Transcriptional Regulation of Human and Rat Hepatic Lipid Metabolism by the Grapefruit Flavonoid Naringenin: Role of PPARα, PPARγ and LXRα

Disruption of lipid and carbohydrate homeostasis is an important factor in the development of prevalent metabolic diseases such as diabetes, obesity, and atherosclerosis. Therefore, small molecules that could reduce insulin dependence and regulate dyslipidemia could have a dramatic effect on public health. The grapefruit flavonoid naringenin has been shown to normalize lipids in diabetes and hypercholesterolemia, as well as inhibit the production of HCV. Here, we demonstrate that naringenin regulates the activity of nuclear receptors PPARα, PPARγ, and LXRα. We show it activates the ligand-binding domain of both PPARα and PPARγ, while inhibiting LXRα in GAL4-fusion reporters. Using TR-FRET, we show that naringenin is a partial agonist of LXRα, inhibiting its association with Trap220 co-activator in the presence of TO901317. In addition, naringenin induces the expression of PPARα co-activator, PGC1α. The flavonoid activates PPAR response element (PPRE) while suppressing LXRα response element (LXRE) in human hepatocytes, translating into the induction of PPAR-regulated fatty acid oxidation genes such as CYP4A11, ACOX, UCP1 and ApoAI, and inhibition of LXRα-regulated lipogenesis genes, such as FAS, ABCA1, ABCG1, and HMGR. This effect results in the induction of a fasted-like state in primary rat hepatocytes in which fatty acid oxidation increases, while cholesterol and bile acid production decreases. Our findings explain the myriad effects of naringenin and support its continued clinical development. Of note, this is the first description of a non-toxic, naturally occurring LXRα inhibitor.


Introduction
The liver is the hub of lipid and carbohydrate homeostasis [1]. Dysregulation of this homeostasis has been implicated in disease processes, such as atherogenesis, insulin resistance, and hypermetabolism [2,3]. Metabolic conditions, such as insulin resistance, may be partly attributable to 'western-style diets' and are associated with medical expenditures and lost productivity totaling over $130 billion annually [4]. Therefore, drugs or dietary supplements that could potentially reduce insulin dependence and regulate dyslipidemia could have a dramatic effect on healthcare expenditures and public health.
One group of compounds previously shown to have hypolipidemic and anti-inflammatory properties both in vivo and in vitro are citrus flavonoids [5,6]. The abundant flavonoid aglycone naringenin, which is responsible for the bitter taste in grapefruits, has been extensively studied in recent years. In vivo studies have demonstrated its potential as a normolipidemic agent: in a recent clinical trial, naringenin was shown to reduce circulating levels of low-density lipoprotein (LDL) by 17% in hypercholesterolemic patients [7]. Similarly, the cholesterol-lowering effects of naringenin have been demonstrated in rabbits [8,9] and rats [10]. In HepG2 cells, naringenin was shown to reduce the secretion of VLDL [11,12] through the inhibition of ACAT2 [11] and MTP [13,14], enzymes critical for VLDL assembly. Naringenin was also shown to induce LDL-R transcription through PI3K activation upstream of SREBP-1a [11,14]. Other studies demonstrated that naringenin inhibited HMG CoA reductase (HMGR), while activating enzymes important in fatty acid oxidation such as CYP4A1 [15]. Naringenin's myriad effects suggest that the flavonoid may be targeting transcriptional regulation of metabolism through nuclear receptors (NRs), a family of ligand-activated transcription factors, which play a critical role in the regulation of lipid metabolism. Strengthening this hypothesis is the anecdotal report that naringenin binds to LXRa [14] and more recently, that the flavonoid induces PPRE activity in U-2OS cells [16].
