Induction of the phospholipid transfer protein gene accounts for the high density lipoprotein enlargement in mice treated with fenofibrate.

Fibrate treatment in mice is known to modulate high density lipoprotein (HDL) metabolism by regulating apolipoprotein (apo)AI and apoAII gene expression. In addition to alterations in plasma HDL levels, fibrates induce the emergence of large, cholesteryl ester-rich HDL in treated transgenic mice expressing human apoAI (HuAITg). The mechanisms of these changes may not be restricted to the modulation of apolipoprotein gene expression, and the aim of the present study was to determine whether the expression of factors known to affect HDL metabolism (i.e. phospholipid transfer protein (PLTP), lecithin:cholesterol acyltransferase, and hepatic lipase) are modified in fenofibrate-treated mice. Significant rises in plasma PLTP activity were observed after 2 weeks of fenofibrate treatment in both wild-type and HuAITg mice. Simultaneously, hepatic PLTP mRNA levels increased in a dose-dependent fashion. In contrast to PLTP, lecithin:cholesterol acyltransferase mRNA levels in HuAITg mice were not significantly modified by fenofibrate despite a significant decrease in plasma cholesterol esterification activity. Fenofibrate did not induce any change in hepatic lipase activity. Fenofibrate significantly increased HDL size, an effect that was more pronounced in HuAITg mice than in wild-type mice. This effect in wild-type mice was completely abolished in PLTP-deficient mice. Finally, fenofibrate treatment did not influence PLTP activity or hepatic mRNA in peroxisome proliferator-activated receptor-alpha-deficient mice. It is concluded that 1) fenofibrate treatment increases plasma phospholipid transfer activity as the result of up-regulation of PLTP gene expression through a peroxisome proliferator-activated receptor-alpha-dependent mechanism, and 2) increased plasma PLTP levels account for the marked enlargement of HDL in fenofibrate-treated mice.

a number of epidemiological studies. This finding has generated much interest in elucidating HDL structure, function, and metabolism. To date, the structural heterogeneity of HDL is well established especially in terms of size distribution. This heterogeneity may be of physiopathological relevance with HDL size distribution being more strongly related with the presence of coronary heart disease than absolute HDL levels (1). Several factors have been reported to influence HDL particles such as lecithin:cholesterol acyl transferase (LCAT) (2), hepatic lipase (HL) (3), cholesteryl ester transfer protein (4), and phospholipid transfer protein (PLTP) (5).
Interestingly, PLTP by facilitating surface phospholipid transfer between lipoproteins (6) favors the continuous remodeling of HDL in the blood stream. As shown by in vivo and in vitro studies, PLTP can modulate the levels, the size, and the composition of HDL (7)(8)(9), and incubation of HDL with PLTP induced the emergence of both large ␣HDL and small pre-␤HDL (10). In support of the physiological relevance of the latter observations, increases in ␣HDL and pre-␤HDL were observed in human PLTP and apoAI transgenic mice (11).
Fibrates constitute a class of frequently used normolipidemic drugs that efficiently decrease triglycerides and increase HDL plasma levels in humans (12,13). They exert their effect, at least in part, through alterations in transcription of genes encoding for proteins that control lipoprotein metabolism (14). Indeed, fibrates activate specific transcription factors termed peroxisome proliferator-activated receptors (PPARs), a subfamily of the nuclear receptors (15). Particularly, the PPAR␣ form mediates fibrate action on HDL cholesterol levels via transcriptional induction of the synthesis of the major HDL apolipoproteins, i.e. apoAI and apoAII (16,17).
In human apoAI transgenic (HuAITg) mice, treatment with fenofibrate led to large increases in human apoAI and HDL cholesterol and large qualitative modifications of HDL with the appearance of large and cholesteryl ester-enriched ␣HDL (16). The mechanisms of the structural modification of HDL in HuA-ITg mice treated with fenofibrate are, however, not yet understood. Because HDL structure and metabolism partly depend on a few plasma enzyme and lipid transfer activities, we assumed that fenofibrate may modulate these activities, which would account for the changes in HDL structure. The goals of the present study were to evaluate the effect of fenofibrate on the expression and activity of PLTP, LCAT, and HL. In a second part of the study, the relative contribution of PLTP to the fenofibrate-mediated alterations in HDL size distribution was specifically addressed in PLTP-deficient mice. Finally, be-cause PPAR␣ has been shown to mediate the effects of fenofibrate on several genes, we sought to test the hypothesis that PPAR␣ is involved in the observed changes using PPAR␣deficient mice.

