Coordinate Regulation of the Expression of the Fatty Acid Transport Protein and Acyl-CoA Synthetase Genes by PPARα and PPARγ Activators*

Intracellular fatty acid (FA) concentrations are in part determined by a regulated import/export system that is controlled by two key proteins, i.e. fatty acid transport protein (FATP) and acyl-CoA synthetase (ACS), which respectively facilitate the transport of FAs across the cell membrane and their esterification to prevent their efflux. The aim of this investigation was to analyze the expression pattern of FATP and ACS and to determine whether their expression was altered by agents that affect FA metabolism through the activation of peroxisome proliferator-activated receptors (PPAR) such as the fibrates and thiazolidinediones. FATP mRNA was ubiquitously expressed, with highest levels being detected in adipose tissue, heart, brain, and testis. Fibrate treatment, which is known to preferentially activate PPARα, induced FATP mRNA levels in rat liver and intestine and induced ACS mRNA levels in liver and kidney. The antidiabetic thiazolidinedione BRL 49653, which is a high-affinity ligand for the adipocyte-specific PPARγ form, caused a small induction of muscle but a robust induction of adipose tissue FATP mRNA levels. BRL 49653 did not affect liver FATP and had a tendency to decrease heart FATP mRNA levels. ACS mRNA levels in general showed a similar pattern after BRL 49653 as FATP except for the muscle where ACS mRNA was induced. This regulation of FATP and ACS expression by PPAR activators was shown to be at the transcriptional level and could also be reproduced in vitro in cell culture systems. In the hepatocyte cell lines AML-12 or Fa 32, fenofibric acid, but not BRL 49653, induced FATP and ACS mRNA levels, whereas in the 3T3-L1 preadipocyte cell line, the PPARγ ligand induced FATP and ACS mRNA levels quicker than fenofibric acid. Inducibility of ACS and FATP mRNA by PPARα or γ activators correlated with the tissue-specific distribution of the respective PPARs and was furthermore associated with a concomitant increase in FA uptake. Most interestingly, thiazolidinedione antidiabetic agents seem to favor adipocyte-specific FA uptake relative to muscle, perhaps underlying in part the beneficial effects of these agents on insulin-mediated glucose disposal.

Transmembrane transport of FAs 1 is still poorly understood despite intense investigation. Uncharged molecules and weak acids such as fatty acids can cross membranes rapidly thanks to their lipid solubility. The rate of movement is controlled by mass action and can be enhanced by proteins such as fatty acid-binding protein that act as a cytoplasmic "sink." Recently, however, several studies provided evidence that in addition to these nonfacilitated systems, facilitated transport also contributes to FA transport (1). Several proteins were hypothesized to be acting as FA transporters. Among these, three deserve further attention. First, plasma membrane fatty acid-binding protein, a protein related to the mitochondrial isoform of aspartate aminotransferase, has been suggested to increase FA uptake in cells (2). Since this protein has not yet been cloned, it is difficult to determine its exact role in FA transport processes. The second protein, fatty acid translocase is an 88-kDa membrane protein that has been cloned in mouse and is homologous to the human CD36 cell surface antigen (3). Although CD36 has been shown to bind FAs and might be involved in signal transduction after binding of a specific ligand (long chain fatty acids), it is until now not clear whether it is a transport protein. The only candidate for a long chain FA transporter for which functionality has been directly demonstrated is the fatty acid transport protein (FATP) (4). FATP is a 63-kDa plasma membrane protein with six predicted membrane-spanning domains that has been cloned using a functional expression cloning technique. It increased oleic acid uptake in FATP-transfected 3T3-L1 cells by Ͼ3-fold. Interestingly FATP is suggested to act in concert with acyl-CoA synthetase (ACS), an enzyme that prevents efflux of the incorporated fatty acids by their conversion into acyl-CoA derivatives and hence rendering the FA uptake process unidirectional. Furthermore, FATP shows a limited region of homology at the protein level (11 amino acids), with ACS leading to the hypothesis that this common region might reflect a common function, such as a binding site (4). These 11 amino acids residues have also been found to be conserved in the rat and the yeast FATP homologues (5).