In this study, we demonstrate that naringenin is an agonist of PPARa and PPARc, and a partial agonist of LXRa. We show that naringenin induces the activation of PPARa and PPARc ligandbinding domain (LBD) in GAL4-fusion protein reporters and induces PPRE activity in Huh7.5 human hepatoma cells. Using an in vitro TR-FRET assay we demonstrate that this interaction does not change the binding of PGC1a co-activator peptide to recombinant PPARa ligand binding domain.
Concomitantly, naringenin inhibits the activation of the LXRa LBD in a GAL4-fusion protein reporter in the presence of the LXRa agonist TO901317. Using an in vitro TR-FRET assay, we demonstrate that this effect is mediated by the inhibition of the binding of the Trap220/Drip-2 co-activator peptide to recombinant LXRa LBD. Expectedly, naringenin also inhibits LXRE activity in Huh7.5 cells. We show that the induction of PPARa and inhibition of LXRa induces the expected transcriptional changes in hepaotcytes, upregulating genes important in fatty acid oxidation and down-regulating cholesterol and fatty acid synthesis. These effects result in the induction of a fasted-like state in primary hepatocytes, in which production of triglycerides and bile acids is inhibited and ketone body generation increases.
To further characterize the interaction between PPARa and naringenin, a LanthaScreen time-resolved fluorescence resonance energy transfer (TR-FRET) assay was performed. This cell-free system measures the ability of a compound to enhance the binding of a recombinant PPARa LBD to a PGC1a co-activator peptide, as measured by an increase in TR-FRET signal. While GW7647 showed a clear dose-dependent increase (EC 50 = 2.5nM) in the binding of PGC1a to PPARa as expected (Fig. 1d), the binding of PGC1a to PPARa did not increase in the presence of naringenin (Fig. 1c), suggesting that naringenin's ability to activate PPARa does not directly involve enhancement of PPARa LBD binding to PGC1a.
One possibility is that naringenin induces the transcription of PGC1a itself, an effect that cannot be seen in the cell-free TR-FRET assay. Indeed, stimulation of Huh7 cells with 380 mM naringenin for 24 hours increased PGC1a mRNA abundance by 14-fold (p = 0.001) compared to DMSO-treated controls.
To test the ability of naringenin to inhibit LXRa activity in hepatocytes, we quantified the activation of LXR response element (LXRE)-reporter in Huh7 cells. Naringenin treatment significantly and dose-dependently decreased LXRE acitivty, reaching a 50.3%62.6% inhibition at 150 mM (p,0.001; Fig 3c). By comparison, a recently published LXRa-specific antagonist, 5CPPSS-50 failed to inhibit LXRE activity under the same conditions (Supp. Fig. 1) and led to significant toxicity at higher doses.
Interestingly, Huff and coworkers previously demonstrated that naringenin activated SREBP1a-dependent LDLR expresion [11,14]. As SREBP is regulated by LXRa we studied the gene expression of SREBP1/2 regulated LDLR and HMGCS promoters in Huh7 cells using reporter constructs. We show that naringenin increases LDLR transcription by 26%611%, but decreases HMGCS transcription by 13%63% (Fig 4d). HMGCS is regulated by SREBP2 rather than SREBP1 and like HMGR plays a role in cholesterol synthesis.
ApoB100 is the structural protein of VLDL whose production is blocked by naringenin [21]. As our results suggest that naringenin acts through PPARa induction, we examined whether PPARa and PPARc agonists, affected ApoB100 secretion. Huh7 cells were stimulated with 200 mM naringenin, 10 mM WY14,643, or 10 mM ciglitazone for 24 hours. Predictably, naringenin led to a 73%69% (p,0.001) reduction in ApoB production ( Fig. 4a)   Naringenin treatment led to a 73%69% (p,0.001) reduction in ApoB production, while WY14,643 led to a 33%612% (p,0.01) reduction. Treatment with cigilitazone did not lead to a significant change in VLDL production. (b) Primary rat hepatocytes were stimulated with 200 mM naringenin or 10 mM WY14,643. Naringenin treatment led to a 61% (p,0.001) reduction in triglyceride production and 17% increase in ketone body formation, not different from WY14,643. However, naringenin treatment led to a 32%611% (p = 0.005) reduction in bile salt production, while WY14,643 did not. Urea accumulation in the media did not change significantly. (c) Intracellular levels of triglycerides in primary rat hepatocytes stimulated with naringenin. A slight decrease is observed. (d) Naringenin effect on SRE-driven gene expression. We show that naringein induces LDLR transcription by 26% (p = 0.02) while inhibiting HMGCS transcription by 13% (p = 0.001). It is thought that each promoter is regulated by a different SREBP isoform. doi:10.1371/journal.pone.0012399.g004 compared with a 33%612% (p,0.01) reduction by the PPARa agonist WY14,643. Cigilitazone did not lead to a significant change in ApoB secretion.