Animals
Mice used in this study were nontransgenic C57BL/6 mice, heterozygous HuAITg mice, PLTP-deficient mice (PLTPϪ/Ϫ) and their wild-type controls (PLTPϩ/ϩ), and homozygous PPAR␣-deficient mice (PPAR␣Ϫ/Ϫ) and their wild-type controls (PPAR␣ϩ/ϩ). HuAITg mice contained 21 copies of an 11-kilobase pair human apoAI genomic DNA fragment and the liver-specific enhancer of the human apoAI gene promoter necessary to drive hepatic apoAI expression (18). PLTPϪ/Ϫ were obtained by deletion of part of the PLTP gene containing exon 2. Because this exon contains the translation initiation codon and the signal peptide, mice were null PLTP animals (19). PPAR␣Ϫ/Ϫ mice were obtained by disrupting the ligand-binding domain of the PPAR␣ by homologous recombination (20,21). 10 -12-week-old mice were caged in an animal room with alternating 12-h light (7 a.m. to 7 p.m.) and dark (7 p.m. to 7 a.m.) cycles and were fed either the control or fenofibrate (Sigma) diet ad libitum for 2 weeks. Fenofibrate was mixed in mouse chow at 0.02 and 0.2% (w/diet w), two dosages commonly used in mice (16). Body weight and food intake of the mice were regularly checked during the treatment period. No toxic side effect of fenofibrate treatment as checked by alterations in body weight was observed.

Plasma Lipid and Lipoprotein Analysis
Blood was taken from the retroorbital plexus of the mice after a 2-h fast. Plasma was separated by centrifugation at 3000 rpm for 15 min. Lipoproteins were purified by sequential ultracentrifugation using a TL 100.4 rotor with a tabletop TL 100 centrifuge (Beckman Instruments). HDL was purified between the densities 1.063 and 1.210 g/ml in all animals with the exception of HDL from HuAITg mice treated with the higher dose of fenofibrate (0.2%), which was purified between 1.019 and 1.210 g/ml because analysis of lipoproteins by density gradient ultracentrifugation showed that apoAI-containing lipoproteins fell within these density limits (data not shown).
Lipoprotein size was determined by gradient gel electrophoresis (4 -20% polyacrylamide gels, Novex gel, Invitrogen Corp., Carlsbad, CA) under nonreducing conditions. The migration was carried out for 12 h, and the gels were stained with Coomassie Brilliant Blue R250 and destained with a solution of methanol/acetic acid/water (30:58:12, v/v/ v). The gels were then scanned with a densitometer (GS-300 scanning densitometer, Hoefer Scientific Instruments, San Francisco, CA). The size of lipoproteins was determined from a curve built from standards of known size (high molecular weight standards, Amersham Pharmacia Biotech) that were migrated simultaneously.

Enzymatic Assays of PLTP, LCAT, and HL Activities
Plasma PLTP Activity-PLTP activity was measured in total plasma as the transfer of radiolabeled phosphatidylcholine from [ 14 C]phosphatidylcholine liposomes to an excess of exogenous human HDL 3 (22). Briefly, 10 mol of [ 14 C]L-␣-phosphatidylcholine (PerkinElmer Life Sciences) and 0.1 mol of butylated hydroxytoluene were mixed, and the lipids were dried under nitrogen, suspended in 1 ml of 150 mM NaCl, 10 mM Tris-HCl, 1 mM EDTA (pH 7.4), sonicated, and centrifuged to remove lipid aggregates and metallic debris. Plasma samples were incubated for 1 h at 37°C with radiolabeled liposomes (50 nmol of phospholipid) and isolated human HDL 3 (250 g of protein) in a final volume of 400 l with 150 mM NaCl, 10 mM Tris-HCl, 1 mM EDTA (pH 7.4). At the end of the incubation, liposomes were precipitated with dextran sulfate. The supernatant (0.5 ml) was used for radioactivity determinations. Plasma volumes were chosen to keep PLTP activity in the linear range of the assay. PLTP activity was expressed as micromoles of phosphatidylcholine transferred from liposomes to exogenous human HDL 3 per milliliter of plasma.
