OSBP-related Protein 8 (ORP8) Suppresses ABCA1 Expression and Cholesterol Efflux from Macrophages*

ORP8 is a previously unexplored member of the family of oxysterol-binding protein-related proteins (ORP). We now report the expression pattern, the subcellular distribution, and data on the ligand binding properties and the physiological function of ORP8. ORP8 is localized in the endoplasmic reticulum (ER) via its C-terminal transmembrane span and binds 25-hydroxycholesterol, identifying it as a new ER oxysterol-binding protein. ORP8 is expressed at highest levels in macrophages, liver, spleen, kidney, and brain. Immunohistochemical analysis revealed ORP8 in the shoulder regions of human coronary atherosclerotic lesions, where it is present in CD68(+) macrophages. In advanced lesions the ORP8 mRNA was up-regulated 2.7-fold as compared with healthy coronary artery wall. Silencing of ORP8 by RNA interference in THP-1 macrophages increased the expression of ATP binding cassette transporter A1 (ABCA1) and concomitantly cholesterol efflux to lipid-free apolipoprotein A-I but had no significant effect on ABCG1 expression or cholesterol efflux to spherical high density lipoprotein HDL2. Experiments employing an ABCA1 promoter-luciferase reporter confirmed that ORP8 silencing enhances ABCA1 transcription. The silencing effect was partially attenuated by mutation of the DR4 element in the ABCA1 promoter and synergized with that of the liver X receptor agonist T0901317. Furthermore, inactivation of the E-box in the promoter synergized with ORP8 silencing, suggesting that the suppressive effect of ORP8 involves both the liver X receptor and the E-box functions. Our data identify ORP8 as a negative regulator of ABCA1 expression and macrophage cholesterol efflux. ORP8 may, thus, modulate the development of atherosclerosis.

Lipid-laden macrophage foam cells are characteristic constituents of the early atherosclerotic lesion and are present at all stages of lesion development (1). Macrophages also play a central role in the inflammatory signaling within the developing plaque (2), and hydrolytic enzymes secreted by macrophages and other inflammatory cells influence plaque structure and stability (3). In addition to cholesterol, 27-carbon oxygenated derivatives of cholesterol, referred to as oxysterols (4), are enriched in atherosclerotic plaques in both humans and in animal models (5,6). Oxysterols have cytotoxic, pro-apoptotic, and pro-inflammatory effects and facilitate the differentiation of monocytes into macrophages. They have, therefore, been suggested to adversely affect lesion development and stability (6,7).
The central apparatus for the transcriptional regulation of sterol metabolism consists of the sterol regulatory elementbinding proteins (8,9) and the liver X receptors (LXRs) 2 (10 -12), both of which are responsive to oxysterols. The Insig proteins that control the intracellular transport and proteolytic activation of sterol regulatory element-binding protein (SREBP) act as 25-hydroxycholesterol receptors in the endoplasmic reticulum (ER) and mediate oxysterol regulation of SREBP maturation (13). The LXRs are activated by oxysterol ligands such as 22(R)-, 24(S)-, and 27-hydroxycholesterol and 24(S), 25-epoxycholesterol (14 -20). Together with the retinoid X receptor, the LXR form obligate heterodimers which facilitate via recruitment of coactivators, the transcription of specific target genes. Genes regulated by LXR are involved in sterol absorption in the intestine, the reverse cholesterol transport process, bile acid synthesis, biliary neutral sterol secretion, hepatic lipogenesis, and synthesis of nascent high density lipoproteins (10,11). Several ATP binding cassette (ABC) transporters are subject to transcriptional control by LXR (11). ABCA1 mediates phospholipid and cholesterol efflux to lipidpoor apolipoprotein A-I, whereas ABCG1 acts in cholesterol efflux to spherical high density lipoprotein particles (21,22). The removal of excess cholesterol from macrophages is regarded as an important anti-atherogenic process (1). Consistent with this, animal model and human genetic/epidemiologic studies support an atheroprotective function of ABCA1 (23)(24)(25)(26)(27), whereas the role of ABCG1 in atherogenesis is more controversial (28 -30).