Several aspects of intracellular lipid and FA metabolism in cells are subjected to transcriptional control by the peroxisome proliferator-activated receptor (PPAR) family. PPARs are members of the superfamily of nuclear hormone receptors that function as ligand-dependent transcription factors. Three receptor subtypes of PPAR termed ␣, ␦ (or ␤), and ␥, have been identified (6 -15). These receptors heterodimerize with the retinoid X receptor and alter the transcription of target genes after binding to peroxisome proliferator response elements (PPREs), which consist of a hexameric nucleotide direct repeat of the recognition motif (TGACCT) spaced by 1 nucleotide (DR-1). Several genes with a crucial role in FA metabolism have been shown to contain a peroxisome proliferator response element in their upstream regulatory sequences (reviewed in Refs. 16 and 17). Interestingly, the transcriptional activity of the PPAR subtypes is enhanced by a multitude of chemical compounds including fatty acids, thiazolidinedione antidiabetic agents, prostaglandins, peroxisome proliferators, and fibrate hypolipidemic drugs. In addition to activating PPARs, some of these compounds have been shown to be direct ligands for them. PPAR␥ directly binds antidiabetic thiazolidinediones (18,19) and prostaglandin derivatives (18,20) but not the other activators, whereas PPAR␣ binds leukotriene B4 and the powerful peroxisome proliferator Wy 14643 (21).
In view of the convergence of FA import and PPARs in lipid and energy metabolism, we investigated the effects of two distinct chemical classes of PPAR activators, i.e. fibrates (PPAR␣-specific) and the antidiabetic thiazolidinediones (PPAR␥-specific) on tissue-specific FATP gene expression. Fibrate treatment induced FATP and ACS expression strongest in liver, whereas BRL 49653, the high affinity ligand for PPAR␥, had no effect on liver but induced adipocyte FATP and ACS expression in adipose tissue. The induction of FATP and ACS by PPAR activators was at the level of transcription and was associated with concomitant changes in cellular FA uptake. Interestingly, the stronger effects of BRL 49653 on fatty acid import in adipose tissue relative to the muscle might limit FA uptake and oxidation in the muscle, an effect associated with an improvement in muscle glucose disposal.
Animal Studies-Animal studies were carried out in compliance with French and European union specifications regarding the use of laboratory animals. Male Wistar rats (90 days old) were treated for 7 days with fenofibrate (Laboratoires Fournier) mixed at the indicated concentrations (by mass) in standard rat chow. The food intake of the rats was recorded every day throughout the treatment period. None of the treatments caused major changes in the amount of food consumed by the animals. Since each rat consumed approximately 20 gm of chow/day, doses of 0.5, 0.05, and 0.005% (by mass mixed in rat chow) correspond to 320, 32, and 3 mg/kg of body weight/day. Adult (95 day) Sprague-Dawley rats were group-housed and accustomed to a 12:12 h day:night ratio illumination cycle (light from 8 am to 8 pm). Rats were divided in groups of a minimum of three animals each and were treated for either 7 or 14 days. The first group received BRL 49653 (5 mg/kg/day) by gavage. The second group of animals received 0.5% w/w of fenofibrate (Ϯ 0.5 g/kg/day) mixed with their food, whereas the third group of animals served as controls and received 10% carboxymethylcellulose (vehicle for gavage) by gavage. In a separate experiment, adult C57Bl6 male mice were either fed during 14 days with a control chow (n ϭ 3) or the same chow containing 0.5% w/w of fenofibrate. At the end of the treatment period, all animals were weighed and sacrificed by exsanguination under ether anesthesia between 8 and 10 a.m. Epididymal adipose tissue (in rats) and liver (in rats and mice) was removed, weighed, rinsed with 0.9% NaCl, and frozen in liquid nitrogen until RNA preparation.
Cell Culture-The mouse hepatoma and preadipocyte cell lines Fa 32 (rat), ob 1771 (mouse) (22), and 3T3-L1 (mouse; ATCC) were maintained in Dulbecco's modified Eagle's minimal essential medium and supplemented with 10% delipidated and charcoal-treated fetal calf serum, L-glutamine, and antibiotics unless stated otherwise. AML-12 mouse hepatocytes (23) were maintained in Dulbecco's modified Eagle's minimal essential medium/Ham's F-12 supplemented with 10% delipidated and charcoal-treated fetal calf serum, insulin, transferrin, and selenium (Collaborative Research), dexamethasone (0.1 M), and gentamycin (50 g/ml). Fenofibric acid and BRL 49653 (both in Me 2 SO) were added to the medium at the appropriate concentrations and times Preparation of Albumin-bound Fatty Acids-Radiolabeled 14 [C]oleate fatty acid was mixed in water at 40°C, albumin (BSA; fraction V, fatty acid-free, Sigma) was then added from a concentrated stock (20 g/100 ml) to give a final molar ratio of 1/1 by gentle mixing. 2 ϫ Hanks' solution was added to obtain a 1 ϫ final solution. Incubation was carried out at 37°C for 45 min.
