Opposing Roles of Peroxisome Proliferator-activated Receptor α and Growth Hormone in the Regulation of CYP4A11 Expression in a Transgenic Mouse Model*

CYP4A11 transgenic mice (CYP4A11 Tg) were generated to examine in vivo regulation of the human CYP4A11 gene. Expression of CYP4A11 in mice yields liver and kidney P450 4A11 levels similar to those found in the corresponding human tissues and leads to an increased microsomal capacity for ω-hydroxylation of lauric acid. Fasted CYP4A11 Tg mice exhibit 2–3-fold increases in hepatic CYP4A11 mRNA and protein, and this response is absent in peroxisome proliferator-activated receptor α (PPARα) null mice. Dietary administration of either of the PPARα agonists, fenofibrate or clofibric acid, increases hepatic and renal CYP4A11 levels by 2–3-fold, and these responses were also abrogated in PPARα null mice. Basal liver CYP4A11 levels are reduced differentially in PPARα−/− females (>95%) and males (<50%) compared with PPARα−/+ mice. Quantitative and temporal differences in growth hormone secretion are known to alter hepatic lipid metabolism and to underlie sexually dimorphic gene expression, respectively. Continuous infusion of low levels of growth hormone reduced CYP4A11 expression by 50% in PPARα-proficient male and female transgenic mice. A larger decrease was observed for the expression of CYP4A11 in PPARα−/− CYP4A11 Tg male mice to levels similar to that of female PPARα-deficient mice. These results suggest that PPARα contributes to the maintenance of basal CYP4A11 expression and mediates CYP4A11 induction in response to fibrates or fasting. In contrast, increased exposure to growth hormone down-regulates CYP4A11 expression in liver.

Cytochrome P-450 4A11 (P450 4A11) 2 constitutes a major fatty acid -hydroxylase in human liver and kidney (1)(2)(3). It catalyzes the -hydroxylation of arachidonic acid as well as of other long chain fatty acids (4). The capacity of family 4 P450s to catalyze -hydroxylation reflects a high degree of regioselectivity for oxygenation of the terminal methyl group rather than the inherently more reactive, adjacent secondary carbons that are typically oxidized by other subtypes of P450s. Subsequent oxidation of the terminal alcohol group can ultimately lead to the generation of a carboxylic acid moiety. The resulting aliphatic carboxylates can be further degraded by peroxisomal ␤-oxidation. This contributes to the degradation of excess, free fatty acids, as well as lipid mediators of inflammation such as leukotrienes and prostanoids (5)(6)(7)(8).
Additionally, arachidonic acid -hydroxylation by CYP4A11 to form 20-hydroxyeicosatetraenoic acid is thought to contribute to the regulation of blood pressure (6). A functional variant of P450 4A11 that carries a phenylalanine to serine substitution at residue 434 exhibits diminished catalytic activity toward lauric and arachidonic acids. This genetic variant was associated with an increased risk for elevated systemic blood pressure in two independent cohorts (9). An association between the F434S variant and an increased risk for hypertension was also found in a larger population of the MONICA Augsburg survey (10) and more recently in a Japanese cohort (11). Moreover, ablation of either the mouse Cyp4a10 (12) or the mouse Cyp4a14 (13) gene leads to distinct hypertensive phenotypes. These findings suggest specific roles for individual CYP4A enzymes in the homeostatic control of blood pressure.
CYP4A11 is the only human CYP4A gene that is significantly expressed in liver and kidney, and it remains unclear whether its expression is regulated by mechanisms that govern the expression of CYP4A genes in other species. Moreover, well characterized species such as rats and mice express a larger number of CYP4A genes than humans, and these genes exhibit different patterns of tissue expression and adaptive responses to physiologic conditions. This includes regulation by growth hormone (14), steroid hormones (15), and environmental chemicals (7).
The nuclear receptor PPAR␣ regulates the expression of specific CYP4A genes in several nonhuman species in response to environmental chemicals and hypolipidemic drugs such as fibrates (16 -18) but not in others such as guinea pig (19). Transcription of CYP4A genes is also elevated during caloric restriction, during fasting, and with streptozotocin-induced diabetes (20 -24) in nonhuman species, and induction of CYP4A gene expression under these conditions is abrogated in mice that are deficient for PPAR␣ (18,20,25). PPAR␣ is activated during fasting and is involved in the increased expression of enzymes comprising the mitochondrial and peroxisomal fatty acid ␤-oxidation pathways, which leads to enhanced ketogenesis (26). The activation of PPAR␣ is thought to occur in response to an increase in circulating free fatty acids during fasting or diabetic ketosis. The increased flux can lead to elevated hepatic free fatty acids that are -hydroxylated by the action of the microsomal family 4 monooxygenases and subsequently converted to dicarboxylic long chain fatty acids. Peroxisomal ␤-oxidation converts the dicarboxylic long chain fatty acids to C6 -C10 dicarboxylic acids. The latter are either excreted or metabolized further by mitochondria to produce succinyl-CoA as the final ␤-oxidation product. This pathway of fatty acid oxidation can account for as much as 15% of hepatic fatty acid oxidation in fasted animals (27). Although this pathway reduces energy conservation and ketogenesis from fatty acid oxidation, the coordinate induction of microsomal fatty acid -hydroxylases with ␤-oxidation contributes to the degradation and elimination of excess and potentially toxic free fatty acids resulting from increased rates of adipocyte lipolysis. Such observations suggest that fasting may also play a role in the regulation of the CYP4A11 gene in humans and that PPAR␣ might mediate this response.
The contribution of PPAR␣ to CYP4A11 regulation is largely derived from experiments with cultured human liver slices, human hepatocytes, or human-derived cell lines, where factors regulating the expression of the CYP4A11 gene may not be optimally maintained (28 -32). Treatment of human hepatocytes with fenofibric acid led to 2-fold increases of CYP4A11 mRNA, and 4-fold increases were evident using a potent, proprietary PPAR␣ agonist (29). Other studies have also described a doubling of CYP4A11 expression in human hepatocytes upon treatment with fenofibrate or clofibrate (31,33). In contrast, treatment of cultured, human liver slices with the PPAR␣ agonist methylclofenapate did not affect P450 4A11 levels as measured by microsomal lauric acid 12-hydroxylation (28) or immunoblots (34).
To investigate the regulation of the human CYP4A11 gene expression in an integrated physiologic model, transgenic mice were developed to examine CYP4A11 gene expression during fasting and to characterize the role of PPAR␣ in this process using PPAR␣-deficient mice. Our results show that human P450 4A11 tissue-selective expression is maintained in mice and that CYP4A11 mRNA and protein induction during fasting or by fenofibrate or clofibric acid is dependent on PPAR␣. Moreover, basal hepatic expression of CYP4A11 declined in PPAR␣ null animals in a sex-dependent manner. Continuous infusion of growth hormone to mimic the female growth hormone pattern reduced male expression of CYP4A11 in the PPAR␣ null mice to near female levels. Treatment of PPAR␣proficient animals with growth hormone decreased the expression of CYP4A11 equally in both males and females.

