Upregulation of hepatic VLDLR via PPARα is required for the triglyceride-lowering effect of fenofibrate.

The liver and the VLDL receptor (VLDLR) play major roles in TG and VLDL metabolism. However, the exact role of liver VLDLR is not well known because of the absence of or difficulty in detecting VLDLR in the liver. In this study, we demonstrate that fenofibrate, a PPARα agonist and widely used TG-lowering drug, markedly upregulated hepatic VLDLR, which is essential for lowering TG. This study also shows that the distinct regulatory roles of PPARα agonists on VLDLR in the liver and peripheral tissues including adipose tissues, heart, and skeletal muscles are due to the pattern of expression of PPARα. The in vivo portion of our study demonstrated that oral fenofibrate robustly increased liver VLDLR expression levels in hyperlipidemic and diabetic mice and significantly reduced the increase in serum TG observed in wt mice after feeding with high-fat diet (HFD) but not in Vldlr(-/-) mice or Pparα(-/-) mice. However, overexpression of mouse VLDLR in livers of Vldlr(-/-) mice significantly prevented the increase in serum TG induced by HFD. The in vitro portion of our study showed that fenofibrate upregulated VLDLR transcriptional activity through PPAR response element binding to the VLDLR promoter. The conclusions of our study provide a novel mechanism for the TG-lowering effects of fenofibrate in the treatment of dyslipidemia.

progression and increased severity of diabetic complications. The dysregulation of hepatic VLDL secretion is believed to be the mechanism underlying increases in TG-related VLDL and is responsible for the pathogenesis of hepatic insulin resistance in type 2 diabetes, representing a principal target for clinical intervention.
Fenofi brate, a fi brate-derived drug, was widely used in the treatment of dyslipidemia in diabetic patients for decades prior to the use of statins ( 3,4 ). Fibrate drugs reduce plasma TG-related VLDLs, raise plasma HDLs ( 5 ), and have a modest effect on LDL cholesterol levels ( 6 ). Fenofi brate acts as a specifi c agonist of PPAR ␣ , a nuclear hormone receptor that functions as a transcription factor and regulates the expression of a number of genes involved in lipid metabolism and insulin resistance. Fenofibrate has been reported to reduce the body weight gain upon high-fat diet (HFD) stimulation ( 7 ). PPAR ␣ lowers serum TG by infl uencing many genes involved in VLDL production, lipid traffi cking, and TG-rich lipoprotein clearance ( 8 ). For example, PPAR ␣ increases the expression of liver microsomal TG transfer protein (MTTP), a protein that mediates the lipidation of apoB100 to form a nascent VLDL ( 9 ); the expression of liver LPL, an enzyme that mediates the clearance of TG-rich VLDL and chylomicrons ( 10 ); and the synthesis of apoC-III ( 11 ), apoA-II ( 12 ), and apoA-V ( 13 ). However, the precise genes underlying the suppressive effect of PPAR ␣ agonists on hepatic VLDL metabolism currently are not well understood.
The VLDL receptor (VLDLR), a member of the LDL receptor (LDLR) family, is widely expressed in the heart, skeletal muscles, adipose tissues, and macrophages ( 14,15 ). However, its levels are barely detectable in the liver under normal conditions. VLDLR mediates the uptake of VLDL by peripheral tissues through LPL-dependent lipolysis or receptor-mediated endocytosis (16)(17)(18)(19), and thus

Immunocytochemical analysis
The procedure for immunocytochemical analysis followed a protocol described previously ( 23 ).

Quantitative real-time RT-PCR
The procedure for quantitative real-time RT-PCR was described previously ( 23 ). Primers were designed from the cDNA sequences spanning >1 kb introns using the Primer3 software.

Generation of adeno-associated virus-mediated mouse VLDLR
The mouse Vldlr gene was amplifi ed from a cDNA clone by PCR. The resulting fragment was sequenced and inserted into pHelper AAV (pAAV)-ALBp . The constructs in the pAAV vectors were transfected into HEK293 cells and purifi ed (Vector Biolabs, Eaglevilla, PA).

