P450CYP2Cepoxygenase andCYP4Aω-hydroxylase mediate ciprofibrate-induced PPARα-dependent peroxisomal proliferation

Peroxisomal proliferators, such as ciprofibrate, are used extensively as effective hypolipidemic drugs. The effects of these compounds on lipid metabolism require ligand binding activation of the peroxisome proliferator-activated receptor (PPAR) α subtype of nuclear receptors and involve transcriptional activation of the metabolic pathways involved in lipid oxidative metabolism, transport, and disposition. ω-Hydroxylated-eicosatrienoic acids (HEETs), products of the sequential metabolism of arachidonic acid (AA) by the cytochrome P450 CYP2C epoxygenase and CYP4A ω-hydroxylase gene subfamilies, have been identified as potent and high-affinity ligands of PPARα in vitro and as PPARα activators in transient transfection assays. Using isolated rat hepatocytes in culture, we demonstrate that specific inhibition of either the CYP2C epoxygenase or the CYP4A ω-hydroxylase abrogates ciprofibrate-induced peroxisomal proliferation, whereas inhibition of other eicosanoid-synthesizing pathways had no effect. Conversely, overexpression of the rat liver CYP2C11 epoxygenase leads to spontaneous peroxisomal proliferation, an effect that is reversed by a CYP inhibitor. Based on these results, we propose that HEETs may serve as endogenous PPARα ligands and that the P450 AA monooxygenases participate in ciprofibrate-induced peroxisomal proliferation and the activation of PPARα downstream targets.

for PPARa (14). These P450 isoforms play important physiological roles in the control of tissue and body homeostasis (15). CYP2C P450s catalyze the regioselective and enantioselective metabolism of AA to epoxyeicosatrienoic acids (EETs) and account for most of the epoxygenase activity in rat and human liver and kidney (16). EETs have vasoactive properties, modulate ion channels (17), and function as second messengers of pathways, such as epidermal growth factor-dependent signaling (18). EETs are excellent substrates for CYP4A isoforms (14), which selectively hydroxylate saturated and unsaturated fatty acids in the v/v-1 position (19)(20)(21). Some of the CYP4A isoforms are under the control of PPARa, which regulates an adaptive CYP4A induction in response to starvation and diabetes (22). Cyp4A is also under the control of other physiological stimuli, including fatty acids and hormones (23).
We report here that inhibition of the CYP2C epoxygenase or the CYP4A v-hydroxylase abrogates ciprofibrateinduced peroxisomal proliferation in isolated rat hepatocytes. Conversely, overexpression of the rat liver CYP2C11 epoxygenase leads to spontaneous peroxisomal proliferation. Our data suggest that HEETs are endogenous PPARa ligands and that the P450 AA monooxygenases participate in ciprofibrate-induced peroxisomal proliferation.

Materials
Ciprofibrate was provided by the Sterling-Winthrop Institute. Williams E medium and FBS were obtained from GIBCO (Grand Island, NY). Indomethacin, lipoxygenase-specific inhibitors, and EET standards were from Cayman Chemical (Ann Arbor, MI). All other drugs and chemicals were obtained from Sigma Chemical Co. (St. Louis, MO). PPARa-specific antibodies were purchased from Affinity BioReagents (Golden, CO). Antibodies against the integral peroxisomal membrane protein (PMP70) and catalase were a kind gift from Dr. M. Santos (Department of Cellular and Molecular Biology, Faculty of Biological Sciences, P. Universidad Catolica de Chile). Antibodies against alkyl-dihydroxyacetone phosphate synthase (ADAPS) were a generous gift from Dr. W. Just (Biochemie-Zentrum, Ruprecht-Karls-Universität, Heidelberg, Germany).