A third protein family that can function as oxysterol sensors consists of oxysterol-binding protein (OSBP) and its homologues. OSBP is a cytoplasmic protein with affinity for several oxysterols (31)(32)(33). It plays a role in the transport of ceramide from the ER to the Golgi apparatus for sphingomyelin synthesis (34) and acts as a sterol-dependent scaffold that regulates the activity of extracellular signal-regulated kinases (35). Proteins displaying sequence homology to the C-terminal sterol binding domain of OSBP are present in most eukaryotic organisms (36,37). In humans the gene/protein family consists of 12 members (38,39). The mammalian OSBP-related proteins (ORP) have been implicated as sterol sensors that regulate a number of cellular functions ranging from sterol and neutral lipid metabolism to vesicle transport and cell signaling (40). However, their connections with the transcriptional control of cellular sterol homeostasis are relatively unexplored.
ORP8 is a previously unexplored member of the ORP family. We now report the expression pattern, the subcellular distribution, and data on the ligand binding properties and the physiological function of ORP8. Our results suggest that ORP8 acts as a sterol sensor that affects the reverse cholesterol transport process via modulation of ABCA1 expression and macrophage cholesterol efflux.

EXPERIMENTAL PROCEDURES
Antibodies and Other Reagents-A glutathione S-transferase (GST) fusion protein carrying amino acid residues 1-60 of human ORP8 was expressed in Escherichia coli BL21(DE3), purified by affinity chromatography on glutathione-Sepharose 4B (GE Healthcare), and used for immunization of New Zealand White rabbits according to a standard protocol. The ORP8 antibodies were bound on a glutathione-Sepharose 4B column carrying covalently coupled GST-ORP8 (1-60) and eluted with 0.2 M glycine, pH 2.8, neutralized, dialyzed against phosphatebuffered saline, 20% glycerol, and stored at Ϫ20°C.
Rabbit antibodies against calnexin were a kind gift from Prof. Ralf Pettersson (Ludwig Institute for Cancer Research, Stockholm, Sweden), monoclonal anti-␤-actin from the Developmental Studies Hybridoma Bank (University of Iowa), anti-protein disulfide isomerase from Stressgen (San Diego, CA), anti-ABCA1 from Novus Biologicals (Littleton, CO), and anti-CD68 from Dako (Glostrup, Denmark). Lipid-free human apoA-I was kindly provided by Dr. Peter Lerch (Swiss Red Cross, Bern, Switzerland). HDL 2 was purified from human plasma by ultracentrifugation (41 cDNA Constructs-Full-length human ORP8 cDNA (accession number NM_001003712) was inserted into the XbaI site of pcDNA4HisMaxC (Invitrogen) to obtain a construct fused with an N-terminal Xpress epitope tag. A truncated cDNA encoding ORP8 that lacks the 19-amino acid C-terminal transmembrane span (ORP8⌬C) was engineered by PCR. Furthermore, a GST fusion of ORP8 ligand binding domain (ORP8-(242-828)) was created in pGEX-1T (GE Healthcare) for protein production in E. coli. Sequence changes were verified by sequencing with a cycle-sequencing kit (BIGDYE, Applied Biosystems, Foster City, CA) and an automated ABI3730 sequencer (Applied Biosystems).
Western Blotting-Protein samples for SDS-PAGE were prepared by homogenizing cultured cells or mouse tissues in 250 mM Tris-HCl, pH 6.8, 8% SDS, protease inhibitor mixture (Roche Diagnostics). The crude extracts were cleared by centrifugation at 16,000 ϫ g for 3 min, and the protein concentration of the supernatant was determined by the D C assay (Bio-Rad). The proteins were electrophoresed on Laemmli gels and electrotransferred to Hybond-C Extra nitrocellulose (GE Healthcare). Nonspecific binding of antibodies was blocked with, and all antibody incubations were carried out in 5% fatfree powdered milk in 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 0.05% Tween 20. The bound primary antibodies were visualized with horseradish peroxidase-conjugated goat anti-rabbit or goat anti-mouse IgG (Bio-Rad) and the enhanced chemiluminescence system ECL (GE Healthcare).
