Sterol-responsive element-binding protein (SREBP) 2 down-regulates ATP-binding cassette transporter A1 in vascular endothelial cells: a novel role of SREBP in regulating cholesterol metabolism.

ATP-binding cassette transporter A1 (ABCA1) is a pivotal regulator of cholesterol efflux from cells to apolipoproteins, whereas sterol-responsive element-binding protein 2 (SREBP2) is the key protein regulating cholesterol synthesis and uptake. We investigated the regulation of ABCA1 by SREBP2 in vascular endothelial cells (ECs). Our results showed that sterol depletion activated SREBP2 and increased its target, low density lipoprotein receptor mRNA, with a concurrent decrease in the ABCA1 mRNA. Transient transfection analysis revealed that sterol depletion decreased the ABCA1 promoter activity by 50%, but low density lipoprotein receptor promoter- and the sterol-responsive element-driven luciferase activities were increased. Overexpression of the N terminus of SREBP2 (SREBP2(N)), an active form of SREBP2, also inhibited the ABCA1 promoter activity. Functionally adenovirus-mediated SREBP2(N) expression increased cholesterol accumulation and decreased apoA-I-mediated cholesterol efflux. The conserved E-box motif was responsible for the SREBP2(N)-mediated inhibition since mutation of the E-box increased the basal activity of the ABCA1 promoter and abolished the inhibitory effect of SREBP2(N). Furthermore sterol depletion and SREBP2(N) overexpression induced the binding of SREBP2(N) to both consensus and ABCA1-specific E-box. Chromatin immunoprecipitation assay demonstrated that serum starvation enhanced the association of SREBP2 and the ABCA1 promoter in ECs. To correlate this mechanism pathophysiologically, we found that oscillatory flow caused the activation of SREBP2 and therefore attenuated ABCA1 promoter activity in ECs. Thus, this SREBP-regulated mechanism may control the efflux of cholesterol, which is a newly defined function of SREBP2 in ECs in addition to its role in cholesterol uptake and biosynthesis.

Epidemiological studies have shown an inverse relationship between levels of high density lipoprotein-cholesterol and risk of coronary artery disease. High density lipoprotein promotes reverse cholesterol transport by facilitating the transfer of cholesterol from peripheral tissues to the liver for disposal (1). ATP-binding cassette transporter A1 (ABCA1), 1 a 254-kDa cytoplasmic membrane protein, is a pivotal regulator of lipid efflux from cells to apolipoproteins (2). ABCA1 apparently plays an important role in reverse cholesterol transfer. Mutations in the ABCA1 gene, discovered in patients with Tangier disease, cause impaired efflux of lipids to apoA-I with a near absence of plasma high density lipoprotein (3)(4)(5). Study of ABCA1 heterozygotes provides direct evidence that the impaired cholesterol efflux is associated with reduced plasma high density lipoprotein-cholesterol and increased risk of coronary artery disease (6). Furthermore, under a high cholesterol diet, ABCA1 transgenic mice showed an atheroprotective lipoprotein profile with decreased atherosclerotic lesions, demonstrating the antiatherogenic effect of the ABCA1 transporter in vivo (7). ABCA1 can be regulated at both the transcriptional and post-transcriptional levels (8). The most studied transcriptional regulation of the ABCA1 gene is the binding of liver X receptor (LXR)/retinoic X receptor (RXR) heterodimers to an imperfect direct repeat spaced by four nucleotides (DR4) to up-regulates the gene (9). Mutation of the DR4 element strongly reduced the LXR agonist-induced ABCA1 gene activation (9,10). However, significant activity is still retained in DR4-deleted constructs, indicating the existence of regulation other than by LXR/RXR (11). Although a silencing regulatory element in ABCA1 promoter has been mapped to a region containing the E-box motif (12), the transcription factors responsive for the negative regulation of the ABCA1 promoter remain elusive.
