PPARα: A Master Regulator of Bilirubin Homeostasis

Hypolipidemic fibrates activate the peroxisome proliferator-activated receptor (PPAR) α to modulate lipid oxidation and metabolism. The present study aimed at evaluating how 3 PPARα agonists, namely, fenofibrate, gemfibrozil, and Wy14,643, affect bilirubin synthesis and metabolism. Human umbilical vein epithelial cells (HUVEC) and coronary artery smooth muscle cells (CASMC) were cultured in the absence or presence of the 3 activators, and mRNA, protein, and/or activity levels of the bilirubin synthesizing heme oxygenase- (HO-) 1 and biliverdin reductase (BVR) enzymes were determined. Human hepatocytes (HH) and HepG2 cells sustained similar treatments, except that the expression of the bilirubin conjugating UDP-glucuronosyltransferase (UGT) 1A1 enzyme and multidrug resistance-associated protein (MRP) 2 transporter was analyzed. In HUVECs, gemfibrozil, fenofibrate, and Wy14,643 upregulated HO-1 mRNA expression without affecting BVR. Wy14,643 and fenofibrate also caused HO-1 protein accumulation, while gemfibrozil and fenofibrate favored the secretion of bilirubin in cell media. Similar positive regulations were also observed with the 3 PPARα ligands in CASMCs where HO-1 mRNA and protein levels were increased. In HH and HepG2 cells, both UGT1A1 and MRP2 transcripts were also accumulating. These observations indicate that PPARα ligands activate bilirubin synthesis in vascular cells and metabolism in liver cells. The clinical implications of these regulatory events are discussed.


Introduction
Bilirubin is an endogenous bile pigment produced from heme degradation by the sequential action of the heme oxygenase (HO) and biliverdin reductase (BVR) enzymes. In humans, 2 active isoforms of heme oxygenase, namely, HO-1 and HO-2, convert heme into carbon monoxide, free iron, and biliverdin. This reaction is considered as the ratelimiting step in heme to bilirubin catabolic process [1]. BVR subsequently reduces biliverdin into bilirubin. While HO-2 is constitutively expressed, HO-1 is encoded by a highly inducible gene activated by a vast variety of endogenous and exogenous stimuli [2]. Actually, HO-1 induction is considered as a major component of the cellular response to oxidative stress, particularly in the vasculature [2]. In humans, HO-1 deficiency is related to many dangerous side effects, including injury of vascular endothelium and cardiovascular diseases [2]. Genetic polymorphisms causing low HO-1 protein expression are positively associated with increased risk for coronary events [3,4].
Following synthesis, bilirubin binds albumin into the blood to reach the liver, where it sustains additional catabolic reactions before its elimination into the bile. The UDP-glucuronosyltransferase (UGT) 1A1 enzyme conjugates bilirubin into hydrophilic mono-and diglucuronide derivatives, which are excreted into the bile through the canalicular multidrug resistance-associated transporter (MRP) 2 protein.
MRP2 is a member of the ATP-binding cassette (ABC) transporters family and is essential for bilirubin-glucuronide secretion into bile [5]. Genetic defects in the human UGT1A1 gene are associated with unconjugated hyperbilirubinemia, which can be either asymptomatic as in individuals with 2 PPAR Research Gilbert syndrome [6] or severe as in the case of Crigler-Najjar syndrome types I and II [7], depending on the remaining UGT1A1 activity. The moderate bilirubin elevation observed in Gilbert's syndrome both lowers the risk of developing coronary heart diseases [8] and accelerates the development of neonatal jaundice during the 2 first days of life [9]. Similarly, functional mutations within the MRP2 gene result in conjugated nonhaemolytic hyperbilirubinemia, also called Dubin-Johnson syndrome [10].
