Expression of Putative Fatty Acid Transporter Genes Are Regulated by Peroxisome Proliferator-activated Receptor α and γ Activators in a Tissue- and Inducer-specific Manner

Regulation of gene expression of three putative long-chain fatty acid transport proteins, fatty acid translocase (FAT), mitochondrial aspartate aminotransferase (mAspAT), and fatty acid transport protein (FATP), by drugs that activate peroxisome proliferator-activated receptor (PPAR) α and γ were studied using normal and obese mice and rat hepatoma cells. FAT mRNA was induced in liver and intestine of normal mice and in hepatoma cells to various extents only by PPARα-activating drugs. FATP mRNA was similarly induced in liver, but to a lesser extent in intestine. The induction time course in the liver was slower for FAT and FATP mRNA than that of an mRNA encoding a peroxisomal enzyme. An obligatory role of PPARα in hepatic FAT and FATP induction was demonstrated, since an increase in these mRNAs was not observed in PPARα-null mice. Levels of mAspAT mRNA were higher in liver and intestine of mice treated with peroxisome proliferators, while levels in hepatoma cells were similar regardless of treatment. In white adipose tissue of KKAy obese mice, thiazolidinedione PPARγ activators (pioglitazone and troglitazone) induced FAT and FATP more efficiently than the PPARα activator, clofibrate. This effect was absent in brown adipose tissue. Under the same conditions, levels of mAspAT mRNA did not change significantly in these tissues. In conclusion, tissue-specific expression of FAT and FATP genes involves both PPARα and -γ. Our data suggest that among the three putative long-chain fatty acid transporters, FAT and FATP appear to have physiological roles. Thus, peroxisome proliferators not only influence the metabolism of intracellular fatty acids but also cellular uptake, which is likely to be an important regulatory step in lipid homeostasis.

addition to their ability to biosynthesize LCFAs, cells can utilize plasma LCFAs released by lipoprotein lipase-catalyzed hydrolysis of triglycerides from circulating chyromicrons and very low density lipoproteins (1). Thus, uptake of LCFAs into cells can be an important step in energy metabolism. However, the mechanism and regulation of uptake of extracellular LCFAs into mammalian cells is not well understood.
Because of their hydrophobic properties, it has been suggested that LCFAs are transported across the plasma membrane by simple diffusion (2,3). However, other studies have suggested that this process is mediated by proteins. Three independent transport proteins have been identified that may contribute to this process, including fatty acid translocase (FAT)(4), mitochondrial aspartate aminotransferase (mAspAT) (5), and fatty acid transport protein (FATP) (6). The precise role of these proteins in mediating LCFA uptake is not known and the mechanisms of LCFA uptake into various mammalian cells may not be the same. Nevertheless, there is evidence that all three may, at least in part, contribute to LCFA cellular uptake.
Peroxisome proliferator-activated receptors (PPARs) have unique roles in lipid homeostasis. PPARs are part of the nuclear hormone receptor superfamily, and there are three subtypes that have been described, ␣, ␤(␦), and ␥. Each is encoded by a separate gene and have unique tissue distribution patterns. Furthermore, their roles in mediating changes in gene expression appear to be cell-and tissue-specific. For example, PPAR␣ mediates fibrate and dietary polyunsaturated fatty acid induction of hepatic peroxisomal lipid-metabolizing enzymes, including acyl-CoA oxidase a key enzyme in regulating peroxisomal lipid catabolism (7,8). Furthermore, PPAR␣ modulates hepatic gene expression of apolipoprotein A-I and C-III in response to the peroxisome proliferator Wy 14,643 (9). It is of interest to note that recently, polyunsaturated fatty acids Wy 14,643 and leukotriene B4 have been shown to bind to PPAR␣ (10 -12). Upon binding, PPAR␣-dependent gene transcription is activated, and these effects are most pronounced in the liver, a tissue with a high capacity for ␤-oxidation of fatty acids. Compounds that bind to another subtype, PPAR␥, have also been identified. The antidiabetic drugs, thiazolidinediones and prostaglandin J2 derivatives, have been shown to preferentially bind to and activate PPAR␥, and these effects occur predominantly in adipose cells (11)(12)(13)(14)(15). In summary, roles for PPAR␣ include those involved in fatty acid catabolism, while those for PPAR␥ include those involved in adipogenesis.