The vascular endothelium forms a barrier between the vessel wall and lipoproteins in the circulation. It plays an important role in maintaining vascular integrity. The disturbance or injury of endothelium can lead to cardiovascular impairments such as atherosclerosis. ABCA1 is expressed in vascular endothelial cells (ECs) and transcriptionally up-regulated by LDL and cholesterol, suggesting that ABCA1 in ECs plays an important role in cholesterol homeostasis in the vessel wall (19). However, the regulation of ABCA1 and its role in lipid trafficking in ECs remain largely unknown. We previously reported that shear stress activated SREBPs and hence up-regulated their targeting genes in ECs (20,21). Given the important role of SREBP2 and ABCA1 in cholesterol homeostasis and the significance of cholesterol traffic in the vessel wall, we investigated the regulation of ABCA1 by SREBP2 in ECs. Our results showed that, by binding to the E-box, SREBP2 could inhibit ABCA1 transcription. Importantly oscillatory flow caused the activation of SREBP2 and inhibition of ABCA1 promoter activity in ECs. Thus, this SREBP-regulated mechanism controlling cholesterol efflux is a newly defined function of SREBP2 in the vascular wall.

EXPERIMENTAL PROCEDURES
Reagents-Cholesterol was purchased from Avanti, Inc. (Alabaster, AL). [␥-32 P]ATP, [␣-32 P]dCTP, and [ 3 H]cholesterol were from MP Biomedicals (Irvine, CA). The DECApriming II DNA labeling kit was from Ambion (Austin, TX). All of the DNA-modifying enzymes and PCR enzymes were obtained from Promega Corp. (Madison, WI). LXR agonist TO901317 was purchased from Cayman Chemical (Ann Arbor, MI). Human apoA-I was from Biodesign International (Saco, ME). Anti-ABCA1 antibody was purchased from Novus Biologicals (Littleton, CO), and antibodies against SREBP2, upstream stimulatory factor 1 (USF1), USF2, c-Myc, Max, HA tag, and ␣-tubulin were from Santa Cruz Biotechnology (Santa Cruz, CA). LDL was isolated from non-frozen human plasma as described previously (19). Recombinant human fibroblast growth factor was a generous gift from Dr. J. A. Thompson at the University of Alabama. All other reagents, purchased from Invitrogen or Sigma, were all of tissue culture or molecular biology grade.
Cell Culture-Human umbilical vein endothelial cells (HUVECs) were isolated and maintained as described previously (22). Briefly HUVECs were maintained in medium 199 supplemented with 20% fetal bovine serum (FBS), 5 ng/ml fibroblast growth factor, and 90 g/ml heparin (EC medium). In experiments involving serum-free treatment, human endothelial serum-free medium (basal growth medium, Invitrogen) supplemented with 5 ng/ml fibroblast growth factor (SFM), was used. Bovine aortic endothelial cells were cultured under standard culture conditions. All experiments were performed with HUVECs up to passage three, and all cells were cultured to confluence before treatment.
Northern Blotting, Quantitative Real Time (RT)-PCR, and Western Blotting-Total RNA isolation and Northern blotting for hABCA1, LDL receptor (LDLR), and von Willebrand factor mRNA was carried out according to standard protocols. The cDNA probes for LDLR and hABCA1 were generated via RT-PCR, and the probes were labeled with [␣-32 P]dCTP by DECApriming (Ambion). For quantitative RT-PCR, total RNA was converted into cDNA by using reverse transcriptase with oligo(dT) as the primer. The obtained cDNAs were then used as the templates for quantitative RT-PCR with the use of Brilliant SYBR Green QPCR Master Mix (Stratagene, La Jolla, CA). The relative amount of ABCA1 mRNAs was calculated using the comparative method with the ␤-actin mRNA as internal control. The nucleotide sequences of the primers were: hABCA1, 5Ј-GCTGCCTCCTCCACAA-AGAAAAC-3Ј and 5Ј-GCTTTGCTGACCCGCTCCTGGATC-3Ј; ␤-actin, 5Ј-TGACCGGGTCACCCACACTGTGCCCATCTA-3Ј and 5Ј-CTAGAA-GCATTTGCGGTGGACGATGGAGGG-3Ј.