Under normal circumstances, circulating levels of total, direct (i.e., conjugated), and indirect (unconjugated) bilirubin are, respectively, <17, 2-5, and 3-12 M [11]. While moderate increases (17 to 20 M) are associated to a reduction of cardiovascular events risk, bilirubin is a neurotoxic molecule at high concentration [12]. Bilirubin's atheroprotective properties relate to bilirubin's ability to efficiently scavenge reactive oxygen species (ROS), and by so to reduce low-density lipoproteins (LDL) oxidation in the vasculature [13]. On the other hand, bilirubin-induced neurotoxicity is mainly observed in newborns, where brain accumulation of unconjugated bilirubin provokes neuronal cell death and causes permanent neurologic sequel (a situation called bilirubininduced neurologic dysfunction, BIND) (reviewed in [14]). The dual role of bilirubin as an atheroprotective agent and neurotoxic molecule renders essential a tight control of its metabolism.
The peroxisome proliferator-activated receptor (PPAR) belongs to the PPAR family of lipid sensors. With 2 other members, PPAR and , these ligand-activated transcription factors regulate lipid and fatty acid homeostasis, as well as energy storage and expenditure [15]. Upon ligand activation, these receptors form an active heterodimer with their partner retinoic X receptor (RXR) and bind to the promoter regions of target genes on specific DNA sequences called PPAR response elements (PPRE) [16]. PPAR target genes play key roles in lipid transport, fatty acid -oxidation, lipogenesis, lipoprotein uptake, and metabolism, as well as in cholesterol transport and elimination [17]. PPAR is mainly expressed in the liver and heart, where it is activated by endogenous activators such as fatty acid derivatives (i.e., eicosanoids, palmitic, oleic, and linoleic acids) or exogenous ligands, such as the Wy14,643 compounds or fibrate drugs (i.e., gemfibrozil, clofibrate, ciprofibrate, and fenofibrate) [18]. These fibric acid derivatives have been used in clinics since the mid-1960s to lower plasma triglyceride (TG) levels in patients with atherogenic dyslipidemia [19].
Several investigations identified PPAR as an important modulator for genes controlling bilirubin synthesis (HO-1) and metabolism (UGT1A1 and MRP2) [20][21][22], leading to the hypothesis that fibrates coordinately control the synthesis and metabolism of this bile pigment. However, all these studies were performed in different experimental settings, using variable ligands, doses, experimental models, and analytical tools. Considering the agonist-and/or cell type-dependent manner in which PPAR agonists regulate their target genes [23,24], we comprehensively and comparatively analyzed the ability of gemfibrozil, fenofibrate, and Wy14,643 compound to regulate HO-1, BVR, UGT1A1, and/or MRP2 expression in relevant hepatic and vascular cell models.

Materials.
Wy14,643 (pirinixic acid, 4-Chloro-6-(2,3xylidino)-2-pyrimidinylthioacetic acid) and fenofibrate were from Sigma (St. Louis, MO) and ICN Pharmaceuticals, Inc. (Montréal, Canada), respectively. Gemfibrozil was from Pfizer Canada (Kirkland, Canada). Fetal bovine serum (FBS) and other cell culture reagents were from Invitrogen (Burlington, Canada). The SYBR Green PCR mix was purchased from Applied Biosystems (Foster City, CA). Protein assay reagents were obtained from Bio-Rad Laboratories Inc. (Marnesla-Coquette, France). The anti-HO-1 antibody was from Santa Cruz (Santa Cruz, CA). The anti-actin antibody was purchased from Sigma and the anti-rabbit IgG antibody was from GE Healthcare (Piscataway Township, NJ). cryopreserved human hepatocytes (HH) from 2 individual donors were from Celsis: In Vitro Technologies (Baltimore, MD). Donor 1 was an African-American man at the age of 46 (cause of death: anoxia), while donor 2 was a Caucasian woman at the age of 40 (cause of death: drug overdose). HepG2 cells were grown as described [26]. Cells (3 × 10 5 per well) were plated in 12-well  of Superscript II (Invitrogen, Burlington, Canada) with up to 1 g of total RNA and 7.5 ng of random hexamere (Roche, Laval, Canada) at 42 ∘ C for 50 min, as described in [26]. The real-time PCR reactions were performed using an ABI Prism 7500 instrument from Applied Biosystems (Foster City, CA). For each reaction, the final volume of 20 L was comprised of 10 L of SYBR Green PCR Mix, 2 L of each primer (Table 1), and 6 L of the 1/500 dilution of RT products. Conditions for real-time PCR were 95 ∘ C for 20 sec, 95 ∘ C for 30 sec, and annealing temperature for 20 sec for 40 cycles. The specific amplification was ensured by direct sequencing of PCR products. Threshold cycle (Ct) values were analyzed using the comparative Ct (ΔΔCt) method as recommended by the manufacturer (Applied Biosystems). The amount of target gene (2 −ΔΔCt ) was obtained by normalizing to the endogenous reference 28S and was expressed relatively to vehicle-treated cells set at 1. For each gene, the amplification efficiency and the accuracies of ΔΔCt of target genes compared with 28S were tested using 2 to 5 log of concentrations of cDNA produced from cell purified mRNA.