Lastly, we characterized the metabolic changes induced by naringenin on primary hepatocytes. Primary rat hepatocytes were stimulated with 200 mM naringenin or 10 mM WY14,643 for 24 hours and culture media was analyzed for changes in urea, triglycerides, bile acid, and ketone bodies (Fig. 4b). As could be expected, primary hepatocytes showed no change in urea production. However, both naringenin and WY14,643 led to a 61% (p,0.001) and 41% (p,0.05) reduction in triglyceride production, respectively (Fig. 4b). Ketone body production was only slightly elevated by 17% and 23%, respectively. Importantly, no increase in intracellular levels of triglycerides were found (Fig. 4c) suggesting this inhibition was a result of increased fatty acid oxidation in primary hepatocytes. Interestingly, while PPARa agonist WY14,643 did not have an effect on hepatic bile acid production, naringenin cased a significant 32%611% (p = 0.005) reduction in bile salt production (Fig. 4b), possibly due to inhibition of cholesterol synthesis, through suppression of LXRa [24,25].

Discussion
Dysregulation of lipid homeostasis is associated with multiple disease states, including metabolic, inflammatory, and infectious disorders [26]. Metabolic regulation is achieved in mammals through an intricate transcriptional mechanism responding to physiological cues. In recent years, a family of ligand-activated transcription factors called nuclear receptors emerged as key regulators of cellular metabolism [25,27,28]. Previously defined as orphan receptors, key metabolites were shown to be the natural ligands of many nuclear receptors, including liver X receptors (LXRs) which respond to oxysterols and glucose [29,30], farnesoid X receptor (FXR) which responds to bile acids [31], and the peroxisome proliferator-activated receptors (PPARs) which respond to fatty acids [32].
The PPAR family includes PPARa, PPARc, and PPARd. The prevalence of these receptor subtypes varies in different tissues, with PPARa being the most prevalent subtype in the liver, and PPARc the most abundant in adipose tissue [18]. PPARa is activated by fatty acids released in a physiological fasting state, leading to increased b-oxidation and gluconeogenesis [33,34]. In clinical practice, PPARa agonists (fibrates) are used to treat hyperlipidemia, whereas PPARc agonists (TZDs) are used to increase insulin sensitivity in muscle and adipose tissue [19,35]. The LXR family includes both LXRa and LXRb [29,30,36]. The latter is ubiquitously expressed, while the former is found primarily in the liver, adipose tissue, and macrophages and is activated by glucose and sterols [37], typical of a physiological fed state. In the liver, following activation by its ligands, LXRa activates lipogenic and glycolytic genes partly through activation of SREBP [38,39]. HMGR, the target of statins, is regulated through this pathway controlling cholesterol availability for bile acid synthesis in hepatocytes.