Primary cultures
Rat hepatocytes were isolated as described (24). Cell viability was 80-90%, as determined by trypan blue exclusion. Isolated hepatocytes were seeded at 7.5 3 10 5 cells in 3.5 cm plastic dishes coated with collagen and cultured in Williams E medium supplemented with 5% FBS, as described previously (5). At 8 h, the medium was replaced by drugs containing medium. Inhibitor drugs were dissolved in DMSO, yielding a final DMSO concentration of 0.01%. For [1-14 C]AA metabolic experiments, cells were incubated for various times in medium containing 4 mM AA and 0.25 mCi/dish [1-14 C]AA (53 mCi/mmol; Perkin-Elmer Life and Analytical Science, Boston, MA). Cells were fixed with 1 ml of methanol containing 0.1% acetic acid and triphenylphosphine (0.1 mM) and extracted twice with 2 ml of chloroform, according to published methods (25). The combined organic phases were evaporated under argon and dissolved in 300 ml of 50% acetonitrile containing 0.1% acetic acid. Separation of [1-14 C] AA from its radioactive metabolites was performed by reverse-phase HPLC with a 5 m C18 symmetric 300 column (4 3 250 mm; Waters Corp., Milford, MA) and a linear solvent gradient from initially acetonitrile-water-acetic acid (49.9:49.9:0.1, v/v/v) to acetonitrile-acetic acid (99.9:0.1, v/v) over 40 min at 1 ml/min (26). Fractions of 0.8 ml were collected, and their radioactivity was determined by liquid scintillation counting. Chemical synthesis of [1-14 C]14,15-EET was performed according to published methods and purified by reverse-phase HPLC (25,27

Reporter expression assays, plasmids, and adenoviral infections
A reporter plasmid containing three tandem repeats of the peroxisome proliferator response element (PPRE) from the AOX gene, fused to the herpes virus thymidine kinase promoter upstream of the coding sequence for luciferase, and an expression vector containing murine PPARa were a kind gift of Dr. R. M. Evans (Howard Hughes Medical Institute, Gene Expression Laboratory, Salk Institute for Biological Studies, La Jolla, CA). The expression vector containing only the coding sequence for luciferase under the control of the herpes virus thymidine kinase promoter (Clontech, Palo Alto, CA) was used as a control. Transient transfections of CHO cells were carried out according to the manufacturer's instructions, using LipofectAmine 2000 (Life Technologies, Carlsbad, CA). CHO cells were grown in DMEM, supplemented with 4% FBS until 70-80% confluent, and transfected with 0.3 mg of reporter plasmid or cotransfected with the reporter plasmid, 0.2 mg of the PPARa plasmid and 40 ng of a CMV-b-Gal vector (Clontech, Palo Alto, CA), for normalization. The recombinant adenovirus containing the predominant rat liver arachidonate epoxygenase (CYP2C11), fused to rat P450 oxidoreductase (Ad-EPOX) (29) and the bacterial b-galactosidase (Ad-bgal) transgenes, was under the control of the cytomegalovirus promoter. Antibodies against CP2C11 and b-galactosidase, as well as CYP2C and CYP4A P450 inhibitors, were kindly donated by Dr. Jorge Capdevila (Department of Medicine, Vanderbilt University Medical School, Nashville, TN). Large-scale production of recombinant adenoviruses was performed from HEK 293 infected cells, as described (30). For adenoviral infections, primary rat hepatocyte cultures were plated on 3.5 cm tissue culture dishes (200,000 total). Twenty-four hours after plating, cells were infected with 7 3 10 7 viral particles/cm 2 . After 2 h of infection, unbound virus was removed and replaced with fresh medium. The cells were incubated at 37jC in 5% CO 2 for 48 h. Cells were then harvested and processed for blotting or immunocytochemistry.

RESULTS
Cytochrome P450 general inhibitors abrogate ciprofibrate-induced peroxisomal proliferation in cultured rat hepatocytes As stated previously, eicosanoid products of the cyclooxygenase, lipoxygenase, and cytochrome P450 pathways have been proposed as endogenous PPARa activators (11,(31)(32)(33)(34). On this basis, we first determined whether inhibitors of these pathways had any effect on ciprofibrateinduced peroxisomal proliferation in isolated rat hepatocytes. We found that the CYP P450 general inhibitors, clotrimazole and ketoconazole (35), strongly decrease ciprofibrateinduced mRNA expression of the peroxisomal proliferation markers AOX and the peroxisomal membrane protein PMP70 (36) in a concentration-dependent manner, using GAPDH as a control (Fig. 1). Although GAPDH increases its expression upon long-term treatment of rats with some PPs (37), we did not observe significant changes in its expression in our short-term (12 h) experiments. Under the same experimental conditions, neither the specific lipoxygenase inhibitors, caffeic acid [a 5-and a 12-lipoxygenase inhibitor (38)], baicalein [a 12-lipoxygenase inhibitor (39)], or sculetin [a 15-lipoxygenase inhibitor (40)], nor the selective cyclooxygenase inhibitor indomethacin (41), all in the 10-100 mM concentration range, had any effect on the ciprofibrate-induced increase of peroxisomal proliferation markers (data not shown).