EXPERIMENTAL PROCEDURES
Generation of CYP4A11 Transgenic Mice-All handling of mice was in accordance with study protocols that were approved by the Institutional Animal Care and Use Committee of the Scripps Research Institute and the University of California, San Diego. Bacterial artificial chromosome (BAC) clone RP11Ϫ346m5, originating from the RPCI-11 human BAC library constructed from a male blood donor (35), was obtained from the BACPAC Resource Center at Children's Hospital Oakland Research Institute and used for the generation of CYP4A11 transgenic mice. The 123,921-bp clone was sequenced by the Wellcome Trust Sanger Institute, and the sequence can be retrieved with accession number AL731892. The production of transgenic mice utilized standard techniques and was carried out at the University of California, San Diego Superfund Transgenic Core Facility. BAC DNA was isolated using a Large Construct kit (Qiagen) and verified by PCR amplification of CYP4A11 exons 1-12 including 0.6 kb of 5Ј-flanking DNA segments and 0.8 kb of 3Ј-flanking DNA segments using primers listed in supplemental Table S1. DNA from the entire BAC clone was microinjected into fertilized CB6F 1 mouse eggs (f 1 -hybrids derived from matings of Balb/c with C57Bl/6 mice) and transplanted into the oviducts of pseudo-pregnant C57BL/6 mice. Genotyping by PCR was conducted to identify founder mice. Tail DNA (100 ng) was used in a reaction mixture that contained 200 nM of each primer, 2 mM dNTPs, 3 mM MgCl 2 , and 1.25 units of Taq polymerase. The cycling conditions were 35 repeats of 15 s at 95°C followed by annealing for 30 s at 58°C and extension for 45 s at 72°C. PCR analyses identified six CYP4A11 transgenic mice. Each potential founder was used for mating with C57Bl/6 wild-type mice. Roughly 50% of progeny from each founder carried the transgene and were designated CYP4A11 Ϫ/ϩ f 1 . A male CYP4A11 Ϫ/ϩ f 1 derived from each founder was then mated with a female C57Bl/6 wild-type mouse, and the progeny of each founder were designated CYP4A11 Tg-A through CYP4A11 Tg-F. A 50% occurrence of the transgene was seen among each progeny. Transgenic mice used in these experiments were heterozygous for the human CYP4A11 transgene ( Ϫ/ϩ ) and of mixed C57Bl/6-CB6F 1 background (CB6F 1 mice are hybrids from crosses of the C57Bl/6 and Balb/c strains). Littermates that lacked the CYP4A11 gene were designated nontransgenic or CYP4A11 ( Ϫ/Ϫ ).
Generation of CYP4A11 Transgenic Mice with PPAR␣ Null Background-A PPAR␣-deficient SV129 breeding pair was purchased from the Jackson Laboratory (Bar Harbor, ME), strain 129S4/SVJae-ppar-tm1Gonz (18), and female offspring were used for mating with male CYP4A11 Tg-B or CYP4A11 Tg-F mice. The resulting PPAR␣ heterozygous offspring consisted of a 1:1 ratio of mice that were heterozygous for CYP4A11 ( Ϫ/ϩ ) and mice that lacked CYP4A11 ( Ϫ/Ϫ ). Male offspring carrying a CYP4A11 Ϫ/ϩ and PPAR␣ Ϫ/ϩ genotype were then mated with SV129 PPAR Ϫ/Ϫ females to generate mice carrying a CYP4A11 Ϫ/ϩ genotype in either a PPAR␣ heterozygous ( Ϫ/ϩ ) or PPAR␣ null ( Ϫ/Ϫ ) background. In all instances, the CYP4A11 genotype was determined by PCR from tail DNA as described above. The PPAR␣ genotype was verified by PCR using primers specific for PPAR␣ (oIMR1199 and oIMR1200; Jackson Laboratory) and specific for the phosphoribosyltransferase II gene (oIMR0013 and oIMR0014) to confirm the presence of the neo gene (Jackson Laboratory). The phosphoribosyltransferase II gene product confers neomycin resistance and was used by Lee et al. (18) for the construction of the PPAR␣ targeting vector used for homologous recombination and generation of the PPAR␣ null mice. Primers oIMR1199 and oIMR1200 were 5Ј-CCATCCAGATGACAC-CTTCC and 5Ј-TCTCTTGCAACATGTGGTGC. Primers oIMR0013 and oIMR0014 were 5Ј-CTTGGGTGGAGAGGC-TATTC and 5Ј-AGGTGAGATGACAGGAGATC. Conditions for amplification were as described by the Jackson Laboratory.
Animal Treatments-All mice were housed in an animal facility with an alternating 12-h light (7 a.m. to 7 p.m.) and 12-h dark cycle and were fed ad libitum with a standard rodent chow (LM-485 rodent diet with a 5% fat content; Teklad, Madison, WI).
For fasting experiments, mice 12-16 weeks of age were provided with unlimited access to water but were deprived of food for a period of 24 h (21), whereas control littermates had free access to both water and food. The mice were sacrificed by exposure to an overdose of 2-bromo-2-chloro-1,1,1-trifluoroethane, and tissues were harvested, flash frozen, and stored at Ϫ80°C until preparation of RNA and microsomes.
To assess the effects of fibrates, the mice were fed chow that contained fenofibrate (0.2% w/w) or clofibric acid (0.5% w/w). Fenofibrate was dissolved in ethanol, and the solution was applied evenly to the surface of the chow. The corresponding control diet was prepared by applying ethanol to the chow. The solvent was allowed to evaporate completely before use. Clofibric acid solution was prepared in acetone, and the chow was prepared as described above. The mice were fed either the PPAR␣ agonist-containing or corresponding control diet and sacrificed after 10 days (18,36). The tissues were collected and frozen until preparation of RNA and microsomes.
For continuous growth hormone infusion, osmotic minipumps, model 1007D-7, 7-day pumps, were obtained from Durect Corp. (Cupertino, CA), and highly purified mouse recombinant growth hormone was purchased from Dr. A. F. Parlow from the National Hormone & Peptide Program of UCLA Medical Center. Growth hormone was dissolved in a buffer that contained 30 mM NaHCO 3 , pH 8.3, 150 mM NaCl 2 and 100 g/ml rat albumin (37). The pumps were filled with 100 l of GH solution or with buffer only for sham treatments and were implanted subcutaneously under isoflurane anesthesia. Growth hormone was infused at a continuous rate of 20 ng/g of body weight/h for 7 days. The mice were sacrificed on day 7, and the livers were collected for RNA preparations.
Human Tissue Specimens-Human liver samples were obtained from the Liver Tissue Procurement and Distribution System (University of Minnesota, Minneapolis, MN). Specimens were frozen in liquid nitrogen within 10 h of death, shipped overnight in dry ice, and stored at Ϫ80°C until preparation of microsomes or RNA. The source of the kidney samples has been described elsewhere (3).