Western Blot Analysis.
Control and treated cells were washed in ice-cold PBS and harvested in ice-cold PBS containing 0.5 mM dithiothreitol. Total proteins (20-25 g) were size-separated on 10% SDS-polyacrylamide gels and immunoblotted with an anti-HO-1 antibody (1 : 400). An anti-rabbit IgG donkey antibody (1 : 10,000) conjugated with peroxidase was used as the second antibody. Immunocomplexes were visualized on hyperfilm. The same membranes were then rehybridized with an anti-actin (1 : 2,000) antibody as a loading control assessment.

Bilirubin Determination in Culture
Media. Culture media were analyzed for bilirubin through liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). Briefly, 500 L media were added to a mixture composed of 100 L butylated hydroxytoluene (BHT, 1% w/v methanol), 100 L ascorbic acid (0.4% w/v methanol), 500 L H 2 O, 25 L HCl (0.2 M), and 100 L mesobilirubin (100 ng/L, internal standard, Sigma-Aldrich Inc.). After liquid : liquid extraction with 2 mL of chloroform, the organic phase was evaporated under nitrogen, and analytes were dissolved in 200 L of a methanol : chloroform : water (81 : 10 : 9) solution. The chromatographic separation was achieved with an Alliance 2690 LC apparatus (Waters, Milford, MA) equipped with an ACE-3 C18 HL 100 × 4.6 mm (3 m particles) column, using a 5 mM ammonium formate methanol gradient. Bilirubin and mesobilirubin detection was performed through tandem mass spectrometry with an API4000 instrument (Applied Biosystems-Sciex) mass spectrometer. The limit of quantification (LOQ) was 1 ng/mL. Bilirubin concentration values were calculated as the ratio of areas under curve for bilirubin versus mesobilirubin.
2.6. Statistical Analyses. All data are presented as mean ± standard deviation (SD). Comparisons between two groups were performed using the two-tailed Student's -test with the JMP V4.0.2 software (SAS Institute, Cary, NC).

Gemfibrozil Stimulates HO-1, UGT1A1, and MRP2
mRNA Expression. We first tested whether gemfibrozil affects mRNA expression of the bilirubin synthesizing HO-1 and BVR enzymes in vascular endothelial cells. After 24 H, the drug caused a 3-time ( < 0.001) accumulation of HO-1 mRNA levels (Figure 1(a)). Similar inductions were also observed in CASMCs (Figure 1(c); 1.4-fold; < 0.01). By contrast, expression of the biliverdin to bilirubin converting BVR enzyme was not affected in both cell models ( Figures  1(b) and 1(d)). We next investigated whether UGT1A1 and MRP2 expression was also responding to gemfibrozil exposure in 2 human liver cell lines. In both human hepatocytes and hepatoma HepG2 cells, a 24 H treatment with gemfibrozil resulted in 2 Overall, these results indicate that both clinically relevant (gemfibrozil, fenofibrate) and high affinity (Wy14,643) PPAR ligands act as inducers of genes controlling bilirubin synthesis and metabolism.