HUVECs were solubilized in a buffer containing detergent, and cellular proteins were separated by SDS-PAGE. The Western blotting analysis was performed with antibodies against ABCA1, SREBP2, SREBP1, USF1, USF2, c-Myc, Max, HA, or ␣-tubulin.
Plasmid Construction and Transient Transfection-Plasmid pABCA1(E-boxmut)-luc was generated using the QuikChange site-directed mutagenesis kit (Stratagene) and pABCA1(Ϫ928)-luc construct (10) as template. The mutagenic primer for E-box mutation was: 5Ј-G-GGCCCCGGCTCCACGgaCTTTCTGCTGAG-3Ј. The sequences underlined indicate the E-box motif, and the bases in small letters are the designated mutations. Mutagenesis was performed according to the manufacturer's protocol. pCMV5-HA-SREBP2(N), the expression plasmid of the N terminus of human SREBP2, was generated as described previously (21). For the promoter activation study, the reporter plasmids of the hABCA1 promoter, 4XSRE-luc, LDLR-luc, and SCAP-C (20), were transiently transfected into HUVECs with the use of the Targefect F2 (Targeting Systems, San Diego, CA). pRSV-␤-gal was co-transfected as a transfection control.
Adenovirus Construction and Infection-Recombinant adenovirus encoding Ad-HA-SREBP2(N) was created, amplified, and titered as reported previously (23). For adenovirus infection, the virus mixture with multiplicity of infection as indicated was added to confluent HUVECs in culture and incubated for 12 h. Ad-␤-gal was used as an infection control. The infected cells were then incubated in fresh growth medium for 24 h followed by RNA or protein extraction.
Assessment of Cholesterol Efflux-The cholesterol efflux was assessed as described previously (19) with modification. Briefly HUVECs in 12-well plates were infected with Ad-HA-SREBP2(N) or control virus for 12 h. The cells were then treated with LXR agonist TO901317 (10 M) in EC medium or SFM for 18 h and labeled with [ 3 H]cholesterol (0.2 Ci/ml) for 6 h. After washing with phosphate-buffered saline-bovine serum albumin, the cells were incubated with fresh SFM with or without apoA-I (10 g/ml) for 2 h. The aliquots of medium and cell lysates were assayed by liquid scintillation.
Electrophoretic Mobility Shift Assay-Confluent HUVECs were treated with EC medium or SFM for 6 h or infected with Ad-HA-SREBP2(N) for 24 h. The cells were then lysed, and nuclear extracts were prepared. Double strand oligonucleotides containing the divergent E-box sequence in the ABCA1 promoter (5Ј-GGGCCCCGGCTCCACG-AGCTTTCTGCTGAG-3Ј) or consensus E-box oligos (Santa Cruz Biotechnology) were end-labeled with [␥-32 P]ATP. Electrophoretic mobility shift assays were performed as described previously (24). To test the specificity of binding, a 100-fold molar excess of unlabeled ABCA1-Ebox or irrelevant AP-1 and SP-1 probes were used for competition experiments. In supershift experiments, the nuclear extracts were incubated with anti-HA antibody for 3 h on ice before the addition of the labeled probes.
Chromatin Immunoprecipitation (ChIP) Assay-The ChIP assays were performed as described previously (23). In brief, cells were fixed with 1% formaldehyde and quenched prior to harvest and sonication. Goat anti-SREBP2 or normal goat IgG and single strand salmon sperm DNA saturated with Protein A-Sepharose 4B beads were added to sheared samples for immunoprecipitation. Immunoprecipitates were pelleted by centrifugation, and the supernatant of the control group was collected as an input control. The immunoprecipitates were eluted from Sepharose 4B beads, and proteinase K solution was added and incubated at 60°C for 8 h. DNA was extracted, purified, and then used to amplify target sequences by PCR. The ABCA1 promoter containing the E-box element was amplified by use of primer set 5Ј-CTGCACCGAGC-GCAGAGGTTA-3Ј and 5Ј-CAACTCCCTAGATGTGTCGTG-3Ј.