Fatty Acid Uptake Assay-The measurement of uptake of 1-14 Clabeled oleate (about 50 mCi/mmol, NEN Life Science Products) was carried out in 24-or 6-well plates with 10 6 cells/ml of medium. Before treatment, the cells were washed with 1 ϫ Hanks' solution. BRL 49653 (100 -250 nM) and fenofibric acid (100 -250 M) were added in fresh Dulbecco's modified Eagle's minimal essential medium containing 10% fetal calf serum. After 48 h of treatment, cells were washed with Hanks' solution and incubated for 1 additional h in serum-free, glucose-free medium. Cells were then washed once at 37°C and twice at 23°C with 1 ϫ Hanks' solution containing BSA. Hanks' solution without BSA was then added before the assay. A volume corresponding to 1 Ci of [ 14 C]oleate albumin-bound solution was added in each well, and cells were incubated for 1 min at room temperature. The incubation was stopped after 1 min with 3 washes of ice-cold 1 ϫ Hanks' solution without BSA. A complementary experiment has been performed to verify whether aspecific cell surface binding of [ 14 C]oleate could interfere with the assay. For this second assay, the cells were washed under more stringent conditions in 1 ϫ Hanks' solution containing 2% BSA. Cells were then lysed in 400 l of 0.1% SDS solution. The lysate was counted for 5 min with 4 ml of scintillation solution. The assay was performed on triplicate points.
RNA Analysis-RNA preparation, Northern blot hybridizations, and quantification of total cellular RNA were performed as described previously (24). A mouse FATP cDNA probe was obtained after cloning a reverse transcription-polymerase chain reaction fragment from mouse adipose tissue RNA using the primers 382 (ATGCGGGCTCCTGGAG-CAGGACAGCC) and 399 (CTGCGTGTCAGGCAGGATGCTCTCAG-GCCC) into pBS-KS. The insert was sequenced and found to be identical to the reported mouse FATP sequence. The rat ACS probe corresponds to the EcoRV restriction fragment of the rat ACS cDNA. The human acidic ribosomal phosphoprotein 36B4 (25) was used as a control probe.

RESULTS
FATP mRNA Is Ubiquitously Expressed-To determine whether FATP expression was ubiquitous or restricted to certain tissues, we hybridized both a mouse (Fig. 1A) and a rat ( Fig. 1B) multiple tissue Northern blot with a radiolabeled FATP probe. In both rat and mouse, adipose tissue, heart, brain, and testis showed the highest level of expression. Intestine and muscle show intermediate levels of expression, and low levels are expressed in the liver, kidney, lung, and spleen.
Fenofibrate, a PPAR␣ Activator, Induces FATP mRNA in Vivo-In addition to being building blocks and energy substrates, fatty acids are also important signaling molecules. Besides being activated by peroxisome proliferators and certain thiazolidinediones, the transcriptional activity of PPARs can be activated by fatty acids (reviewed in Refs. 16 and 17). Therefore, we were interested in analyzing whether activation of these PPARs would affect FA uptake in general and FATP expression in particular. To address this issue we first assessed the effect of fibrates, potent PPAR␣ activators, on the expression of the fatp gene in rats. Rats were hence treated with different doses of fenofibrate (14 days treatment at the doses 0.005, 0.05, and 0.5% by mass) mixed in food. Next, RNA was isolated from various organs and analyzed by Northern blot hybridization. In liver, FATP mRNA levels increased gradually starting from 0.05% fenofibrate and reached a maximal 4.2-fold induction at the highest dose of 0.5% fenofibrate (Fig. 2, A and  B). A representative Northern blot showing the induction of FATP mRNA in the liver is depicted in Fig. 2B. Next, the response of FATP mRNA levels to fibrates was studied in intestine, skeletal muscle, heart, and kidney. Only intestinal FATP mRNA expression was slightly induced (2-fold) by the highest dose of fibrate treatment, whereas muscle, kidney, and heart FATP mRNA expression remained unchanged. Also in mice, administration of fenofibrate (0.5%) induced FATP mRNA levels in liver (15-fold) (Fig. 2C).