Selective inhibitors of cytochrome P450 isoforms CYP4A and CYP2C abrogate ciprofibrate-induced AOX and PMP70 mRNA expression without affecting PPARa transcriptional activity or its mRNA expression and with minor or no effect on CYP isoform protein expression Next, we used 6-(2-propargyloxyphenyl) hexanoic acid (PPOH) and N-methylsulfonyl-12,12-dibromododec-11enamide (DDMS), specific inhibitors of CYP4A-dependent v-hydroxylation and CYP2C-dependent epoxidation, respectively (42), to dissect the P450 pathways involved in the inhibition of peroxisomal proliferation marker induction by ciprofibrate. Both inhibitors abrogate the ciprofibrateinduced expression of AOX and PMP70 in a concentrationdependent manner ( Fig. 2A, B). To evaluate whether these inhibitors could act as direct antagonists of PPARa, Fig. 1. Cytochrome P450 general inhibitors abrogate ciprofibrate-induced peroxisomal proliferation in cultured rat hepatocytes. Control and ciprofibrate-treated hepatocytes (100 mM) were cultured for 12 h in the absence or presence of increasing concentrations of the general P450 inhibitors ketoconazole (KE) and clotrimazole (CO). The mRNA expression levels of acyl-coenzyme A oxidase (AOX) and peroxisomal membrane protein (PMP70) were evaluated by Northern blot. A: A representative Northern blot is presented, with duplicates from different culture dishes in each case. The expression of GAPDH was used as a control. B: Northern blots were quantified by scanning densitometry. Data represent means 6 SD of three independent experiments, corrected using GAPDH expression and expressed relative to mRNA expression in the absence of ciprofibrate and inhibitors. Asterisks indicate statistically significant differences from ciprofibrate-treated control cells (P , 0.05 or less).
we transiently cotransfected CHO cells with a PPREluciferase reporter plasmid and a mouse PPARa expression plasmid (see Materials and Methods), and the transfected cells were treated for 24 h with 100 mM ciprofibrate in the presence or absence of increasing concentrations of DDMS or PPOH. No effect on the relative ciprofibrate-induced luciferase activity was observed for either drug (Fig. 2C), showing that they do not affect PPARa transcriptional activation. We used CHO cells because of the low efficiency of cultured hepatocytes in transfection experiments with the reporter Fig. 2. Selective inhibitors of cytochrome P450 isoforms CYP4A and CYP2C abrogate ciprofibrate-induced AOX and PMP70 mRNA expression without affecting peroxisome proliferator-activated receptor a (PPARa) transcriptional activity or its mRNA expression and with minor or no effect on CYP isoform protein expression. Hepatocytes were treated with or without 100 mM ciprofibrate for 12 h in the absence or presence of increasing concentrations of either 6-(2-propargyloxyphenyl) hexanoic acid (PPOH) or N-methylsulfonyl-12,12dibromododec-11-enamide (DDMS), specific inhibitors of v-hydroxylation and epoxidation, respectively. A: A representative Northern blot is presented. B: Northern blots from three independent experiments were quantified by scanning densitometry for AOX (upper panel) and PMP70 (lower panel), corrected by the expression of GAPDH, and each one expressed relative to the corrected respective mRNA expression in ciprofibrate-untreated cells in the absence of inhibitors. Data represent means 6 SD. Asterisks indicate statistically significant differences from ciprofibrate-treated cells in the absence of inhibitors (P , 0.05 or less). C: CHO cells were transiently transfected with the peroxisome proliferator response element (PPRE)-luciferase reporter, or cotransfected with the PPRE reporter and mouse PPARa (PPRE 1 mPPARa), and treated for 24 h with 100 mM ciprofibrate in the absence or presence of 5 and 20 mM DDMS or PPOH. Luciferase activity was standardized relative to b-galactosidase (see Materials and Methods). Data represent means 6 SD of corrected luciferase activity from three independent experiments, expressed