Immunohistochemical Analysis of Human Tissue Specimens-Formalin-fixed, paraffin-embedded 5-m-thick human coronary artery sections were stained for ORP8, smooth muscle cell ␣-actin, CD31, and CD68 as described (42). Specificity of the ORP8 immunostaining was verified by simultaneous staining of adjacent coronary sections with antibody aliquots preincubated overnight at ϩ4°C with GST-ORP8 (1-60) or GST. In the single stainings the bound ORP8 antibody was detected using the avidin-biotin system (Vectastain ABC Elite rabbit kit, Vector Laboratories, Burlingame, CA) and 3,3Ј-diaminobenzidine (Sigma-Aldrich) as the chromogen. Sections were counterstained with Mayer's hematoxylin. For immunofluorescence double staining, sections were double-stained with rabbit anti-ORP8 and monoclonal anti-CD68. The bound antibodies were visualized using goat-anti rabbit IgG (HϩL) Alexa fluorescein isothiocyanate and goat-anti mouse Alexa 594 (Invitrogen), and the sections were counterstained with 4Ј,6-diamidino-2-phenylindole (Sigma-Aldrich). The stained samples were photo-graphed with a digital camera (Spot RT color operated with Spot advanced software, Version 4.6, Diagnostic Instruments, Sterling Heights, MI) attached to a Nikon Eclipse E600 microscope. Colocalization of ORP8 and CD68 in low magnification images was analyzed using Image-Pro Plus Version 6.1.0.346 for Windows XP Professional (Media Cybernetics, Bethesda, MD).
ORP8 mRNA Analysis of Human Coronary Artery Specimens-Samples from human left anterior descending coronary artery were obtained from recipient hearts at the time of cardiac transplantation or from the hearts of organ donors that could not be used for transplantation due to size or tissue type mismatch or coronary atherosclerosis. Coronary artery samples snap-frozen and stored at Ϫ80°C were dissected free of periarterial tissues, and the degree of atherosclerosis was evaluated under a stereo microscope. Areas without visible signs of lipid accumulation (corresponding to American Heart Association classes I-II) and areas with advanced atherosclerotic lesions (corresponding to AHA classes V-VI) were selected for RNA analysis. Total RNA was isolated, and the ORP8 mRNA was quantified by quantitative real-time reverse transcription-PCR as described below. The sample collection was conducted in accordance with local legislation and World Medical Association Declaration of Helsinki. The institutional ethical review board of Helsinki University Central Hospital and The National Authority for Medicolegal Affairs of Finland has approved the study protocol. All patients gave a written informed consent before sample collection.
In Vitro Assay for Oxysterol Binding by ORP8-Binding of [ 3 H]25OHC, 24(S)OHC, and 7-ketocholesterol to GST-ORP8-(242-828), purified from E. coli BL21(DE3) on glutathione-Sepharose 4B (GE Healthcare) according to the manufacturer's instructions, was assayed as described previously (43). Plain GST was used as a negative control. Briefly, the protein at ϳ0.1 M was incubated overnight with 5, 10, 20, and 40 nM 3 H-labeled oxysterols in the absence or presence of a 40-fold excess of the corresponding unlabeled sterols. The free sterol was thereafter removed with charcoal-dextran, and the proteinbound [ 3 H]sterol remaining in the supernatant was analyzed by liquid scintillation counting.

Quantitative Real-time Reverse Transcription-PCR-Cells
were homogenized in RLT buffer (Qiagen), and total RNA was isolated with RNeasy Mini kit according to the manufacturer's instructions. The total RNA (2 g) was treated with DNase I (Promega, Madison, WI) in the presence of RNase Inhibitor (Promega) and reverse-transcribed by using Superscript II (Invitrogen) and random hexamer primers (Applied Biosystems). Each RNA sample was amplified in triplicate for the genes of interest and ␤-actin as a housekeeping marker on a 7000 Sequence Detection System (Applied Biosystems) by using a SYBR-green kit (ABgene, Surrey, UK). The primers used are listed in Table 1. The threshold was set in the linear range of fluorescence, and a threshold cycle (Ct) was measured for each well. The data were analyzed according to Pfaffl (44).