PPAR Ligands Differentially Affect HO-1 Protein Levels and/or Bilirubin Secretion in Vascular Cells.
To further evaluate the consequences of the fibrate-dependent activation of HO-1 mRNA expression, HUVECs were subsequently investigated for HO-1 protein levels and bilirubin secretion in culture media (Figure 4). While HUVECs exposed to 250 M gemfibrozil for 48 H displayed no major changes in HO-1 protein contents (Figure 4(a)), homogenates from cells cultured for the same duration but in the presence of Wy14,643 (75 M) or fenofibrate (250 M) displayed convincing accumulation of the heme oxygenase protein when compared to   samples from control cells (vehicle) (Figures 4(b) and 4(c)). Interestingly, an agonist-specific pattern was also observed when analyzing bilirubin secretion (Figure 4(d)). While gemfibrozil and fenofibrate led to 2.3-and 3.0-fold increases in culture media bilirubin concentration, Wy14,643 had the opposite effect since the bile pigment was actually 1.6-fold less abundant than in media from control cells.
To evaluate whether such discrepancies could reflect an inadequacy of the cell model, similar experiments were then performed with CASMCs (Figures 4(e)-4(g)). In these cells, all 3 PPAR ligands caused HO-1 proteins accumulation. Interestingly, while similar amounts of cell homogenates and antibody were used for Western blotting, detection of the HO-1 protein in HUVEC homogenates required a longer film exposure than those from CASMCs, suggesting that, in baseline, HO-1 is more abundant in the smooth muscle model than in the endothelial ones. However, the opposite was observed when measuring bilirubin formation since this molecule was only detected at concentrations below the limit of quantification in media from untreated CASMCs (data not shown). Interestingly, exposing these cells to gemfibrozil, Wy14,643 and fenofibrate resulted in an unquantifiable but convincing increase of the bilirubin's AUC in LC-MS/MS analyses (data not shown).
In summary, even when considering technical limitations, these observations support the idea that PPAR ligands are positive regulators of HO-1 expression and activity in HUVECs and CASMCs, when taken together.

Discussion
The present investigations provide a comprehensive analysis of PPAR agonists' impacts on bilirubin synthesis and metabolism. Our observations indicate that each tested PPAR activator regulates genes controlling bilirubin synthesis in vascular cells and its metabolism in hepatic cells. Considering the antioxidant properties of bilirubin, it can be envisioned that this coordination allows beneficial local antiatherosclerotic effects, while avoiding toxic accumulation of the bile pigment in the systemic circulation ( Figure 5).