relative to the activity in the absence of ciprofibrate and inhibitors. D: Expression of PPARa mRNA was assessed by Northern blot in hepatocytes treated for 12 h with 100 mM ciprofibrate in the absence (control) or presence of 20 mM DDMS or PPOH, as indicated. GAPDH mRNA expression was used as a control (Ctr). No effect of the drugs was found in PPARa expression in two independent experiments. E: The expression of CYP2C and CYP4A was determined by Western blot in hepatocytes treated with 20 mM DDMS, 20 mM PPOH, or 10 mM ketoconazole (KE) in the absence and presence of ciprofibrate. A representative Western blot is presented. F: Western blots from three independent experiments were quantified by scanning densitometry for AOX CYP4A and CP2C11, corrected by the expression of tubulin, and each one expressed relative to the corrected protein expression in ciprofibrate-untreated cells in the absence of inhibitors. Data represent means 6 SD. Asterisks indicate statistically significant differences from control cells. No significant changes were induced by the various treatments for CYP2C expression, whereas ciprofibrate induced CYP4A expression by 1.5-to 1.8-fold, an effect that was reverted to control values by CYP inhibitors. plasmid. Similarly, no effect of the drugs was found on PPARa mRNA expression in isolated hepatocytes at the maximal concentration used (Fig. 2D).
Because both CYP4A and CYP2C are also regulated by PPARa activators in opposite directions (43), we determined the levels of these proteins in control and ciprofibrate-treated cells in the absence and presence of CYP inhibitors ( Fig. 2E; quantified in Fig. 2F). No effect of inhibitors was found in control cells for both proteins; however, in ciprofibrate-treated cells, CYP4A increased by 1.5-to 1.8-fold. Consistent with previous data, ciprofibrateinduced CYP4A upregulation was reversed by CYP inhibitors. Ciprofibrate and the inhibitors were without effect on CYP2C expression, either in the absence or presence of ciprofibrate. The lack of ciprofibrate effect on CYP2C11 level is in agreement with published data showing that downregulation of CYP2C occurs only after 48 h of exposure of hepatocytes to the strong PP WY-14,643 (43). A short (12 h) exposure to ciprofibrate was used in our experiments, which also explains the low, although significant, effect of ciprofibrate on CYP4A expression. These results suggest that CYP2C and CYP4A activities are important for ciprofibrate-induced peroxisomal proliferation and that HEETs, the products of the sequential action of CYP2C and CYP4A on unsaturated fatty acids, may be the physiological activators of PPARa in rat liver.
Putative hydroxy acid metabolites of AA, but not EETs, can be detected in control or ciprofibrate-treated hepatocytes metabolizing [1- 14

C]AA
To assess whether ciprofibrate induces measurable changes in AA oxidized metabolites, hepatocytes were incubated with [1-14 C]AA to label only oxidized AA metabolites and not chain-shortened fatty acids, which are formed from AA as well as from EETs (44). Hepatocytes were incubated with [1-14 C]AA for 1 and 6 h in the presence or absence of 100 mM ciprofibrate. The labeled oxidized metabolites and the remaining [1-14 C]AA were extracted with chloroform and resolved by reverse-phase HPLC, according to published methods (25, 26) (see Materials and Methods). AA is rapidly metabolized by hepatocytes, and only 30-35% of the radioactivity remained in the organic phase as [1-14 C]AA after 1 h of incubation; however, after 6 h, only ?8% of radioactivity was still present as AA, as assessed by reverse-phase HPLC of the organic extracts (data not shown). No radioactivity was detected in the retention time corresponding to EETs, either in the presence or absence of ciprofibrate (Fig. 3), suggesting that EETs are very rapidly metabolized. For both incubation times, radioactivity was recovered mainly as [1-14 C]AA and also in metabolites eluting earlier than EETs (peak X in Fig. 3A-C).
