Coactivator PGC-1α regulates the fasting inducible xenobiotic-metabolizing enzyme CYP2A5 in mouse primary hepatocytes

Toxicology and Applied Pharmacology 232 (2008) 135–141

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y t a a p

Coactivator PGC-1α regulates the fasting inducible xenobiotic-metabolizing enzyme CYP2A5 in mouse primary hepatocytes Satu Arpiainen a, Sanna-Mari Järvenpää a, Aki Manninen b, Pirkko Viitala a, Matti A. Lang c, Olavi Pelkonen a, Jukka Hakkola a,⁎ a b c

Department of Pharmacology and Toxicology, University of Oulu, Oulu, Finland Biocenter Oulu, Department of Medical Biochemistry and Molecular Biology, University of Oulu, Oulu, Finland Division of Pharmaceutical Biochemistry, Uppsala Biomedical Center, Uppsala University, Uppsala, Sweden

a r t i c l e

i n f o

Article history: Received 15 April 2008 Revised 5 June 2008 Accepted 8 June 2008 Available online 17 June 2008 Keywords: cAMP Cytochrome P450 Drug metabolism Fasting HNF-4α PGC-1α

a b s t r a c t The nutritional state of organisms and energy balance related diseases such as diabetes regulate the metabolism of xenobiotics such as drugs, toxins and carcinogens. However, the mechanisms behind this regulation are mostly unknown. The xenobiotic-metabolizing cytochrome P450 (CYP) 2A5 enzyme has been shown to be induced by fasting and by glucagon and cyclic AMP (cAMP), which mediate numerous fasting responses. Peroxisome proliferator-activated receptor γ coactivator (PGC)-1α triggers many of the important hepatic fasting effects in response to elevated cAMP levels. In the present study, we were able to show that cAMP causes a coordinated induction of PGC-1α and CYP2A5 mRNAs in murine primary hepatocytes. Furthermore, the elevation of the PGC-1α expression level by adenovirus mediated gene transfer increased CYP2A5 transcription. Co-transfection of Cyp2a5 5′ promoter constructs with the PGC-1α expression vector demonstrated that PGC-1α is able to activate Cyp2a5 transcription through the hepatocyte nuclear factor (HNF)-4α response element in the proximal promoter of the Cyp2a5 gene. Chromatin immunoprecipitation assays showed that PGC-1α binds, together with HNF-4α, to the same region at the Cyp2a5 proximal promoter. In conclusion, PGC-1α mediates the expression of CYP2A5 induced by cAMP in mouse hepatocytes through coactivation of transcription factor HNF-4α. This strongly suggests that PGC-1α is the major factor mediating the fasting response of CYP2A5. © 2008 Elsevier Inc. All rights reserved.

Introduction The cytochrome P450 (CYP) enzymes, particularly in the subfamilies CYP1–3, play key roles in the phase I metabolism of xenobiotics, such as drugs, environmental toxins and dietary compounds (Nebert and Dalton, 2006). Similar to many other liver processes, the metabolism of xenobiotics is affected by nutritional homeostasis. Fasting and energy balance related diseases, including diabetes and obesity, alter the expression of many CYP enzymes, and this, consequently, changes the metabolism of drugs and other xenobiotics (Kim and Novak, 2007).

Abbreviations: CYP, cytochrome P450; PGC-1α, peroxisome proliferator-activated receptor γ coactivator 1α; cAMP, cyclic adenosine monophosphate; HNF-4α, hepatocyte nuclear factor 4α; PPAR, peroxisome proliferator-activated receptor; COH, coumarin 7-hydroxylase; Q-PCR, quantitative polymerase chain reaction; CREB, cAMP-response element binding protein; CAR, constitutive androstane receptor; PXR, pregnane X receptor; db-cAMP, dibutyryl cAMP; PEPCK, phosphoenolpyruvate carboxykinase. ⁎ Corresponding author. Department of Pharmacology and Toxicology, University of Oulu, POB 5000 (Aapistie 5B), 90014 University of Oulu, Oulu, Finland. Fax: +358 8 537 5247. E-mail address: jukka.hakkola@oulu.fi (J. Hakkola). 0041-008X/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.taap.2008.06.001