TG measurement
The measurement of TG concentrations in plasma followed a manufacture's procedure (Triglyceride Determination Kit; Sigma-Aldrich). Briefl y, blood was collected and centrifuged at 700 g for 10 min at 4°C. The plasma in the supernatant was collected without disturbing the white buffy layer. The TG in the plasma was initiated enzymatically by adding lipase and incubated at 30°C for 10 min to convert TG to free FA and glycerol. The released glycerol is subsequently measured by a coupled enzymatic reaction system with a colorimetric readout at 540 nm. To determine the total TG level, the glycerol is continuously catalyzed at 37°C for 15 min by a reconstituted TG reagent including ATP, glycerol kinase, glycerol phosphate oxidase, and perioxidase; the released quninoneimine dye, directly proportional to TG concentration of the sample, was measured and recorded at 540 nm . The concentration of total TG and true TG in the sample was calculated as follows: Total TG Concentration = (FA sample Ϫ FA blank )/(FA standard Ϫ FA blank ) × Concentration of Standard Glycerol Concentration = (IA sample Ϫ IA blank )/(IA standard Ϫ IA blank ) × Concentration of Standard plays an important role in VLDL metabolism. In addition, VLDLR has multiple protective roles against obesity, insulin resistance, premature heart disease, tumor growth, infl ammation, and angiogenesis. In mammals, VLDLR is highly conserved in humans and rodents ( 14,15 ). The deletion of VLDLR in mouse models usually does not directly lead to dyslipidemia; however, under stress conditions or after feeding with a HFD, these mice have signifi cantly increased serum TG levels ( 20,21 ). The reintroduction of VLDLR in VLDLR KO mice signifi cantly increases atherosclerotic lesion development, indicating that the VLDLR expressed on macrophages is a pro-atherogenic factor ( 22 ). These characteristics make VLDLR KO mice a good model for the investigation of the pathological mechanisms underlying dyslipidemia. There are many potential mechanisms whereby PPAR ␣ agonists may reduce serum VLDL, including the decreased synthesis of VLDL and increased lipolysis of VLDL by lipase. However, the regulatory effect of fenofi brate on hepatic VLDLR, the key receptor for VLDL catabolism, has not been thoroughly elucidated.
In the present study, we investigated the lipid-regulatory role of fenofi brate in lipid metabolism and the tissue-specifi c relationships of VLDLR, PPAR ␣ , and LPL, thereby providing a novel mechanism by which fenofi brate inhibits the progression of metabolic diseases.

Primary liver cell culture
Liver tissue (10 g) from Ppar ␣ Ϫ / Ϫ and age-matched wt mice was weighted, minced, and digested in Hank's buffered salt solution with 0.25% trypsin (Sigma-Aldrich) at 37°C for 20 min. The digested material was fi ltered and centrifuged at 1,000 g for 10 min. The resulting pellets were washed and resuspended in MEM with 10% FCS and seeded in a 24-well plate.

Statistical analysis
The quantitative data of body weight in different diet groups of mice were analyzed by repeated measures ANOVA. All other quantitative data were analyzed and compared using the Mann-Whitney test. Statistical signifi cance was set at P < 0.05.

Fenofi brate increased VLDLR expression in hepatic cells but not in adipose cells
LPL and VLDLR play crucial roles in the uptake and clearance of TG-rich VLDL. We investigated the effects of fenofi brate on VLDLR and LPL expression in liver and adipose cells.
Confl uent HepG2 cells and 3T3 cells were treated with different doses of fenofi brate for different durations, and the VLDLR levels were determined by Western blot analysis. At 24 h after treatment with fenofi brate, VLDLR had increased by 1.8-, 3.8-, and 5.5-fold over the basal level at concentrations of 25, 50, and 100 M, respectively (n = 3, P < 0.05) ( Fig. 1A, B ). At a concentration of 50 M, fenofi brate increased VLDLR by 1.1-, 3.6-, and 3.8-fold over the basal level at 6, 12, and 24 h, respectively (n = 3, P < 0.05) ( Fig. 1C, D ). Unlike the dose-and time-dependent increases in HepG2 cells, the VLDLR level was not significantly changed by fenofi brate in the 3T3 cells ( Fig. 1E-H ). Additional evidence for fenofi brate-regulated increases in VLDLR was obtained from immunohistochemistry staining of the HepG2 cells with the antibodies of VLDL and VLDLR. As shown in Fig. 1I , compared with the control, both VLDL and VLDLR signals were markedly increased in the hepatic cells treated with fenofi brate, consistent with the VLDL lowering effect of fenofi brate.
We also investigated whether fenofi brate regulates LPL expression in liver and adipose cells. In accordance with the elevation of VLDLR in liver cells, fenofi brate significantly increased the cellular LPL levels in HepG2 cells to approximately 1.8-fold the basal levels at concentrations of 50 and 100 M (n = 3, P < 0.05 vs. the group not treated with fenofi brate) ( Fig. 1C, D ). Fenofi brate did not significantly alter the cellular levels of LPL in the 3T3 cells at the same concentrations ( Fig. 1E, F ). These results suggest that fenofi brate regulates VLDLR and LPL differently in liver and adipose cells.