Assays for Cholesterol Efflux-THP-1 cells on 24-well plates were transfected with control or ORP8 siRNAs as described above. At 48 h after transfection, the cells were labeled for 24 h with 0.5 Ci/ml 1,2-[ 3 H]cholesterol (GE Healthcare). They were then washed 3 times with phosphate-buffered saline and incubated for 2 h at 37°C with Eagle's minimal essential medium, 0.2% fatty acid free bovine albumin (Eagle's minimal essential medium/bovine serum albumin). The medium was replaced with fresh Eagle's minimal essential medium/bovine serum albumin with or without pure lipid-free human apoA-I (10 g/ml) or HDL 2 (25 g protein/ml) and incubated at 37°C for 6 h. The media were then collected and analyzed by liquid scintillation counting. Cells were dissolved in 0.2 N NaOH, and the cellular radioactivity and protein were determined. The efflux is expressed as % of the total [ 3 H]cholesterol radioactivity present in the cells ϩ the efflux medium.
Luciferase Reporter Assays-HEK293 cells were seeded on 48-well plates and grown to 20% confluence in Dulbecco's modified Eagle's medium/high glucose medium containing 10% fetal calf serum. siNT or siORP8.1 duplexes (5 pmol) were transfected in serum-containing medium for 48 h using Lipofectamine 2000 (0.8 l/well). Promoter analysis was then performed using a ϳ1-kilobase fragment (from Ϫ928 to ϩ 101 bp) of the human ABCA1 promoter that was linked to the firefly luciferase reporter gene (45). Mutations in the DR4 element, the E-box, or the GnT-box were generated using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) (46). Mutagenesis was performed according to the manufacturer's protocol, and sequences were verified. Reporter plasmids (125 ng/well) were transiently transfected for 24 h into cells using Lipofectamine 2000 (0.8 l/well) together with a reapplication of the siNT or siORP8.1 duplexes (5 pmol). The phRL-TK Renilla internal control plasmid (12.5 ng/well) was also co-transfected for normalization of transfection efficiency. Incubations were carried out in the presence and absence of the LXR agonist T0901317 at 1 M. After transfection (24 h), cells were washed and resuspended in 50 l of 1ϫ passive lysis buffer

Tissue Distribution and Subcellular
Localization of ORP8-To analyze the tissue distribution of ORP8, a rabbit polyclonal antibody was generated against the variable N-terminal end of the protein and affinity-purified. In Western blot analysis of cultured human cell lines the antibody detected a protein at a position corresponding to a molecular mass of ϳ100 kDa (shown for HEK293 cells and THP-1 macrophages in Fig. 1A). The protein appeared as a closely spaced doublet of bands, which most likely corresponds to the two known ORP8 splice variants (variant 1, NM_020841; variant 2, NM_001003712), of 101 and 97 kDa. Western blotting of mouse tissues revealed that ORP8 is most abundant in liver, spleen, kidney, brain, and adipose tissue. In the other tissues analyzed only weak signals were detectable (Fig. 1B). Interestingly, ORP8 was found to be present at high levels also in mouse peritoneal macrophages as well as in the human THP-1 macrophages differentiated using phorbol 12-myristate 13-acetate and in human primary monocyte-macrophages (Fig. 1C). The highest signals were detected in the human macrophages, probably due to the fact that the antibody was raised against the human ORP8 sequence.
The affinity-purified ORP8 antiserum was then used to study localization of the endogenous ORP8 abundant in THP-1 macrophages. A reticular staining coinciding with that of protein disulfide isomerase suggested that the endogenous ORP8 is associated with ER membranes (Fig. 2, A-C). The ORP8 immunofluorescence was completely abolished by preincubation of the antiserum with the immunizing fusion protein, demonstrating specificity of the staining (Fig. 2, D and E). The subcellular localization of ORP8 was also investigated in THP-1 macrophages and HEK293 cells transfected with ORP8/pcDNA4HisMaxC. A reticular and nuclear membrane staining pattern that coincided with the ER marker protein calnexin was observed in both cell types (Fig. 2, G-L), demonstrating ER localization of the Xpress epitope-tagged ORP8. To determine whether the ER localization depends on the predicted C-terminal 19-amino acid transmembrane segment present in ORP8, a truncated construct lacking this segment (ORP8⌬C) was created. The truncated protein displayed an even cytosolic distribution (Fig. 2, M-O) quite different from that of full-length ORP8, demonstrating that the C-terminal segment is responsible for the observed ER targeting. Impor- tantly, Western blotting of the transfected cells confirmed that ORP8⌬C is stable and has the expected mobility on SDS-PAGE (Fig. 2F).