Although it is well established that alterations in the nutrition balance change the expression pattern of drug metabolizing enzymes, the mechanisms behind this effect are still largely unknown. Nutrition related hormonal changes, predominantly in circulating insulin and glucagon levels, are known to affect the regulation of several CYP enzymes (Kim and Novak, 2007). Insulin has been shown to downregulate CYP2E1 expression both by transcriptional and posttranscriptional mechanisms and through several signal transduction pathways. Studies with chemical kinase inhibitors suggest that PI3, Akt, mTOR and p70S6 kinases may be involved in CYP2E1 suppression by insulin (Kim and Novak, 2007, Woodcroft et al., 2002). Glucagon regulates various drug metabolizing CYP enzymes differently. It down-regulates CYP2C11, CYP2B1 and CYP2B2 while it up-regulates CYP2A5 and CYP2E1 (Iber et al., 2001, Sidhu and Omiecinski 1995, Viitala et al., 2001, Woodcroft and Novak 1999). The effects of glucagon are mediated by a specific membrane bound G protein-coupled receptor that activates adenylate cyclase and increases the intracellular cAMP level (Mayo et al., 2003). All the known effects of glucagon on CYP enzyme expression appear to be mediated by cAMP and predominantly through transcriptional regulation. However, the molecular constituents of the signal transduction pathways and the transcription factors involved remain unknown.

136

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

Peroxisome proliferator-activated receptor γ coactivator (PGC)-1α is involved in numerous biological responses related to energy metabolism, including thermal regulation and glucose homeostasis. The PGC1α level is elevated by a number of external stimuli and different signaling pathways are involved in PGC-1α regulation [see review by Handschin and Spiegelman, 2006]. cAMP signaling is a key activator of PGC-1α transcription in many tissues. PGC-1α activates nearly all known hepatic fasting responses, including gluconeogenesis, fatty acid β-oxidation, ketogenesis and bile acid homeostasis (Finck and Kelly 2006). PGC-1α induces the expression of the key gluconeogenic enzymes phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6phosphatase through interactions with the transcription factors hepatocyte nuclear factor (HNF)-4 and forkhead box O1 (FOXO-1) (Boustead et al., 2003, Puigserver et al., 2003, Rhee et al., 2003, Yoon et al., 2001). Both of these transcription factors are constitutively active and do not require an exogenous ligand for activation. HNF-4α maintains the transcription of several genes encoding for important hepatic xenobiotic-metabolizing enzymes, including CYP enzymes (Jover et al., 2001, Tirona and Kim 2005, Ulvila et al., 2004). Furthermore, a lack of PGC-1α down-regulates the expression of several HNF-4α transactivated CYP genes (Martinez-Jimenez et al., 2006). Bauer et al. (2004) have studied the starvation response in mouse liver by microarray analysis. The expression of several CYP genes was shown to be changed after 24 or 48 h of fasting. CYP2A5 was one of the CYPs most strongly induced by starvation. We have previously shown that glucagon is an efficient inducer of CYP2A5 expression and that this induction is mediated by cAMP (Salonpaa et al., 1994; Viitala et al., 2001). We hypothesized that PGC-1α is involved in the nutritional regulation of xenobiotic metabolism and mediates the induction of CYP2A5 by cAMP. In the current study, we present evidence indicating that elevated cellular levels of PGC-1α activate Cyp2a5 transcription through interaction with transcription factor HNF-4α. These results establish transcriptional coactivation by PGC-1α as a novel mechanism that mediates nutrition balance induced changes in the metabolism of xenobiotics. Methods Fasting experiments and coumarin 7-hydroxylase assay. Male DBA2/J mice, 10 weeks old, were caged in groups of five. The mice were kept at the Uppsala University animal house facilities in a 12 h light–dark cycle. The control groups had access to standard animal chow and water whereas the experimental groups had access only to water. The mice were sacrificed after 24, 48 or 72 h of fasting, and the livers were removed. The catalytic activity of CYP2A5 was determined by measuring coumarin 7-hydroxylase (COH) activity from the livers as described previously (Aitio, 1978). Preparation of primary cultures of hepatocytes. Hepatocytes were isolated from male DBA/2 (OlaHsd) mice (Center for Experimental Animals, University of Oulu, Finland) aged 8 to 10 weeks. Livers were perfused with collagenase solution (Worthington Biochemical Co., Lakewood, NJ, USA) as described previously (Salonpaa et al., 1994). After filtration and centrifugation, the isolated hepatocytes were dispersed in William's medium E (Sigma Chemical Co., St. Louis, MO, USA) containing dexamethasone (Sigma) 20 ng/ml, ITS (insulin 5 mg/l, transferrin 5 mg/l, sodium selenate 5 μg/l) (Sigma), gentamicin (Invitrogen, Paisley, Scotland) 50 μg/ml, and 10% fetal bovine serum (Invitrogen) at a density of 2 × 107 cells/175 cm2 flask, 1 × 107 cells/75 cm2 flask, 1 × 106 cells/one well in six-well plates, and 3 × 105 cells/one well in twelve-well plates (FALCON Polystyrene Cell Culture Dish, BD Biosciences, San Jose, CA, USA). The cultures were maintained at 37 °C in a humidified incubator for 1 to 2 h, after which non-attached cells were discarded by aspiration, and the medium was replaced by serum-free William's E medium. The cultures were maintained for an additional 24 h before dibutyryl