Long PCR-To assess the extent of integration of upstream regulatory and downstream intergene DNA flanking the CYP4A11 gene, PCR was conducted with primers that were designed to amplify ϳ10-kb segments with an overlap of 50 -200 bp (supplemental Table S2). BAC DNA (ϳ50 copies) or mouse genomic DNA (200 ng) were used as templates for amplification in a reaction mixture that contained 300 nM of each primer, 500 M of each dNTP, 2.75 mM MgCl 2 , 1ϫ reaction buffer 2 (Expand Long Template PCR system; Roche Applied Science), and 3.75 units of a Taq DNA polymerase and Tgo DNA polymerase mixture (Expand Long Template PCR system; Roche Applied Science). The cycling conditions were as follows: one cycle of 93°C for 2 min, followed by 10 cycles of denaturation at 93°C for 10 s, annealing at 60°C for 30 s, and extension for 8 min at 68°C, followed by 20 cycles of 93°C for 15 s, annealing at 60°C for 30 s, and extension for 9 min at 68°C. The reaction was concluded by an extension cycle of 15 min at 68°C. The reaction products were analyzed on agarose gels (0.5%) and visualized by ethidium bromide staining. The authenticities of PCR products were verified by partial sequencing of gel-purified DNA fragments.
Quantitative PCR (qPCR)-RNA from tissues was prepared using TRIzol reagent (Invitrogen) and subjected to RNase-free DNase I digestion for removal of residual genomic DNA. Total RNA (5 g) was reverse transcribed using AffinityScript multiple temperature reverse transcriptase (Stratagene, Cedar Creek, TX). CYP4A11 primers (5Ј-TTGCCCAAAGGTATCA-TGGTC, 5Ј-GTTTCCCAATGCAGTTCCTTGAT; accession L04751, coordinates 1245-1412), Cyp4a14 (5Ј-TGAATTGCT-GCCAGATCCCACCAGGATC, 5Ј-GTTCAGTGGCTGGT-CAGA; accession NM_007822, coordinates 1451-1780) were used for qPCR. Mouse ribosomal protein L27 primers (5Ј-CTGTCGAGATGGGCA AGTTCAT, 5Ј-TTGTGGGCATG-AGGTGGTTGTA; accession BC090395, coordinates 11-268) served as the housekeeping gene for normalization of RNA contents (21). The PPAR␣ mRNA levels were assessed by using primers oIMR1199 and oIMR1200 described above. Prior to real time PCR, gene-specific fragments were generated by conventional PCR, and each was subcloned into the EcoRI site of the pCRII-TOPO vector (Invitrogen) to generate plasmid containing each specific target cDNA. This resulted in three plasmids containing 167-bp (CYP4A11), 329-bp (Cyp4a14), and 257-bp (L27) fragments. Serial dilutions of these plasmids were used in real time PCR (IQ SYBR Green Super Mix; Bio-Rad) to generate standard curves for calibrating the specific mRNA content in each tissue sample as described previously (32). Relative PPAR␣ levels were determined from standard curves using serial dilutions of the L27 plasmid as template. The cycling reactions were initiated by denaturing for 5 min at 95°C. This was followed by 50 repeats of denaturation at 94°C for 15 s and annealing/extension at 60°C for 30 s. In addition, real time PCR primer sets were used for conventional RT-PCR of CYP4A11, Cyp4a14, and L27.
Adsorption of Anti-human P450 4A11 Antibody to Remove Cross-reactivity with Mouse P450 4A Proteins-Liver microsomes were prepared using conventional methods from C57BL/6 mice that had been treated with clofibrate or fenofibrate. Microsomes (43 mg of protein) were diluted to 5.0 mg of protein/ml with 250 mM NaHCO 3 buffer, pH 8.5, containing 1.0% sodium cholate, 1.0% Lubrol PX, and 1 mM EDTA. The protein mixture was stirred for 30 min at 4°C and then centrifuged at 105,000 ϫ g for 70 min to remove insoluble protein.
The detergent-solubilized microsomes were combined with 3 mg of purified, recombinant mouse P450 4A14, and then reacted overnight at room temperature with preactivated Affi-Gel 10 resin (Bio-Rad) with end-over-end mixing. After coupling, the resin was poured into an empty glass column to a final size of 1.5 ϫ 8.5 cm and washed at room temperature with 5 volumes each of the buffer used for protein coupling, 250 mM Tris-HCl buffer, pH 8.5, containing 0.5 M KCl to block any remaining amino-reactive functional groups, and IgG buffer (100 mM potassium phosphate buffer, pH 7.4, containing 150 mM KCl). The column was placed into a cold box, and 10 mg of purified anti-human CYP4A11 IgG (38), diluted to 0.7 mg/ml with IgG buffer, was recirculated over the Affi-Gel 10 backadsorbent resin for 5 h at 4°C using a flow rate of 60 ml/h (equivalent to 20 passes over the column). Upon recirculation, the column was washed with IgG buffer to remove nonbound antibody. The protein content of back-adsorbed CYP4A11 antibody solutions was assessed from the absorbance at 280 nm using an extinction coefficient of 13.8 mM Ϫ1 cm Ϫ1 for IgG. Western blotting with anti-P450 4A11 IgG, performed as described elsewhere (38), was used to assess the efficiency of antibody back-adsorption.
Immunoblotting-Mouse hepatic or renal microsomes were adjusted to contain 2 mg of protein/ml using the BCA protein assay reagent (Pierce). The microsomes were then diluted 2-fold in 2ϫ SDS-PAGE loading buffer, and 2.5 or 5 g/lane were separated on gels containing 8% acrylamide containing Tris-glycine gels (Invitrogen). Baculovirus-expressed P450 4A11 Supersomes TM (Gentest, Woburn, MA) and human liver microsomes were used as references. The proteins were transferred onto nitrocellulose, and the membranes were stained with the MemCode TM reversible protein stain (Pierce). The membranes with protein stain were scanned, and their images were used to determine loading uniformity by quantifying proteins across all lanes using ImageQuant software (Molecular Dynamics). Stained membranes that exhibited uniform protein loading were subjected to immunoblotting with anti-P450 4A11. The protein stain was removed prior to immunoblotting using MemCode Destain TM , and the membranes were blocked in Tris-buffered saline Tween 20 (TBST) containing 5% nonfat dry milk for 1 h at room temperature. The membranes were incubated with the adsorbed P450 4A11 antibody at 1 g/ml in TBST for 2 h at room temperature, followed by incubation with a horseradish peroxide-conjugated monoclonal anti-rabbit IgG antibody (clone RG-96; Sigma-Aldrich) at a 1:30,000 dilution for 1 h. Immunoreactive proteins were visualized by ECL using the Western Lightning Chemiluminescence TM reagents (PerkinElmer Life Sciences). MemBlots and x-ray films were scanned, and the intensity for each band was analyzed using ImageQuant software, Molecular Dynamics (GE Healthcare). Varying amounts of P450 4A11 supersomes were also subjected to immunoblotting for the generation of standard curves that were used to determine P450 4A11 content in each microsomal sample and also to assure that the P450 4A11 signals obtained from the immunoblots were within the linear range of the standard curves.