Fenofi brate upregulated hepatic VLDLR expression and the transcription-enhancing activity of the VLDLR gene in HFD-induced hyperlipidemic and diabetic animals
Our in vitro study established that fenofi brate regulates VLDLR differently in liver and adipose cells. We next examined how fenofi brate regulates VLDLR in vivo. Adult wt mice were fed a HFD either with or without fenofi brate

Electrophoretic mobility shift assay
The PPAR ␣ proteins extracted from Sf9 cells were mixed with purifi ed retinoid X receptor . The mixture was preincubated with nonlabeled wt or mutant competitor oligonucleotides (20or 100-fold) for 30 min and then incubated with biotinylated labeled PPAR response element (PPRE) oligonucleotides. The sequences of double-stranded oligonucleotides were as follows: mouse VLDLR PPRE, 5 ′ -GATTTCAGTTTACAGGTCAGAT-GGCAGGCACAG-3 ′ ( 23 ); nonspecifi c competitor DNA, Poly (deoxyinosinic-deoxycytidylic). Double-stranded oligonucleotides were end-labeled with biotin using the Biotin 3 ′ End DNA labeling kit (Pierce Biotechnology, Rockford, IL). All of the electrophoretic mobility shift assay (EMSA) reactions were conducted according to the manufacturer's instruction of LightShift® Chemiluminescent EMSA kit (Pierce Biotechnology, Rockford, IL). DNA-protein complexes were resolved by electrophoresis through 6% polyacrylamide gels.

VLDLR promoter luciferase reporter activity
HepG2 cells (1 × 10 4 cells/well of a 24-well dish) were cotransfected with 0.1 g pGL-mouse VLDLR promotor Luc and 20 ng Renilla luciferase expression vector pGL4.75 (Promega, Madison, WI) as an internal control by lipofection . Twenty-four hours after achieving optimal effi ciency, cells were exposed to a medium containing high glucose (30 mM) and/or VLDL (5 M) in the presence or absence of fenofi brate (50 M) and GW6471 (10 M) for 24 h. Luciferase activities were assayed using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI).

Protein-DNA binding assay
Nuclear protein was extracted from freshly prepared liver from mice (Millipore, Billerica, MA) and precipitated with rabbit polyclonal anti-PPAR ␣ antibody (H100, Santa Cruz Biotechnology) or normal rabbit IgG (Santa Cruz Biotechnology) and subjected to protein-DNA binding assay with double-stranded oligonucleotide probes according to the manufacturer's instructions (Epigentek, Brooklyn, NY). The binding activity was measured at 520 nm on a Tecan Genios Pro microplate reader. The background absorbance with a blank was subtracted from the sample's reading. The sequences of double-stranded oligonucleotides were as follows (only one strand is shown, and the half-site of the putative PPRE and the mutated PPRE are underlined):

