Altered Constitutive Expression of Fatty Acid-metabolizing Enzymes in Mice Lacking the Peroxisome Proliferator-activated Receptor α (PPARα)*

Peroxisome proliferator-activated receptor α (PPARα) is a member of the steroid/nuclear receptor superfamily and mediates the biological and toxicological effects of peroxisome proliferators. To determine the physiological role of PPARα in fatty acid metabolism, levels of peroxisomal and mitochondrial fatty acid metabolizing enzymes were determined in the PPARα null mouse. Constitutive liver β-oxidation of the long chain fatty acid, palmitic acid, was lower in the PPARα null mice as compared with wild type mice, indicating defective mitochondrial fatty acid catabolism. In contrast, constitutive oxidation of the very long chain fatty acid, lignoceric acid, was not different between wild type and PPARα null mice, suggesting that constitutive expression of enzymes involved in peroxisomal β-oxidation is independent of PPARα. Indeed, the PPARα null mice had normal levels of the peroxisomal acyl-CoA oxidase, bifunctional protein (hydratase + 3-hydroxyacyl-CoA dehydrogenase), and thiolase but lower constitutive expression of the D-type bifunctional protein (hydratase + 3-hydroxyacyl-CoA dehydrogenase). Several mitochondrial fatty acid metabolizing enzymes including very long chain acyl-CoA dehydrogenase, long chain acyl-CoA dehydrogenase, short chain-specific 3-ketoacyl-CoA thiolase, and long chain acyl-CoA synthetase are also expressed at lower levels in the untreated PPARα null mice, whereas other fatty acid metabolizing enzymes were not different between the untreated null mice and wild type mice. A lower constitutive expression of mRNAs encoding these enzymes was also found, suggesting that the effect was due to altered gene expression. In wild type mice, both peroxisomal and mitochondrial enzymes were induced by the peroxisome proliferator Wy-14,643; induction was not observed in the PPARα null animals. These data indicate that PPARα modulates constitutive expression of genes encoding several mitochondrial fatty acid-catabolizing enzymes in addition to mediating inducible mitochondrial and peroxisomal fatty acid β-oxidation, thus establishing a role for the receptor in fatty acid homeostasis.

Peroxisomes are single membrane-bound subcellular organelles that contain a variety of enzymes involved in a number of metabolic processes (1). The most well characterized reactions carried out by peroxisomes are those that catalyze in fatty acid ␤-oxidation. Since plants lack mitochondria, peroxisomes are solely responsible for their fatty acid ␤-oxidation. The peroxisomal fatty acid ␤-oxidation pathway produces hydrogen peroxide through the activity of acyl-CoA oxidase, thus historically accounting for the name "peroxisomes." Typically, H 2 O 2 is decomposed to molecular oxygen and water by catalase and glutathione peroxidase. Human genetic deficiencies in peroxisome biogenesis and individual peroxisomal enzymes have been described that result in accumulation of long chain fatty acids (2). The most severe of the peroxisome deficiencies causes neurological and anatomical abnormalities.
In addition to fatty acid oxidation, peroxisomes also carry out ␤-oxidation of the cholesterol side chain during the synthesis of bile acids and participate in the biosynthesis of cholesterol (3), ether glycolipids, and dolichols. Catabolism of purines, polyamines, glyoxylate and certain amino acids have been attributed to peroxisome-localized enzymes. Thus, peroxisomes are essential organelles for maintaining cellular and organismal homeostasis.
Peroxisome proliferation is most pronounced in liver, kidney, and heart. In liver, the number of peroxisomes increases from about 500 -600/cell to Ͼ5,000/cell after exposure to peroxisome proliferators (1). This is accompanied by an increase in cell volume and cell number, resulting in hepatomegaly. Coincident with an increase in the number of peroxisomes, several peroxisomal enzymes are induced by transcriptional activation (4). Transcription of genes encoding the key ␤-oxidation enzymes acyl-CoA oxidase, enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (bifunctional enzyme), and thiolase are markedly elevated as a result of treatment with peroxisome proliferators (5). Genes encoding the microsomal cytochrome P450 in the CYP4A family are also activated by these agents (6). Changes in peroxisomal and microsomal gene expression induced by peroxisome proliferators are mediated by the peroxisome proliferator-activated receptor (PPAR), 1 a member of the nuclear receptor superfamily. Three distinct PPARs have been found, designated PPAR␣, ␦ (also called NUC-1 and ␤), and ␥. Tissue distribution of each receptor is different, suggesting that each has unique functions. In rodents, PPAR␣ is abundant in the liver, kidney, and heart, all of which display peroxisome proliferation in response to PPAR␣ activators and have high rates of lipid metabolism (7). Expression of PPAR␦ is ubiquitous and is highly expressed in the central nervous system (7). The role of PPAR␦ is not known. PPAR␥ and PPAR␥2, resulting from differential mRNA splicing, are present predominantly in adipose tissue and spleen. PPAR␥ is responsible in part for adipocyte differentiation and regulation of adipocyte-specific genes (8). It is also the target for the thiazolidinedione drugs that increase insulin sensitivity of target tissues (9).