cAMP (db-cAMP) treatment, transient transfection or adenovirus infection. All animal experiments were approved by the local committees for laboratory animal welfare. Preparation of the Ad-PGC-1α virus and infection of hepatocytes. A recombinant adenovirus expressing mouse PGC-1α (Ad-PGC-1α) was prepared as follows: PGC-1α cDNA was amplified from the pcDNA3PGC-1α expression vector (a kind gift from Dr. J. K. Kemper) with the following primers: forward 5′ CCG CTC GAG CCA TGG CGT GGG ACA TGT GC 3′, where italicized nucleotides code for the XhoI restriction site and bases in bold indicate the Kozak sequence, and reverse 5′ GGC CTC GAG TTACCT GCG AAG CTT C (XhoI and STOP codon) [based on Bhalla et al. (2004)]. The cDNA was subcloned into the XhoI sites of the adenoviral shuttle vector, pShuttle-CMV (Qbiogene Inc., Illkirch, Cedex-France). An adenoviral vector was prepared according to the manufacturer's instructions with some modifications. Briefly, the shuttle vector was linearized with PmeI and transformed into BJ5183AD-1 competent cells using Gene Pulser Transfection Apparatus (BioRad, Hercules, CA, USA). The transformants were selected for kanamycin resistance, and the homologous recombination into the pAdEasy-1 vector was confirmed by PacI digestion. The resulting pAdEasy-1-PGC-1α vector was then linearized with PacI and transfected into QBI-293A cells (Qbiogene Inc.) using Lipofectamine 2000 (Invitrogen). Adenoviruses carrying the PGC-1α gene (Ad-PGC1α) were expanded into high-concentration stock by extracting the produced virus using freeze–thaw cycles and infecting fresh QBI-293A cells with the extracted virus. After four amplification cycles the viruses were concentrated and purified using 15%:30%:40% iodixanol (OPTIPREP, Axis-Shield PoC AS, Oslo, Norway) density gradient centrifugation (100 000 g, at 4 °C overnight). The purified viruses were diluted in glycerol buffer (15 mM Tris pH 8.0, 150 mM NaCl, 0.15% BSA and 50% glycerol) and in 90% glycerol (1:1:1, virus gradient: buffer: glycerol) and stored at −70 °C. The multiplicity of the infection for Ad-PGC-1α virus was titrated using the AdEasy™ Viral Titer Kit according to the manufacturer's instructions. Mouse primary hepatocytes were infected with Ad-PGC-1α or a control adenovirus expressing a membrane-targeted GFP (Ad-GFP-GLGPI) in Opti-MEM I medium (Invitrogen). The infected cultures were maintained at +37 °C in a humidified incubator for 1 h, after which the medium was replaced by serum-free William's E medium. The cultures were maintained for additional 12–72 h before RNA isolation and 48 h before protein extraction. Transduction efficiency of the hepatocytes with the used amounts of adenovirus is near 100% based on the observation of the cells infected with the GFP virus. RNA preparation and quantitative PCR. Murine hepatocytes (six-well plates) were treated with 25 μM dibutyryl cAMP (Sigma) or vehicle (dimethyl sulfoxide) only, for 0.5 – 12 h, after which the total RNA was isolated using the Tri-Reagent (Sigma) according to the manufacturer's protocol for monolayer cells. 1 μg of each RNA sample was reverse transcribed to produce cDNA using a First-Strand cDNA Synthesis Kit (Amersham Biosciences, Little Chalfont, UK) as suggested in the manufacturer's instructions. The quantitative PCR reactions for PGC-1α mRNA were done with an ABI 7700 Sequence Detection System using TaqMan chemistry. The sequences for the primers and probes used were as follows: mPGC-1α-FW 5′CAGTCTCCCCGTGGATGAA-3′, -RV 5′-GTGGTCACGGCTCCATCTG-3′, and -Tamra 5′-ACGGATTGCCCTCATTTGATGCACTG-3′, 18S-FW 5′TGGTTGCAAAGCTGAAACTTAAAG-3′, -RV 5′-AGTCAAATTAAGCCGCAGGC-3′ and -Tamra 5′-CCTGGTGGTGCCCTTCCGTCA-3′. For the measurement of PGC-1α mRNA levels from Ad-PGC-1α infected cells (6-well plates) and for all CYP2A5 and PEPCK mRNA measurements, 1 μg of each RNA sample was reverse transcribed to produce cDNA using p(dN)6 random primers (Roche) and M-MLV reverse transcriptase (Promega). The AmpliQ Universal Real Time PCR