Following a ligand binding event, both PPARs and LXRs become activated and heterodimerize with the retinoid X receptor (RXR) [27]. This heterodimer then binds conserved response elements such as PPRE or LXRE, while recruiting other coregulatory molecules, such as the co-activators PGC1a [40] and Trap220 [41] for PPARa and LXRa, respectively. The requirement of a RXR binding partner leads to competitive inhibition at the level of receptor activation, offering a transcriptional layer of control over fasted-to-fed transition [42,43,44]. The existence of both a PPAR response element (PPRE) and an LXR response element (LXRE) in the regulatory region of LXRa [45,46] suggests further levels of cross-regulation. Lastly, other coactivators, corepressors and kinases, such as PI3K and ERK, can regulate nuclear receptor activity by non-transcriptional mechanisms [47,48,49].
Naringenin is an aglycone of the grapefruit flavonoid naringin, which is responsible for the bitter taste in grapefruit. Naringenin has been reported to be an antioxidant with hypolipidemic, anticarcinogenic and anti-inflammatory properties both in vivo and in vitro [5,7,8,9,10].The flavonoid was shown to reduce VLDL secretion [50,51] through inhibition of ACAT2 and MTP [50,52], critical enzymes for VLDL assembly. Allister et al. demonstrated that this inhibition is regulated through the MAPK/ERK pathway [52]. In addition, naringenin was shown to upregulate SREBPdependent LDLR through PI3K activation [23]. Naringenin has also been shown to inhibit SREBP-dependent HMGR [53], while activating enzymes important in fatty acid oxidation such as CYP4A1 [54]. These myriad effects suggest that the flavonoid's target might be at the nuclear receptor level. Strengthening this hypothesis is the anecdotal report that naringenin binds to LXRa [23] and, more recently, that it induces PPRE activity in U-2OS cells [55].
In this work we demonstrate that naringenin activates the LBD of both PPARa and PPARc using a reporter cell line over expressing GAL4 fusion proteins to either PPARa LBD or PPARc LBD [20]. Activation of PPAR LBD releases the complex and allows it to bind the UAS G response element, expressing luciferase. This reporter system demonstrates that naringenin acts on the LBD of both PPARa and PPARc (Fig. 1), suggesting it serves as a natural ligand. However, the TR-FRET assay suggests that naringenin does not induce a conformational change in PPARa LBD like other ligands, such as GW7647, failing to increase its binding to the PGC1a co-activator (Fig. 1). One possibility is that naringenin induces a different conformational change in the PPARa LBD that recruits another co-activator, not found in the cell-free TR-FRET assay. However, a more likely scenario is that naringenin induces PPARa phosphorylation or alternately, PGC1a expression. Indeed our data shows that naringenin stimulation increases the mRNA abundance of PGC1a in Huh7 cells by 14-fold. Regardless of the exact nature of the interaction, naringenin-induced PPARa activation, lead to increased PPRE activity in human hepatocytes (Fig. 3a) and the expression of PPARa-regulated genes (Fig. 3b).
Concomitantly with PPARa activation, we show that naringenin inhibits the activity of LXRa. Using a similar reporter cell line over expressing the GAL4 fusion protein with LXRa LBD, we show a significant inhibition of LXRa LBD in the presence of TO901317, a classical agonist (Fig. 2). In contrast to the PPARa findings, we show that naringenin specifically increases the interaction of the Trap-220 co-activator with LXRa LBD in the cell-free TR-FRET assay. Interestingly, in the presence of LXRa agonist TO901317, naringenin actually decreased the interaction of Trap-220 with the LXRa LBD, demonstrating it is a partial agonist of LXRa naturally leading to a competitive inhibition of LXRa activity. This conclusion is further supported by the decrease in LXRE activity in human hepatocytes (Fig 3c) and the down-regulation of LXRa target genes (Fig. 3d).