To determine the function of PPAR␣ and its role in peroxisome proliferation and hepatocarcinogenesis, a PPAR␣ null mouse was generated (10). These animals exhibit a normal phenotype and normal basal levels of hepatic peroxisomes. However, the PPAR␣ null mouse is nonresponsive to peroxisome proliferation. Compared with wild type mice, administration of peroxisome proliferators to PPAR␣ null mice does not cause an increase in the number of peroxisomes, hepatomegaly, nor increases in mRNA encoded by target genes. Furthermore, these mice do not display physiological, toxicological, or carcinogenic responses induced by peroxisome proliferators (11)(12)(13). Interestingly, abnormal hepatic lipid accumulation was initially reported in the PPAR␣ null mice, suggesting an alteration in lipid metabolism (10). To investigate the biochemical basis for altered lipid metabolism, constitutive levels of peroxisomal and mitochondrial fatty acid-metabolizing enzymes were examined in the PPAR␣ null mouse. Animals and Wy-14,643 Treatment-PPAR␣ null mice on an Sv/129 genetic background were produced as described (10). Wild type Sv/129 were used as controls in all experiments. Mice were fed either a control diet or one containing 0.1% Wy-14,643 for 2 weeks.
Fatty Acid ␤-Oxidation Activity-Fatty acid ␤-oxidation activity was measured by the method of Shindo et al., (14). Briefly, unfrozen livers were homogenized in four volumes of 0.25 M sucrose containing 1 mM EDTA in a Potter-Elvehjem homogenizer using a tight-fitting teflon pestle. Approximately 500 g of homogenate was incubated with the assay medium in 0.2 ml of 150 mM potassium chloride, 10 mM HEPES, pH 7.2, 0.1 mM EDTA, 1 mM potassium phosphate buffer, pH 7.2, 5 mM Tris malonate, 10 mM magnesium chloride, 1 mM carnitine, 0.15% bovine serum albumin, 5 mM ATP, and 50 M each fatty acid (5.0 ϫ 10 4 cpm of radioactive substrate). The reaction was run for 30 min at 25°C and stopped by the addition of 0.2 ml of 0.6 N perchloric acid. The mixture was centrifuged at 2,000 ϫ g for 10 min, and the unreacted fatty acid in the supernatant was removed with 2 ml of n-hexane using three extractions. Radioactive degradation products in the water phase were counted. In some experiments, 20 M POCA or 2 mM potassium cyanide was added to the incubation mixture to inhibit mitochondrial ␤-oxidation activity. Fatty acid ␤-oxidation activity was expressed as nmol/min/liver.

Analysis of Fatty
Acid-metabolizing Enzymes-To identify specific fatty acid-metabolizing enzymes that were influenced by PPAR␣, antibodies were used to measure protein levels on immunoblots ( Fig. 1, Table I). Constitutive expression of several enzymes (VLCAD, LCAD, LACS, and T1) were lower by 30 -60% in untreated PPAR␣ null mice as compared with untreated wild type mice. Curiously, constitutive expression of the SCHAD was higher by about 4-fold in the PPAR␣ null mouse liver as compared with wild type mice. Other mitochondrial enzymes examined were expressed at similar levels in untreated PPAR␣ null and wild type mice. The expression of all mitochondrial, microsomal, and cytosolic fatty acid-metabolizing enzymes except for MH were increased in wild type mice fed Wy-14,643 compared with controls, with levels of induction ranging from 1.7-fold for T2 to 4.7-fold for SCHAD. The expression of CTE II was totally dependent on Wy-14,643 treatment in wild type mice. In the PPAR␣ null animals, there was no increase in expression of these enzymes after feeding Wy-14,643 for 2 weeks (Fig. 1, Table I).
