Expression of the peroxisome proliferator-activated receptor alpha gene is stimulated by stress and follows a diurnal rhythm.

Peroxisome proliferator-activated receptors (PPARs) are nuclear hormone receptors that can be activated by fatty acids and peroxisome proliferators. The PPARα subtype mediates the pleiotropic effects of these activators in liver and regulates several target genes involved in fatty acid catabolism. In primary hepatocytes cultured in vitro, the PPARα gene is regulated at the transcriptional level by glucocorticoids. We investigated if this hormonal regulation also occurs in the whole animal in physiological situations leading to increased plasma corticosterone levels in rats. We show here that an immobilization stress is a potent and rapid stimulator of PPARα expression in liver but not in hippocampus. The injection of the synthetic glucocorticoid dexamethasone into adult rats produces a similar increase in PPARα expression in liver, whereas the administration of the antiglucocorticoid RU 486 inhibits the stress-dependent stimulation. We conclude that glucocorticoids are major mediators of the stress response. Consistent with this hormonal regulation, hepatic PPARα mRNA and protein levels follow a diurnal rhythm, which parallels that of circulating corticosterone. To test the effects of variations in PPARα expression on PPARα target gene activity, high glucocorticoid-dependent PPARα expression was mimicked in cultured primary hepatocytes. Under these conditions, hormonal stimulation of receptor expression synergizes with receptor activation by WY-14,643 to induce the expression of the PPARα target gene acyl-CoA oxidase. Together, these results show that regulation of the PPARα expression levels efficiently modulates PPAR activator signaling and thus may affect downstream metabolic pathways involved in lipid homeostasis.

The peroxisome proliferator-activated receptors (PPARs) 1 are orphan nuclear hormone receptors (1,2). To date, three different subtypes, PPAR␣, PPAR␤ or -␦ (also named FAAR or NUC1), and PPAR␥, have been cloned in amphibians (3), rodents (4 -10), and man (11)(12)(13). Within the nuclear hormone receptor superfamily, PPARs belong to the subfamily that comprises the thyroid hormone receptors and retinoic acid receptors (14). The first PPAR cDNA cloned was isolated from a mouse liver library and corresponds to the PPAR␣ subtype (4). This receptor was shown to be activated by peroxisome proliferators (3,4), a class of compounds that have characteristic pleiotropic effects, especially in rodents. In hepatocytes, peroxisome proliferators cause a dramatic increase in the number and the size of peroxisomes, an effect associated with a concomitant induction of the activity of several enzymes of the peroxisomal ␤-oxidation pathway (15). Definitive proof that PPAR␣ is the major mediator of these effects was provided by the absence of the typical hepatic response to peroxisome proliferators in PPAR␣-deficient mice generated by targeted disruption of the PPAR␣ gene (16). Whereas little is known about the mechanisms that lead to peroxisome proliferation, the associated increase in the level of expression of several genes is better understood. Indeed, several PPAR␣ target genes were identified that contain one or several PPAR-responsive elements in their promoter (1,2). To date, the known PPAR␣ target genes code for enzymes involved in the following metabolic pathways: (i) activation of fatty acids to acyl coenzyme A (CoA) derivatives (acyl-CoA synthetase, Ref. 17), (ii) peroxisomal ␤-oxidation (acyl-CoA oxidase (ACO), Refs. 3, 18, and 19, and bifunctional enzyme, Ref. 20), (iii) mitochondrial ␤-oxidation (medium chain acyl-CoA dehydrogenase, Ref. 21), (iv) microsomal -oxidation (CYP 4A6, Ref. 22) and (v) ketogenesis (hydroxymethylglutaryl-CoA synthase, 23). In addition, the genes coding for apolipoproteins AI (24), AII (25), and CIII (26) are also regulated by PPAR␣, suggesting an involvement of these receptors in the regulation of the extracellular transport of lipids.