The metabolic effect of PPARa induction and LXRa inhibition by naringenin are shown on gene expression (Fig. 3) and functional levels (Fig. 4). The mRNA abundance of PPARatarget genes that control fatty acid oxidation, such as CYP4A11, ACOX, and UCP1 significantly increases in human hepatoma cells. As lipid metabolism of hepatoma cell lines is dramatically lower than that of primary hepatocytes, we studied the metabolic aspects of PPARa and LXRa regulation in primary rat hepatocytes. As could be expected, both naringenin and PPARa agonist, WY14,643 led to a similar decrease in triglyceride production and an increase in ketone body secretion (Fig. 4b). Intracellular levels of hepatic triglycerides were also slightly reduced (Fig. 4c). Interestingly, naringenin caused a much steeper 73% decrease in VLDL secretion compared to 33% decrease by WY14,643 (Fig. 4a). This difference was significant (p = 0.006), and could possibly be due to inhibition of cholesterol synthesis through LXRa.
Indeed, the mRNA abundance of LXRa-target genes that regulates fatty acid and cholesterol synthesis, such as ABCA1, ABCG1, HMGR, and FASN decreses (Fig. 3d). While cholesterol production could not be detected in our system (data not shown), cholesterol serves as the percuror of hepatic bile acids. Interestingly, LXRa activation was shown to drive bile synthesis in rats [24,25]. Therefore, the 32% decrease in bile acids prodcution following naringenin stimulation (Fig. 4b) serves as a surrogate measure of choelsterol production. WY14,643 which upregulates PPARa without effecting LXRa, showed no such change. Regretfully, no reliable LXRa inhibitor is commercially available, and 5CPPSS-50 showed significant toxicity in our hands (Fig. S1). Preliminary results using siRNA to LXRa show some inhibition of bile acid and VLDL production, although results were inconclusive (data not shown).
We note that the GAL4 fusion reporter data suggests that inspite of the well known cross-regulation between PPARa and LXRa [42,43,44], naringenin appears to acts independently on each of these nuclear receptors. This is another indication of the nuclear receptor family promiscuity, and suggests that complex metabolic programs could be induced by relatively few compounds. Indeed, dual PPARa and PPARc agonists have recently been investigated as normoglycemic and antiatherogenic agents [56]. Naringenin activation of both PPARa and PPARc suggests a similar ability to regulate insulin sensitivity and LDL levels. However, in contrast to other dual PPAR agonists, such as Aleglitazar, our work shows naringenin is also an LXRa inhibitor. The metabolic program provoked by naringenin, appears to be a fed-to-fasted transition in the lipid metabolism of primary hepatocytes. Naringenin not only increases fatty acid oxidation but also inhibit fatty acid and cholesterol synthesis.
The potential of using a naturally occurring dietary supplement to regulate lipid metabolism is appealing as this by product of the grapefruit juice industry is non-toxic, cheap, and has demonstrated anti-inflammatory properties. This is especially important in the context of the rising costs of cardiovascular care, estimated by the AHA to rise above $500 billion this year. Naringenin ability to inhibit HMGR, the target of statins, while upregulating PPARa, the target of fibrates, suggest it can naturally find its place in the routine treatment of hyperlipidemia.
Finally, our group and other have shown that the Hepatitis C Virus (HCV) is critically dependent on host lipid metabolism [21,57,58]. Similar interplays were shown for the Hepatitis B Virus (HBV) [59,60]. Therefore, compounds that modulate hepatic lipid metabolism could have significant antiviral effect.
And indeed, our work shows that naringenin blocks HCV production from Huh7.5.1/JFH1 infected cells [21]. These findings form the basis of a currently conducted clinical trial to explore naringenin inhibition of HCV production in nonresponding patients. Interestingly, the anti-inflammatory properties of naringenin could be readily explained in the context of PPAR activation. Such properties could have a significant effect on liver inflammation, preventing or delaying the development of hepatosteatosis and cancer [61].

Cell culture
Huh7 cells were a kind gift of Prof. Raymong Chung, Massachusetts General Hospital. The cells were cultured in DMEM supplemented with 10% FBS, and 200 units/mL penicillin and streptomycin in a 5% CO 2 -humidified incubator at 37uC. Huh7 cells were passaged every 3 days and used at passage ,15.