ORP8 Is Present in the Macrophages of Human
Atherosclerotic Lesions-Immunohistochemical analysis using the affinity-purified anti-ORP8 revealed prominent ORP8 staining in human coronary artery atherosclerotic lesions (Fig. 3A). The ORP8-positive cells were mainly found in the shoulder regions of the lesions. Most of these cells were macrophages, as revealed by immunofluorescence double staining for CD68 (Fig. 3B). No significant colocalization was detected between ORP8 and ␣-actin or CD31 (data not shown), suggesting that, in contrast to macrophages, the lesion smooth muscle and endothelial cells did not express detectable levels of ORP8. The protein was also detectable in CD68(ϩ) cells in a variety of other human tissues, such as lymph nodes, thymus, liver, caecum, and in rheumatoid synovia (data not shown). To investigate whether the abundant ORP8 expression in coronary artery lesions can also be detected at the mRNA level, we dissected the intima-media from healthy artery wall (AHA classes I-II) and from advanced lesions (AHA classes V-VI) and quantified the ORP8 mRNA. The ORP8 message was elevated 2.7-fold in the advanced lesions (Fig. 3C).
Colocalization of ORP8 and CD68 staining in the same cells of the coronary artery wall was quantified using low magnification images of double-stained human coronary arteries (Fig.  4A). In apparently healthy parts of the arterial wall colocalization of the two antigens was relatively poor (Pearson's correlation coefficient r ϭ 0.28), whereas in the region close to the lesion lipid core ORP8 was markedly concentrated in the CD68(ϩ) cells (r ϭ 0.51; Fig. 4, B and C).
ORP8 Is an Oxysterol-binding Protein-To test whether ORP8 is capable of oxysterol binding, the ORP8 ligand binding domain (ORP8-(242-828)) was produced as a GST fusion in E. coli and used for in vitro oxysterol binding assays (43).   no specific binding signal with the oxysterols used (data not shown). ORP8 Modulates ABCA1 mRNA and Protein Expression-To explore the possibility that ORP8 might act as an oxysterol sensor involved in macrophage cholesterol efflux, we silenced ORP8 in THP-1 macrophages followed by real-time reverse transcription-PCR quantification of the mRNAs for ABCA1 and ABCG1, central mediators of macrophage cholesterol efflux. The two ORP8 siRNAs reduced the amount of cellular ORP8 protein by 75% (Fig. 6A). Concomitantly, the ORP8 siRNA treatments increased ABCA1 mRNA abundance 2.1-fold as compared with cells transfected with the control siRNA but did not significantly affect the levels of ABCG1 mRNA (Fig. 6B). We then confirmed by Western blotting that the increased ABCA1 mRNA levels resulted in an increased amount of total cellular ABCA1 protein. The analysis revealed a 141% increase of ABCA1 (p Ͻ 0.05) in the cells subjected to ORP8 silencing (Fig. 6C).
Silencing of ORP8 Enhances Cholesterol Efflux to Apolipoprotein A-I-To study the functional significance of the observed ABCA1 induction, we measured the efflux of [ 3 H]cholesterol from THP-1 macrophages treated with control or ORP8 siRNA using either lipid-free apoA-I or spherical HDL 2 as acceptors. Consistent with the increase in cellular ABCA1 protein, cholesterol efflux from siORP8.1-treated cells to apoA-I was increased by 80% (p Ͻ 0.01) as compared with cells treated with control siRNA (Fig. 6D). However, no change was observed in the efflux to HDL 2 , in keeping with the observation that the ABCG1 mRNA level was not significantly affected by ORP8 silencing.