Although peroxisome proliferators are now used as prototypical activators of PPAR␣, there is still no evidence that these compounds bind directly to the receptor. The only PPAR ligands identified so far are antidiabetic agents of the thiazolidinedione family, which bind with high affinity to PPAR␥ (27). The discovery that several fatty acids, such as arachidonic acid or linoleic acid, activate PPAR␣ suggests that fatty acids could represent biological activators (5,18). According to this hypothesis, PPAR␣ could function as a fatty acid sensor, which would allow the fatty acids to regulate their own metabolism (1,2).
PPAR␣ mRNA is predominantly expressed in tissues capable of oxidizing fatty acids, such as brown adipose tissue, liver, heart, kidney, and muscle (8). Absence of PPAR␣ expression in knockout mice prevents the inducibility of several target genes in liver, including ACO and bifunctional enzyme, by peroxisome proliferators, suggesting that the level of expression of PPAR␣ is important for the proper regulation of these genes in vivo (16). In fact, PPAR␣ expression in adult rat liver is subject to marked interindividual variations for so far unknown reasons. 2 We and others (28,29) have shown recently that PPAR␣ expression is directly regulated at the transcriptional level by glucocorticoids in rat hepatocytes or hepatoma cell lines cultured in vitro. This regulation is mediated by the glucocorticoid receptor and does not involve stabilization of the mRNA (28). These findings suggested that PPAR␣ expression in liver could be subject to hormonal regulation in vivo in situations of high circulating glucocorticoid levels.
In the present study, we show that PPAR␣ mRNA levels increase in rat liver during an immobilization stress situation and follow a diurnal rhythm. The impact of variations in the levels of PPAR␣ expression on the regulation of a prototypical PPAR␣ target gene, the ACO gene, is furthermore analyzed using hepatocytes in primary culture.

MATERIALS AND METHODS
Animals and Treatment-Eight-week-old male Fisher 344 rats (BRL, Basel, Switzerland) were group-housed and had free access to water and food. The animals were kept on a 12-h light-dark cycle (light from 8:00 a.m. to 8:00 p.m. in the stress and dexamethasone experiments; light from 7:30 a.m. to 7:30 p.m. in the diurnal variation experiment). All animals were accustomed to this cycle during at least 14 days. Rats were stressed by immobilization in transparent plastic tubes (5 cm diameter, 20 cm long) with small holes in the front allowing breathing and a hole at the back for the tail. Immobilization started at 9:00 a.m. The unstressed control animals were not manipulated during the duration of the experiment. Stressed (n ϭ 3) and control (n ϭ 3) rats were sacrificed at the same time. In the RU 486 experiments, both the stressed (n ϭ 3) and control (n ϭ 3) animals were injected subcutaneously with either vehicle or 30 mg/kg body weight (BW) of RU 486 dissolved in 200 l of propylene glycol. In the dexamethasone treatment experiments, animals were injected intraperitoneally, under mild Forene (Abbott, Cham, Switzerland) anesthesia, with either vehicle (n ϭ 3) or 40 g/kg BW of dexamethasone dissolved in 500 l of saline containing 1% ethanol (n ϭ 3). In the diurnal variation experiment, the rats (12 animals) were sacrificed at precise time points (9:00 a.m., 3:30 p.m., 5:00 p.m., and 6:30 p.m.). For each time point, the animals (n ϭ 3) were sacrificed on three consecutive days. Since, in this experiment, the rats were housed in groups of five animals in four independent cages, the last animals killed were never alone, thus preventing an isolation stress. Furthermore, the rats were picked randomly among the four cages. The rats used in the dexamethasone and diurnal expression experiments were sacrificed by decapitation under Forene anesthesia, whereas the rats in the stress experiments were sacrificed by exsanguination under ether anesthesia. Blood samples were collected in heparinized tubes (Milian, Geneva) and centrifuged. Plasma was frozen until analyzed. The tissues were rapidly dissected and frozen in liquid nitrogen.