GAL4-nuclear receptor activation assays
Activation of PPAR LBD was quantified using the previously described HGLN5 PPARa and PPARc cell line [20]. Briefly, HeLa cells were stably transfected with the p(GAL4RE)5-bGlob-Luc-SVNeo plasmid, encoding the firefly luciferase gene driven by a pentamer of yeast activator GAL4 binding sites in front of bglobin promoter [20]. Cells were subsequently stably transfected with either pGAL4-PPARa-puro, or pGAL4-PPARc-puro, encoding amino acids 1-147 of GAL4, followed by a short linker and the LBD of either PPARa or PPARc, respectively [20]. HGLN5 cells were seeded at a density of 100,000 cells/cm 2 , test compounds were added 8 hours later and incubated for 16 hours. Following treatment, cells were washed with PBS and lysed in 25 mM Tris buffer (pH 7.8). Protein concentration was calculated using the Bradford assay and used to normalize the luciferase activity. Finally, activation of PPARa and PPARc reporters is presented as percent of maximal activation by the known agonists GW7647 and BRL49653, respectively.
LXRa activation was investigated using the GeneBLAzer Betalactamase reporter technology (Invitrogen SelectScreen Cell-Based Nuclear Receptor Profiling Service, Madison, WI). LXR-alpha-UAS-bla HEK 293T cells were thawed and resuspended in Assay Media (DMEM phenol red free, 2% CD-treated FBS, 0.1 mM NEAA, 1 mM sodium pyruvate, 100 units/mL penicillin and streptomycin) to a concentration of 312,500 cells/mL. The control agonist TO901317 at the pre-determined EC 80 concentration (5 nM) was added to wells containing variable concentrations of naringenin. The plate was incubated for 16-24 hours at 37uC and 5% CO 2 in a humidified incubator. Substrate loading solution was added to each well and the plate is incubated for 2 hours at room temperature. The plate is read on a fluorescence plate reader. Results for each concentration (n = 4) are reported as percent activation of TO901317-stimulated, naringenin-free controls.

TR-FRET Assays
LanthaScreen TR-FRET Coactivator Assays, purchased from Invitrogen (Madison, WI), were used to identify agonists and antagonists of PPARa and of LXRa. In these cell-free assays, ligands are identified by their ability to bind the recombinant LBD of the respective receptor and induce a conformational change that results in recruitment of a fluorescein-labeled coactivator peptide. A purified, glutathione S-transferase (GST)tagged PPAR alpha or LXRa LBD is indirectly labeled using a terbium-labeled anti-GST tag antibody. Recruitment of fluorescein-labeled coactivator peptide -PGC1a for PPARa or Trap220 for LXRa -is measured by monitoring fluorescence resonance energy transfer (FRET) from the terbium-labeled antibody to the fluorescein on the peptide, resulting in a high TR-FRET ratio (520/490 nm emission). Test compounds were diluted in DMSO, and assays were run per the manufacturer's instructions. Briefly, to test the ability of a molecule to function as an agonist, increasing concentrations of naringenin or control agonist were added to LBD and co-activator peptide solutions. To test the ability of a molecule to function as an antgonist, a similar protocol was followed, but 250 nM TO901317 (EC80 of the agonist, measured in this assay) was added to all wells. In both agonist and antagonist modes, following 1 to 2 hour incubation at room temperature, the 520/490 TR-FRET ratio was measured with a PerkinElmer Envision fluorescent plate reader with TRF laser excitation using the following filter set: excitation 330 nm, emission 495 nm, and emission 520 nm. A 100 msec delay followed by a 200 msec integration time was used to collect the time-resolved signal. Results are displayed as percent activation compared to maximal activation of positive control.