Radioactivity in peak X, which presents a retention time corresponding roughly to that of v/v-1 hydroxy acids, such as HEETs (26), increased 10-15% in ciprofibratetreated cells. However, v/v-1 hydroxylation is one of the known routes for EET metabolism (45), and peak X probably contains other oxidized metabolites of AA. To indirectly assess whether this fraction may contain HEETs, hepatocytes were incubated for 1, 2, and 6 h in the pres-ence of 14,15-[1-14 C]EET in the absence or presence of 20 mM of the CYP4A inhibitor PPOH. Labeled EET disappeared from the incubation medium even faster than AA in both conditions, and only 20% of the initial radioactivity eluted in the retention time of 14,15-EET after 1 h of incubation; after 6 h, it was ,5%, as assessed by reverse-phase HPLC of the organic extracts (data not shown). After 1 h of incubation, no differences were induced by PPOH in the amount of radioactivity eluting with the EET peak. However, the radioactivity eluting in fractions around the retention time of peak X was reduced by almost 50% in the presence of PPOH (Fig. 3C). Quantification of the time course of radioactivity eluting in peak X in hepatocytes metabolizing 14,15-[1-14 C]EET is presented in Fig. 3D. In all cases, PPOH induced a significant decrease in the radioactivity eluting in peak X, showing that fractions eluting in the peak X retention time are indeed sensitive to the CYP4A1 inhibitor and that most of them probably contain HEETs.
The development of mass spectral methods for HEET quantification in biological samples and cultured cells, not yet available, will be necessary to directly assess and demonstrate that HEETs are present in peak X and whether ciprofibrate indeed increases HEET levels in hepatocytes after short-term incubation periods. Nevertheless, the results in this section partially answer two relevant points. First, they show that under our experimental conditions, EETs produced from AA are rapidly metabolized and their physiological concentration is probably maintained very low, making it unlikely that they act as PPARa physiological ligands. This observation is in agreement with published data showing that in the liver free EETs represent ,1% of total endogenous EETs (46). Second, although indirectly, the PPOH experiments suggest that HEETs are present in peak X and that PPOH indeed inhibits the formation of metabolic products from 14,15-[1-14 C]EET, supporting the hypothesis that these eicosanoids are involved in PPARa activation.

Adenovirus-mediated CYP2C11 overexpression induces peroxisomal proliferation in cultured rat hepatocytes
To further substantiate our previous conclusion, we infected cultured hepatocytes with a viral vector containing CPY2C11, the predominant liver epoxygenase, fused to rat P450 oxidoreductase (see Materials and Methods). Infected hepatocytes express the bifunctional protein (Ad-EPOX), as assessed by Western blot (Fig. 4A). Note that the antibody also recognizes endogenous CP2C11 at the expected molecular mass, which is present in both uninfected and infected cells. Hepatocytes infected with the same viral vector but expressing b-galactosidase (Ad-bgal), to be used as controls, also showed increased expression of the b-galactosidase protein (data not shown). The levels of AOX and PMP70 mRNAs, as assessed by Northern blot, were increased by 4-and 6-fold, respectively, in hepatocytes overexpressing Ad-EPOX ( Fig. 4B; quantified in Fig. 4C) compared with uninfected cells or cells infected with the Ad-bgal. A similar increase was observed in ciprofibrate-treated cells used as the positive control.