Since PPARs are known to be key transcription factors of different genes participating in lipid homeostasis (16), we examined the effects of specific activators of PPARs on mRNA levels of the three putative fatty acid transporters. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture-The Fao cell line, a subclone of rat hepatoma HII4E, were cultured under conditions as reported previously (19). When a PPAR activator was added, a concentrated solution in Me 2 SO was added to medium before adding serum and sonicated to dissolve completely prior to adding to the cells. Treatment was initiated by changing the medium to a prewarmed drug-containing medium.
RNA Preparation and Northern Blot Analysis-Total RNA was prepared from the liver, intestine, adipose tissue, or the cultured hepatoma cells by the acid guanidium thiocyanate-phenol-chloroform extraction method (20). Northern blot analysis was carried out essentially as described previously (17). Among the cDNAs used for probes included peroxisomal HD, ␣2u-globulin, apolipoprotein E (apoE), and acyl-CoA oxidase, which have been described previously (17). The remaining cDNAs used for probes were obtained by cloning of PCR products of cDNA synthesized from poly(A) RNA isolated from the liver of Wy 14,643-fed mice. Their identities were confirmed by sequencing from both ends to 300 -400 bases inside after cloning into the SmaI site of pUC18. The synthetic oligonucleotides used to amplify respective cDNA sequences were 5Ј-TCTGACATTTGCAGGTCCATCTATGCTG and 5Ј-ATCTCAACCAGGCCCAGGAGCTTTATTT for FAT (corresponding to nucleotide number from 873 to 1410 of the published rat sequence (4) (GeneBank TM accession number L19658); 5Ј-CTTACGTGCTCCCCAG-TGTCTGGAAG and 5Ј-GGAGGAGGACACTCTGCTCTGGGATT for mAspAT (corresponding to nucleotide number from 134 to 538 of the published mouse sequence (5) (GeneBank TM accession number U82470)); 5Ј-GCCATTGTGGTGCACAGCAGGTACTA and 5Ј-TCGTG-TCCTCATTGACCTTGACCAGA for FATP (corresponding to nucleotide number from 775 to 1285 of the published mouse sequence (6)(U-15976)); 5Ј-GGGAGTTTGGCTCCAGAGTTTGACCG and 5Ј-GGAACA-CTTTGTAGGGCATCTGAGAGCG for lipoprotein lipase (corresponding to nucleotide number from 8 to 1000 of the published mouse sequence (21) (GeneBank TM accession numbers J03302 and J02740); 5Ј-AGGCT-TTTCTGAAAGGGTGAGGCATTTT and 5Ј-ACCATAGGAGTGGATGC-TAATGTGCCCT for leptin (corresponding to nucleotide number from 1422 to 2280 of the published sequence (22) (GeneBank TM accession number U18812); and 5Ј-CCTACAGATGTGGTAAAGGTCCGCTTCC and 5Ј-GAGTCATCAGTACAGAGGCACAGGGAGG for uncoupling protein 2 (corresponding to nucleotide number from 701 to 1368 of the reported mouse sequence (GeneBank TM accession number U69135). The cDNA clones for fatty acid-binding proteins were obtained by reverse transcription-PCR after adding oligo(dC) tails to the cDNA synthesized as above by terminal deoxynucleotidyl tansferase and dCTP. The PCR primers were oligo(dG) and three specific primers: 5Ј-GTAC-CAAGTGCAGAGCCAAGAGAACTTT for rat liver FAB corresponding to nucleotide number from 353 to 380 of the published sequence (23) (GeneBank TM accession number J00732), 5Ј-CCAAGCTTCCTCTTCAT-CACATTAATGCCCATTTT for mouse intestinal FAB corresponding to nucleotide number from 96 to 130 of the published sequence (24) (GeneBank TM accession number M65034), and 5Ј-TCATGTAATCATC-GAAGTTTTCACTGGAGACAAGC for mouse adipocyte FAB corresponding to nucleotide number from 604 to 638 of the published genomic sequence (25) (GeneBank TM accession number M13385). After hybridization, membranes were washed and autoradiographed with an intensifying screen at Ϫ80°C. For quantitative analysis, a BAS2000 image analyzer (Fujix, Japan) was used. [␣-32 P]dCTP (3000 Ci/mmol) and a random prime labeling kit were purchased from Muromachi Kagaku (Tokyo, Japan) and Takara (Osaka, Japan).