Constitutive levels of hepatic peroxisomal fatty acid ␤-oxidation enzyme expression of AOX, PH, PT, and VLACS was not significantly different between wild type and PPAR␣ null mice (Fig. 1, Table I), although DBF was 36% lower in the PPAR␣ null animals. All of the peroxisomal enzymes were induced from 3-7-fold in the wild type mice, and this effect was not observed in the PPAR␣ null mice.
Expression of mRNAs-To determine whether the lower expression of fatty acid-metabolizing enzymes and ME is due to altered gene expression, hepatic mRNA levels were analyzed by Northern blots (Fig. 2). Constitutive levels of VLCAD, LCAD, ME, and LACS mRNA were lower in the PPAR␣ null mice compared with controls, consistent with the protein measurements. It is noteworthy that hepatic levels of mRNA for SCHAD were not different between untreated PPAR␣ null and wild type mice even though the protein levels were increased by 4-fold in the null mouse. Levels of the mRNAs encoding, LCAD, and ME were significantly induced in wild type mice by Wy-14,643. These data are also consistent with the Western blot analysis. In contrast, the mRNAs encoding LACS and SCHAD were not significantly increased by the drug. In the PPAR␣ null mice, there was no difference in mRNA levels for any of the enzymes after treatment with Wy-14,643.
Analysis of Overall Fatty Acid ␤-Oxidation Activity-As shown in Table I, six enzymes involved in fatty acid ␤-oxidation had lower constitutive expression in the PPAR␣ null mice. Four of the six (VLCAD, LCAD, LACS, and DBF) have highest catalytic activities with long chain fatty acid substrates (25)(26)(27)34), whereas the other two (SCHAD and T1) are more active with short and medium chain fatty acids (21,30). To evaluate the significance of the altered fatty acid ␤-oxidation enzymes, total hepatic ␤-oxidation was measured using lauric acid (C-12), palmitic acid (C-16), and lignoceric acid (C-24). Compared with wild type controls, the PPAR␣ null mice basal levels of total fatty acid ␤-oxidation was lower with palmitic acid as a substrate; there was no difference in metabolism of lauric acid and lignoceric acid (Fig. 3). Wy-14,643 feeding caused a significant increase in metabolism of all three fatty acids in wild type animals. No induction was observed in PPAR␣ null mice, consistent with the results found with the enzymes levels ( Fig. 1 and Table I). Results were identical whether the data were calculated per liver protein or per liver. These data provide evidence that the lower constitutive expression of several long chain-specific fatty acid ␤-oxidation enzymes in the PPAR␣ null mice compared with wild type mice (Table I)

TABLE I Western immunoblot quantitation of hepatic fatty acid-metabolizing enzymes in wild type (ϩ/ϩ) and PPAR␣-null (Ϫ/Ϫ) mice
Mice were fed either a control diet or 0.1% Wy-14,643 for 14 days. Total liver cell extract was subjected to electrophoresis, and proteins were transferred to nitrocellulose membranes and screened with specific antibodies. The signals were quantified by scanning densitometry, and the values from (ϩ/ϩ) mice fed control diets were assigned the number 1.0. Results are the means Ϯ S.D. of three determinations. ND, not detected. affects long chain fatty acid oxidation. The lack of a difference in constitutive oxidation of the very long chain-specific fatty acid, lignoceric acid, carried out by peroxisomal enzymes supports the finding of no effect of PPAR␣ on expression of these enzymes except DBF in untreated null mice (Table I).
To confirm that palmitic acid is preferentially metabolized by mitochondrial enzymes, the inhibitors KCN, an inhibitor of the mitochondrial respiratory chain, and POCA, a potent inhibitor of carnitine palmitoyl-CoA transferase I, were employed (Table  III). Both compounds inhibited palmitic acid oxidation by 70 -93% in either mouse line, irrespective of Wy-14,643 administration. These data confirm that the contribution of peroxisomal enzymes to palmitic acid ␤-oxidation was minimal.