Effect of ORP8 Silencing on a Luciferase Reporter Driven by ABCA1 Promoter-To further confirm the effect of ORP8 on ABCA1 at the transcriptional level, we used ABCA1 promoter-luciferase reporter constructs (Fig. 7A) transfected into HEK293 cells treated with control or ORP8 siRNAs. ORP8 silencing resulted in a significant, 125% induction of the reporter (p Ͻ 0.01), an effect similar to that observed after treatment with the synthetic LXR agonist T0901317 (Fig. 7B). To study whether the impact of ORP8 silencing is mediated by LXR, we carried out similar experiments with a reporter construct in which the LXR response element (DR4) was inactivated. Destruction of the DR4 completely abolished the induction of the reporter by T0901317 but also reduced significantly the induction by ORP8 silencing. Interestingly, combination of ORP8 silencing and T0901317 resulted in markedly enhanced (580% increase) reporter induction, providing further evidence for a functional interaction between ORP8 and the LXR (Fig. 7C). This interpretation is supported by the fact that ORP8 silencing also had a significant effect (90% induction; p Ͻ 0.01) on a reporter carrying a minimal promoter controlled by three tandem repeats of the LXR response element (Fig. 7D). Also with this reporter, ORP8 silencing and T0901317 showed a synergistic effect. To investigate if the ORP8 effect could also involve the E-box or GnT-box elements present in the ABCA1 promoter (47), we mutagenized these elements as well. Inactivation of the E-box and ORP8 silencing acted in a synergistic fashion; with the E-box mutant reporter the luciferase activity in the siORP8.1-treated cells was 4-fold as compared with that in cells treated with the control siRNA (Fig. 7E), whereas the wild-type reporter was induced only 2-fold. Also, the GnTbox mutant reporter was induced by ORP8 silencing. In this case, however, no synergistic effect was detected. To conclude, the reporter experiments strongly suggested that the suppressive effect of ORP8 on the ABCA1 promoter involves both LXR function and that of the E-box in the promoter.

DISCUSSION
The mammalian ORP have been implicated as sterol sensors that regulate a number of cellular functions ranging from sterol and neutral lipid metabolism to vesicle transport and cell signaling (40). OSBP, ORP1, ORP2, and ORP4 as well as Saccharomyces cerevisiae Osh4p are known to bind oxysterols (32, 33, 48 -51). We here identify ORP8 as a mammalian oxysterolbinding protein that localizes to the ER via a C-terminal 19amino acid segment predicted to traverse the membrane bilayer. The protein is relatively abundant in tissues rich in cholesterol, such as brain, or in tissues actively involved in sterol and lipid metabolism, such as liver, kidney, and adipose tissue.
ORP8 is also expressed at high levels in macrophages, including those in human coronary atherosclerotic lesions.
Silencing of ORP8 in THP-1 macrophages induced the expression of ABCA1 but had hardly any effect on ABCG1. The functional consequences of ORP8 silencing were verified with assays for cellular cholesterol efflux from THP-1 macrophages. Consistent with the mRNA and protein changes observed, ORP8 silencing induced [ 3 H]cholesterol efflux to lipid-free apoA-I but not to spherical HDL. These results strongly suggest that the endogenous macrophage ORP8 acts to suppress ABCA1 expression and cholesterol efflux to lipid-poor apoA-I.
The increase of the ORP8 mRNA in advanced atherosclerotic lesions and the presence of ORP8 protein in the lesion macrophage foam cells raises the question of whether ORP8 could be involved in lesion development. Human subjects homo-or heterozygous for Tangier disease mutations have an increased cardiovascular disease (CVD) risk (52), and polymorphisms in the ABCA1 gene are associated with either increased or decreased CVD incidence (24,27). Moreover, studies with mouse models suggest that macrophage ABCA1 executes a cardioprotective function (23,25,26). Therefore, down-regulation of ABCA1 by ORP8 and decreased cholesterol efflux from macrophages to apoA-I could facilitate the progression of atherosclerotic lesions. ORP8 was also found to be present in CD68(ϩ) macrophages in tissues other than atherosclerotic plaques. Therefore, the increase of the ORP8 mRNA in lesions relative to healthy arterial wall could merely reflect the increase of macrophage number in the lesion tissue. However, the findings that ORP8 immunofluorescence staining concentrated in the CD68(ϩ) cells in the vicinity of the lesion lipid core and that treatment of primary human macrophages with certain oxysterols resulted in a modest but significant elevation of ORP8 mRNA and protein 3 indicate that ORP8 could be specifically induced under the conditions present in oxysterol-enriched atherosclerotic lesions. Alternatively, the CD68(ϩ) cell subtype that differentiates during the inflammatory process in these regions expresses ORP8 abundantly.