Chromatin immunoprecipitation
Freshly prepared retina from mice treated with or without fenofi brate were chopped into small pieces. The DNA and protein were cross-linked with 1.5% formaldehyde for 15 min and stopped by 0.125 M glycine in a fume hood. Soluble chromatin was prepared using a chromatin immunoprecipitation (ChIP) assay kit (Upstate Biotechnology, Lake Placid, NY). After sonication, lysates from the retina were precipitated with rabbit polyclonal anti-PPAR ␣ antibody (H100) and normal rabbit IgG (Santa Cruz Biotechnology) according to the manufacturer's instructions. Primers used for ChIP PCR were as follows: VLDLR-PPRE forward, 5 ′ -TGAGGCCACAGATGATTTTG-3 ′ ; VLDLR-PPRE reverse, 5 ′ -GGCTCTACACTCAACCTGGTG-3 ′ ; aP2-PPRE forward, 5 ′ -ATGTCACAGGCATCTTATCCACC-3 ′ ; aP2-PPRE Western blots were performed from 60 g of lysate proteins (mean ± SD, n = 6; * P < 0.05 vs. control). E-H: Representative blots (E) and relative expression of VLDLR (F) in 3T3-L1 cells treated with 50 M fenofi brate for the indicated time. Representative blots (G) and relative expression of VLDLR and LPL (H) in 3T3-L1 cells treated with fenofi brate for 24 h at the indicated doses. Western blots were performed from 60 g of lysate proteins (mean ± SD, n = 6; * P < 0.05, ** P < 0.01). I: After 48 h lipoprotein deletion, primary cultured liver cells were cultured in a medium containing DiI-VLDL (5 g/ml) with or without fenofi brate (100 M) for 24 h and subjected to immunohistochemistry with an antibody against VLDLR. Green, VLDLR; red, VLDL. Scale bar, 20 M. (120 mg/kg/day) for 8 weeks; subsequently, the expression of VLDLR in the liver, adipose tissues, heart, and skeletal muscles was determined using Western blotting . Compared with the control mice, a HFD did not lead to a signifi cant change in hepatic VLDLR levels, but feeding with fenofi brate resulted in a robust elevation in hepatic VLDLR levels by approximately 4-fold (n = 3, P < 0.05 vs. high-TG chow without fenofi brate) ( Fig. 2A ). However, in the adipose tissues, heart, or skeletal muscles, high-fat stimulation alone increased the VLDLR level, which was signifi cantly decreased by fenofi brate administration ( Fig. 2B ).

Fenofi brate reduced TG levels induced by a HFD in wt mice but not in Vldlr Ϫ / Ϫ mice
Nascent VLDL is assembled from TGs in the liver and redistributed into peripheral tissues via VLDLR. To determine whether the upregulation of VLDLR plays an essential role in decreasing TG levels, we compared the serum TG levels in wt mice and Vldlr Ϫ / Ϫ mice fed a HFD either with or without fenofi brate. The body weights were recorded weekly. Serum was collected at week 8 for the measurement of TG levels. In wt mice, a HFD induced a marked increase in serum TG levels and body weight, and these variables were signifi cantly reduced by fenofi brate ( P < 0.05) ( Fig. 4A, B ). However, fenofi brate did not signifi cantly alter the serum TG levels in Vldlr Ϫ / Ϫ mice ( Fig. 4C ). Interestingly, although high-fat stimulation signifi cantly increased body weight, this increase the VLDLR gene or a decrease in VLDLR protein degradation, we examined the effect of fenofi brate on the hepatic VLDLR mRNA levels. As shown in Fig. 2C , consistent with the protein VLDLR levels, a high-TG diet did not result in any alterations in liver VLDLR mRNA, but fenofi brate administration led to signifi cantly increased liver VLDLR mRNA levels (6-fold increase, P < 0.01 vs. HFD without fenofi brate). In contrast to the upregulation effect of fenofibrate on hepatic VLDLR, the adipose VLDLR mRNA levels were not altered by fenofi brate administration ( Fig. 2D ).