Time Course of Induction-The kinetics of VLCAD and ME mRNA and protein expression were determined after administration of Wy-14,643. VLCAD mRNA and protein were rapidly induced within 1 day after Wy-14,643 treatment in wild type mice (Fig. 4A). No induction of VLCAD was found in PPAR␣ null mice after 14 days of feeding Wy-14,643. Levels of VLCAD mRNA and protein decreased to about 2-fold after 5 days of feeding yet remained elevated up to 14 days of feeding. A similar time course of induction was also observed in protein levels of several mitochondrial fatty acid ␤-oxidation enzymes (LCAD, SCAD, TP␣, and TP␤). Induction of ME mRNA and enzyme activity reached maximal levels of 4-fold after 7 days of Wy-14,643 feeding in the wild type mice; neither mRNA nor activity were induced in the PPAR␣ null animal (Fig. 4B). Similar time course of induction was observed in protein levels of several other enzymes (LACS, AOX, PH, PT, DBF, VLACS, MCAD, and CTE II). Thus, the kinetics of the increase in ME expression is slower than that of VLCAD. The reason for this differential induction is not presently known. DISCUSSION Constitutive expression of VLCAD, LCAD, SCHAD, T1, LACS, DBF, and ME is regulated by PPAR␣ since their abundance was significantly altered in the absence of PPAR␣ as compared with wild type controls. With the exception of SCHAD, which was up-regulated in the PPAR␣ null mice, all of these proteins were found at lower levels in the PPAR␣ null mice. This shows that PPAR␣ has an important role in regulating basal levels of these enzymes involved in fatty acid ␤-oxidation and ME that participates in fatty acid synthesis. Although the mechanism for this peroxisome proliferator-independent mechanism is not known, it may be a result of altered gene expression since PPAR␣ is known to control transcription through interaction with peroxisome proliferator response elements (4). The down-regulation is selective since constitutive expression of other enzymes including MCAD, SCAD, TP␣, TP␤, MH, T2, CPT II, and MTE I appear to be unaffected by loss of the receptor. This is similar to the peroxisomal enzymes AOX, PH, PT, and VLACS, where there was no difference in expression levels between the untreated wild type and PPAR␣ null mice. The data on enzyme levels are supported by the results of total fatty acid metabolism in liver where oxidation of long chain fatty acid, palmitate, which is reflective of mitochondrial metabolism, was lower in the PPAR␣ null mice, whereas oxidation of the very long chain fatty acid, lignocerate, was not different between the wild type and PPAR␣ null mice. The lack of difference in metabolism of this very long chain fatty acid is almost certainly due to similar levels of the peroxisomal enzymes in the two genotypes. Lower constitutive expression of fatty acid-metabolizing enzymes and mitochondrial palmitic acid ␤-oxidation suggests that PPAR␣ controls gene expression in the absence of exogenous ligands for the receptor, and mice lacking PPAR␣ have an impaired ability to metabolize lipids.
To further elucidate the role of PPAR␣ in lipid metabolism, the effect of the prototypical peroxisome proliferator Wy-14,643 in PPAR␣ null mice was investigated. Indeed, PPAR␣ was shown to transactivate genes in the presence of peroxisome proliferators (40 -42). In addition to the peroxisomal ␤-oxidation enzymes and microsomal fatty acid hydroxylase P450, liver fatty acid binding protein and the genes encoding MCAD (43), 3-hydroxy-3-methylglutaryl-CoA synthase (44), and ME (45) are also activated by PPAR␣ as indicated by transactivation

FIG. 2. Northern analysis of hepatic mRNAs.
Representative samples from three separate mice were used. Diet and genotype are indicated. Total RNA (5.4 g) from three representative mice from each group were electrophoresed on a denaturing gel and probed using cDNAs for VLCAD (panel A), LCAD (panel B), ME (panel C), LACS (panel D), and SCHAD (panel E). The blots were exposed to autoradiographic film for 5 days (panels A-D) and 1 day (panel E).
assays. The present study extends these observations by confirming that expression of VLCAD, LCAD, SCAD, TP␣, TP␤, MTE I, CTE II, SCHAD, T1, T2, LACS, and CPT II are all higher as a result of Wy-14,643 feeding in wild type mice but not in PPAR␣ null mice. Northern analysis of mRNA encoding some of these enzymes revealed that induction is most likely due to increases in mRNA. These observations demonstrate that changes in gene expression of proteins involved in lipid metabolism are mediated by PPAR␣ after exposure to peroxisome proliferators. Indeed, peroxisome proliferator response elements have been found and shown to be functionally active in the MCAD (43) and LACS (46) genes. A peroxisome proliferator response element has not been found in the LCAD (47), even though it is induced in wild type mice by Wy-14,643 feeding. Expression of ME is also elevated at the mRNA and protein level in agreement with a role of PPAR␣ in its regulation (45).