RNA Preparation and Analysis-Total RNA was prepared using either the acid guanidinium thiocyanate-phenol-chloroform extraction method (31) or using the TRIzol Reagent (Life Technologies, Inc.). Quantification of PPAR␣ and L27 mRNA levels by RNase protection assay were performed as described (28). ACO mRNA was detected by RNase protection using a rat ACO probe corresponding to the 447 nucleotide long SalI-SacI fragment of the full-length cDNA (32).
Antibodies-A polyclonal antibody raised against the AB domain of mPPAR␣ was generated as follows. A cDNA fragment corresponding to the 101 first amino acids of mPPAR␣ was amplified by PCR from the full-length cDNA (Ref. 4, upstream primer: 5Ј-CCGGATCCATGGTG-GACACAGAGAGCC-3Ј, downstream primer: 5Ј-GCGCCCGGGATGT-TCAGGGCACTGCCGG-3Ј) and digested with BamHI and SmaI. This fragment was cloned into the bacterial expression pQE-9 vector (Qiagen) using the BamHI and Klenow-filled HindIII sites. The same fragment was inserted into the pGEX-2T vector (Pharmacia Biotech Inc.) using the BamHI and SmaI sites. The pQE-9 construct was used to overexpress the mPPAR␣/AB domain fused to a 6 ϫ His tag in XL-1 bacteria (Stratagene). The resulting soluble polypeptide was purified from bacterial extract by affinity chromatography on a nickel-nitrilotriacetic acid-agarose column under native conditions according to the manufacturer's instructions (Qiagen, Hilden, Germany). The purified polypeptide was injected subcutaneously into KOBU rabbits (one primary and four booster injections of 200 g). Serum was collected 10 days after the final antigen injection. To affinity purify the serum, an antigen-coupled column was prepared. The pGEX-2T construct described above was used to overexpress a GST-PPAR␣/AB domain fusion protein, which was purified onto a glutathione-Sepharose column (Pharmacia) and coupled to a N-hydroxysuccinylimide-activated Hi-Trap column (Pharmacia). The resulting affinity column was then used to purify the immune serum. The resulting polyclonal antibody crossreacts with rat PPAR␣, which is 98% conserved at the amino acid level in the AB region. Preimmune serum IgGs were purified using a protein G-Sepharose 4 fast flow column (Pharmacia).
Nuclear Extracts and Western Blotting-Nuclei were prepared as follows. Liver samples (0.5 g) or cultured hepatocytes (2.4 ϫ 10 7 cells) were homogenized in 0.5 M sucrose, 50 mM Tris-Cl, pH 7.5, 1 mM EDTA, 25 mM KCl (0.5 M sucrose TEKS). Cells were lysed with 0.5% Triton X-100 for 30 min at 4°C. The homogenate was then layered on a 0.9 M sucrose TEKS cushion and centrifuged at 2000 ϫ g for 20 min. Nuclei were resuspended in 40% glycerol, 50 mM Tris-Cl, pH 8, 5 mM MgCl 2 , 0.1 mM EDTA and stored at Ϫ70°C. The concentrations of the nuclei were determined by measuring A at 260 nm in 5 M NaCl. Nuclei were lysed directly in SDS-gel loading buffer and loaded onto a 10% polyacrylamide-SDS gel. After electrotransfer on nitrocellulose, equal loading was checked by staining the blots with 0.2% Ponceau S red. The blots were blocked 1 h at 25°C with 5% non-fat dry milk in 25 mM Tris-Cl, pH 8.3, 140 mM NaCl, 2 mM KCl, and 0.05% Tween 20 (NFDM TBS-Tween) and incubated overnight at 4°C with the primary antibody at a dilution of 1:1000 in 5% NFDM TBS-Tween. Six 10-min washes at 25°C with 5% NFDM TBS-Tween were performed. The filters were subsequently incubated 2 h at 25°C with the secondary antibody, horseradish peroxidase-conjugated goat anti-rabbit IgG (Cappel, Turnhaut, Belgium), at a dilution of 1:1000 in 5% NFDM TBS-Tween and washed six times 10 min in TBS-Tween. Signal detection was achieved by chemiluminescence with the ECL system (Amersham) and 15-s to 5-min exposure to x-ray films. Signals were quantified using an Elscript 400-AT/SM densitometer (ATH, Neuried, Switzerland).