Enzyme Assays-Metabolism of lauric acid was assessed in 50 mM KPO 4 buffer (0.5 ml) that contained either supersomes (3 pmol P450 4A11), liver microsomes (25 g) from transgenic mice, or liver microsomes (100 g) from nontransgenic mice. The samples were preincubated with 20 M 14 [C]lauric acid (55 mCi/mmol; American Radiolabeled Chemicals Inc.) for 10 min at 37°C. The reactions were initiated by the addition of an NADPH regeneration system consisting of 1 mM NADPH, 0.5 units/ml isocitrate dehydrogenase, 5 mM isocitrate, and 5 mM MgCl 2 at final concentrations. After 13 min of incubation at 37°C, the reactions were stopped by the addition of 1 N HCl to a final concentration of 0.2 N. Metabolites were extracted with water-saturated ethyl acetate as described (39). The reaction products were resolved by silica gel IB2-F TLC plates (J. T. Baker, Phillipsburg, NJ) with a solvent system composed of diethyl ether/light petroleum ether/formic acid (70:30:1) as described (40). TLC plates were analyzed using a PhosphoImager (Model SI) and quantified by ImageQuant software. The -hydroxylase activity was expressed as 12-OH lauric acid formation (nmol/min/mg microsomal protein) after correction for background generated from the same sample in the absence of the NADPH regeneration system.
Statistics-The values are expressed as the means Ϯ S.D. Comparisons of control versus drug administered or fed versus fasted mice were made by employing unpaired one-tailed Student's t test. Comparisons of multiple groups were achieved by two-way analysis of variance (ANOVA) with Bonferroni posthoc test using GraphPad Prism 5 software. p values Ͻ0.05 were considered statistically significant.

Production of CYP4A11 Transgenic Mice-Human CYP4A11
Tg mice were generated using BAC clone RP11Ϫ346m5 that harbors a 123-kb segment of human chromosome 1. Sequencing of RP11Ϫ346m5 by the Wellcome Trust Sanger Institute indicates that it contained the complete CYP4A11 gene including 58 kb of upstream DNA and 28 kb of downstream DNA spanning the CYP4A11-CYP4Z2P intergenic segment (Fig. 1A). Prior to generation of transgenics, the authenticity of RP11Ϫ346m5 was verified by PCR amplification of portions of the CYP4A11 gene using specific primer sets listed in supplemental Table S1. Analysis of PCR products on agarose gels verified the presence of the complete CYP4A11 gene in RP11Ϫ346m5 as judged by the amplification of all exons and a 5Ј-upstream segment (data not shown). DNA sequencing of a 917-bp fragment encompassing CYP4A11 exon 2 and exon 3 confirmed the predicted sequence differences between the CYP4A11 gene and the highly similar CYP4A22 gene. PCR amplification identified six potential transgenic mice designated CYP4A11 Tg-A through Tg-F. Amplification of all CYP4A11 exons revealed the presence of the complete gene in each founder as judged by amplification of all 12 exons and a portion of the 5Ј-flanking sequence (supplemental Fig. S1). The founder mice were bred into a C57Bl/6 background to produce CYP4A11 Ϫ/ϩ f 1 mice.
The extent of continuously integrated flanking upstream and downstream DNA was estimated by long PCR conducted on genomic DNA from all progenitors as well as examples of their f 1 progeny. The primers were designed to amplify 8 -10-kb genomic fragments (supplemental Table S2). Five overlapping segments encompassing 49.5 kb of 5Ј-CYP4A11 and three overlapping 3Ј-CYP4A11 segments spanning 28.7 kb of DNA were amplified (supplemental Fig. S1). Partial sequence obtained from DNA sequencing of these amplicons provided additional verification. Construction of the BAC library, which produced RP11Ϫ346m5, utilized the pBACe3.6 vector (35). An additional primer set was used to amplify upstream of position Ϫ49 kb to the cloning site in pBACe3.6, and this resulted in the amplification of a 9-kb fragment bridging the gap to the cloning site (data not shown). This result indicated the most complete integration of RP11Ϫ346m5 into the mouse genome in founder B, which contained roughly 58 kb of upstream and at least 28 kb of CYP4A11 downstream DNA. Founder F contained at least 49 kb of the 5Ј and at least 10 kb of the 3Ј intergenic regions in a continuous segment with the CYP4A11 gene.
P450 4A11 Is Expressed in Hepatic and Renal Microsomes from CYP4A11 Tg Mice-Each founder was bred into a C57Bl/6 background, and heterozygous CYP4A11 mice were utilized to assess the expression of P450 4A11 in liver and kidney microsomes, because these are identified sites of CYP4A11 gene expression in humans. Immunoblot analysis performed using a polyclonal antibody to human P450 4A11 demonstrated hepatic and renal P450 4A11 expression in heterozygous f 1 progeny of each founder at levels similar to levels seen in human liver and kidney (Fig. 1, B and C). Immunoreactive proteins of a molecular weight indicative of the endogenous mouse P450 4A proteins were not seen in CYP4A11 Ϫ/Ϫ microsomes (Fig. 1). The identity of the low molecular weight band recognized by the antibody in liver samples is unknown.
Expression of Functional P450 4A11 in Transgenic Micef 1 mice from two progenitors, founder B and F, were used for all subsequent experiments and are referred to as CYP4A11 Tg-B and CYP4A11 Tg-F. Comparison of CYP4A11 mRNA expression in males versus females was performed by qPCR. A twoway ANOVA with a Bonferroni post-test revealed the lack of significant differences in CYP4A11 expression between sexes in CYP4A11 Tg-B and CYP4A11 Tg-F mice (p Ͼ 0.05) but indicated a significant, roughly 3-fold difference in relative CYP4A11 mRNA expression level of CYP4A11 Tg-B versus CYP4A11 Tg-F mice (p Ͻ 0.05). To test the functionality of the CYP4A11 transgene product, liver microsomes from CYP4A11 Tg females were used to measure the capacity to catalyze lauric acid hydroxylation. As shown in Table 1, liver microsomes from nontransgenic littermates of CYP4A11 Tg-B or CYP4A11-Tg-F mice exhibited lauric acid -hydroxylation of 1.1 Ϯ 0.6 or 0.9 Ϯ 0.3 nmol/min/mg of protein, respectively. This activity is consistent with laurate -hydroxylation values reported for female C57Bl/6 mouse liver microsomes, which exhibited a specific 12-OH hydroxylation rate of 1.2 Ϯ 0.1 nmol/min/mg (41). The rate of 12-OH lauric acid formation was increased by 3.6 nmol/min/mg protein in liver microsomes from female CYP4A11 Tg-B and by 1.7 nmol/min/mg protein in microsomes from CYP4A11 Tg-F mice when compared with nontransgenic female littermates. The larger increase in laurate -hydroxylation in the CYP4A11 Tg-B animals is consistent with the higher level of expression of P450 4A11 in these mice relative to the CYP4A11 Tg-F mice. The increase in -hydroxylation of an exemplary fatty acid in microsomes from transgenic versus nontransgenic mice demonstrates that the transgene encodes a functional P450 4A11 protein and that the level of expression contributes to constitutive -hydroxylation capacity in transgenic mice.