We observed that the 3 PPAR agonists assayed with HUVECs and/or CASMCs were able to upregulate HO-1 mRNA expression, protein content, and/or activity levels. Similar inductions were previously reported with fenofibrate and Wy14,643 in HUVEC and human vascular smooth muscle cells (VSMCs) [20]. In addition, fenofibrate also prevents the reduction of HO-1 transcript levels caused by exposure of human pulmonary artery endothelial cells to serum from patients with stable advanced chronic heart failure [27]. However, to the best of our knowledge, the present study provides the first experimental evidences that gemfibrozil also positively regulates HO-1 expression in cell models of the human vasculature. Thus, our investigations further support the idea that PPAR agonists act as positive regulators of the human HO-1 gene in cells from the vascular wall. These regulatory events are thought to participate to the therapeutic benefits observed with fibrate drugs in the context of several pathological situations. Indeed, a number of experimental evidences demonstrate the contribution of the HO-1 induction to the positive effects exerted by PPAR agonists against: (i) the hepatotoxicity caused by iron deposition in the liver [28]; (ii) renal injury caused by ischemia/reperfusion [29]; (iii) iron-induced cardiomyopathy [30]; and (iv) carboplatininduced nephrotoxicity [31]. Thus, it is tempting to speculate that in vascular wall cells, HO-1 induction also participates to the antiatherosclerotic effects of fibrate drugs. This idea is supported by the recent observation that HO-1 inhibitors reduce the anti-inflammatory and antiproliferative effects of PPAR activators in human VSMCs [20]. Actually, the atheroprotective role played by HO-1 has been deeply documented through gene transfer experiments involving a series of relevant animal models [32][33][34]. These studies   Figure 5: Coordinated regulation of bilirubin synthesis and metabolism by PPAR agonists has antiatherosclerotic effects in the vasculature and normobilirubinemic consequences in the liver. Results presented here suggest that PPAR agonists, such as fenofibrate, gemfibrozil, and Wy14,643, activate the synthesis of bilirubin in the vasculature and its elimination in the liver. In vascular cells (HUVEC and CASMC), fibrates activate PPAR , which in turn binds to PPREs located in the promoter region of the HO-1 gene [20]. This leads to an increased HO-1 expression and heme-to-bilirubin conversion. As a potent antioxidant, bilirubin scavenges reactive oxygen species, and by so exerts atheroprotective effects. In liver cells (hepatocytes), fibrates also activate PPAR , which in turn binds to PPREs located in the promoter region of the UGT1A1 [25] and potentially MRP2 genes (this remains to be established). These regulatory events lead to increased bilirubin-glucuronide production and biliary secretion and by so contribute to reduce systemic accumulation of bilirubin. BVR: biliverdin reductase; CASMC: coronary artery smooth muscle cells; G: glucuronide; HO-1: heme oxygenase-1; HUVEC: human umbilical vein endothelial cells; MRP2: multidrug resistanceassociated transporter protein 2; PPAR: peroxisome proliferator-activated receptor; PPRE: PPAR response element; ROS: reactive oxygen species; and UGT1A1: UDP-glucuronosyltransferase 1A1.
established that HO-1 overexpression reduces the vascular inflammatory response in rat VSMCs [33] and attenuates atherosclerosis development in apoE deficient mice [34] and the rat aortic transplant model [32]. Thus, the PPARdependent induction of HO-1 in vascular wall cell models may also generate antiatherosclerotic effects ( Figure 5). In addition to the bile pigment, the heme-to-bilirubin conversion system also generates carbon monoxide and free iron [35]. The relative contribution of each of these endproducts in HO-1 mediated atheroprotection still remains to be clarified [35]. However, an impressive number of clinical studies illustrate the inverse relationship linking serum total bilirubin concentration to the cardiovascular risk, at least in stable coronary conditions (reviewed in [36]). To quote only a few recent examples, serum bilirubin is inversely correlated with: (i) the severity of disease in patients with stable coronary artery disease [37]; (ii) the carotid artery intima/media thickness in nondiabetic and type 2 diabetic subjects [38]; and (iii) the clinical outcomes at the time of the 5-year follow-up of patients with cardiac syndrome X [39]. The protective properties of bilirubin relate to its powerful antioxidant capabilities, and its efficiency at scavenging peroxyl radicals [40]. This characteristic is of particular interest in the vasculature where bilirubin prevents a precursor event in atherogenesis, namely, the oxidation of LDL particles [40]. Accordingly, high serum bilirubin levels are associated with reduced circulating levels of oxidized LDL in healthy and Gilbert syndrome individuals [41,42]. Thus, in patients with stable coronary conditions, high bilirubin levels are associated with reduced oxLDL formation and favorable endothelial function [43]. Here, we observe that both clinically relevant fibrates (gemfibrozil and fenofibrate) caused a significant increase in bilirubin production in culture media of vascular endothelial cells. Such an accumulation of unconjugated bilirubin (UCB) in media has been previously associated with intracellular content of UCB in a linear manner and diminution of lipid peroxidation susceptibility in UCB-treated rat brains [44]. It can therefore be postulated that fibrate treatment increases bilirubin production in the vascular wall and by so contribute to reduce LDL oxidation and atherosclerotic plaque formation ( Figure 5). One can argue that such a mechanism could be minimized by the low BVR modulation observed in PPAR agonists treated cells. However, it is well established that HO-1 catalyzes the rate-limiting process in heme degradation (reviewed in [1]). Actually, the absence of BVR modulation in our experiments is consistent with the fact that BVR is a noninducible protein [45], which mainly sustains posttranslational regulatory processes, such as autophosphorylation, instead of transcriptional gene controls [45].