Parallels between the Fibrate Effects on FATP and ACS-Several proteins are hypothesized to enhance fatty acid uptake into cells. In contrast to FATP, which acts as an FA transport protein, ACS prevents the efflux from the imported FAs by converting them into acyl-CoA derivatives, which can subse-quently be used in both anabolic and catabolic pathways. Therefore, we next analyzed whether there was a parallel between the induction of FATP and ACS after fibrate treatment (Fig. 3). Fenofibrate induced liver and kidney ACS mRNA expression, whereas no change in ACS expression was observed in heart and intestine. Therefore both FATP and ACS mRNA levels seem to be coordinately regulated in liver and heart, since fibrates affect both parameters in a similar fashion. The regulation of ACS and FATP in the kidney and intestine seems to be divergent, since in these tissues only one of the respective mRNAs is regulated by fibrate treatment.
PPAR␥ Activators Induce FATP mRNA in Adipose Tissue-In addition to the well established effects of peroxisome proliferators such as the different fibrates on PPAR␣ activity, we next tested the effects of the PPAR␥-selective ligand BRL 49653 on FATP and ACS expression in various rat tissues after administration of these compounds. Fenofibrate (0.5% (w/w), Ϯ0.5 g/kg/day) induced FATP and ACS mRNA in rat liver (Fig.  4, A and B), confirming our previous observations (26). By contrast, treatment with fenofibrate did not change FATP and ACS mRNA levels significantly in adipose tissue, skeletal muscle, or heart (Fig. 4, A and B). Administration of 5 mg/kg/day of BRL 49653 was associated with the expected decrease in serum triglyceride levels (from 167 to 88 mg/dl). Furthermore, this treatment with BRL 49653 resulted in a significant induction of adipose tissue FATP (7-fold) and ACS (7-fold) mRNA levels (Fig. 4). This induction of FATP and ACS mRNA by BRL 49653 was observed in epididymal (Fig. 5) and omental (data not shown) adipose tissue. In perirenal adipose tissue, however, only FATP but not ACS mRNA was induced (Fig. 5). We observed a 1.6-and 3.1-fold induction of respective levels of FATP and ACS mRNA in skeletal muscle after BRL 49653 administration. BRL 49653 did not significantly influence the expression of FATP or ACS in liver, whereas FATP mRNA levels showed a tendency to decrease in the heart after BRL 49653 treatment.
The Induction of FATP and ACS Expression by Fenofibrate Is at the Transcriptional Level-To analyze whether the induction of FATP and ACS mRNAs occurred at the transcriptional level, a nuclear run-on assay was performed on liver nuclei obtained from fenofibrate-treated rats (Fig. 6). In comparison with control liver nuclei, the rate of FATP and ACS transcription was respectively 3-and 3.5-fold higher in nuclei from fenofibrate-treated animals. The transcription rate of acyl-CoA oxidase, a key enzyme in the peroxisomal ␤ oxidation pathway, a positive control for fibrate action, was induced (5-fold), whereas the glyceraldehyde phosphate dehydrogenase gene, a negative control, did not change.
