y, Joseph SB, et al. A PPAR gamma-LXR-ABCA1 pathway in macrophages is involved in cholesterol efflux and atherogenesis. Mol Cell 2001;7:161–71.

150

H. Ozasa et al. / Atherosclerosis 219 (2011) 141–150

[21] Ogata M, Tsujita M, Hossain MA, Akita N, Gonzalez FJ, Staels B, et al. On the mechanism for PPAR agonists to enhance ABCA1 gene expression. Atherosclerosis 2009;205:413–9. [22] Quinet EM, Savio DA, Halpern AR, Chen L, Schuster GU, Gustafsson JA, et al. X receptor (LXR)-beta regulation in LXRalpha-deficient mice: implications for therapeutic targeting. Mol Pharmacol 2006;70:1340–9. [23] Bradley MN, Hong C, Chen M, Joseph SB, Wilpitz DC, Wang X, et al. Ligand activation of LXR beta reverses atherosclerosis and cellular cholesterol overload in mice lacking LXR alpha and apoE. J Clin Invest 2007;117:2337–46. [24] Kennedy MA, Barrera GC, Nakamura K, Baldan A, Tarr P, Fishbein MC, et al. ABCG1 has a critical role in mediating cholesterol efflux to HDL and preventing cellular lipid accumulation. Cell Metab 2005;1:121–31.

[25] Langmann T, Porsch-Ozcurumez M, Unkelbach U, Klucken J, Schmitz G. Genomic organization and characterization of the promoter of the human ATP-binding cassette transporter-G1 (ABCG1) gene. Biochim Biophys Acta 2000;1494:175–80. [26] Jakobsson T, Venteclef N, Toresson G, Damdimopoulos AE, Ehrlund A, Lou X, et al. GPS2 is required for cholesterol efflux by triggering histone demethylation, LXR recruitment, and coregulator assembly at the ABCG1 locus. Mol Cell 2009;34:510–8. [27] Umesono K, Murakami KK, Thompson CC, Evans RM. Direct repeats as selective response elements for the thyroid hormone, retinoic acid, and vitamin D3 receptors. Cell 1991;65:1255–66.