LCAT Activity-LCAT activity was determined using an exogenous substrate according to the method described by Chen and Albers (23). Mouse plasma (50 l) was incubated for 1 h at 37°C with a proteoliposome substrate containing apoAI, [ 14 C]cholesterol, and 1-palmitoyl-2oleoyl-sn-glycero-3-phosphocholine. After incubation, lipids were extracted with chloroform/methanol (2:1, v/v) and separated by thin-layer chromatography on silica gel plates using a solvent system composed of petroleum ether/diethyl ether/acetic acid (90:10:5, v/v/v). Cholesteryl ester and free cholesterol radioactivity were measured on a Phospho-rImager. LCAT activity was expressed as the percentage of free cholesterol esterified during incubation.
Hepatic Lipase Activity-HL activity was measured according to the method described by Henderson et al. (24) in postheparin plasma. Postheparin plasma was obtained 5 min after an intravenous bolus injection of 60 IU of sodium heparin/kg of body weight. Blood was immediately centrifuged at 4°C and frozen at Ϫ80°C until assay.
The assay mixture for HL was prepared with 5 Ci of tri-[1-14 C]oleoylglycerol, 5.68 ml of 90 g/liter Arabic gum in 0.2 M Tris-HCl, pH 8.5, 100 mg of unlabeled triolein, and 2.75 ml of 200 g/liter bovine serum albumin. After sonication, 2.58 ml of 3.24 M NaCl in 0.2 M Tris buffer, pH 8.5, was added to the mixture. Triglyceride hydrolysis was started by adding postheparin plasma and allowed to proceed for 1 h at 37°C. Triglyceride hydrolysis was stopped by adding 3.25 ml of chloroform/ methanol/heptane (50:56:40, v/v/v) and 1 ml of 0.1 M carbonate borate buffer, pH 10.5. Reaction tubes were shaken for 5 min and centrifuged. A 0.5-ml aliquot of the upper phase was removed and counted in a scintillation counter. After correction for recovery, results were expressed in micromoles of fatty acids liberated per milliliter of plasma per hour.

RNA Analysis
Total cellular RNA was extracted using the guanidinium thiocyanate/phenol/chloroform method (25). Northern blot hybridizations were performed as previously described (26) using the cDNA fragments of LCAT (27) and PLTP and 28 S rRNA as probes. PLTP cDNA was amplified by reverse transcription-polymerase chain reaction using the following primers: 5Ј-TCG GCG GAG GGT GTG TCC AT-3Ј and 5Ј-CAT FIG. 1. Size distribution of HDL in HuAITg mice before and after treatment with fenofibrate. Animals were treated for 15 days with fenofibrate (0.02 and 0.2%), and plasma was collected. Lipoproteins were isolated by sequential ultracentrifugation and separated by polyacrylamide gradient gel electrophoresis (4 -20%). Gels were stained with Coomassie Brilliant Blue R-250. Profiles were obtained with a densitometer (GS-300 scanning densitometer), and the migration distance of each absorption maximum (corresponding to a stained band) was determined. The molecular diameter corresponding to each of the peaks was calculated from a calibration curve using standards of known diameters. AU, absorbance units; FF, fenofibrate.
GGC AGA GTC AAA GAA GA-3Ј (fragment size 477 base pairs). Quantification was performed by phosphorimaging with the Molecular Analyst TM software (Bio-Rad). The relative levels of PLTP and LCAT were determined by normalizing to the level of 28 S rRNA.

Effect of Fenofibrate on HDL Size Distribution in HuAITg
Mice-As expected (16), fenofibrate treatment induced marked alterations in the plasma HDL profile in HuAITg mice. As measured by native polyacrylamide gradient gel electrophoresis, the mean apparent diameter of the most prominent HDL subpopulation rose from 11.7 nm in untreated HuAITg mice toward 15.0 and 18.0 nm in animals treated with 0.02 and 0.2% fenofibrate, respectively (Fig. 1).
PLTP, LCAT, and HL Activities in HuAITg Mice-To determine whether the modifications observed in HDL physicochemical characteristics induced by fenofibrate were linked to those of activities of lipoprotein modifying enzymes, we measured PLTP, LCAT, and HL in the plasma of HuAITg mice. PLTP activity measured in treated and nontreated mice is shown in Fig. 2a. Fenofibrate significantly increased PLTP activity in a dose-dependent fashion from 7 mol/ml/h in nontreated mice to 22 and 37 mol/ml/h in mice fed with 0.02 and 0.2% fenofibrate, respectively.