BRL 49653 Induces FATP mRNA Specifically in Preadipocyte Cells, whereas Fibrates Induce FATP mRNA in Cells of Hepatic Origin-To study the cellular mechanism of this induction, we investigated the regulation of the FATP gene expression by fibrates and BRL 49653 in hepatocyte (Fig. 7), adipocyte (Fig. 8), and muscle cells cell lines. FATP and ACS mRNA were measured in mouse AML-12 and rat Fa 32 liverderived cell lines. A strong induction of expression of both FATP and ACS mRNA levels was seen after treatment of these liver-derived lines with fenofibric acid. Fenofibric acid induced both mRNAs optimally within 24 h (Fig. 7A) at a dose of 250 M (Fig. 7B). The results of dose response and time course of FATP and ACS induction after treatment with fibrates seem to show an apparent difference between Fa 32 and AML-12 cells. The reason for this apparent difference in induction of FATP and ACS in the two cell lines is most likely caused by the difference in basal levels of FATP, which in Fa 32 cells is barely detectable. In contrast, under basal conditions, AML-12 expresses

FIG. 4. Tissue-selective induction of FATP (A) and ACS (B)
mRNA in rat liver, adipose tissue, skeletal muscle, and heart by fenofibrate and BRL 49653, respectively. A, expression of FATP mRNA in liver, epididymal adipose tissue, skeletal muscle, and heart of animals treated with fenofibrate (FF, 0.5% (w/w) during 7 days, approximately 0.5 g/kg/day) or BRL 49653 (5 mg/kg/day during 7 days). The blots were stripped and rehybridized with the human acidic ribosomal phosphoprotein 36B4 control cDNA. B, expression of ACS mRNA in liver, epididymal adipose tissue, skeletal muscle, and heart of animals treated with fenofibrate (FF, 0.5% (w/w) during 7 days, approximately 0.5 g/kg/day) or BRL 49653 (5 mg/kg/day during 7 days). The blots were stripped and rehybridized with the human acidic ribosomal phosphoprotein 36B4 control cDNA. Animal treatment and preparation and analysis of RNA is described under "Experimental Procedures." C, bar graph summarizing the regulation of ACS and FATP mRNA in liver, epididymal adipose tissue, skeletal muscle, and heart of animals treated with or without BRL 49653. R.A.U., relative absorbance units. Values statistically significant from controls (Mann-Whitney; p Ͻ 0.05) are indicated by an asterisk. much higher levels of FATP transcript. This difference between these cells will result in an overestimation of the induction of the FATP in Fa 32 cells, explaining the discordance between relative levels of induction between FATP and ACS in the two cell lines.
To examine FATP and ACS regulation in adipocyte-like cell lines, 3T3-L1 preadipocyte cells were used. First, we analyzed the effects of BRL 49653 on nondifferentiated 3T3-L1 preadipocyte cells. As shown in Fig. 8A, a limited effect of BRL 49653 was observed in undifferentiated 3T3-L1 cells. In differentiated 3T3-L1 adipocytes, FATP and ACS mRNA levels were induced 5-and 9-fold after 4 days of treatment with BRL 49653. The ob 1771 preadipocyte cell line (22) was also analyzed (Fig. 8). The addition of BRL 49653 also induced FATP and ACS mRNA levels in this cell line (Fig. 8), whereas fenofibric acid had only a weak effect (data not shown).
Finally the effects of both BRL and fenofibric on L6 muscle cells were analyzed. Unlike in adipocyte or hepatocyte cell lines, no change in L6 ACS and FATP mRNA levels were detected upon treatment with either fenofibrate or BRL 49653 (data not shown).
Induction of FATP or ACS mRNA Levels Results in a Change in FA Uptake into Cells-To verify whether changes in mRNA levels of FATP and ACS were correlated with alterations in fatty acid uptake, we analyzed [ 14 C]oleic acid uptake in Fa 32 and AML-12 hepatic cells, L6 muscle cells, and 3T3-L1 preadipocytes. As shown in Fig. 9, fatty acid uptake of [ 14 C]oleic acid significantly increased after treatment of the liver-derived AML-12 cells with fenofibric acid and after treatment of the differentiated adipocyte-like 3T3-L1 cells with BRL 49653. In Fa 32 cells, fatty acid uptake was also increased (data not shown). The increase, although statistically significant, was however less pronounced than in AML-12 cells. As expected in view of the absence of a major regulation of ACS and FATP mRNA in muscle cells, no effects of either fenofibric acid nor BRL 49653 were observed on FA uptake in L6 muscle cells. The regulation of FA uptake was hence completely consistent with the regulation of respective mRNA levels of FATP and ACS in the various cell models. DISCUSSION Both ACS and FATP have been suggested to play a crucial role in the transport of fatty acids into the cell (4). FATP acts as a fatty acid transport protein, whereas ACS prevents efflux of the newly imported fatty acids by their esterification with coenzyme A. Fatty acids are important cellular components that can function both as metabolic substrates or as signaling molecules, by functioning as second messengers and triggering signal transduction pathways or by directly activating transcription factors such as the PPAR family of nuclear receptors. Since FATP and ACS control, in part, the intracellular availability of FAs, important PPAR activators, the aim of the present investigation was to perform detailed analysis of FATP and ACS expression and to establish whether FATP and ACS expression themselves might be subject to control by PPARs.