Fenofi brate upregulated hepatic VLDLR expression in type 1 and type 2 diabetic mice
The liver VLDLR levels were also measured in both type 1 and type 2 diabetic mice. Compared with a regular diet, feeding with fenofi brate signifi cantly increased the expression of liver VLDLR. The protein levels were increased 6.22-fold (n = 5, P < 0.01) in Akita mice ( Fig. 3A, B ) and 12-fold (n = 5, P < 0.01) in db/db mice ( Fig. 3C, D ); while the mRNA levels were increased 19-fold in the age-matched Fig. 3. Fenofi brate upregulates hepatic VLDLR in both T1DM and T2DM mice. Eight-week-old Akita mice and db/db mice were fed with 8 weeks of fenofi brate (120 mg/kg/day). Western blot analysis of VLDLR (A) and relative protein expression (B) in liver tissue from C57 mice (non-diabetes mellitus, NDM model) and Akita mice (T1DM model) fed with normal chow (R) and chow with 0.06% fenofibrate (Feno), respectively. Representative blots (C) and relative protein expression (D) of VLDLR in liver tissue from C57 mice and db/db mice (T2DM model) fed with normal chow and chow with 0.06% fenofi brate, respectively. All Western blots were performed from 60 g of lysate proteins (mean ± SD, n = 4; ** P < 0.01). E-G: Total hepatic RNA was isolated from Akita mice, db/db mice, and their age-matched (16 weeks old) nondiabetic mice treated with or without fenofi brate chow (120 mg/kg/day, 8 weeks). The levels of VLDLR mRNA were determined by quantifi ed real-time PCR (mean ± SD, n = 5; * P < 0.05, ** P < 0.01).

Fenofi brate reduced TG levels and upregulated liver VLDLR levels in a PPAR ␣ dose-dependent manner
Fenofi brate is a PPAR ␣ -specifi c agonist. To determine whether fenofi brate upregulates VLDLR through PPAR ␣ activation, we tested the effects of different PPAR ␣ agonists on VLDLR expression in liver cells. Our data demonstrated that in the absence of activator, knockdown of PPAR ␣ alone reduced the VLDLR expression level. Overexpression of PPAR ␣ by transfecting the cells with cDNA upregulated VLDLR levels. Compared with the cells' absence of PPAR ␣ agonists, the presence of fenofi brate or GW54047 remarkably increased VLDLR levels ( Fig. 5A, B ), was only slightly reduced by fenofi brate (n = 5, P < 0.05) ( Fig. 4D ).
To further confi rm that upregulating liver VLDLR plays an essential role in decreasing TG levels, we overexpressed liver VLDLR by delivery of AAV-mVLDLr into Vldlr Ϫ / Ϫ mice and then fed the mice with a HFD. As shown in Fig.  4E, F , feeding a HFD in Vldlr Ϫ / Ϫ mice injected with AAV-GFP induced a 2.3-fold increase in total TG and a 3.9-fold increase in true TG, whereas overexpression of mVLDLR decreased total TG by 34% and true TG by 60.7 %. These results indicate that upregulating liver VLDLR plays an essential role in TG reduction.
To further confi rm that PPAR ␣ is required for fenofibrate upregulation of VLDLR expression, we examined the VLDLR expression in primary cultured mouse liver cells . As shown in Fig. 5C , the activation of PPAR ␣ with fenofi brate enhanced the expression of VLDLR in wt hepatic cells, whereas blocking PPAR ␣ activation by GW6471 returned the VLDLR expression to basal levels ( Fig. 5C, D ). However, in liver cells lacking PPAR ␣ , neither fenofi brate nor GW6471 altered the VLDLR levels ( Fig. 5E, F ), indicating the essential role of PPAR ␣ in fenofi brate-regulated liver VLDLR expression.