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

137

Western blot. 30 μg of the 13 000 g supernatant proteins were subjected to SDS-polyacrylamide gel (10% polyacrylamide) electrophoresis. The proteins were transferred on to a Hybond ECL nitrocellulose membrane (Amersham Biosciences). The membrane was then incubated with primary rabbit polyclonal anti-PGC-1α antibody (SC-13067, Santa Cruz Biotechnology, Santa Cruz, CA, USA) (1:500 dilution) and secondary HRP anti-rabbit IgG (A6154, Sigma) (1:2000 dilution). After washing, the immunoreactive bands were visualized with the Chemiluminescent Peroxidase Substrate (CPS) 1 reaction (Sigma). The immunoreactive bands were quantitated using Quantity One software (Bio-Rad, Hercules, CA, USA). Plasmids and transient transfection assays. The preparation of Cyp2a5 5′-Luc reporter constructs and the site-directed mutation of them have been described previously (Ulvila et al., 2004). The pcDNA3-PGC-1α expression vector was kindly provided by Dr. J. K. Kemper (University of Illinois at Urbana-Champaign, Urbana; Illinois, USA). The reporter gene constructs were transfected into mouse hepatocytes (12-well plates) together with Renilla luciferase reporter vector (pRL3-TK) (Promega), which was used as an internal control. 0.5 μg of Cyp2a5 5′-Luc, 0.1 μg of pRL3-TK and 0.1 μg of expression vector DNA were transfected per 3 × 105 hepatocytes using Tfx-20 reagent (Promega) according to the manufacturer's protocol in OptiMEM I medium (Invitrogen). 48 h after the transfection, the cells were assayed for luciferase activity with the Dual-Luciferase Reporter Assay System (Promega).

Fig. 1. Effect of cAMP on PGC-1α and CYP2A5 mRNA expression in hepatocytes. Hepatocytes were treated with 25 μM db-cAMP or vehicle (dimethyl sulfoxide) only and RNAs were extracted from the cells at 0.5, 2, 6 and 12 h. Quantitative PCR analysis of PGC-1α (A), CYP2A5 (B) and PEPCK (C) mRNAs. Bars represent relative RNA levels normalized with 18S control using the comparative CT method. Values are means of 3–4 individual culture samples. The means + SD of four normalized samples were compared to cells treated with dimethyl sulfoxide. The difference to untreated cells is ⁎⁎⁎P b 0.001 and ⁎P b 0.05 (Student's t test). The experiments were repeated, and these independent experiments gave similar results.

Master Mix Kit (Ampliqon, Copenhagen, Denmark) was used for realtime quantitative PCR (Q-PCR) together with the following primers: mPGC-1α-FW 5′-CTGCTCTGGTTGGTGAGGA-3′ and -RV 5′-GCAGGCTCATTGTTGTACTG-3′, CYP2A5-FW 5′-CCAAGAAAGTGGAACACAATCA -3′ a n d -RV 5 ′- G G G GT TC TT C TT CT C CT CC A - 3 ′ , PE PCK-FW 5′ GGTGTTTACTGGGAAGGCATC and -RV 5′-CAATAATGGGGCACTGGCTG or m18S-FW 5′-CGCCGCTAGAGGTGAAATTC-3′ and -RV 5′-CCAGTCGGCATCGTTTATGG-3′, in an Mx3000P Q-PCR system (Stratagene). The fluorescence values of the Q-PCR products were corrected with the fluorescence signals of the passive reference dye (ROX). The RNA levels of PGC-1α, CYP2A5 and PEPCK were normalized against the 18S control levels using the comparative CT (ΔΔCT) method, as described in the Methods and Application Guide, Introduction to Quantitative PCR (Stratagene, IN #70200-01/Revision #105002).