Tissue Distribution of CYP4A11 mRNA in Transgenic Mice-Examination of CYP4A11 tissue distribution in CYP4A11 Tg-B mice by RT-PCR revealed abundant expression in liver and kidney, but not in other examined tissues (Fig. 2). CYP4A11 tissue  -F) and from a nontransgenic littermate ( Ϫ/Ϫ ) were resolved using SDS-PAGE and subsequently transferred to nitrocellulose membranes. The membranes were incubated with an antibody to P450 4A11 that did not cross-react with mouse P450 4A proteins in the nontransgenic mouse. Lanes marked liv1 and liv2 contain liver microsomes (5 g of protein) from two different human subjects, whereas those marked kid1 and kid2 contain kidney microsomes (5 g of protein) from two different human subjects.

TABLE 1 -Hydroxylation of lauric acid by hepatic microsomes from CYP4A11 transgenic mice
Lauric acid 12-hydroxylation was determined in hepatic microsomes from female nontransgenic (CYP4A11 Ϫ/Ϫ ) and CYP4A11 Tg-B or CYP4A11 Tg-F ( Ϫ/ϩ ) mice as described under "Experimental Procedures." Enzyme activity is expressed as nmol of 12-hydroxylaurate formed per min/mg microsomal protein after correction for background values generated from the same sample in the absence of the NADPH regeneration system. The values represent the means Ϯ S.D. for the number of animals (n) within a group. Statistically significant changes are indicated by p values determined by Student's t test.

Genotype
distribution was also assessed in CYP4A11 Tg-F mice and was consistent with tissue expression in line B. The results obtained for CYP4A11 mRNA at 30 PCR cycles was consistent with findings observed at lower cycles as shown in Fig. 2. The RT-PCR results are consistent with the gene atlas gene expression profiling of 79 human tissues (42,43). Concordant with these findings, hybridization of a CYP4A11 cDNA to a human tissue RNA membrane that included but was not limited to RNA from lung, pancreas, skeletal muscle, heart, brain, mammary, prostate, testis, ovaries, stomach, and other sections of the digestive system revealed abundant CYP4A11 mRNA transcripts in adult as well as fetal liver and kidney. 3 Skeletal muscle RNA exhibited a weak hybridization to the CYP4A11 cDNA. The expression of CYP4A11 in human skeletal muscle was re-evaluated in two additional skeletal muscle specimens by PCR and was not detected after 30 cycles of amplification (data not shown). Consistent with these findings, expression of CYP4A11 was not seen in skeletal muscle RNA of CYP4A11 transgenics at 30 PCR cycles (n ϭ 5). Fenofibrate treatment elevated CYP4A11 transcripts in liver and kidney and, despite induction by fenofibrate, remained below detection in other examined tissues (Fig. 2). The tissue distribution of the endogenous Cyp4a14 was evaluated in the same sample sets as an example of a PPAR␣-responsive mouse gene. Cyp4a14 exhibited a broader overall tissue distribution and in addition to liver and kidney could also be detected in stomach, muscle, heart, and ovaries and was inducible in liver, kidney, spleen, stomach, and muscle (Fig. 2). Under conditions that revealed expression of Cyp4a14 in tissues other than liver and kidney, CYP4A11 was not detectable under the same experimental conditions. Overall, these data show that tissue specificity of CYP4A11 expression described in humans is maintained in transgenic mice and demonstrate different overall tissue distribution for human CYP4A11 and mouse Cyp4a14.

Fasting
Increases Hepatic CYP4A11 Transcripts in CYP4A11 Transgenic Mice-To test whether the CYP4A11 gene can be regulated by fasting, CYP4A11 mRNA levels were measured in animals that were starved for 24 h. Fasting produced a 2.7-fold increase in CYP4A11 mRNA in CYP4A11 Tg-B livers compared with fed mice ( Table 2). Fasted CYP4A11 Tg-F mice exhibited nearly the same increase (3-fold) in CYP4A11 mRNA levels.
In these experiments, measurement of murine liver Cyp4a14 mRNA levels served as a positive control to monitor the activation of gene expression associated with fasting. Compared with the fed state, large increases of 73-91-fold were noted for murine Cyp4a14 mRNA after 24 h of fasting (Table 2), thus providing validation for the fasting model used here.
PPAR␣ Agonists Elevate Hepatic CYP4A11 mRNA in CYP4A11 Transgenic Mice-Feeding a clofibric acid-containing diet to CYP4A11 Tg-B mice resulted in a 3.2-fold increase of hepatic CYP4A11 mRNA levels. The PPAR␣ agonist fenofibrate also produced a 2-fold increase in CYP4A11 transcripts relative to control levels (Table 3). Similarly, clofibric acid or fenofibrate increased liver CYP4A11 mRNA levels by at least 2-fold in CYP4A11 Tg-F mice ( Table 3). As expected, large increases in hepatic Cyp4a14 mRNA expression were seen after the use of either PPAR␣ agonist.
Fasting and PPAR␣ Agonists Elevate P450 4A11-Immunoblot analysis was conducted to examine whether the induction of CYP4A11 mRNA was associated with an increase in the corresponding protein. For these studies, it was first necessary to develop an antibody that recognized only human P450 4A11 and not endogenous and/or inducible mouse P450 4A proteins. The lack of immunoreactivity between anti-human P450 4A11 IgG and liver microsomes from untreated mice is shown in Fig. 1. However, upon treatment with PPAR␣ agonists or after fasting, the human P450 4A11 antibody now displayed extensive reactivity with the enhanced levels of mouse P450 4A proteins. This species cross-reactivity of the P450 4A11 antibody prevented an accurate assessment of P450 4A11 protein levels (supplemental Fig. S2). To minimize such cross-reactions, the human P450 antibody was adsorbed against hepatic microsomes from fibrate-treated CYP4A11 Ϫ/Ϫ mice and purified P450 4A14. This adsorption process eliminated antibody recognition of murine P450 4A proteins while preserving reactivity toward human P450 4A11 (supplemental Fig. S2).