LCAT activity was measured in the same animals using an exogenous substrate. The low dose of fenofibrate did not significantly modify LCAT activity as compared with controls, whereas enzyme activity was 3-fold lower with the high dose of fenofibrate as compared with the nontreated animals (Fig. 2a). HL activity measured in postheparin plasma was not modified by fenofibrate treatment (results not shown).
Effect of Fenofibrate on PLTP and LCAT mRNA Levels in Livers of Human ApoAI Transgenic Mice-Phospholipid transfer and cholesterol esterification activities measured in the present study reflect plasma PLTP and LCAT concentrations, respectively. We thus checked whether the modifications of the enzyme plasma levels induced by fenofibrate were due to changes in their gene expression levels. As shown in Fig. 2b, the levels of hepatic PLTP mRNA significantly rose in a dosedependent manner in fenofibrate-treated HuAITg mice with approximately 4-and 10-fold increases in PLTP mRNA levels observed in animals treated with 0.02 or 0.2% fenofibrate, respectively. Unlike for the PLTP gene, fenofibrate did not produce significant changes in the levels of LCAT mRNA in the livers of HuAITg treated mice (Fig. 2b).
Effect of Fenofibrate on Plasma Lipoproteins in Nontransgenic Mice-Given that the expression of both human apoAI and murine PLTP genes are increased concomitantly in HuA-ITg animals, the contribution of PLTP induction to the fenofibrate-induced changes in HDL size was assessed using nontransgenic mice. In nontreated wild-type C57BL/6 mice, the analysis of HDL size showed a monodisperse population of HDL with a mean apparent diameter of 11.0 nm (Fig. 3). As observed in HuAITg mice (see above), fenofibrate treatment of wild-type mice produced a shift of plasma HDL toward the large size range. One new discrete subfraction of 12.9-nm diameter appeared in mice treated with 0.02% fenofibrate, and FIG. 2. Effect of fenofibrate on plasma (a) and liver mRNA levels (b) of PLTP and LCAT activities in HuA-ITg mice. PLTP activity was determined by incubating total plasma with [ 14 C]phosphatidylcholine liposomes and exogenous human HDL 3 , and LCAT activity was determined by incubating total plasma with apoAI proteoliposomes containing [ 14 C]cholesterol as described under "Materials and Methods." Total RNA was extracted from liver, and PLTP and LCAT mRNA were quantified in control and treated mice by Northern blot analysis using specific probes. Bars represent the mean Ϯ S.D. of five animals (*, p Ͻ 0.05; ***, p Ͻ 0.001). R.A.U., mRNA arbitrary unit.
Effect of Fenofibrate on PLTP, LCAT, and HL Nontransgenic Mice-In nontransgenic C57BL/6 mice, fenofibrate significantly increased PLTP activity in a dose-dependent fashion from 5 mol/ml/h in nontreated mice to 15 and 25 mol/ml/h in mice fed with the 0.02 and 0.2% doses of fenofibrate, respectively (Fig. 4a). LCAT activity decreased significantly in C57BL/6 mice treated with 0.02% fenofibrate and even to a greater extent at the highest dose tested (0.2%) (Fig. 4a). HL was not significantly modified by treatment with fenofibrate (data not shown).
As shown in Fig. 4b, the levels of PLTP mRNA in mouse liver rose significantly in a dose-dependent manner in fenofibratetreated C57BL6 mice as compared with nontreated controls with approximately 7-and 10-fold increases at the 0.02 and 0.2% doses, respectively (Fig. 4b). Unlike for the PLTP gene, the levels of LCAT mRNA in the liver from C57BL6 mice were only marginally altered after treatment with fenofibrate with approximately 20 and 30% decreases at the 0.02 and 0.2% doses, respectively (Fig. 4b).