Chromatin immunoprecipitation (ChIP) and re-ChIP. Immunoprecipitation of protein–DNA complexes was done from murine primary hepatocytes (in 175 cm2 flasks). Chromatin immunoprecipitations were carried out according to Abu-Bakar et al. (2007) with the addition of re-ChIP experiments. 1 μg of rabbit polyclonal anti-HNF-4α antibody (H-171, sc-8987, Santa Cruz Biotechnology), rabbit polyclonal anti-PGC-1 antibody (H-300, sc-13067, Santa Cruz Biotechnology) or rabbit IgG (Zymed Laboratories, South San Francisco, CA, USA) as a negative control were used for the first chromatin immunoprecipitations. For the second immunoprecipitation (re-ChIP), the immuno-complexes, collected and washed after the first immunoprecipitation, were eluted by adding 200 μl of 10 mM dithiothreitol and incubating at 22 °C for 30 min with rotation. The beads were spun down and the supernatants were diluted 1:40 in dilution buffer. 2 μg of antibody against the second protein of interest was added and incubated at 4 °C overnight. The new immunocomplexes were collected by incubating with 120 μl of protein A agarose slurry at 4 °C for 2 h and then washing as for the first ChIP. DNA fragments from both ChIP- and re-ChIP samples were extracted and resuspended in 60 μl of H2O. A dilution series was prepared from an input sample containing the total fragmented chromatin. 5 μl of precipitated DNA as a template, the following PCR primers (Cyp2a5 5′ HNF-4 RE −131FW: 5′-CAGTGTTGGCAATGTCCCAA-3′ and Cyp2a5 5′ HNF-4 RE +10RV: 5′-GATAGACAGACAGTGATGGC-3′) and the AmpliQ Universal Real Time PCR Master Mix Kit (Ampliqon, Copenhagen, Denmark) were used for real-time PCR in an Mx3000P Q-PCR system (Stratagene). The fluorescence values of the real-time PCR products were corrected with the fluorescence signals of the passive reference dye (ROX). The specificity of the PCR products was confirmed with melting curve analysis and by size as determined by agarose gel electrophoresis. Chromatin immunoprecipitation method is based on the affinity of the antibodies used and therefore the results obtained with different antibodies cannot be compared with each other. Furthermore, the re-ChIP samples have gone through an additional immunoprecipitation and washing steps compared with the ChIP samples with one antibody. Therefore the re-ChIP values should not be directly compared with those of the ChIP samples with one antibody.

138

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

Fig. 2. Effect of PGC-1α adenovirus infection on the CYP2A5 expression in hepatocytes. (A) Recombinant adenoviruses expressing PGC-1α or GFP (MOI 2) were infected into murine primary hepatocytes. After 48 h, the cells were harvested and PGC-1α protein and mRNA and CYP2A5 mRNA levels were measured. 30 μg of 13 000 g supernatant protein/lane was subjected to immunoblotting and the blots were stained with anti-PGC-1α antibody. PGC-1α and CYP2A5 mRNAs were detected by quantitative PCR analysis. Bars represent relative mRNA levels normalized with 18S control using the comparative CT method. The values are the means + SD of three individual culture samples and are presented as fold induction compared to untreated control. (B) Ad-PGC-1α at the MOIs of 2 and 8 were infected into murine primary hepatocytes. After 0, 12, 24 and 72 h the cells were harvested and three individual samples in each time point were pooled. PGC-1α protein level was determined with immunoblotting with anti-PGC-1α antibody. The values are presented as fold induction compared to untreated control in each time point. Immunoblot at 72 h time point is shown as an example. (C) From the same samples as in figure B PGC-1α, CYP2A5 and PEPCK mRNAs were measured by quantitative PCR analysis by using a pool of three individual culture samples. Bars represent (mean + range of two measurements) relative mRNA levels normalized with 18S control using the comparative CT method. Data is presented relative to the untreated sample in 0 h time point. The fold induction compared to the untreated sample in each time point is indicated above each bar. The difference to the control samples is statistically significant ⁎⁎⁎P b 0.001 or ⁎⁎P b 0.01 (one-way ANOVA followed by LSD post hoc test).

Statistical analysis. Student's t test was used for comparisons between two groups. Comparisons of several groups were done with one-way ANOVA followed by the least significant difference post hoc test. Differences were considered significant when P b 0.05. Results Fasting induces CYP2A5 activity in mouse liver CYP2A5 mRNA has been shown to be induced in mouse liver after fasting. Bauer et al. (2004) measured the starvation responses of 129/ SV male mice, aged 8–15 weeks using microarray analysis. 1.84- and 2.18-fold inductions in the CYP2A5 mRNA level were detected after 24 and 48 h of fasting, respectively. We measured the COH activity,

catalyzed specifically by CYP2A5, in DBA2/J male mouse livers after different periods of fasting to find out if the induction could also be seen at the enzyme activity level. We could detect 1.3-, 1.4-, and 2.0fold induction in COH activity at 24, 48 and 72 h after fasting, respectively (data not shown). These results indicate that CYP2A5 activity is also regulated by the nutritional status of the animal. cAMP causes coordinated up-regulation of PGC-1α and CYP2A5 mRNA levels Activation of the cAMP signaling pathway by glucagon is the main mechanism causing the induction of PGC-1α in fasted liver. We next studied the temporal pattern of CYP2A5 and PGC-1α induction by cAMP. Primary hepatocytes were treated with 25 μM db-cAMP and