Hepatocytes Primary Cultures-Rat hepatocytes were isolated by collagenase perfusion (33) of livers from 200 -250 g rats (cell viability higher than 85% by trypan blue exclusion test). The hepatocytes were cultured in monolayer (1.5 ϫ 10 5 cells/cm 2 ) in Williams' E medium (Life Technologies, Inc.) supplemented with 5% fetal calf serum and antibiotics, at 37°C in a humidified atmosphere of 5% CO 2 /95% air. Treatments with WY-14,643 (100 M in ethanol) and dexamethasone (1 M in ethanol) were started immediately after seeding.

RESULTS
Stress Induces PPAR␣ mRNA Expression-Physical and psychological stress triggers a multihormonal response that mainly comprises the release of catecholamines by the sympathetic nervous system and glucocorticoids by the adrenal cortex (34,35). The glucocorticoid levels in blood are elevated during experimental stress situations such as swimming, heat or cold exposure, photic or acoustic stimuli, and forced immobilization (36 -38). Since the PPAR␣ gene is under direct control of glucocorticoid hormones in rat hepatocytes cultured in vitro (28,29), we used stress as an in vivo paradigm to study the regulatory effects of circulating glucocorticoids on the expression of the PPAR␣ gene. In this study, an immobilization protocol was used because it is not associated with an increase of physical activity, in contrast to swimming for example. The reason to avoid experimental protocols requiring physical activity was that PPARs are involved in energy homeostasis (1). PPAR␣ expression was analyzed in rats stressed by forced immobilization during 4 h. All stress experiments were started at 9 a.m. to circumvent interference with the diurnal variations of plasma corticosterone levels (see below). For the same reason, unstressed control animals were sacrificed at the same time as the stressed animals. The 4-h immobilization led to a 3-fold increase in the plasma levels of corticosterone, which is the major active glucocorticoid in rats (Fig. 1C). Total RNA was extracted from liver and hippocampus, a structure in the central nervous system that contains significant amounts of PPAR␣ (39) and that has been described as one of the regions of the brain most sensitive to stress and glucocorticoids (40). Rat PPAR␣ mRNA levels, as well as the levels of the mRNA of the large ribosomal subunit 27-kDa protein (L27) as a control, were assayed by RNase protection. After a 4-h immobilization, stressed animals displayed a 4.5-fold increase in the PPAR␣ mRNA levels in liver relative to the unstressed animals (Fig. 1,  A and B). Stress is therefore a potent physiological inducer of PPAR␣ expression in liver. In contrast to the liver, no significant variation of PPAR␣ mRNA levels could be detected in the hippocampus (Fig. 1, A and B).
To test whether glucocorticoids are indeed involved in the stress-dependent stimulation of PPAR␣ expression, animals were treated before immobilization with the specific glucocorticoid antagonist RU 486 (30 mg/kg BW), or saline as control, and the effect on PPAR␣ mRNA levels was analyzed. In the saline-injected animals, the 4-h immobilization produced a significant 3-fold increase of PPAR␣ mRNA levels in liver (Fig. 2). In marked contrast, the stress-dependent induction of PPAR␣ expression was inhibited in the rats injected with RU 486 (Fig.  2). These results demonstrate that glucocorticoids are the major endocrine mediators of the induction of PPAR␣ expression by stress.