Flow Experiments-The flow experiments were performed as described previously (25). In brief, ECs were seeded on glass until confluence. Cells were then subjected to a parallel plate flow channel to impose an oscillatory flow (mean shear stress, 0 dyne/cm 2 ; amplitude of pulsatility, Ϯ3 dyne/cm 2 ; 1 Hz) or a steady laminar flow (12 dyne/cm 2 ) for 12 h. The flow system was kept at 37°C and ventilated with 95% humidified air with 5% CO 2 .
Statistics-Quantitative data were expressed as mean Ϯ S.D. Statistical significance of the data was evaluated by analysis of variance or Student's t test. p values less than 0.05 were considered significant. For nonquantitative data, results were representative from at least three independent experiments.

RESULTS
The Reciprocal Response of ABCA1 and SREBP2 Transcripts to Serum Deprivation-To study the regulation of ABCA1 in ECs under serum-free condition, we cultured confluent HUVECs in SFM in the presence or absence of LDL for 24 h. Under this condition, HUVECs had normal morphology, and no cell death was detected. As shown in Fig. 1A, SFM greatly decreased ABCA1 mRNA in HUVECs, and the inclusion of LDL in SFM reversed such an inhibition. In control experiments, cells in EC medium containing 20% FBS were exposed to LDL or cholesterol. LDL and cholesterol increased the level of ABCA1 mRNA. The level of mRNA encoding LDLR, an SREBP2 target gene, was much lower in cells in FBS than those in SFM. Supplementing SFM with LDL inhibited the expression of LDLR mRNA. Consistent with the SFM-up-regulated LDLR, SREBP2 was activated as revealed by the in-creased SREBP2 cleavage at 6 h after exposure to SFM and the subsequent increase in both precursor and cleavage forms of SREBP2 (Fig. 1B). Furthermore results from transient transfection assays showed that SFM decreased the luciferase reporter driven by the ABCA1 promoter by 49 Ϯ 6% compared with cells under 20% FBS (Fig. 1C). However, SFM increased the luciferase reporter driven by the LDLR promoter or SRE by 5-6 times. Thus, ABCA1 and SREBP2 transcripts in ECs responded reciprocally to SFM.
SREBP2 Inhibits the Expression of ABCA1-To ascertain whether SREBP2 up-regulation inhibits the expression of ABCA1, we compared the level of ABCA1 protein in HUVECs infected with recombinant Ad-HA-SREBP2(N) encoding the N terminus of SREBP2 with those infected with Ad-␤-gal. Western blotting analysis revealed that ABCA1 in HUVECs was decreased by SREBP2(N) overexpression. This inhibitory effect was similar to that by SFM ( Fig. 2A). Because ABCA1 is a pivotal regulator of cholesterol efflux from cells to apolipoproteins such as apoA-I, we explored cholesterol efflux to apoA-I as a functional consequence of the SREBP2-suppressed ABCA1. Compared with the Ad-␤-gal-infected controls, the infection of Ad-HA-SREBP2(N) increased the [ 3 H]cholesterol uptake and/or accumulation by 20 Ϯ 8% (Fig. 2B). Furthermore cholesterol efflux in the presence of apoA-I was significantly decreased in Ad-HA-SREBP2(N)-infected cells (Fig. 2C). As an LXR target gene, ABCA1 is up-regulated by TO901317, a selective LXR agonist (26,27). Treating Ad-␤-gal-infected HU-VECs with TO901317 promoted the apoA-I-mediated cholesterol efflux when the cholesterol accumulation was unchanged.
SREBP2 Down-regulates the ABCA1 Transcription via E-box Motif-To study whether the inhibition of ABCA1 by serum starvation or SREBP2 is at the level of transcription and to determine the involved transcriptional element within the ABCA1 promoter, we performed transient transfection assays using ABCA1 promoter-driven reporter constructs. Fig. 3A shows that overexpression of SREBP2(N) by the co-transfected pCMV5-HA-SREBP2(N) decreased the activities of promoter constructs pABCA1(Ϫ928)-luc and pABCA1(Ϫ156)-luc by 50% compared with controls transfected with vector plasmid pCMV5. Further deletion of a segment between Ϫ156 and Ϫ116 containing the E-box site, a silencer of the ABCA1 promoter (28,29), abolished the inhibitory effect of SREBP2(N). Indeed mutation of the E-box in pABCA1(Ϫ928 E-boxmut)-luc not only increased the basal activity but also reversed the inhibition of ABCA1 promoter by SREBP2(N). Furthermore we studied the effect of serum starvation on the ABCA1 promoter by incubating the various transfected HUVECs with SFM or 20% FBS. As shown in Fig. 3B, SFM and SREBP2(N) overexpression exhibited a similar pattern of regulation on ABCA1 promoter constructs.