RESULTS AND DISCUSSION
All Three Fatty Acid Transporters, FAT, FATP, and mAspAT, Are Induced by the Peroxisome Proliferator Wy 14,643-FATP mRNA is highly expressed in skeletal muscle, heart, and fat, but only found at low levels in liver of normal mice (6). We observed that FATP mRNA is expressed significantly higher in the liver of Wy 14,643-treated mice as compared with controls during efforts to isolate the cDNA by reverse transcription-PCR. To examine the possibility that fatty acid transporters mRNAs are induced by the peroxisome proliferator Wy 14,643, we measured the time course effect of the drug on the levels of FAT, FATP, and mAspAT mRNAs in the liver and intestine (Fig. 1). Feeding a diet containing Wy 14,643 resulted in a large induction of peroxisomal HD mRNA in liver, reaching maximal level within a day as reported previously (26). In the intestine HD mRNA had not yet reached a plateau by 5 days. Intestinal fatty acid-binding protein mRNA was also gradually induced by Wy 14,643 and expressed only in the intestine. During these periods, the levels of apolipoprotein AI mRNA did not change significantly both in the liver and intestine. mRNAs for FAT, FATP, and mAspAT all responded to Wy 14,643 administration, but with different patterns. FAT, FATP, and mAspAT mRNAs were all gradually induced in the liver with a different time course than that of HD mRNA. FAT and mAspAT mRNA were also detected in higher levels in the intestine. Longer exposure of autoradiographic film revealed that FATP was also induced in the intestine with a similar time course as in the liver, but this increase by day 5 was less than 10% of the liver level (not shown). It should be pointed out that for FAT, we did observe two different sizes of mRNA as reported previously (4). Northern blot analysis showed that all three putative LCFA transport proteins FAT, FATP, and mAspAT were induced in a tissue-specific manner by the peroxisome proliferator Wy 14,643.

FAT and FATP mRNAs Are Induced in Mouse Liver by
Various PPAR␣ Activators-To examine whether PPAR␣ activators other than Wy 14,643 induce FAT and FATP in the liver, the effects of three other peroxisome proliferators were compared. The mRNA levels in liver from mice treated with these compounds for 5 days were determined by Northern blot analysis as shown in Fig. 2. All four PPAR␣ activators induced liver fatty acid binding protein compared with controls. In addition, peroxisomal HD mRNA was markedly increased in response to all 4 PPAR␣ activators consistent with previous work (not shown; see Ref. 17). Similarly, FAT mRNA was markedly induced by all four PPAR␣ activators compared with controls. Interestingly, mRNA of FATP and lipoprotein lipase were higher, and ␣2u-globulin was lower, as a result of exposure to Wy 14,643, clofibrate, or DEHA, while these effects were less pronounced or absent in mice treated with DEHP. This could be due to the fact that DEHP is less effective at activating PPAR␣dependent processes, including increasing reporter gene activity in transient co-transfections (27), peroxisome proliferation (28), increasing replicative DNA synthesis (29), and formation of hepatic tumors (29). In contrast, mAspAT mRNA levels were not significantly different in response to any of the PPAR␣ activators, although Wy 14,643 did cause a small increase in this mRNA compared with controls. These results suggest that expression of FAT and FATP, but not mAspAT, mRNA in the liver is under the control of PPAR␣.

FAT and FATP mRNAs Are Induced by Various PPAR␣ Activators but Not by a PPAR␥ Activator in Rat Hepatoma
Cells-To further examine the effects of a wide variety of PPAR activators on the induction of FAT and FATP mRNAs, we used the rat hepatoma Fao cells line that is responsive to peroxisome proliferators (19). Fao cells were treated with various PPAR activators, including Wy 14,643, ciprofibrate, troglitazone, carbacyclin, indomethacin, ibuprofen, and perfluorooleate for 6 or 24 h, and the levels of FAT, FATP, and mAspAT mRNAs, together with control mRNAs, were measured by Northern blots as shown in Fig. 3. Fao cells responded to most PPAR␣ activators examined (Wy 14,643, ciprofibrate, indomethacin, ibuprofen, and perfluorooleate) as assessed by the increased levels of peroxisomal acyl-CoA oxidase and HD mRNAs. Time courses of induction was different among the activators, and indomethacin was the least effective. FAT and FATP mRNAs were increased by PPAR␣ activators similarly as the peroxisomal mRNAs were while the effect of indomethacin on these mRNAs was negligible. Thus the relative abilities of various PPAR␣ activators to induce FAT and FATP mRNAs were of similar magnitude to those observed with peroxisomal enzyme mRNAs. Levels of mAspAT and apoE mRNA were not affected by any treatment. The results obtained using cultured hepatoma cells are consistent with those observed in the liver of mice treated with PPAR␣ activators. In contrast to the PPAR␣ activators, a typical PPAR␥ activator Troglitazone did not cause any change in the levels of all the mRNAs examined. Combined, these results suggest that both the FAT and FATP genes are PPAR␣ target genes.