To further confi rm the essential role of PPAR ␣ in the hepatic VLDLR activity in vivo, and thus to determine the possible function of fenofi brate for the treatment of dyslipidemia, we examined the effects of oral fenofi brate on hepatic VLDLR and serum TG levels in Ppar ␣ Ϫ / Ϫ mice.
Neither a HFD nor fenofi brate affected the liver VLDLR levels in Ppar ␣ Ϫ / Ϫ mice ( Fig. 5G, H ). Additionally, we measured the liver VLDLR mRNA. Unlike the induction effect of fenofi brate on liver VLDLR mRNA levels, fenofi brate had no effect on the liver VLDLR mRNA levels ( Fig. 5I ). Feeding a HFD resulted in a signifi cant elevation of serum TG in Ppar ␣ Ϫ / Ϫ mice, but this increase was not altered by the administration of fenofi brate ( Fig. 5J ). Moreover, there were no signifi cant changes in body weight among Ppar ␣ Ϫ / Ϫ mice fed a regular diet, a HFD, or a HFD with fenofi brate ( Fig. 5K ). These results confirm that the fenofibrateregulated increase in hepatic VLDLR is a PPAR ␣ -dependent mechanism.
Fenofi brate upregulated transcription-enhancing activity of VLDLR by activating PPAR ␣ binding with PPRE in the VLDLR promoter region Next, to further confi rm that fenofi brate upregulates liver VLDLR through PPAR ␣ -regulated increases in VLDLR transcriptional activity, we measured mouse VLDLR promoter activity using a luciferase reporter assay. This assay was performed in the presence or absence of fenofi brate in HepG2 cells in which the PPAR ␣ was either activated or inactivated. As shown in Fig. 6A , activation of PPAR ␣ signifi cantly increased the transcriptional activity of the VLDLR promoter, whereas inactivation of PPAR ␣ significantly lowered the transcriptional activity of the VLDLR promoter, indicating that fenofi brate increases mouse VLDLR promoter transcriptional activity through PPAR ␣ activation.
The function of PPAR ␣ is activated by ligand-binding. Because fenofi brate was shown to increase VLDLR mRNA and VLDLR promoter activity through PPAR ␣ activation, we studied whether fenofi brate directly enhances VLDLR gene transcription via a putative PPRE in the VLDLR promoter region. A region at Ϫ 2307 to Ϫ 2288 bp PPRE motif in the mouse VLDLR promoter region was investigated to determine whether activated PPAR ␣ directly binds to this PPRE motif . The nuclear extracts from wt and Ppar ␣ Ϫ / Ϫ mice fed with fenofi brate were precipitated with control IgG or PPAR ␣ antibody prior to the addition of wt or mutated FITC-labeled PPRE motif probes. As depicted in Fig. 6B , the binding activity in wt with wt PPRE motif probes was 30-fold higher than in the control samples precipitated without the PPAR ␣ antibody, 8-fold higher than in the samples treated with a probe containing a mutant PPRE motif, and 8-fold higher than the activity in Ppar ␣ Ϫ / Ϫ mice. These results indicate that endogenous PPAR ␣ specifi cally binds to the PPRE motif in the mouse VLDLR promoter region and that the activation of PPAR ␣ by fenofi brate directly enhances VLDLR gene transcription via binding with the PPRE motif. To further investigate whether fenofi brate regulates the endogenous VLDLR gene transcription, we performed a ChIP assay using a primer set spanning the PPRE motif in the VLDLR promoter . As shown in Fig. 6C , treatment with fenofi brate in wt mice, but not in Ppar ␣ Ϫ / Ϫ mice, increased VLDLR association with PPRE.