SREBP2(N) Binds to the E-box of the ABCA1
Promoter-Given the possibility that E-box is an SREBP-responsive element, which down-regulates the ABCA1 expression, we studied the binding of SRBEP2(N) to E-box by electrophoretic mobility shift assay. As shown in Fig. 4A, SREBP2(N) overexpression in ECs because of Ad-HA-SREBP2(N) infection increased the binding of nuclear extracts to the ABCA1-specific E-box compared with control cells infected with Ad-␤-gal. The supershifted band that resulted from the inclusion of anti-HA antibody revealed the specific binding of SREBP2 to the ABCA1-E-box motif. Furthermore SFM treatment for 6 h increased the bindings of nuclear extracts to both consensus and ABCA1specific E-box (Fig. 4B). Addition of LDL or cholesterol to SFM attenuated such a binding. We also performed ChIP assay to ascertain whether SREBP2(N) binds to the ABCA1 promoter in vivo. As shown in Fig. 4C, although SREBP2(N) associated

FIG. 1. Effect of SFM on ABCA1 in human vascular ECs. A,
HUVECs were incubated with 180 mg/dl LDL, 10 g/ml cholesterol (CHL) in 20% FBS medium or SFM for 24 h. RNA was isolated, and samples of 15 g of total RNA were resolved by gel electrophoresis and then hybridized with ␣-32 P-labeled ABCA1, LDLR, or von Willebrand factor (vWF) cDNA. B, HUVECs were incubated with LDL in 20% FBS medium or in SFM for different periods of time. Cells were lysed, and the proteins were then resolved by 6% SDS-PAGE, transferred to a nitrocellulose membrane, and visualized with anti-SREBP2 and anti-␣-tubulin antibodies. C, HUVECs were transfected with hABCA1luc(Ϫ928), LDLR-luc, or 4XSRE-luc for 24 h. Then cells were incubated with medium with or without serum (SFM) for 24 h. Promoter activities were measured by use of the reporter luciferase and normalized with ␤-galactosidase. Data are mean Ϯ S.D. of the relative luciferase activities in three independent experiments, each performed in triplicate. Results are representative from three independent experiments. with the ABCA1 promoter in cells in 20% FBS, SFM greatly enhanced the association. Thus, the inhibition of ABCA1 transcription by SREBP2(N) is most likely through the binding of SREBP2 to the E-box site in the ABCA1 promoter.
Serum Starvation Does Not Increase Other E-box-related Transcription Factors-In addition to SREBPs, E-box element also binds to several other nuclear proteins, including c-Myc, Max, and c-Myc-related regulatory factors such as USF1 and USF2. To explore the possible regulation of these E-box-binding proteins, the induction of Myc, Max, USF1, and USF2 by SFM was detected in the nuclear extract. Western blotting showed that SFM increased SREBP1 and SREBP2. However, levels of nuclear c-Myc and USF2 were decreased, and those of Max and USF1 remained unchanged (Fig. 5). This result suggested that Myc, Max, or USFs were unlikely to be involved in the binding of the E-box of ABCA1 promoter, leading to the suppression of the ABCA1 in response to SFM.