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

Fig. 3. Effect of PGC-1α co-transfection on the function of the Cyp2a5 5′ promoter in hepatocytes. The expression vector of PGC-1α was co-transfected with the Cyp2a5 5′ promoter-Luc reporter plasmids into murine primary hepatocytes. After 48 h, the cells were harvested and the luciferase activities measured. The activities produced by the studied constructs were normalized against the co-transfected control plasmid (pRLTK) activities. The values represent the means + SD of 3–4 individual samples. ⁎⁎⁎The difference to the control with the empty expression vector is statistically significant P b 0.001 (Student's t test). The experiment was repeated twice with similar results.

PGC-1α, CYP2A5 and PEPCK mRNA levels were measured at several time points. The PGC-1α mRNA reached the maximal level only 2 h after treatment, after which the expression level started to decline. Both the CYP2A5 and PEPCK mRNA levels were induced after 2 h and remained at elevated level still after 12 h (Fig. 1). Temporal induction pattern of CYP2A5 was thus similar to a well-established PGC-1α target gene PEPCK. Ad-PGC-1α induces the expression of endogenous CYP2A5 in hepatocytes To study the direct role of PGC-1α in the induction of CYP2A5 we produced a recombinant adenovirus expressing mouse PGC-1α. The expression of PGC-1α in infected cells was first studied by Western blot and by quantitative reverse transcriptase PCR. After 48 h both the PGC-1α mRNA and protein levels were increased (Fig. 2A). The level of CYP2A5 mRNA was measured from the same samples and parallel upregulation of CYP2A5 and PGC-1α was seen while the GFP expressing control virus had no effect on either gene (Fig. 2A). To further establish the relationship between PGC-1α and CYP2A5 expression levels we next performed a detailed time course experiment with two different amounts (2 and 8 MOIs) of Ad-PGC-1α. PGC-1α protein expression was found to be dose dependently increased 24 h after infection and the expression was further increased up to 9.7-fold compared with the control in the latest time point after 72 h (Fig. 2B). From the same samples we also measured PGC-1α, CYP2A5 and PEPCK mRNA expression (Fig. 2C). CYP2A5 and PEPCK mRNA levels (but not PGC1α mRNA) were decreased in untreated cells during the culturing. In Ad-PGC-1α treated cells PGC-1α mRNA was elevated after 12 h and kept increasing still in the latest time point of 72 h. CYP2A5 mRNA was modestly, but significantly induced after 24 h with the higher Ad-PGC1α dose. Moreover, after 72 h CYP2A5 was induced dose dependently up to 12.4-fold compared with the control. PEPCK mRNA was upregulated already after 12 h and the highest induction (350-fold) was detected after 24 h. Thus CYP2A5 mRNA level appears to be less sensitive to and more slowly regulated by increased PGC-1α expression than PEPCK mRNA level. However, CYP2A5 mRNA level correlation to PGC-1α protein level was actually better than that of PEPCK mRNA. PGC-1α up-regulates CYP2A5 promoter function via the HNF-4 response element To localize promoter regions involved in Cyp2a5 regulation by PGC-1α we co-transfected the Cyp2a5 5′-Luc reporter gene together

139

with the PGC-1α expression vector into mouse primary hepatocytes. PGC-1α activated the Cyp2a5 promoter regulated luciferase activity about 3-fold and the construct with only 271 bp of the proximal promoter was activated equally to the longer constructs (Fig. 3). We have previously shown that HNF-4α and NF-I are the key factors that activate the Cyp2a5 proximal promoter (Ulvila et al., 2004). To investigate if these factors were involved in the PGC-1α mediated regulation of Cyp2a5, we mutated the HNF-4 and NF-I binding sites in the Cyp2a5 5′ promoter constructs. As expected, the HNF-4 and NF-I binding site mutations decreased the basal promoter activity to 7% and 8% of the activity of a non-mutated construct, respectively (results not shown). The induction of the luciferase activity by PGC1α was totally abolished by the HNF-4 binding site mutation. In contrast, mutation of the NF-I site did not affect PGC-1α induction (Fig. 3). HNF-4α and PGC-1α bind to the same regulatory complex in the Cyp2a5 proximal promoter The binding of PGC-1α to the endogenous Cyp2a5 5′ promoter was studied by chromatin immunoprecipitation. HNF-4α and PGC-1α antibodies were used to immunoprecipitate the fixed protein–DNA complexes from murine primary hepatocytes. Cyp2a5 proximal promoter fragment was found in the complexes immunoprecipitated by both of these antibodies (schematic presentation in Fig. 4A). Also, in the re-ChIP experiment, sequential immunoprecipitation first by the HNF-4α antibody and then by the PGC-1α antibody produced the same DNA fragment. The re-ChIP shows that the proteins bound simultaneously (directly or indirectly) to the same DNA fragment. As a coactivator is not able to bind to DNA itself, this suggests (however it