PPAR and LXRa response element luciferase reporter assays
Activation of PPRE and LXRE was quantified by transiently transfecting Huh7 cells with previously described firefly luciferase reporter plasmids, pACOX(62)luc and pDR4(62)luc, respectively [44,63]. The pRL-TK plasmid (Promega, Madison, WI), constitutively expressing renilla luciferase, was co-transfected as positive control. pACOX(62)luc was transfected into Huh7 cells cultured in OptiMEM. After 22 hours of culture, cells were stimulated with naringenin, WY14,643, or ciglitizone for 24 hours in standard culture medium. To quantify LXRE activity, cells were similarly transfected and treated with naringenin, 5CPPSS-50, or TO901317. Ratio of firefly to renilla luciferase luminescence was quantified using a Dual Luciferase Assay kit (Promega) following the manufacturer's instructions. DMSO levels were equal in all samples and never exceeded 0.5%. Results are reported as percent activation compared to DMSO-only controls.

Quantitative Real Time Polymerase Chain Reaction (qRT-PCR)
Following a 24-hour stimulation, cells were lysed with RLT Plus buffer containing b-mercaptoethanol and RNA was isolated using RNeasy Mini Kit on a QIACube device (Qiagen, Valencia, CA). Total RNA was quantified on a ND-1000 spectrophotometer (NanoDrop Technologies, Rockland, Del.) and mRNA transcript abundance was measured on a MyiQ Real-Time PCR Detection System using iScript One-Step RT-PCR Kit With SYBR Green (Bio-Rad, Hercules, CA), according to the manufacturers' instructions. Primers used in these reactions (Integrated DNA Technologies, Coralville, IA) were designed using the PRIMER-BLAST program and appear in Table 1.

Human ApoB Enzyme-Linked Immunosorbent Assay (ELISA)
Huh7-secreted ApoB-100 was detected using ALerCHEK, Inc. (Portland, ME), total human ApoB-100 ELISA kit. The medium was diluted 1:10 with the specimen diluent, and the assay was carried out according to the manufacturer's directions.

Analysis of metabolic changes in primary rat hepatocytes
Primary rat hepatocytes were harvested from adult female Lewis rats purchased from Charles River Laboratories, as previously described Hepatocyte viability was greater than 90% and purity above 95%. [64]. All animals were treated in accordance with National Research Council guidelines and approved by the Subcommittee on Research Animal Care at the Massachusetts General Hospital (IACUC #2005N000109). Cells were seeded on collagen-coated dishes at a density of 150,000 cells/cm 2 under serum-free conditions, using 100 mL/mL soluble collagen type-I as attachment factor. Serum-free hepatocyte culture medium was purchased from Lonza (Walkersville, MD). Cells were stimulated with naringenin or WY14,643 for 24 hours, and cell culture medium was collected for metabolic analysis. Cell pellet was collected for intracellular triglyceride and total protein determination.
Urea concentration was measured using diacetylmonoxime methodology using a commercial available Blood Urea Nitrogen kit (Stanbio Labs, Boerne, TX). Triglycerides, in the culture medium and cell extracts, were quantified using a commercial kit (Sigma Chemical, St.Louis, MO) based on enzymatic hydrolysis by lipase to glycerol. Ketone bodies, were measured based on the appearance of NADH in conversion to acetoacetate in presence of b-hydroxybutyrate dehydrogenase (Zupke et al.1998). Total cholesterol was measured by a commercial available kit (StandBio Labs) based on the reaction of free cholesterol and cholesterol esters with cholesterol oxidase. Bile acids were determined through the formation of NADH in presence of the enzyme 3-a-hydroxysteroid dehydrogenase (Bio-Quant, San Diego, CA).  Statistics Data are expressed as the mean 6 standard deviation. Statistical significance was determined by a one-tailed Student's t-test. A P-value of 0.05 was used for statistical significance. Figure S1 5CPPSS-50 led to no change in LXRE activity. In all experiments, Renilla luciferase was used to account for variability in transfection efficiencies. Found at: doi:10.1371/journal.pone.0012399.s001 (3.87 MB TIF)