The fibrate class of drugs can also lead to suppression of gene expression of numerous proteins involved in lipid metabolism (4). Levels of apolipoprotein A-I, apolipoprotein C-III, apolipoprotein A-IV, hepatic lipase, and lecithin cholesterol acyltransferase are all lowered by treatment of mice with fibrate drugs (4,48,49). The alterations of these proteins in addition to altered gene expression of peroxisomal fatty acid ␤-oxidizing enzymes are thought to contribute to the lipid-lowering effects of hypolipidemic drugs (4). In addition, it was recently shown that the down-regulation of apolipoprotein C-III mRNA and protein that contributes to the triglyceride-lowering effect of Wy-14,643 is mediated by PPAR␣ (50). The results presented here extend these observations by demonstrating that the peroxisome proliferator Wy-14,643 induces significant changes in many fatty acid ␤-oxidizing enzymes and total hepatic ␤-oxidation, which in turn are likely to further contribute to the triglyceride-lowering effect of the fibrate class of hypolipidemic drugs. Combined, these results establish that PPAR␣ functions in the control of lipid homeostasis in mice by regulating constitutive and inducible expression of fatty acid catabolism. Since nuclear receptors usually require a ligand for gene activation, the constitutive control of the enzymes would suggest that an endogenous ligand exists in liver.
The lower expression of mRNAs encoding several fatty acidmetabolizing enzymes in PPAR␣ null mice suggests either the presence of an endogenous ligand that preferentially controls genes encoding fatty acid-metabolizing enzymes and ME or that a ligand-independent mechanism is involved. Indeed, phosphorylation of a Ser-112 in PPAR␥ through mitogen-activated protein kinase has been shown to modulate its activity (51,52). This kinase recognition site is conserved between PPAR␥ and PPAR␣. Irrespective of the mechanism of PPAR␣ activation, these results suggest that it differentially activates genes in the absence of exogenous ligands.
The identification of endogenous ligands has recently been addressed. It was shown that PPAR␣ participates in the control of the inflammatory response involving leukotriene B 4 (53). These studies also established that leukotriene B 4 can directly bind to recombinant PPAR␣. Possible direct binding of peroxisome proliferators was shown by induced conformational changes as detected by protease sensitivity of in vitro translated PPAR␣ (54). Other indirect transactivation experiments suggest that PPAR␣ may mediate the action of 8(S)-hydroxyeicosatetraenoic acid (55). It is likely that other ligands for PPAR␣ exist, including fatty acid metabolites, as first suggested by the ability of fatty acids to transactivate the receptor (56,57). Evidence exists for the presence of endogenous PPAR␣ activators in cultured cells used for transactivation studies. High background levels of reporter gene activation are usually  found in these types of studies (41,58). Further support for the existence of endogenous ligands was provided by the demonstration that unsaturated fatty acids can bind to PPAR␣ (59). Thus, constitutive regulation of genes encoding fatty acid-metabolizing enzymes and ME may be mediated by levels of one or more fatty acid metabolites. These metabolites may be important endogenous ligands that function in the control of fatty acid metabolism. Support for this idea was provide by the observation that dietary polyunsaturated fatty acids induce AOX and CYP4A P450 in a PPAR␣-dependent mechanism (11). Taken together, these studies suggest that PPAR␣ may have several endogenous ligands, which upon binding to the receptor result in the activation of fatty acid catabolism including the oxidative degradation of leukotrienes, arachidonic acid epoxides, and other fatty acid derivatives. Among the important issues that need to be addressed is the species differences in response to peroxisome proliferators (60). Mice and rats are highly susceptible to peroxisome proliferation and hepatocarcinogenesis, whereas nonhuman primates and humans appear to be resistant. The mechanism of this species differences is not presently known, but it might be due to lower hepatic levels of PPAR␣ (61,62). Despite the lack of demonstratable peroxisome proliferation, the fibrate drugs are highly effective lipid-lowering agents in humans (63). Thus, PPAR␣ may differentially regulate expression of genes that cause peroxisome proliferation and genes encoding enzymes that are responsible for fatty acid mobilization, transport, and catabolism. For example, high cellular levels of receptor may activate genes encoding peroxisomal, mitochondrial, and microsomal fatty acid metabolism in addition to genes that directly or indirectly control the cell cycle and peroxisome proliferation.