Fig. 4. In vivo binding of HNF-4α and PGC-1α to the Cyp2a5 5′ promoter. In chromatin immunoprecipitation assays HNF-4α and PGC-1α antibodies were used to precipitate fixed DNA–protein complexes from murine primary hepatocytes. The extracted DNA fragments were amplified with specific primers using real-time PCR and the relative amounts of the DNA copies were counted by comparing the sample fluorescence to the fluorescence values measured from an input dilution series. (A) Schematic picture of the proximal Cyp2a5 5′ promoter and the site of the amplified DNA fragment. (B) The relative amounts of DNA copies amplified with HNF-4 response element (RE) primers. The values represent the means + range of two samples. The experiment was repeated with similar results.

140

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

does not prove) that PGC-1α and HNF-4α are in the same protein complex (Fig. 4B). Discussion Cyp2a5 gene expression has been reported to be induced in liver by fasting (Bauer et al., 2004). Our current results show that this is reflected also in the enzyme activity level. Our previous studies have indicated that glucagon, a major hormonal mediator of the fasting response, directs the Cyp2a5 induction through cAMP signaling (Salonpaa et al., 1994; Viitala et al., 2001). The current study presents evidence that establishes PGC-1α as the key factor controlling the Cyp2a5 expression through coactivation of HNF-4α. We were able to show that cAMP increases both PGC-1α and CYP2A5 mRNA expression in primary hepatocytes and that induction of CYP2A5 resembles that of a well-established PGC-1α target gene PEPCK (Yoon et al., 2001). Direct evidence that PGC-1α upregulation indeed induces CYP2A5 expression was obtained by overexpression of PGC-1α in hepatocytes and consequent induction of CYP2A5 mRNA expression. We have previously reported that HNF-4α plays a major role in Cyp2a5 transcription (Ulvila et al., 2004). We now show by chromatin immunoprecipitation that PGC1α interacts together with HNF-4α with the Cyp2a5 proximal promoter. Furthermore, PGC-1α co-transfection induced reporter gene activity under the control of the Cyp2a5 promoter indicating transcriptional regulation by PGC-1α. Finally, the PGC-1α response was abolished by mutation of the HNF-4α response element, but not by mutation of the NF-I binding site in the Cyp2a5 promoter. Altogether, these findings demonstrate that cAMP induces CYP2A5 expression via induced expression of PGC-1α and consequently increased coactivation of HNF-4α. Together with the previous evidence, indicating that PGC-1α is strongly induced by fasting through glucagon and cAMP mediated mechanism (Yoon et al., 2001) and that CYP2A5 is regulated by glucagon/cAMP pathway (Salonpaa et al., 1994; Viitala et al., 2001), these findings suggest that PGC-1α is the key factor mediating the fasting induction of Cyp2a5. Recently, Ding et al. (2006) showed that fasting and cyclic AMP induce the expression of nuclear receptor constitutive androstane receptor (CAR) through HNF-4α and PGC-1α interaction. They also suggested that CAR could mediate the fasting response of several xenobiotic-metabolizing enzymes. CAR is expressed mainly in the liver and possesses a strong constitutive, ligand independent transactivation domain. However, CAR is normally retained in the cytosolic complex with cytoplasmic CAR retention protein (CCRP) and 90-kDa heat shock protein (Hsp90), and it will only move to the nucleus and activate its target genes, including certain CYP enzymes, in response to its xenobiotic ligands or other signals (Kobayashi et al., 2003). Thus, it would seem likely that CAR induction by fasting may augment gene regulation by phenobarbital and other activators of CAR. Indeed, we have previously shown that phenobarbital and cAMP have an additive inductive effect on CYP2A5 expression (Salonpaa et al., 1994). Nevertheless, it is unclear if CAR in fact can affect constitutive gene regulation in the absence of a chemical activator of the receptor. Our current results suggest that direct coactivation of HNF-4α by an increased amount of PGC-1α regulates Cyp2a5 induction by fasting. This could be a feasible mechanism explaining regulation of also other HNF-4α regulated CYP genes by fasting. The shift from the fed to the fasted state involves major metabolic changes. The fasting response attempts to secure the energy balance of the body by stimulating fatty acid oxidation and gluconeogenesis in the liver (Finck and Kelly, 2006). In addition to up-regulation of gluconeogenesis, PGC-1α also facilitates other liver functions such as heme biosynthesis, for which the physiological significance is not as obvious (Handschin et al., 2005). Similarly, the rationale for the increased expression of xenobiotic-metabolizing enzymes during