Induction of FAT and FATP in the Liver Is Mediated by PPAR␣-To directly examine the role of PPAR␣ in the induction of FAT and FATP mRNAs, we utilized the PPAR␣-null mouse (7). Northern blot analyses of liver RNA from (ϩ/ϩ) and (Ϫ/Ϫ) mice fed either a control diet or one containing 0.1% Wy 14,643 are shown in Fig. 4. FAT and FATP mRNAs as well as HD mRNA were higher in the wild-type mice (ϩ/ϩ) fed Wy 14,643, and this effect was not observed in the (Ϫ/Ϫ) mice fed Wy 14,643. In addition, ␣2u-globulin mRNA was significantly lower in (ϩ/ϩ) mice fed Wy 14,643 but unaffected by Wy 14,643 feeding in (Ϫ/Ϫ) mice as reported previously (17). These results indicate that PPAR␣ has an obligatory role in Wy 14,643 induction of FAT and FATP mRNAs in the mouse liver. This could be due to direct interaction of PPAR␣ with peroxisome proliferator responsive elements, although peroxisome proliferator responsive elements have not been indentified in these genes to date. Alternatively, the increase in gene transcription of FATP as a result of PPAR␣ activators (30) could be the result of other PPAR␣-dependent events that indirectly result in altering gene expression of fatty acid transport proteins to compensate for the increase in fatty acid catabolism.
FAT and FATP Are Induced by PPAR␥ Activators in White Adipose Tissue-Although FAT and FATP mRNAs were induced by PPAR␣ activators in hepatoma cells and in a PPAR␣dependent manner in the liver, the possible involvement of PPAR␥ in other tissues where the ␥-type receptor is predominant over the ␣-type cannot be excluded (31). Therefore, we examined the effect of PPAR␥ activators in adipose tissue of mice. We used KKAy obese mice (18) for this purpose to facilitate adipose tissue preparation. We utilized two white adipose tissue stores, interintestinal and subcutaneous, and brown adipose tissue. Northern blots from the three adipose tissues are shown in Fig. 5. Changes in the levels of several mRNAs were different among the three adipose stores. Brown adipose did not respond to the PPAR␣ or PPAR␥ activators compared with controls. Basal levels of FAT, FATP, A-FAB (adipose-type fatty acid-binding protein), lipoprotein lipase, and leptin mRNAs were lower in interintestinal white adipose than those in subcutaneous white adipose tissue. The greatest effect induced by PPAR activators was observed in interintestinal adipose. In both white adipose tissues, FAT and FATP were induced most efficiently by pioglitazone followed by troglitazone and then clofibrate, just as other characterized PPAR␥ target genes such as A-FAB, lipoprotein lipase, and leptin. Thus, both FAT and FATP were induced by PPAR␥ activators in white adipose tissue.
Comparison of FAT and FATP-The responses of both putative LCFA transporter mRNAs to PPAR activators were similar, but not exactly coordinated. The differences were: 1) FATP was not induced as much in the intestine as FAT (see Fig. 1); 2) most peroxisome proliferators induced both mRNAs in the liver, but DEHP only induced FAT (see Fig. 2); and 3) FATP responded by rapid or transient induction, but FAT did not (see the lanes of RNA samples from the cells treated for 6 h with Wy 14,643, carbacyclin, and perfluorooleate in Fig. 3). These results suggest a complexity in the PPAR-mediated transcriptional activation (32) and/or differences in their physiological roles.
During the completion of this manuscript, another paper concerning FATP and PPAR activators was published (30). Their conclusion on regulation of the expression of FATP is essentially the same as ours. We extend these observations by suggesting that FAT, in addition to FATP, may have an important role in fatty acid uptake and lipid homeostasis. Furthermore, we demonstrate that in liver, PPAR␣ is required for the induction of FAT and FATP, since these effects were absent in PPAR␣-null mice fed Wy 14,643. Functional analysis of both proteins will be necessary to better understand their roles in lipid homeostasis.