DISCUSSION
Fenofi brate has been shown to lower TG-rich VLDL through the activation of LPL and to reduce the synthesis of TG-rich particles. However, the association between fenofi brate and liver VLDLR levels has not been reported previously. For the fi rst time, our study demonstrated that PPAR ␣ activation substantially upregulates liver VLDLR, but not the VLDLR in adipose, heart, or skeletal muscle under various conditions, including in dyslipidemic and diabetic animal models. Our results also suggest that fenofi brate lowers TG at least partially through the upregulation of liver VLDLR. We also characterized the mechanism of the PPAR ␣ upregulation of VLDLR through binding with PPRE at the VLDLR promoter region. Our fi ndings provide a novel mechanism of fenofi brate regulation of lipid metabolism in metabolic syndrome and diabetes.
VLDLR is believed to deliver VLDL-derived FAs to the peripheral tissues. FAs are primarily utilized by the muscles and heart, whereas adipose tissue acts as a reservoir for FA storage. Thus, VLDLR is most abundantly expressed in the heart, muscle, and adipose tissues, and VLDLR is located on the endothelial surface and the smooth muscle cells of vessels ( 24 ). Although the liver is one of the most important organs for lipid metabolism, the exact role of hepatic VLDLR remains largely unknown, possibly because of the absence or low levels of VLDLR in the liver. Fenofi brate effectively upregulated hepatic VLDLR levels in both diabetic animals and mice with HFD-induced hyperlipidemia; this conclusion is consistent with the observation that liver VLDLR is downregulated in PPAR ␤ / ␦ Ϫ / Ϫ mice in the fed and fasted state ( 25 ). VLDLR upregulation was correlated with a reduction in serum TG. A lack of VLDLR signifi cantly blunted the TG-lowering effect of fenofi brate in mice with HFD-induced hyperlipidemia, confi rming that the upregulation of liver VLDLR is essential and necessary for VLDL-TG metabolism. This fi nding is in agreement with the results of a study by Tacken et al. ( 20 ), which showed that the overexpression of VLDLR was associated with decreased VLDL-rich TG levels in VLDLR steatosis in rats receiving orotic acid and in mice with spontaneous hepatic steatosis ( 26,27 ). These studies support the theory that fenofi brate-regulated increases in the expression of liver VLDLR might not cause excess TG uptake but will improve TG clearance in the liver. Although some consequences of the upregulation of liver VLDLR are already well-described, further and more advanced studies investigating the mechanism of action of liver VLDLR are warranted. transgenic mice, whereas the absence of VLDLR was associated with increased TG levels. Thus, VLDLR may be one of the major mediators for fenofi brate-regulated TGrich VLDL clearance. As a result, the fenofi brate-regulated decrease in TG levels is most likely caused by an increase in TG clearance by VLDLR in the liver. However, increases in liver VLDLR may also potentially cause increased lipid uptake by the liver. Studies by Ferreira and Harano revealed that fenofi brate prevents hepatic Fig. 6. PPAR ␣ activation upregulation of VLDLR transcriptional activity through binding to PPRE. A: 293T cells were cotransfected with a vector containing VLDLR promoter and PRL-TK to normalize the transfection effi ciency. Twenty-four hours later, after achieving quiescence in a freshly prepared medium, the transfected cells were then exposed to a cultured medium in the presence or absence of fenofi brate (Feno) (50 M) with or without 10 M GW6471 for 16 h; the same amount of vehicle served as control. The transcriptional activity of the VLDLR promoter was measured by luciferase assay and normalized by Renilla luciferase activity (mean ± SD, n = 3; * P < 0.05). B: Double-stranded oligonucleotides corresponding to the mouse VLDLR PPRE and mutant PPRE were incubated with the nuclear extract from the mouse tissues and precipitated with or without anti-mouse PPAR ␣ antibody. Lane 1, FITC-labeled VLDLR-PPRE oligonucleotides with nuclear extract from wt mice without being precipitated with PPAR ␣ antibody; lane 2, FITC-labeled VLDLR-PPRE oligonucleotides with PPAR ␣ antibody precipitated nuclear extract from wt mice; lane 3, FITC-labeled VLDLR-PPRE mutant oligonucleotides with PPAR ␣ antibody precipitated nuclear extract from wt mice; lanes 4 and 5, FITC-labeled oligonucleotides of VLDLR-PPRE (lane 4) and VLDLR-PPRE mutant (lane 5) with PPAR ␣ antibody precipitated nuclear extract from PPAR ␣ Ϫ / Ϫ mice. C: ChIP analysis of the association of PPAR ␣ with the VLDLR promoter in the wt mice and Ppar a Ϫ / Ϫ mice treated with fenofi brate using primers located between ‫ف‬ 10 bp upstream and downstream of putative PPRE in the VLDLR promoter region, and control with primers located ‫ف‬ 2,000 bp upstream of the PPRE in the VLDLR promoter region .