In the liver, one of the major organs susceptible to peroxisomal proliferation, FATP gene transcription is strongly induced upon fibrate treatment. This strong induction is not surprising if one takes the strong induction of peroxisomal ␤ oxidation into account after fibrate treatment. FATP is likely to be responsible in part for the increased FA import necessary to sustain this increased ␤ oxidation. Furthermore, a striking parallelism exists between the induction by fibrates of a number of genes involved in fatty acid import in the liver. In fact, the mRNAs for lipoprotein lipase (LPL) (27), ACS (this paper and Refs. 27 and 28), and FATP genes are all induced after fibrate treatment in the liver. This coinduction of genes seems to prime the cells for more efficient ␤ oxidation. Induced liver LPL expression will increase lipolysis in the vascular bed of the liver, generating more fatty acids, which are then avidly taken up by the cells thanks to the higher levels of FATP expression. Efflux of these imported fatty acids is prevented by induced ACS levels, which in addition, primes them for subsequent metabolism. Therefore it seems that fibrates not only induce ␤ oxidation but also induce genes important for supplying the cells with the extra fatty acids they need to sustain this increase in ␤ oxidation.
In the intestine, FATP was also induced by fibrates, albeit to a lesser extent. The FATP induction in this tissue shows a striking parallel to the induction of CD36 after fibrate treatment in this tissue (29). In contrast to the liver and intestine, heart FATP mRNA levels did not vary substantially under fibrate treatment. The basal levels of FATP expression were, however, very high in heart, which almost exclusively uses fatty acids as energy source. In this tissue, fatty acids are, however, constitutively metabolized to provide energy necessary for the contraction of the heart muscle. Unresponsiveness of FATP expression in the heart to hormonal control could hence be physiologically significant, since the continued function of the heart is far too crucial to allow any form of major regulation of a transporter vital to its energy supply. This would suggest that in the heart, the FATP promoter is maximally active, resulting in a high level of constitutive FATP gene expression, which would be consistent with the high basal levels of FATP mRNA in this tissue. Similar to the heart, less extreme changes were observed in adipose tissue FATP mRNA after fibrate treatment. The absence of an effect of fibrates on FATP expression in adipose tissue is most likely due to the lower levels of PPAR␣ relative to PPAR␥. For kidney and intestine, the regulation of ACS and FATP by fibrates is discordant. This is consistant with the less crucial functions lipids play in kidney and intestinal metabolism. Kidney expressed only low levels of FATP mRNA, and its expression was furthermore refractory to induction by fibrates. Relative to heart and liver, the kidney utilizes relatively little fatty acids, and therefore a coordinate import mechanism is of lesser importance. In intestine, fatty acids are primarily absorbed, but they are less actively metabolized than in heart and liver. Therefore, intestine apparently has an actively regulated transport mechanism, as evidenced by regulation of the expression of both FATP and FAT, another transport protein that is also expressed and highly regulated in this organ (29). Since fatty acids are less actively metabolized and rather resecreted under the form of lipoproteins, there is less need for their conversion to acyl-CoA derivatives and hence less need for coordinated regulation of ACS together with these transport proteins suggested to be implicated in fatty acid transport.