This improved 4A11 antibody was then used to assess hepatic 4A11 protein levels (Fig. 3). Basal hepatic P450 4A11 in CYP4A11 Tg-B or Tg-F mice were similar to levels seen in human liver and were increased by treatment with fenofibrate. In contrast, no cross-reactive bands, indicative of 3 Ü Savas and E. F. Johnson, unpublished observations. FIGURE 2. Tissue distribution of CYP4A11 mRNA in CYP4A11 Tg-B mice. The animals were fed a control (Ϫ) or fenofibrate-containing (ϩ) diet for 10 days as described under "Experimental Procedures." RNA was isolated from the indicated tissues and was used in RT-PCR with specific CYP4A11, Cyp4a14, and L27 primers. The reaction products were analyzed on 0.8% containing agarose gels and visualized by ethidium bromide staining. The results for CYP4A11, Cyp4a14, and L27 are shown after 22, 20, and 25 cycles of amplification, respectively. In addition to the examined tissues, white adipose was evaluated for the CYP4A11 transgene expression, and CYP4A11 mRNA could not be detected after 25 cycles of PCR amplification (data not shown). endogenous P450 4A proteins, were observed in hepatic microsomes from nontransgenic littermates fed with a fenofibrate containing diet. Basal hepatic P450 4A11 content in CYP4A11 Tg-B mice was 62 Ϯ 20 pmol/mg microsomal protein and increased to 152 Ϯ 20 pmol/mg microsomal protein after fasting (Table 4). Similar increases were also seen in CYP4A11 Tg-F mice, where P450 4A11 contents in fed and fasted mice were 42 Ϯ 8 and 104 Ϯ 41 pmol/mg. These results are consistent with the increases seen in hepatic CYP4A11 mRNA levels during fasting. Treatment with fenofibrate also led to a 2-3-fold increase in hepatic P450 4A11 protein (Table 4). In the CYP4A11 Tg-B mouse kidney, a 1.8-fold increase in P450 4A11 was noted upon fenofibrate administration (81 Ϯ 31 versus 148 Ϯ 38 pmol of CYP4A11/ mg). In CYP4A11 Tg-F mice, basal renal P450 4A11 content was 85 Ϯ 26 pmol/mg and increased 2-fold upon treatment with fenofibrate (Table 4).

PPAR␣ Deficiency Abrogates CYP4A11 Induction by Fenofibrate and Leads to Reductions in Constitutive CYP4A11 Expression-To
test the role of PPAR␣ in the regulation of the human CYP4A11 gene by fenofibrate, CYP4A11 Tg-B and Tg-F mice were generated that carried a PPAR␣ Ϫ/Ϫ genotype. The expression of P450 4A11 was assessed by immunoblotting of liver microsomes and compared with the results for PPAR␣ Ϫ/ϩ littermates. As shown in Fig. 4, P450 4A11 levels were substantially diminished in livers of PPAR␣-deficient female CYP4A11 Tg-B mice, whereas expression was largely retained in male transgenics with the PPAR␣ null genotype.
Quantitation of the immunoblots and analysis of P450 4A11 protein levels using two-way ANOVA indicated the absence of a significant effect of treatment (p Ͼ 0.05) for either males or females with PPAR␣ Ϫ/Ϫ genotype but a significant difference between sexes in the expression of P450 4A11 (p Ͻ 0.05). As expected, a two-way ANOVA indicated a significant induction of P450 4A11 by fenofibrate treatment in PPAR␣ Ϫ/ϩ transgenic males and females (p Ͻ 0.05), but a significant difference between males and females within each treatment group was not evident (p Ͼ 0.05). The decreased constitutive P450 4A11 levels in livers of female PPAR␣ Ϫ/Ϫ mice were less than 5% of the P450 4A11 levels seen in livers of female PPAR␣ Ϫ/ϩ littermates (Table 5). In PPAR␣-deficient male CYP4A11 Tg-B and Tg-F mice, P450 4A11 was expressed at 60 -80% of contents The animals were fed a control (Ϫ) or fenofibrate-containing (ϩ) diet for 10 days as described under "Experimental Procedures." Hepatic microsomes from control or fenofibrate fed CYP4A11 Tg-B and CYP4A11 Tg-F mice were prepared from liver, and 2.5 g were subjected to SDS-PAGE. Microsomes from fenofibrate-fed nontransgenic littermates ( Ϫ/Ϫ ) and human liver (h-liv) were included for comparison. Resolved proteins were transferred to nitrocellulose membranes, and a CYP4A11 antibody was used to detect CYP4A11 as described under "Experimental Procedures" (upper panel). The transferred proteins were visualized with MemCode TM protein stain prior to immunochemical staining (lower panel).

TABLE 2 Effect of fasting on CYP4A11 mRNA in CYP4A11 transgenic mice
Quantitative PCR was performed on liver RNA from CYP4A11 Tg-B and CYP4A11 Tg-F mice using CYP4A11, Cyp4a14, or L27 primers. The murine Cyp4a14 represented an example of a PPAR␣ target gene that is up-regulated during fasting and served as a positive control for the treatment protocol. The CYP4A mRNA levels were normalized by determining the CYP4A11/L27 and Cyp4a14/L27 ratios. The values represent the means Ϯ S.D. for the number of animals within each group. Statistically significant differences between fed and fasted groups were analyzed using Student's t test and are indicated by the p values.  and it also appears to be necessary for maintenance of normal levels of hepatic CYP4A11 expression in the absence of treatment.
The changes in P450 4A11 levels were paralleled by changes seen for CYP4A11 transcripts (Fig. 5). CYP4A11 mRNA levels were also diminished in female PPAR␣-deficient CYP4A11-Tg mouse liver and accounted for Ͻ5% of mRNA levels seen in PPAR␣ Ϫ/ϩ female livers. This was in contrast to male PPAR␣deficient CYP4A11-Tg mice, where hepatic CYP4A11 mRNA still accounted for approximately half of the mRNA levels seen in PPAR␣ Ϫ/ϩ males in CYP4A11 Tg-B mice. Line B and F exhibited ϳ2-fold induction by fenofibrate in PPAR␣ heterozygotes over control littermates. No induction of CYP4A11 mRNA by fenofibrate was noted in either male or female PPAR␣-deficient CYP4A11 transgenics. Similar to CYP4A11, constitutive mRNA levels for Cyp4a12 were decreased by ϳ50% in male PPAR␣ null mice in our study (data not shown).
A significant effect of fenofibrate on Cyp4a12 mRNA expression was not evident irrespective of the PPAR␣ genotype; however, induction of Cyp4a10 and Cyp4a14 mRNAs by fenofibrate could not be achieved in PPAR␣-deficient males or females (data not shown). Similar to the protein data, these results demonstrate that CYP4A11 mRNA induction by fenofibrate requires the presence of PPAR␣ and further suggest a contribution of PPAR␣ to the maintenance of normal hepatic CYP4A11 expression.
Modulation of Hepatic CYP4A11 mRNA Levels by Continuous Growth Hormone Infusion-Although a sex difference in basal CYP4A11 expression was not evident in PPAR␣-proficient mice, a CYP4A11 sexual dimorphism was observed in CYP4A11 liver expression in PPAR␣-deficient transgenics. Sexually dimorphic expression of the Cyp4a12 gene in mouse liver is thought to reflect sex differences in the pattern of growth hormone secretion by the pituitary gland. Females exhibit a low and relatively continuous pattern of secretion, whereas males exhibit a highly pulsatile pattern of secretion

Effects of fasting or administration of fenofibrate on P450 4A11 protein expression in CYP4A11 transgenic mice
The standard curves were generated by applying defined amounts of baculovirus-expressed P450 4A11 to Western blots. Blotting of samples used for standard curves as well as sample microsomes were performed in parallel, and the membranes were exposed for the same length of time. The standard curves were then used to quantitate microsomal P450 4A11 content by comparison of the immunochemical staining intensities obtained with microsomes to those obtained with the P450 4A11 standards. Statistically significant differences are indicated by the p values; n denotes the number of animals used. ND, not determined.  The animals were fed a control (Ϫ) or fenofibrate-containing (ϩ) diet for 10 days as described under "Experimental Procedures." Hepatic microsomes from heterozygous (CYP4A11 Ϫ/ϩ ) female (upper panel) and male (lower panel) CYP4A11 Tg-B mice that carried either a PPAR␣ Ϫ/ϩ or PPAR␣ Ϫ/Ϫ genotype were subjected to immunoblotting using an antibody for CYP4A11.