Beyond HO-1 expression in vascular wall cells, our investigations also demonstrate that gemfibrozil, fenofibrate, and Wy14,643 positively regulate UGT1A1 and MRP2 expression in human cell models. In humans, UGT1A1 is the unique glucuronosyltransferase enzyme catalyzing the conversion of bilirubin into highly hydrophilic and easily excretable glucuronide derivatives [7], and MRP2 ensures the export of these derivatives into the bile ( Figure 5) [5]. The present observations therefore indicate that PPAR activators coordinately activate the 2-step bilirubin detoxification system in liver cells. These observations are fully consistent with previous investigations revealing that gemfibrozil [23], fenofibrate [23], and Wy14,643 [25] cause significant accumulations of UGT1A1 transcripts in hepatocytes. By contrast, the PPARdependent regulation of MRP2 expression sustained some controversy in the recent years. Although bezafibrate exposure caused significant increase in MRP2 mRNA in human hepatic HepaRG cells [22] as well as in liver cells from treated mice [46], it was unclear whether such effects were PPAR -dependent [22]. Other in vivo investigations using ciprofibrate also revealed conflicting results. While Kok and colleagues [47] and Aleksunes and Klaassen [48] reported PPAR -dependent induction of Mrp2 mRNA expression in livers from ciprofibrate-treated mice, other studies involving clofibric acid and ciprofibrate revealed nonsignificant reduction of the Mrp2 proteins in rat liver [49] and mRNA in mice liver [50], respectively. In summary, results of the current study confirm the positive effects that PPAR agonists exert on UGT1A1 expression and support a similar response of the MRP2 gene, at least in human hepatocytes and hepatoma HepG2 cells.
As for HO-1, the PPAR -dependent induction of UGT1A1 expression is pharmacologically relevant for the treatment of hyperbilirubinemia. Indeed, the improved glucuronidation activity resulting from these regulatory events favors bilirubin elimination [51]. Unconjugated hyperbilirubinemia (i.e., jaundice) is a common clinical problem during the neonatal period which may result in brain damage, even in healthy full-term newborns [52]. While the main treatment for hyperbilirubinemic neonates remains phototherapy [53], the UGT1A1 inducing effects of fibrate drugs encouraged their evaluation for bilirubin elimination. These investigations suggested that clofibrate is efficient in reducing both total serum bilirubin and phototherapy duration (reviewed in [54]). However, such a therapeutic benefit appears to be fibrate-dependent since, in similar studies performed with preterm and term neonates with moderate jaundice, gemfibrozil failed to reduce the phototherapy duration or peak bilirubin levels [55]. Furthermore, as highlighted in a recent Cochrane evaluation [54], most of clinical studies performed with clofibrate were conducted in Iran, and the benefits of the therapy remain to be ascertained through larger trials performed in different countries.

Conclusion
In conclusion, the present study demonstrates that PPAR activators, such as gemfibrozil, fenofibrate, and Wy14,643, coordinately stimulate the gene expression of HO-1 in cell models of the human vasculature and of UGT1A1 and MRP2 in human liver cells. These observations support the idea that, by modulating both bilirubin synthesis and catabolism, fibrate drugs exert antiatherosclerotic effects in the vasculature and antihyperbilirubinemic properties in the liver ( Figure 5).