Next, we determined, by Western blot, the expression levels of peroxisomal proliferation marker proteins. We used ADAPS and catalase, together with PMP70 and CYP4A antibodies, as markers of peroxisomal proliferation. AOX was not used as a marker because of the unavailability of antibodies. All three peroxisomal proteins (36,47,48), as well as CYP4A, showed increased expression in Ad-EPOXinfected cells, ranging from 3-fold (catalase and ADAPS) to ?5to 6-fold (PMP70 and CYP4A), compared with control (no virus) or Ad-bgal-transfected cells ( Fig. 5A; quantified in Fig. 5B). Ciprofibrate treatment also induced a significant increase in peroxisomal markers, as expected, but was lower than in Ad-EPOX-expressing cells for PMP70 and catalase. Furthermore, immunocytochemistry against PMP70 and catalase showed the distinctive punctate pattern of peroxisomal markers (36), indicating a clear increase in peroxisome number in hepatocytes overexpressing Ad-EPOX compared with cells infected with Ad-bgal (Fig. 5C). Finally, we determined whether a CYP inhibitor could reverse the effect of CYP2C11 overexpression. As shown in Fig. 5D (quantified in Fig. 5E), 1 mM ketoconazole partially reversed the increased expression of PMP70 and CYP4A induced by CYP2C11 overexpression, supporting a role of the CYP isoforms in PPARa activation. We did not use higher ketoconazole concentrations, because some decrease in cell viability was observed after the 48 h culture period at concentrations of .1 mM ketoconazole. ?30-35% of the initial radioactivity was present in the AA peak at the end of incubation, as assessed by reverse-phase HPLC. The retention times of EET standards and AA are presented in gray. No radioactivity eluting in the retention time of EET standards was detected after 1 h. A peak of radioactivity was detected at a retention time corresponding roughly to that of v/v-1 hydroxy acids (peak X; see text), which was increased slightly in ciprofibrate-treated cells. A representative chromatogram is presented. Similar results were obtained in three independent experiments. B: Cells were incubated for 6 h with [1-14 C]AA, a condition at which only ?8% of the radioactivity remained as the [1-14 C]AA peak at the end of the incubation. As in the 1 h experiment, no radioactivity eluted at the EET retention time, and ciprofibrate increased the radioactivity eluting in peak X. A representative chromatogram is presented. Similar results were obtained in two independent experiments performed with duplicates. C: Hepatocytes were incubated for 1 h in the presence of 4 mM 14,15-EET and 0.1 mCi/dish [1-14 C]14,15-EET in the absence or presence of the CYP4A1 inhibitor PPOH. After organic solvent extraction, labeled 14,15-EET was resolved from its metabolites by reversephase HPLC. Approximately 20% of the initial radioactivity was recovered in the 14,15-EET peak at the end of incubation. No changes were induced by PPOH in the radioactivity recovered with the 14,15-EET peak. However, PPOH induced a decrease in the radioactivity eluting with a retention time similar to peak X in experiments with [1-14 C]AA. D: Time course of the radioactivity eluting in peak X in the absence or presence of PPOH. Data represent means 6 SD of the radioactivity eluting with peak X in triplicates of one experiment. In all cases, PPOH induced a significant decrease of radioactivity (* P , 0.05 or less). Similar results were obtained in a second independent experiment.

DISCUSSION
This report provides evidence that CYP2C and CYP4A P450 isoforms play a role in the induction of peroxisomal proliferation and that their products are most likely endog-enous PPARa ligands in rat liver cells. The abrogation of ciprofibrate-induced peroxisomal proliferation by general and specific CYP2C and CYP4A inhibitors, but not of other eicosanoid-generating pathways, and their lack of effect on PPARa expression or transcriptional activity suggest that HEET products of the sequential action of CYP2C and CYP4A on unsaturated fatty acids participate in, and apparently are necessary for, ciprofibrate-induced peroxisomal proliferation. Only minor or no changes in the expression of these isoforms were observed after short-term (12 h) exposure to ciprofibrate, a time period in which mRNA expression of PPARa target genes is already increased. Moreover, the spontaneous peroxisomal proliferation observed in cells overexpressing CYPC11 suggests that these metabolites, by themselves, induce PPARa activation.
This proposal is further supported by a previous report showing that HEETs are high-affinity PPARa ligands, that they transactivate PPARa in gene reporter assays, and that their precursors, the EET products of CYP2C, are excellent substrates for CYP4A (14). In addition, studies in the peroxisomal AOX null mouse, which exhibits massive peroxisomal proliferation, including CYP4A induction, show that increased intracellular concentration of fatty acid metabolites can induce peroxisomal proliferation (49). This view is further supported by studies in mouse nullizygous for both PPARa and AOX that suggest a function of CYP4A isoforms in PPARa signaling (50). The fact that we were unable to detect EET formation from [1-14 C]AA under our experimental conditions suggests that EETs are rapidly metabolized and probably do not accumulate as free fatty acids in the liver. This result is consistent with previous data showing that endogenous liver EETs are present mainly as esters in phospholipids, which is a major pathway of EET metabolism (45,46,51), and that ,1% is present as free fatty acids (46).