Flow Inhibits ABCA1 Promoter Activation in ECs-We previously reported that steady laminar flow caused a transient activation of the SREBP-regulated genes, whereas disturbed flow patterns caused a sustained activation of SREBPs and their targeting genes in ECs (20). To determine whether the oscillatory flow-activated SREBP2, like serum depletion, can down-regulate ABCA1, HUVECs were subjected to an oscillatory flow (0 Ϯ 3 dyne/cm 2 ), and the levels of ABCA1 mRNA were determined by quantitative RT-PCR. As shown in Fig. 6A, oscillatory flow indeed decreased the level of ABCA1 mRNA compared with static controls or cells subjected to laminar flow. Furthermore ECs were transiently transfected with pABCA1(Ϫ156)-luc, and the transfected cells were then subjected to different patterns of flow for luciferase induction assays. As shown in Fig. 6B, oscillatory flow decreased the ABCA1 promoter activity compared with static controls or laminar flow. The oscillatory flow-decreased ABCA1 promoter activity was reversed if E-box site was mutated or by co-transfection with SCAP-C, which encodes a truncated C terminus of SCAP and has been shown to block SREBP translocation from the endoplasmic reticulum to the Golgi (Fig. 6C) (30). These data suggest that oscillatory flow down-regulates ABCA1 in ECs, which is mediated through the up-regulation of SREBP2. DISCUSSION In the present study we reported that the binding of SREBP2(N) to E-box motif within the hABCA1 promoter is responsible for the repressive effect of serum deprivation and oscillatory flow on the expression of ABCA1 in ECs. Evidence supporting such a thesis includes the following. 1) The adenovirus-mediated overexpression of SREBP2(N) could bind to the E-box motif of the ABCA1 promoter, hence inhibiting its activity, and the mutation of the E-box abolished the inhibitory effect of SREBP2(N). 2) Electrophoretic mobility shift and ChIP assays showed the association of SREBP2 and the ABCA1 promoter in ECs. 3) Functionally the SREBP2-down-regulated ABCA1 increased cholesterol accumulation and decreased the apoA-I-mediated cholesterol efflux. 4) Pathophysiologically disturbed flow patterns caused sustained activation of SREBP2 (21) with consequent decrease in ABCA1 expression in ECs. Thus, this SREBP-regulated mechanism controls the efflux of cholesterol, which is a newly defined function of SREBP2 in the vascular wall.
Apparently ABCA1 plays an important role in reverse cholesterol transfer and atherogenic protection (2). Most in vitro experiments involving ABCA1 function and regulation have been performed with the use of cultured macrophages, fibroblasts, and hepatocytes. Because it is exposed to the lipoproteins in the circulation, the vascular endothelium plays an important role in maintaining vascular integrity. We and others reported that ABCA1 is expressed in human aortic ECs, HUVECs, porcine brain capillary ECs, and rat liver ECs (19,31,32). Although the dysfunction of lipid metabolism and the ensuing oxidation and deposition in vascular cells impose important pathophysiological consequences, there have been only a few documented reports on ABCA1 regulation in ECs. The most studied transcriptional regulation of ABCA1 is the binding of the LXR/RXR heterodimer to the DR4 site of the ABCA1 promoter to up-regulate the expression of the ABCA1 gene (9). Similar regulation has also been observed in ECs (19). The E-box of the ABCA1 promoter has been reported to be a silencing regulatory element (12,29) that binds to various transcrip-  5-7) for 4 h. After cross-linking and sonication, the nuclear proteins were extracted. Using anti-SREBP2 antibody and normal IgG as a control for immunoprecipitation and ABCA1 promoter-specific primers for PCR, ChIP assays were performed to detect the binding of SREBP2 to the ABCA1 promoter. The data presented are representative from three independent experiments. CHL, cholesterol.

FIG. 5. Effect of SFM on E-box-related binding proteins.