Effect of Fenofibrate on HDL Size Distribution in PLTPdeficient Mice-Because fenofibrate exerts a marked effect on the expression of PLTP in both HuAITg and wild-type mice, the relative contribution of PLTP to the observed alterations in HDL structure was determined by comparing the effect of fenofibrate treatment on HDL size distribution in wild-type and PLTP-deficient mice. In agreement with data presented above, treatment of wild-type mice with 0.2% fenofibrate pro-duced a marked 2-fold rise in plasma PLTP activity (data not shown). As expected (19), PLTP activity was undetectable in homozygous PLTP-deficient mice whether they were treated or not with fenofibrate (data not shown). HDL size was measured in both wild-type and PLTP-deficient mice (n ϭ 4/group). A representative profile of an animal of each group is shown in Fig. 5. In treated wild-type mice, two types of HDL of 12.8 Ϯ 0.5 and 11.4 Ϯ 0.1 nm were observed, whereas HDL was monodisperse in nontreated wild-type mice (11.0 Ϯ 0.5 nm). In contrast, fenofibrate treatment did not produce any detectable effect on the size distribution of plasma HDL (9.7 Ϯ 0.6 and 9.3 Ϯ 0.4 nm in nontreated versus treated PLTP-deficient mice).
Effects of Fenofibrate in PPAR␣-deficient Mice-Because fenofibrate has been shown to exert its effect by modifying the expression of a number of genes via activation of the nuclear receptor PPAR␣, we examined whether the observed change in PLTP was dependent on PPAR␣ activation by carrying out experiments in PPAR␣-deficient mice. Wild-type and PPAR␣deficient mice were treated with the same dose of fenofibrate (0.2%). Treatment of PPAR␣ϩ/ϩ mice with fenofibrate increased PLTP activity from 4 to 13 mol/ml/h, whereas no increase in PLTP activity was observed in treated PPAR␣Ϫ/Ϫ mice (Fig. 6a). As shown in Fig. 6b, basal PLTP mRNA levels were comparable in wild-type and PPAR␣Ϫ/Ϫ mice. By contrast, fenofibrate treatment induced a significant 4-fold increase in PLTP mRNA levels in livers of wild-type mice, whereas no changes in PLTP mRNA levels were observed in PPAR␣Ϫ/Ϫ animals. DISCUSSION HDL structure and metabolism are modulated by fibric acid derivatives in humans as well as in rodents. Whereas most attention has been paid to the effects of fibrates on apoAI and apoAII, their effects on the expression of other proteins involved in HDL metabolism have not been studied previously. LCAT is associated with HDL in that it generates cholesteryl esters and lysophospholipids with phospholipids and free cholesterol as substrates, driving as such the further influx of free cholesterol into HDL. As a result of the LCAT-mediated influx of core lipids, HDL tends to enlarge (28). Conversely, shrinkage of HDL as the consequence of phospholipid and triglyceride hydrolysis by HL leads to the emergence of small HDL (29). PLTP by facilitating phospholipid transfer between lipoproteins plays an important role in HDL remodeling, favoring the appearance of both large and lipid-poor apoAI-containing particles (9). Thus, plasma HDL metabolism in mice (i.e. a cholesteryl ester transfer protein-deficient species) is mainly dependent on LCAT, HL, and PLTP activities. Given that HDL was previously shown to undergo significant structural changes as the result of fenofibrate treatment in HuAITg mice (16), we postulated that one or several enzyme/transfer activities involved in HDL metabolism are modulated by fenofibrate. Therefore, the aim of the present study was to search for potential alterations in LCAT, PLTP, and HL in several mouse lines treated or not with fenofibrate.
The most striking finding of the present study is the marked dose-dependent increase in PLTP activity and hepatic mRNA in HuAITg and wild-type mice treated with fenofibrate. The increase in HDL size was the initial finding associated with fenofibrate treatment in HuAITg mice (16), and consistent observations were made in the present study in both HuAITg and nontransgenic mice (Figs. 1 and 3). Based on previous in vivo and in vitro observations, the increase in PLTP activity due to fenofibrate would predictably have resulted in increased ␣HDL size. In further support of the latter view, incubation of purified PLTP with isolated HDL induced the appearance of a population of larger HDL particles (8), and a positive relation- ship between PLTP activity and HDL size was observed in different inbred mouse strains (30). However, in vivo studies in genetically engineered mice yielded somewhat controversial data with regard to the role of PLTP in determining the size distribution of ␣HDL (11,19,(31)(32)(33). In particular, an increase in PLTP activity in human PLTP transgenic mice or adenovirus-infected mice was associated with either no marked changes or a net increase in mean ␣HDL size (11,(31)(32)(33)(34). This, heterogeneity of experimental data concerning the role of PLTP in HDL structure might relate at least in part to marked differences in the plasma levels and tissue expression of PLTP in the distinct mouse models (11,19,(31)(32)(33). In this respect, stimulation of PLTP expression by fenofibrate in wild-type as well as in HuAITg mice indicates that pharmacological modification of PLTP expression results in pronounced alterations in HDL structure and metabolism.