The demonstration of the inducibility of the FATP and ACS genes by PPAR␥ ligands such as the thiazolidinedione BRL 49653 has important implications for adipocyte physiology. PPAR␥ has been shown to promote preadipocyte determination as well as terminal differentiation (13,30), and its mRNA is itself induced in the earliest steps of adipocyte differentiation before the induction of early marker genes for adipocyte differentiation. Many of these genes induced during adipocyte differentiation encode proteins involved in lipid storage and metabolism. The increase in FATP and ACS expression in differentiated adipocyte-like cells caused by PPAR␥ ligands will result in an increased delivery of fatty acids to the adipocytes, which possibly sustains a positive regulatory feedback loop involving continued PPAR␥ activation of the FATP and ACS (28) genes and aimed at promoting and maintaining the mature adipocyte phenotype. In fact, in addition to the thiazolidinediones, certain fatty acid-derived prostaglandin derivatives, whose delivery to the cell is increased by FATP, bind to and/or activate PPAR␥ (19,20,31). This hypothesis is supported by the observation that fatty acids (including arachidonic acid-derived prostaglandins) and fatty acid analogues induce the expression of adipocyte-specific genes and enhance adipocyte conversion (30,(32)(33)(34)(35). In addition to being potent PPAR activators (7,12,31,36,37), fatty acids will provide the necessary building blocks for triglyceride accumulation, ultimately enhancing adipocyte differentiation. The PPAR-mediated activation of FATP and ACS expression in cells of the adipogenic lineage might furthermore in part be responsible for the previously reported capacities of thiazolidinediones to induce adipocyte differentiation and induce the development of obesity (38 -46). In this context, it is interesting to note that the PPAR␥-mediated effects of BRL 49653 on FATP and ACS expression might act in concert with induced LPL expression and the reduced leptin mRNA and protein levels and the associated increase in caloric intake enhancing energy storage in the adipocytes observed with this compound (47,48). Interestingly, FATP and ACS are not coordinately regulated in perirenal and epididymal adipose tissue stores. This differential regulation is consistent with the distinct metabolic nature of the different adipose tissue depots (49 -51). Further studies are required to determine whether the role of FATP is a consequence of or is causative of the physiologic differences between the adipose tissue depots.
The tissue-selective effects of the various PPAR activators/ ligands are highly intriguing and provide insight in their effects on triglyceride metabolism. Fibrate treatment induced FATP and ACS expression strongest in liver, whereas BRL 49653 had no effect on liver, but strongly induced adipocyte FATP and to a lesser extent ACS expression. The effects of fibrates (PPAR␣ activators) on the liver and PPAR␥ ligands in adipose tissue correlates well with the tissue-specific expression of the respective receptors and suggests that the FATP and ACS genes show a tissue-selective activation similar to the one previously described for the LPL gene (27). In this context, we need however to address the discrepancy between ACS and FATP regulation after BRL 49653 administration in skeletal muscle. Muscle tissue expresses very low levels of PPAR␥, 2 which is consistent with the absence of an important regulatory effect of PPAR␥ activators in this tissue as observed for the LPL (27) or FATP (this study) genes. In this context, the induction of ACS expression by BRL 49653 is however difficult to explain. One must however bear in mind that not all of the effects of the thiazolidinediones are mediated via PPAR␥ activation, and it is has been shown that these agents activate several other signaling pathways (52)(53)(54). Further investigations need to address whether ACS expression, unlike FATP or LPL expression, is subject to such a regulatory circuit.
One remaining question is the relationship between PPAR␥, thiazolidinediones, and insulin resistance. It is tempting to speculate that the increase of LPL, ACS, and FATP activity in adipose tissue is related to the antidiabetic effects of the thiazolidinediones. Due to the enhanced triglyceride clearance in adipose tissue, less triglycerides will become available to be hydrolyzed to fatty acids in the vascular bed of the muscle. Furthermore, relative to the strong induction of FATP and ACS in adipose tissue by BRL 49653, very limited inductions of both genes are observed in skeletal and heart muscle, favoring uptake of fatty acids in adipose tissue relative to muscle. In view of the inhibitory effects of fatty acids on insulin-mediated glucose metabolism (55), the decrease in fatty acids delivered to the muscle cells might be responsible for the improvement in insulin sensitivity of this tissue.
In conclusion, FATP and ACS mRNA levels can be regulated in a tissue-specific fashion by PPAR␣ activators and PPAR␥ ligands. In adipose tissue, the increase in FATP, ACS, and LPL (27) production after treatment with thiazolidinediones will enhance the clearance of plasma triglycerides (27,56) and provide the (pre)adipocytes with additional fatty acids, which can further stimulate the transactivation capacity of PPAR or which can be stored under form of triglycerides. In the liver, the enhanced production of FATP and ACS after fibrates together with the increase in ␤ oxidation and the reduced production of apoCIII (57), may contribute to the hypolipidemic action of these compounds. This tissue-selective induction of FATP and ACS gene transcription by activators of different PPARs, demonstrates the feasibility of the development of highly specific PPAR subtype-specific agonists and antagonists, which can be used as drugs.