If glucocorticoids are indeed directly involved in the stimu-lation of PPAR␣ mRNA expression induced by stress, the acute injection of exogenous glucocorticoids should lead to increased PPAR␣ mRNA levels. Rats were hence injected either with saline or with dexamethasone (40 g/kg BW) and sacrificed 4 h later. As predicted, dexamethasone-injected rats displayed 3.5fold higher PPAR␣ mRNA levels in liver as compared with saline-injected animals (Fig. 3, A and B). Thus, a single injection of the glucocorticoid agonist dexamethasone reproduced the effects of endogenous glucocorticoids secreted in response to stress. Another effect of administration of dexamethasone was the well described blockade of the hypothalamo-pituitary-adrenal axis (37), resulting in a dramatic decrease of the levels of circulating corticosterone (Fig. 3C).
Cycling of PPAR␣ Expression-In rats, similar to the situation in other mammals, the circulating levels of glucocorticoids are subject to diurnal variations. The plasma levels of corticosterone, which are low in the morning, increase in the afternoon to reach a maximum about 2-3 h before the light-dark switch (41). In view of the results presented above, we expected the levels of PPAR␣ mRNA to follow a similar diurnal rhythm. Thus, we compared PPAR␣ mRNA expression in liver in the morning to its expression in the afternoon. The animals were kept on a 12-h light-dark cycle, with the light-dark switch at 7:30 p.m. Under these conditions, the peak of circulating corticosterone is expected to occur approximately at 5:00 p.m., which was indeed observed (Fig. 4A) Hence, liver samples were taken from animals sacrificed in the morning (9:30 a.m.) and at three different time points in the afternoon (3:30 p.m., 5:00 p.m., and 6:30 p.m.). The analysis was performed over three consecutive days to assess the periodic nature of the variations of PPAR␣ expression and to exclude the possibility of an iso-

FIG. 1. Stress stimulates PPAR␣ expression in liver. A 4-h immobilization stress was achieved as described under "Materials and
Methods." Total RNA (15 g) was analyzed by RNase protection assay using a probe for the PPAR␣ mRNA and a probe for the mRNA of the large ribosomal subunit 27-kDa protein (L27) as control. A, in the liver (left panel), the stressed animals (S, n ϭ 3) show higher levels of PPAR␣ mRNA than the unstressed control animals (C, n ϭ 3). In contrast, there is no variation in PPAR␣ mRNA levels in the hippocampus of the same animals (right panel). B, graphic representation of PPAR␣ mRNA levels in liver and hippocampus of the control (C) and stressed (S) animals. PPAR␣ mRNA levels were normalized to those of L27. The PPAR␣ mRNA level in liver was arbitrarily set to 1. C, plasma corticosterone (CS) levels (ng/ml) of stressed (S) and control animals (C). Results are the mean Ϯ S.D. of three animals.

FIG. 2. RU 486 inhibits the stress-dependent stimulation of PPAR␣ expression.
Rats were injected subcutaneously either with vehicle (SHAM) or with 30 mg/kg BW RU 486. Immediately after the injection, the control animals (C) were returned to their cage, whereas the stressed animals were subjected to a 4-h immobilization. Liver total RNA was analyzed as described in Fig. 1 Fig. 4B). These variations of PPAR␣ mRNA levels correlate well with the diurnal variations of plasma corticosterone levels (Fig. 4A), which strongly suggests that the PPAR␣ gene responds to the diurnal variations of circulating corticosterone.
To test whether the diurnal variations of PPAR␣ mRNA levels resulted into changes in PPAR␣ protein expression, the relative levels of the receptor were measured in liver nuclear extracts. PPAR␣ protein was detected on Western blots using an anti-PPAR␣ antibody. This antibody detects a major band at 55 kDa, which corresponds to the predicted size of PPAR␣. This signal is specific, since it is not detected by preimmune IgGs (Fig. 4C, lane 1). Moreover, its intensity is markedly reduced when the antibody is co-incubated with 20 g of the purified antigen (Fig. 4C, lane 6). In nuclear extracts from the liver of the animals analyzed during the second day of the 3-day experiment, PPAR␣ protein levels were low in the morning and 2-, 3-, and 5-fold higher at 3:30 p.m., 5:00 p.m., and 6:30 p.m., respectively (Fig. 4C, lanes 2-5). Similar results were obtained for the two other days of the experiment. Thus, when the levels of PPAR␣ mRNA and protein measured in each individual animal are plotted successively, according to the time at which the animal was sacrificed, both mRNA and protein levels show a striking cyclic pattern of expression over the three consecutive days (Fig. 4D).