Although in the present studies fenofibrate treatment led to increases in HDL size in both HuAITg and nontransgenic mice, clear differences in HDL size distribution appeared between the two mouse lines (Figs. 1 and 3). In HuAITg mice treated with the high dose of fenofibrate, very large HDL, almost of human LDL size, was observed much similar to those recently described in HuAITg mice strongly expressing human PLTP after receiving an adenovirus infection (32). In nontransgenic wild-type mice, fenofibrate led to weaker changes in HDL size distribution. Interestingly and as already observed (11,32), PLTP activity is clearly higher in HuAITg mice than in nontransgenic animals. Thus, higher PLTP activity in HuAITg mice could account for the greater increase in HDL size as compared with control animals (11,32). The amounts of PLTP in HuAITg versus nontransgenic mice may be due to the fact that PLTP displays a higher affinity for large HDL than for small HDL (35). Thus, even though fenofibrate exerts a similar effect on PLTP gene expression in both HuAITg and nontransgenic mice, the overexpression of human apoAI would clearly emphasize the PLTP-mediated HDL changes.
Fenofibrate decreased plasma LCAT activity, but unlike PLTP gene expression LCAT gene expression was only marginally reduced in wild-type mice, and no change in LCAT mRNA levels was observed in HuAITg mice. Because LCAT activity values in the present study mainly reflect LCAT mass concentration, the results indicate that fenofibrate significantly decreased the concentration of LCAT in the plasma of fenofibratetreated animals. Interestingly, decreased LCAT activity has already been noted in mice overexpressing PLTP (32), and it is conceivable that LCAT has a lower affinity for large HDL than for small particles. Although to our knowledge no study has directly examined this point in mice, it is noteworthy that the shift in HDL size toward small HDL in mice transgenic for both human apoAI and cholesteryl ester transfer protein as compared with HuAITg mice was associated with a concomitant increase in LCAT mass (36). All together, these data suggest therefore that a substantial increase in the mean HDL size would secondarily result in lower plasma LCAT levels because of a weaker lipoprotein affinity (and thus a faster catabolism) of the enzyme. In other words and in contrast to increased plasma PLTP levels, decreased plasma LCAT levels would be the consequence and not the cause of the HDL enlargement in fenofibrate-treated mice.
In direct support of a key role of PLTP in the HDL enlargement in treated mice, the present study demonstrated for the first time that the HDL size redistribution induced by fenofibrate is completely abolished in homozygous PLTP-deficient animals. Although fibric acid derivatives are known to affect mouse lipoprotein metabolism in several distinct ways, in particular through the regulation of the expression of a number of proteins, the lack of HDL size redistribution in fenofibratetreated PLTP-deficient mice clearly indicates that PLTP overexpression is the most important if not the sole mechanism that accounts for the observed HDL enlargement. Nevertheless and as demonstrated in the present studies in HuAITg mice, the impact of PLTP on HDL size can be markedly modulated by the level of apolipoprotein gene expression.
Because fibrates exert their effects through activation of the transcription factor PPAR␣ (15), it has been postulated that the increase in hepatic PLTP mRNA was mediated through a PPAR␣-dependent mechanism. It can be concluded from the present study that fenofibrate affects PLTP gene expression via the activation of the nuclear receptor PPAR␣. Several other genes involved in lipoprotein metabolism are transcriptionally induced by PPAR␣ activation. Indeed, fibrates activate PPAR␣, which after heterodimerization with another nuclear receptor, the retinoid X receptor, interacts with specific response elements termed peroxisome proliferator response elements in the regulatory sequences of target genes. The binding of PPAR␣retinoid X receptor heterodimer to a peroxisome proliferator response element leads to altered gene expression (37). It can, therefore, be suggested that the promoter of the mouse hepatic PLTP gene contains a functional peroxisome proliferator response element, but a direct demonstration of its presence remains to be done.
In conclusion, results of the present study established for the first time that fenofibrate induces the expression of murine PLTP in liver through a PPAR␣-dependent mechanism. The resulting rise in plasma PLTP activity was shown to account for the HDL size enlargement in plasma from treated animals. Finally, the absence of HDL size changes in PLTP-deficient mice clearly demonstrated that PLTP is a key factor in determining the size distribution of ␣HDL in vivo.