TABLE 5
Expression of P450 4A11 protein in male and female CYP4A11 transgenic mice with PPAR␣ ؊/؊ or PPAR␣ ؊/؉ genotypes Male and female CYP4A11 Tg-B and Tg-F mice that were either heterozygous ( Ϫ/ϩ ) or homozygous ( Ϫ/Ϫ ) for the disrupted PPAR␣ gene were fed a control or fenofibrate-containing diet for 10 days. Liver microsomes were prepared from each animal, and immunoblot analysis was used to measure P450 4A11 expression as described under "Experimental Procedures." The P450 4A11 content in microsomes is expressed as pmol/mg of protein. The results obtained for each PPAR␣ genotype were analyzed using a two-way ANOVA to examine the effects of treatment and sex on CYP4A11 expression. Differences between the individual means were tested for rejection of the null hypothesis using the Bonferroni post-hoc test with p Ͻ 0.05 considered to be significant. with deep troughs that fall below female growth hormone levels (14). Continuous infusion of growth hormone to male mice to override the pulsatile pattern of male secretion suppresses the expression of the Cyp4a12 gene in male mice (37). This protocol was used to examine whether elimination of the troughs in growth hormone secretion would reduce the expression of CYP4A11 in male mice to the levels seen in PPAR␣ Ϫ/Ϫ females. As expected, this treatment led to near complete suppression of Cyp4a12 expression in both PPAR␣ Ϫ/ϩ and PPAR␣ Ϫ/Ϫ male liver (data not shown). Under these conditions, the hepatic expression of CYP4A11 was suppressed by 5-10-fold in male PPAR␣ Ϫ/Ϫ mice by growth hormone exhibiting CYP4A11 mRNA levels similar to female PPAR␣ Ϫ/Ϫ mice (Fig. 6). The levels of CYP4A11 transcripts were also decreased by half in PPAR␣ Ϫ/ϩ males (Fig. 6). Examination of PPAR␣ homozygous female transgenics also revealed a 50% decline of CYP4A11 transcripts that was seen for Cyp4a14 as well (Fig. 7). These studies demonstrate that growth hormone is a negative regulator of CYP4A11 expression in mice and could act as such in humans. In the absence of PPAR␣, a sex difference in the temporal pattern of growth hormone secretion could underlie differences between males and females in the expression of CYP4A11 in mouse liver.
PPAR␣ Deficiency Abrogates CYP4A11 Induction during Fasting-The role of PPAR␣ in CYP4A11 induction by fasting was evaluated in PPAR␣ Ϫ/Ϫ CYP4A11 transgenic mice. Analogous to induction with fenofibrate, CYP4A11 was induced by fasting in PPAR␣ Ϫ/ϩ males by 1.7-(Tg-B) and 1.5-fold (Tg-F). Increases in females were 1.8-fold for both lines, and the increases were statistically significant in both lines as indicated in Fig. 8. Fasting did not significantly alter the expression of CYP4A11 in PPAR␣deficient females or males compared with the corresponding control, fed, mice (Fig. 8). Similar to the lack of CYP4A11 mRNA induction in the absence of PPAR␣, Cyp4a14 mRNA was not inducible during fasting in PPAR␣-deficient mice (data not shown). Overall the results in PPAR␣-deficient CYP4A11 transgenics suggest the requirement of PPAR␣ for the regulation of the CYP4A11 gene during fasting or by fenofibrate and imply that modulators of PPAR␣ could affect expression of CYP4A11 in humans.

DISCUSSION
In this study, CYP4A11 transgenic mice were generated to examine the effects of complex physiological changes such as fasting on the in vivo regulation of the human CYP4A11 gene. Herein, we successfully produced two transgenic mouse lines from a BAC clone that contained large portions of CYP4A11 Relative mRNA levels are expressed as CYP4A11/L27 ratios. The values represent the means Ϯ S.D. of n ϭ 5. The results were analyzed by two-way ANOVA with a Bonferroni post-hoc test and showed statistically significant differences between the effect of genotype within each treatment group for the expression of CYP4A11 in males and females. Significant differences for the effect of treatment within each genotype are indicated. *, p Ͻ 0.05. FIGURE 6. Effect of growth hormone on CYP4A11 expression in PPAR␣ ؊/؉ and PPAR␣ ؊/؊ CYP4A11 transgenic males. Male mice were continuously infused with recombinant mouse growth hormone (filled bars) over a period of 7 days using osmotic mini-pumps as described under "Experimental Procedures." Liver RNA was subjected to qPCR for CYP4A11 mRNA measurements. The values represent the mean CYP4A11/L27 mRNA ratios Ϯ S.D. (n ϭ 3 for control; n ϭ 4 for growth hormone). The symbols denote statistically significant decrease by growth hormone relative to control within each genotype as determined by Student's t test. *, p Ͻ 0.05; ‡, p Ͻ 0.005; #, p Ͻ 0.001. Mice implanted with mini-pumps containing only buffer, sham treatment, did not exhibit significantly changed levels of CYP4A11 mRNA than mice without any treatment (control, open bars). upstream and downstream DNA and is likely to contain ciselements required for transcriptional regulation. Also, incorporation of large portions of DNA flanking the CYP4A11 gene provides insulation of the transgene from host regulatory elements surrounding the integration site in the mouse genome.
The CYP4A11 transgene is abundantly expressed in mouse liver and kidney, a situation that parallels what is observed for humans, as seen by cDNA hybridization to RNA from multiple human tissues 3 and reported for tissue arrays (42). Moreover, the microsomal concentration of hepatic and renal P450 4A11 in untreated transgenic mice is similar to levels detected in human liver and kidney microsomes. P450 4A11 expressed from the transgene in mouse liver microsomes is catalytically active, as assessed by the increased rates of microsomal laurate -hydroxylation compared with nontransgenic mice. A lauric acid turnover number of 38 nmol/min/nmol of P450 4A11 was obtained from the ratio of the differential increase in activity to microsomal P450 4A11 content. This value is consistent with the turnover number for laurate with P450 4A11 purified from human liver (1). These results suggest that mice support CYP4A11 transgene expression that resembles human liver and kidney expression.
CYP4A11 mRNA contents were elevated ϳ3-fold by the PPAR␣ agonists or fasting in transgenic mice and paralleled the increases seen for P450 4A11 protein under the same conditions. In PPAR␣-deficient CYP4A11 transgenic mice, hepatic CYP4A11 mRNA and protein were no longer inducible by fenofibrates or fasting, indicating a PPAR␣ dependence of regulation by fasting or PPAR␣ agonists. The observations indicate that PPAR␣ is required for induction of CYP4A11 by fasting or by PPAR␣ agonists as has been shown for mouse Cyp4a genes here and previously by Northern blotting (18,25) or qPCR (20,44).