Dexamethasone Potentiates the Induction of the ACO Gene by WY-14,643-Animal treatments by peroxisome proliferators, such as WY-14,643, produce multiple effects. Since specific PPAR antagonists have not yet been identified, it is difficult to distinguish between the PPAR-mediated direct effects of these hypolipidemic drugs and indirect effects involving metabolic or hormonal feedback mechanisms. Thus, we used hepatocytes in primary culture as an in vitro model to test whether the amount of PPAR␣ protein is a limiting factor for the induction of its target genes. Dexamethasone provokes a 4-fold increase of PPAR␣ protein level after 6 h of treatment (Fig. 5A). A similar result was obtained after 24 h of treatment (data not shown). It was thus possible to test if an increase of the amount of PPAR␣ protein could potentiate the stimulation of the ACO gene expression by WY-14,643, a well characterized peroxisome proliferator and activator of PPAR␣ (8). The ACO gene, which encodes the rate-limiting enzyme of peroxisomal fatty acid ␤-oxidation, is controlled by PPAR␣ through a specific response element (3). Hepatocytes cultured in the presence or absence of dexamethasone (1 M) were treated during 24 h with WY-14,643 (100 M) or vehicle. As expected, addition of dexamethasone efficiently stimulated the expression of PPAR␣ mRNA (6-fold, Fig. 5B, lanes 3 and 4). However, dexamethasone alone had no effect on ACO gene expression, presumably because of a lack of PPAR activators (Fig. 5B, lane 3, and C). In contrast, when hepatocytes were treated with WY-14,643 alone, a 3-fold increase of ACO mRNA levels was observed (Fig.  5B, lane 2, and C). Furthermore, simultaneous treatment of the hepatocytes with both WY-14,643 and dexamethasone increased the ACO mRNA levels 6.5-fold (Fig. 5B, lane 4, and C). Thus, dexamethasone provokes a marked potentiation of the induction of ACO gene by WY-14,643. These results strongly suggest that the amount of receptor is a limiting factor for the stimulation of PPAR␣ target genes. DISCUSSION The hormonal response to stress involves essentially the release of catecholamines by the sympathetic nervous system and the secretion of glucocorticoids by the adrenal medulla through the activation of the hypothalamo-pituitary-adrenal axis (35). The onset and the duration of the glucocorticoid component of the stress response are slower and more sustained, respectively, than those of the catecholamine component. Thus, elevated glucocorticoid levels can be considered as a second hormonal wave following the initial peak of catecholamines. Our results demonstrate that, in vivo, the PPAR␣ gene responds mainly to the glucocorticoid component of the hormonal response to stress. Indeed, using the antagonist RU 486 and the agonist dexamethasone, it was shown that glucocorticoids are necessary and sufficient to induce PPAR␣ gene expression in liver during stress situations. Interestingly, stress was unable to modify PPAR␣ expression in the hippocampus, despite the presence of glucocorticoid receptor in this tissue. One hypothesis is that liver-specific factors are required to permit the regulatory action of glucocorticoid re-ceptor on the PPAR␣ gene promoter. Alternatively, brain-specific factors might inhibit this regulation. In liver, the induction of PPAR␣ mRNA is a fast response since it can be observed already 4 h after immobilization or agonist injection. The rapid regulation of PPAR␣ mRNA expression in vivo reported herein argues for a similar direct transcriptional effect of glucocorticoids on PPAR␣ gene expression in liver as in hepatocytes cultured in vitro (28,29).