A commonly accepted paradigm states that oxysterols represent a signal for increased cellular sterol content, resulting in reduced cholesterol biosynthesis and increased cholesterol efflux (53). One may wonder why an oxysterol sensor such as ORP8 should suppress ABCA1 expression. We find it possible that the protein could act as a counterbalancing lever in the oxysterol regulation of ABCA1. ABCA1 and ABCG1, as a rule, respond in concert to regulatory signals from the LXR or other transcription factors (47,54,55). Therefore, the fact that ORP8 silencing had hardly any effect on ABCG1, whereasABCA1 expression was significantly modified, is intriguing. There are several examples of divergent regulation of ABCA1 and ABCG1. Unlike ABCA1, 8-bromo-cAMP fails to induce ABCG1 expression in J774 macrophages (56). In mouse macrophages the two genes respond differentially to retinoic acid receptor-mediated induction (57). Moreover, bacterial lipopolysaccharide was reported to induce ABCA1 but not ABCG1, the effect being independent of the LXR (58).   , siNTϩT). The luciferase activity was determined as specified under "Experimental Procedures." B, effect of ORP8 silencing on the wild-type ABCA1 promoter. The data represent mean Ϯ S.E. (**, p Ͻ 0.01, n ϭ 4 separate experiments, t test). C, comparison of the wild-type (WT) ABCA1 promoter and a mutant with the DR4 element inactivated (DR4 mut). A representative experiment carried out in triplicate is shown. *, p Ͻ 0.05, t test. D, effect of ORP8 silencing on a reporter with a minimal promoter controlled by three tandem repeats of the DR4 element (3ϫDR4). The data represent the mean Ϯ S.E. **, p Ͻ 0.01, n ϭ 3 separate experiments, t test. E, effect of ORP8 silencing on ABCA1 promoter versions where the E-box (E-box mut) or the GnT-box (GnT-box mut) was inactivated. A representative experiment carried out in triplicate is shown.
The present reporter assay data provide strong evidence that the effect of ORP8 silencing is in part mediated by the LXR. This finding represents the first documented functional connection between the OSBP/ORP proteins and the LXR. In the absence of ligand LXR/RXR heterodimers are engaged in transcriptional co-repressor complexes, which are replaced by co-activator complexes upon ligand binding (59 -61). The synergy of ORP8 silencing with a non-sterol LXR agonist suggests that ORP8 does not simply sequester endogenous LXR ligands but, rather, indicates a reduction in co-repressor activity. The fact that the ORP8 silencing effect synergized with inactivation of the E-box in the ABCA1 promoter suggests that the suppressive effect of ORP8 also affects the E-box, a regulatory element that mediates transcriptional repression of ABCA1 by several factors (47,62). The fact that no classical E-box has been identified in the ABCG1 promoter provides one possible explanation for the difference in ABCA1 and G1 responsiveness to ORP8 silencing; if suppression of ABCA1 by ORP8 arises from concerted effects on the DR4 and E-box, lack of one of these in the ABCG1 promoter may result in very weak activity toward this promoter. ORP8 displays no significant intranuclear localization. Therefore, we favor the hypothesis that products arising from a metabolic activity of this protein could stabilize co-repressor proteins associated specifically with the DR4 and E-box elements in the ABCA1 promoter. Alternatively, ORP8 anchored in the ER might capture transcriptional co-activators. These functional schemes will be further explored in future studies.
The present findings identify ORP8 as a new regulator of cellular cholesterol homeostasis and implies that it may play a role in the development of atherosclerotic lesions. Future work with animal models is necessary for a comprehensive understanding of its role in lipid metabolism and atherogenesis.