These considerations make it unlikely that EETs are involved directly in PPARa activation. The ciprofibrateinduced slight increase in [1-14 C]AA metabolites present in peak X and the partial decrease in the radioactivity eluting in that position induced by the CYP4A1 inhibitor PPOH in cells metabolizing [1-14 C]14,15-EET suggest that more polar metabolites, such as HEETs, might be responsible for PPARa activation. However, because sensitive techniques to measure HEETs in cells are not yet available, we cannot exclude the possibility that other metabolites of AA are increased in ciprofibrate-treated hepatocytes. Other derivatives of EETs, products of soluble epoxide hydrolase and hydroxy-eicosatrienoic acids, are also activators of PPARa (52). However, their in vitro affinity for PPARa is in the micromolar range, and most of the hydroxy-eicosatrienoic acids that are formed by cells are released into the extracellular fluid (45). Additionally, the suppression of ciprofibrate-induced peroxisomal proliferation by CYP2C11 and CYP4A inhibitors does not support a physiological role for hydroxy-eicosatrienoic acids in PPARa activation. Nevertheless, soluble epoxide hydrolase, which is also a PPARa target gene (53), may be physiologically relevant in the disposing of excess EETs under conditions in which EET and HEET degradation Because CYP4A is a PPARa target gene, it is conceivable that its upregulation may contribute to the modulation of PPARa-dependent gene transcription and liver lipid homeostasis. Similarly, the downregulation of CYP2C after long-term exposure to PPs might also represent a feedback mechanism to decrease the synthesis of PPARa activators. There is extensive evidence for the function of CYP2C and CYP4A metabolites in cell signaling in extrahepatic tissues, including the modulation of ion channel activity and transporters, mitogenesis, and the activation of tyrosine kinase cascades, mitogen-activated protein kinases, and phosphatidylinositol 3-kinase signaling pathways (51,54), but whether these metabolites may activate similar pathways in liver cells is not known. As noted above, the identity and relative contribution of monooxygenase metabolites to PPARa-dependent peroxisome proliferation will have to await the development of sensitive techniques for HEET quantification in cells and tissues. Nevertheless, the present data add a new perspective from which to experimentally approach unresolved issues regarding the mechanism mediating PP effects as well as on PPARa functionality in . Ad-bgal-transfected cells were used as controls. Representative Western blots of PMP70 and CYP4A, which are highly induced in cells overexpressing the epoxygenase, are presented. Tubulin was used as a control. E: Quantification of three independent experiments as shown in D, corrected by the expression of tubulin, and each one expressed relative to control cells. Data represent means 6 SD. Asterisks indicate statistically significant differences from Ad-EPOX cells (P , 0.01). A partial decrease in the peroxisomal target genes was observed at 1 mM ketoconazole. Higher ketoconazole concentrations were not used because they partially decreased cell viability after the 48 h culture period. peroxisome proliferation-unresponsive species, such as humans (55,56).
Although it is clear that PP-induced sustained PPARa transcriptional activity is essential for PP-induced pleiotropic effects (11,34,57), contradictory observations persist when comparing the relative affinity of various PPs for PPARa with its potency as a PP. The strong peroxisomal proliferator WY-14643 binds to PPARa with a K d in the low micromolar range (13), whereas nafenopin, with a similar potency, does not bind at all (58). Furthermore, only a few PPARa ligands, such as WY-14643, consistently bind to PPARa in in vitro assays and also induce PPAR transactivation in gene reporter assays (11,57). For example, fibrates, such as ciprofibrate and bezafibrate, do not bind to PPARa in coactivator-dependent binding assays, although they induce PPAR transactivation (57). Thus, it is possible that immediate fibrate-induced metabolic effects, preceding the expression of PPARa target genes, might contribute to PPARa sustained activation independently or in addition to PPARa activation through direct binding.