The metabolic response to stress is characterized by energy mobilization. Under the action of lipolytic hormones (catecholamines mainly), fatty acid mobilization occurs in the adipose tissue. In liver, the mobilized fatty acids enter the ␤-oxidation pathway and ketogenesis is stimulated (42). Remarkably, PPAR␣ regulates genes involved in the activation of fatty acids as well as in the ␤-oxidation and ketogenesis pathways (1). Thus, stimulation of PPAR␣ expression in liver by glucocorticoids during stress may potentiate the regulation of these target genes and contribute to the stimulation of the metabolic pathways involved in energy homeostasis.
The expression of the PPAR␣ gene is showing a diurnal cycling pattern in liver, which parallels that of circulating corticosterone. This is consistent with a high sensitivity in liver of the PPAR␣ gene to the levels of circulating corticosterone. The diurnal variations of PPAR␣ mRNA is closely followed by a parallel cycling of PPAR␣ protein suggesting that PPAR␣ mRNA is efficiently translated. Furthermore, the cyling of the PPAR␣ protein levels implies that the half-life of the protein is short enough to allow its levels to significantly decrease after 12 h. Altogether, these results suggest that PPAR␣ signaling pathway may be efficiently modulated by a rapid and transient regulation of receptor levels. Post-translational mechanisms may furthermore exist, since we did not detect clear variations of PPAR␣ protein levels after a 4-h stress (data not shown). Alternatively, this very short time of stimulation may be insufficient to give rise to a detectable increase of PPAR␣ protein levels.
The investigation of the glucocorticoid-dependent regulation of PPAR␣ gene was possible in the whole animal, since physiological situations associated with variations of circulating glucocorticoids are well characterized. Moreover, specific glucocorticoid agonists and antagonists are available. More difficult is the in vivo study of PPAR␣-mediated gene regulation, since the natural ligands of this receptor are still unknown, and specific agonists or antagonists have not been reported so far. The PPAR activators known to date, such as peroxisome proliferators, have pleiotropic effects in vivo, making it difficult to discriminate between the direct and indirect actions of these drugs in the animal. In contrast, hepatocyte primary cultures represent a model close to the in vivo situation, which is, however, isolated from the hormonal and metabolic complexity of the intact animal. Using this system, we show that high levels of expression of PPAR␣ and activation of the receptor are necessary for a maximal stimulation of PPAR␣ target genes. Indeed, the dexamethasone-dependent increase in PPAR␣ expression is associated with a marked potentiation of the effects of the PPAR␣ activator WY-14,643 on the expression of the ACO gene. These experiments provide evidence that the amount of receptor is a limiting factor and thus that the regulation of the level of PPAR␣ expression by glucocorticoids has impact on the regulation of its target genes. In the whole animal, the exact physiological conditions in which such a glucocorticoid-dependent modulation of PPAR␣ signaling occurs remain, however, to be investigated. For example, we did not observe diurnal variations in expression of the ACO gene (data not shown). This may well be due to the absence of significant levels of endogenous PPAR␣ ligand. The same phe-  2 and 4). ACO, PPAR␣, and L27 mRNA were detected by RNase protection as described in the legend to Fig. 1. C, L27-normalized ACO mRNA levels. The ACO mRNA level of the vehicle-treated hepatocytes was arbitrarily set to 1. F.I. ϭ -fold induction. nomenon was observed in primary hepatocytes in which the induction of PPAR␣ expression by dexamethasone is insufficient to stimulate ACO mRNA levels in the absence of an activator of the receptor. These observations suggest that activation of the receptor is a prerequisite for an effect of glucocorticoids on PPAR␣ target genes. Investigation of physiological states in which these conditions are fulfilled, that is in which PPAR activators/ligand are produced, will possibly give important clues for the identification of the natural PPAR␣ ligand.
In conclusion, data in this paper suggest that glucocorticoids have an important regulatory impact on PPAR␣ expression in vivo. Physiological situations, such as stress and the diurnal surge of glucocorticoids, affect PPAR␣ expression in liver. The regulation of PPAR␣ expression provides a control mechanism which, when coupled to activator availability, regulates the PPAR␣ action on its target genes and associated metabolic pathways.