showed that a HFD reduced hepatic LPL levels, whereas treatment with fenofi brate increased LPL levels. In mice with defi cient VLDLR levels, neither a HFD nor fenofibrate affected hepatic LPL levels (data not shown). Moreover, despite a certain amount of hepatic LPL activity in Vldlr Ϫ / Ϫ mice, fenofi brate lost its TG lowering effects in these animals. These fi ndings are in accordance with the results that the presence of VLDLR is essential for LPL regulation. In contrast, this study did not fi nd any evidence that fenofi brate had similar regulation effects on adipose LPL levels. Studies conducted by Schoonjans and colleagues demonstrated that PPAR ␣ agonists inhibit cardiac LPL activity ( 10 ) and that the response of LPL to a PPAR ␣ agonist is tissue-specifi c ( 11 ). It is unclear whether the variance of LPL levels in liver and adipose tissues is the same as the variance in VLDLR levels.
In conclusion, this study resulted in a novel fi nding that hepatic VLDLR upregulation plays an essential role in the TG lowering effect of fenofi brate via the activation of PPAR ␣ . The potential mechanisms linking VLDL metabolism and fenofi brate are multiple, overlapping, and highly correlated. Considerable interest should be raised regarding the tissue-specifi c effects of PPAR ␣ agonists on genes involved in VLDL metabolism and the precise mechanisms underlying the different functions of VLDLR in the liver and in other tissues.
Interestingly, unlike the fact that fenofi brate increases liver VLDLR, VLDLR upregulation does not occur in adipose tissues or adipocyte cells after fenofi brate treatment, suggesting that fenofi brate-regulated increases in liver VLDLR are tissue-specifi c responses. This difference is most likely due to the various expressional patterns of PPAR ␣ in the liver and the adipose tissues. The PPAR ␣ levels are extremely high in the liver but nearly absent in adipose tissues. Treatment with fenofi brate upregulates the liver PPAR ␣ level but has no effect on adipose PPAR ␣ , which is in accordance with the effect of fenofi brate on VLDLR levels. Functional differences may also be responsible for the differences between VLDLR expression in the liver and the peripheral tissues like adipose tissues. Immunostaining of VLDLR in naturally expressed sites, such as adipose tissue, heart, and muscle, revealed intense staining of VLDLR in the capillaries and arterioles but not in the tissues lacking endothelial cells. This fi nding indicates that the role of VLDLR is the transport of VLDL and the plasma constituent from the vascular compartment to adjacent tissues ( 28 ). A similar study conducted by Wyne et al. ( 28 ) showed that the expression of VLDLR on the endothelial lining of the venules in the liver is generally very low, which may explain the low liver VLDLR levels. In the naturally expressed peripheral tissues, the study by Tacken et al. ( 20 ) showed that VLDLR facilitates clearing TG and mainly affects the entry of VLDL TG into these tissues. Perman et al. ( 29 ) reported that heart VLDLR levels were upregulated in ischemic myocardial diseases and were associated with TG accumulation in cardiomyocytes. In the present study, we found that a HFD increased adipose VLDLR levels and body weight, whereas fenofi brate reduced adipose VLDLR levels and body weight. These results all indicate that the function of VLDLR may differ by tissue type. In tissues where VLDLR naturally occurs, such as adipose tissue, macrophage cells, and cardiac tissue, TG clearance is associated with TG uptake through VLDLR into peripheral tissues. Although our study found that activation of PPAR ␣ has no effect on the VLDLR in the peripheral tissue such as adipose tissues, several studies have reported that adipose VLDLR gene expression can be upregulated by PPAR ␥ ( 30-32 ). Interestingly, unlike the low level of PPAR ␥ in the liver and high level of PPAR ␥ in the adipose tissues, PPAR ␣ is extremely abundant in the liver but very low in the adipose tissues. Whether the difference of PPAR ␥ and PPAR ␣ in tissue distribution represents their different impacts on VLDLR functions requires more in-depth study.
We also investigated the relationships among PPAR ␣ , LPL, and VLDLR. LPL is a TG hydrolase and a ligand/ bridging factor for receptor-mediated VLDL uptake ( 33,34 ). Like VLDLR, LPL encoded by LPL is abundantly expressed in endothelial cells ( 34 ). A study done by Yagyu et al. ( 35 ) indicates that the presence of VLDLR is required for normal LPL regulation and that the disruption of VLDLR reduces LPL activity. However, Hu et al. ( 36 ) found that in the absence of VLDLR, LDLR, and apoE, three major apolipoprotein recognizing receptors, hepatic VLDL uptake is regulated by LPL. Our present study