The first detectable event in isolated rat liver cells exposed to certain PPs, such as ciprofibrate, is the formation of nonmetabolizable acyl-CoA thioesters of the drugs, which attain a steady-state concentration after 30 min, amounting to a significant proportion of total cell CoASH and most of cytosolic CoASH (5,59). A high concentration of acyl-CoA thioesters of PPs is also observed in the liver of PP-treated rats (6). The resulting immediate sequestration and unavailability of CoASH may result in a transient increased concentration of v-hydroxylation products, which could contribute to PPARa activation, because, in contrast to b-oxidation, CoASH is not required for these oxidative pathways. Furthermore, incorporation of EETs to phospholipids also requires previous activation to CoASH esters (46), and CoASH unavailability may direct EET metabolism to v-hydroxylation and other oxidative pathways. The subsequent induction of b-oxidation enzymes, of CoASH synthesis (7,8), and of acyl-CoA thioesterases, upon longer PPs exposure (60), probably would contribute to the feedback control of these putative and short-term concentration changes. It is worth noting that HEETs bind to PPARa with a K d in the low nanomolar range (i.e., almost 3 orders of magnitude lower than that of PPs with highest PPARa affinity). Hence, although ligandactivated PPARa appears essential for peroxisome proliferation, it is possible that low concentrations of other signaling molecules, such as HEETs, but with much greater affinity for PPARa than ciprofibrate, might be necessary to initiate PPARa productive binding to response elements in the promoters of target genes.
Although the in vivo involvement of PP-CoA formation in peroxisomal proliferation remains speculative, PPs-CoAs have an essential role in PP-induced hypolipidemic activity. The human liver, which, as already mentioned, is resistant to peroxisomal proliferation, also forms PPs-CoA thioesters (5). There is strong evidence that PP-induced hypolipidemic activity in humans results from the suppression of hepatic nuclear factor-4a (HNF-4a) activity by metabolically formed PPs-CoAs and is PPARa-independent (61). In the rat, nonproductive PPARa binding and displacement of HNF-4a from a response element shared by both transcription factors is thought to be responsible for hypolipidemia (61,62). HNF-4a controls genes encoding proteins involved in the production of lipoproteins, such as the apolipoprotein C-III gene (63). The overall effect of HNF-4a conditional disruption in mouse liver is the reduction of circulating lipid levels (64), consistent with its proposed role as a target of hypolipidemic drugs. Thus, as PP hypolipidemic activity in humans can be explained by a PPARaindependent mechanism, the main difference with the rat is the inability of human PPARa to support peroxisomal proliferation, an observation that is ascribed to low PPARa concentration in human liver (65). However, human PPARa supports the induction of peroxisomal proliferation in PPARa-deficient mouse liver (66), and PPARa concentration does not appear to be limiting for the transcription of PPARa target genes in the mouse, because they are induced even under reduced PPARa expression (64).
On this basis, it is tempting to speculate that the liver's capacity to synthesize endogenous PPARa ligands might be one of the factors responsible for the differences in species sensitivity to peroxisomal proliferation. In mammals, the nature and distribution of P450 isoforms is speciesspecific and may also depend on age, sex, hormonal status, diet, and the organ (51 and references cited therein). In human liver, EETs are products primarily of CYP2C8 epoxygenase (67,68). A single human CYP4A11 has been cloned and shown to catalyze AA and lauric acid v/v-1 hydroxylation (21); however, it is unknown whether this enzyme catalyzes the v/v-1 hydroxylation of EETs. In fact, in vitro studies suggest that CYP4F3s are the main catalyst in the oxidation of fatty acid epoxides in humans, and the human CYP4F family appears to play a similar functional role as the CYP4A family in rats (69). However, little is know about CYP4F3 in vivo function, and it remains to be established whether differences in P450 monooxygenase specificity or activity might account for the apparent inability of human PPARa to support peroxisomal proliferation.
In summary, this is the first report showing that CYP2C and CYP4A are involved in ciprofibrate-induced peroxisomal proliferation in isolated rat liver cells. It suggests that metabolites of these P450 isoforms might contribute to or be responsible for PPARa-dependent peroxisomal proliferation, representing a putative alternative mechanism to the canonical view of PP-induced PPARa activation.