HUVECs were incubated with LDL in EC medium or in SFM for 12 h. Cells were lysed, and the proteins were then resolved by SDS-PAGE, transferred to a nitrocellulose membrane, and visualized with antibodies against SREBP1, SREBP2, c-Myc, Max, USF1, or USF2. Results are representative from three independent experiments. tion factors such as USF1, USF2, and Fra2 for the repression of the human ABCA1 promoter in murine RAW cells. However, overexpression of USF1 and USF2 enhanced activity of the wild-type proximal hABCA1 promoter. In addition, the bindings of USFs were not modulated by any known activators of ABCA1. In this report, we found that serum starvation activated SREBP2(N) bound to both consensus and the ABCA1specific E-box. The mutation of the E-box abolished the inhibitory effect of SREBP2 in the transient transfection assay. Under the same condition, the levels of USF-1 showed little change, and USF-2 was down-regulated. Thus, SREBP-2, but not USFs, was most likely to be involved in the negative regulation of ABCA1 promoter activity in ECs. Since USFs can activate or inhibit transcription by competing with other transcriptional factors, repressors, or activators for the E-box, the possibility that SREBPs replace USFs bound to the E-box under serum-free condition could not be excluded.
In sterol-depleted cells, SREBP(N) is released from membranes and then translocates into the nucleus where it binds to SREs in the promoters of multiple genes encoding enzymes for cholesterol biosynthesis, unsaturated fatty acid biosynthesis, triglyceride biosynthesis, and lipid uptake (for reviews, see Refs. [15][16][17]. SREBPs can bind both the SRE sequence and E-box motif in the promoters of several genes (15). SREBP1 primarily activates the fatty acid triglyceride and phospholipid pathways, whereas SREBP2 is responsible for cholesterol synthesis and uptake. Since the major function of ABCA1 is to mediate intracellular free cholesterol efflux to apolipoproteins, we therefore focused on the regulation of ABCA1 by SREBP2. It was reported that LDL incubation decreases the expression of SREBP2 and its target genes in cultured ECs (33). We found that sterol depletion activated SREBP2 with the concurrent increase in LDLR and decrease in ABCA1 at the transcription level. This regulation was mimicked by the SREBP2(N) overexpression, which inhibited ABCA1 at the protein level with ensuing increase in cholesterol accumulation and decrease in apoA-I-mediated cholesterol efflux.
In peripheral cells, intracellular cholesterol homeostasis is exquisitely regulated and depends on the balance between cholesterol synthesis, degradation, cholesterol ester formation, influx, and efflux (34,35). In principle, the sterol deprivationactivated SREBPs serve as transcriptional factors for lipid/ cholesterol synthesis, uptake, and storage. But whether SREBPs also regulate genes governing cholesterol efflux is not clear. It was postulated that caveolin functions as a regulator of cellular free cholesterol homeostasis in quiescent peripheral cells in which caveolin mediated free cholesterol efflux. Previous study by others showed that SREBPs inhibited the expression of the caveolin gene, which is contradictory to its stimulating effect on other promoters (36). Our result indicates that SREBP2 binds directly to the E-box of the ABCA1 promoter. Whether SREBP2 forms a complex with other transcriptional factors or the binding of SREBP2 recruits other repressors deserves further studies. Recently we found that endoplasmic reticulum stress-induced activating transcription factor 6 suppressed the SREBP-mediated transcription in glucose-deprived HepG2 cells. The attenuated transcriptional activity of SREBP was due to the recruitment of a repressor, histone deacetylase 1, to the ATF6-SREBP complex (23,24). Nevertheless ABCA1 down-regulation by SREBP2 is another example of the coordinated regulation of cholesterol synthesis, influx, and efflux.
In vitro study suggested that SREBPs in peripheral cells were suppressed by high levels of serum or cholesterol (17). The physiological implications of the SREBP2 activation and ABCA1 down-regulation in ECs by SFM would be plausible since endothelium is exposed to full blood in vivo. We observed that oscillatory flow activated SREBPs and up-regulated their targeting genes in ECs in the presence of high levels of LDL and 25-hydroxycholesterol, an agent that is commonly used to suppress SREBP activation (20,21). Consistently we demonstrated in this study that oscillatory flow inhibited the expression of ABCA1 and the ABCA1 promoter activities in ECs. The block of SREBP activation by SCAP-C was able to reverse such an inhibitory effect, indicating the involvement of SREBPs. Thus, the pathophysiological implication for this result is that disturbed flow patterns, causing sustained SREBP2 activation and ABCA1 suppression, may impair the lipid/cholesterol homeostasis in the curves or branches of the arterial tree, promoting atherogenesis.