The near 3-fold induction of CYP4A11 gene expression in transgenic mice by PPAR␣ agonists is similar to the increases noted in CYP4A11 mRNA levels after clofibrate or fenofibrate treatment of human hepatocytes (31,33). Other investigators have reported only modest increases of CYP4A11 mRNA content in human hepatocytes treated with fenofibric acid, although 4-fold or greater increases of CYP4A11 transcripts were reported with a more potent proprietary PPAR␣ agonist (29). Similar to the induction of CYP4A11 by PPAR␣ agonists, 2-3-fold induction were also reported for CYP4A mRNA species in koalas (45) and beagle dogs (46).
Homozygous disruption of the PPAR␣ gene reduces basal expression of the CYP4A11 transgene in liver providing additional evidence that PPAR␣ plays an important role in the CYP4A11 gene expression. Similar results have been reported previously for other PPAR␣ target genes. A study by Corten et al. (20) reported a decline of basal Cyp4a14 FIGURE 7. Effect of growth hormone on CYP4A11 and Cyp4a14 expression in PPAR␣ ؉/؉ CYP4A11 transgenic females. Female PPAR␣ ϩ/ϩ mice were treated with recombinant mouse growth hormone (filled bars) as continuous infusion over a period of 7 days as described under "Experimental Procedures," and liver RNA was subjected to qPCR for evaluation of CYP4A11 and Cyp4a14 mRNA expression. The values represent the mean CYP4A11/L27 or Cyp4a14/L27 mRNA ratios Ϯ S.D. (n ϭ 3). The asterisk denotes statistically significant decrease by growth hormone (p Ͻ 0.005) as determined by a Student's t test.  F (n ϭ 3). The results were analyzed by two-way ANOVA with Bonferroni post-hoc test unless otherwise indicated. Statistically significant differences were noted between the effect of genotype within each treatment group in males and females for CYP4A11. The asterisks denote statistically significant differences (**, p Ͻ 0.05; *, p Ͻ 0.1) between fed versus fasted state in PPAR␣-proficient females or males. ‡, statistically significant increase (p ϭ 0.029) as determined by Student's t test. A statistically significant increase was not seen when the Bonferroni post-hoc test was applied. expression in PPAR␣ Ϫ/Ϫ versus PPAR␣ ϩ/ϩ mice. An important role for PPAR␣ in constitutive expression of enzymes that catalyze fatty acid ␤-oxidation and fatty acid synthesis has also been reported by other investigators. These include a decline in expression of the very long chain and long chain acyl-CoA dehydrogenases, short chain 3-ketoacyl-CoA thiolase, long chain acyl-CoA synthase, D-type peroxisomal bifunctional protein, and malic enzyme (18,47) in PPAR␣ null mice. Although significant differences in CYP4A11 expression between sexes is not apparent for CYP4A11 transgenic mice that are homozygous or heterozygous for PPAR␣, constitutive hepatic CYP4A11 expression is dramatically lower (Ͼ95%) in females in the absence of PPAR␣ expression, whereas male mice exhibited less than a 50% reduction. Differences in the temporal pattern of growth hormone secretion are known to underlie sex differences in the hepatic expression of some growth hormone responsive genes (14). Sex-specific regulation was documented for mouse Cyp4a12, which is expressed constitutively in livers and kidneys of adult males but not of adult females. The male-specific expression of Cyp4a12 in liver (48) is thought to reflect activation of signal transducer and activator of transcription (STAT) 5b in male mice as a result of temporal differences in the secretion of growth hormone between male and female mice. Male mice exhibit a pulsate pattern of growth hormone release that exhibits periods of little or no exposure, whereas female mice display a more continuous and lower pattern of growth hormone release from the pituitary gland. Infusion of growth hormone from osmotic mini-pumps has been used to mimic the more continuous exposure to growth hormone in the troughs seen in female mice and leads to a dramatic suppression of Cyp4a12 expression in the livers of male animals (37) as confirmed in our studies. This treatment also suppressed the expression of the CYP4A11 gene in male PPAR␣ Ϫ/Ϫ transgenic mice to levels similar to that seen in females.
Apart from its role in the regulation of sexually dimorphic hepatic gene expression, growth hormone modulates lipid metabolism by stimulating adipocyte lipolysis and promoting the production of very low density lipoprotein. Growth hormone has also been shown to suppress hepatic expression of PPAR␣ and other transcription factors that regulate lipid metabolism as well as genes involved in microsomal, peroxisomal, and mitochondrial fatty acid oxidation and ketone body production (49,50). As shown here, the infusion of growth hormone also suppressed the expression of CYP4A11 in PPAR␣proficient male and female mice, albeit to a lesser extent (50%), where a significant difference in the expression of CYP4A11 between the sexes is not evident. These results indicate that the effect of growth hormone on CYP4A11 expression is not specific to PPAR␣ null mice. It was speculated that growth hormone elicited reductions in PPAR␣ could underlie the effects of growth hormone on lipid oxidation (51); this could not account for the effects observed in our studies using PPAR␣ null mice. The strong suppression of CYP4A11 by growth hormone in male PPAR␣ Ϫ/Ϫ mice suggests that in the absence of PPAR␣, the effects of growth hormone are more pronounced. Differences in the temporal pattern of growth hormone exposure in male and female mice might contribute to the observed sex difference of CYP4A11 expression in PPAR␣ null mice.
An advantage of the CYP4A11 transgenic mouse model is the ability to examine the effects of changes in physiologic status on the expression of the transgene in an integrated biological model that cannot be mimicked by cell cultures. The mouse Cyp4a genes are up-regulated in response to fasting, and as shown in this study, the human CYP4A11 transgene is also upregulated by food deprivation. The magnitude of the response is similar to that seen for PPAR␣ activation by PPAR␣ agonists. These results suggest that CYP4A11 is regulated concordantly with other genes in the human liver to increase fatty acid oxidation (52). The role of microsomal -hydroxylation in these processes may be largely to protect the liver from excess free fatty acids caused by the increased flux from triglyceride hydrolysis.
Several reports indicate an association between the P450 4A11 F434S variant and the loss of P450 4A11 functionality that could contribute to increased systemic blood pressure in humans (9,10). Mice that are disrupted for the Cyp4a14 gene develop hypertension that is more pronounced in males, and Cyp4a10 null mice, similar to humans, develop hypertension that is increased by dietary salt uptake (12). Thus, it is conceivable that the up-regulation of the P450 4A enzymes may serve to lower the risk for hypertension, and treatment of Cyp4a10 null mice with the PPAR␣ agonist pirinixic acid ameliorated the hypertension in this model (12). Conversely, reduction in PPAR␣ contents as it occurs in aging (53) or by growth hormone (51) could increase this risk by decreasing P450 4A expression. Future efforts will focus on characterizing the role of P450 4A11 in hypertension by breeding CYP4A11 transgenic mice into the Cyp4a14 Ϫ/Ϫ and Cyp4a10 Ϫ/Ϫ backgrounds.