fasting is unclear. Putatively, the specific enzymes increased during periods of food restriction may have specialized functions in the metabolism of endogenous compounds. Fasting is known to induce unconjugated serum bilirubin concentration in normal human individuals, more dramatically in patients with Gilbert's syndrome and also in several experimental animals (Cornelius 1993; Hirschfield and Alexander, 2006). This may be due to increased activity of the hepatic heme oxygenase 1 enzyme (Bakken et al., 1972). Besides being a degradation product of heme, bilirubin is believed to have antioxidant properties important for cellular protection from oxidative stress. However, excess concentrations of bilirubin are toxic and need to be carefully controlled (Kapitulnik, 2004). Bilirubin is normally mainly glucuronidated and subsequently excreted to the bile. However, in cases of impaired glucuronidation or drastic elevation of bilirubin, an oxidative bilirubin metabolism may offer an alternative detoxification pathway (Schmid and Hammaker, 1963). CYP2A5 has been reported to play a role in microsomal bilirubin oxidation in conditions with elevated bilirubin levels due to cadmium treatment and heme oxygenase 1 induction (Abu-Bakar et al., 2005). The physiological purpose of CYP2A5 induction during fasting could thus be to prevent dangerous increases of bilirubin levels and to work as part of the protective, metabolic safety network. PGC-1α is also a powerful regulator of reactive oxygen species (ROS) metabolism, and needed for the induction of many ROSdetoxifying enzymes (St Pierre et al., 2006). CYP2A5 is induced by many compounds and conditions that cause oxidative stress in the cell (Gilmore and Kirby, 2004). We have recently shown that oxidative stress activated transcription factor nuclear factor (erythroid-derived 2)-like 2 (Nrf2) regulates Cyp2a5 (Abu-Bakar et al., 2007). However, PGC-1α could represent a second pathway regulating Cyp2a5 in fasting induced oxidative stress. The functional role of CYP2A5 in response to oxidative stress still remains unexplored, although it has been suggested to be a component of an adaptive response mechanism to altered cellular redox (Gilmore and Kirby, 2004). Recent results from Liu et al. (2007) show that PGC-1α is rhythmically expressed in mouse liver and that PGC-1α itself upregulates the expression of proteins involved in the control of circadian rhythm. Cyp2a5 has also been shown to display circadian expression in mouse liver. Transcription factor D-site binding protein (DBP) is known to participate in the circadian regulation of CYP2A5, but the circadian rhythm in CYP2A5 expression is not totally eliminated in dbp−/− mice (Lavery et al., 1999). This indicates that some other factor, possibly PGC-1α, is also involved in the circadian regulation of Cyp2a5. Changes in feeding times have been shown to influence circadian rhythm in the liver (Damiola et al., 2000). Since PGC-1α expression is highly responsive to nutritional signals it is tempting to speculate that PGC-1α could also cause the changes in the circadian rhythm in response to alterations in feeding times. In conclusion, we have presented evidence that PGC-1α mediates the induced expression of CYP2A5 by cAMP in mouse hepatocytes through coactivation of transcription factor HNF-4α. This strongly suggests that PGC-1α is a major factor mediating the fasting response of CYP2A5. Induction by increased expression of a coactivator represents a novel concept to regulate xenobiotic-metabolizing enzymes. Induction of CYP2A5 expression by fasting suggests a physiological role for this enzyme in the energy balance affected metabolic functions. Furthermore, metabolism of xenobiotic substrates for CYP2A5 is accelerated. Acknowledgments This study was supported by the Academy of Finland (contracts 110591 and 1114330) and the Finnish Technological Research Agency. The skillful technical assistance of Päivi Tyni and Ritva Tauriainen is gratefully acknowledged. We are also thankful to Mika Ilves (Department of Physiology, University of Oulu, Oulu, Finland) and Esa Huusela

S. Arpiainen et al. / Toxicology and Applied Pharmacology 232 (2008) 135–141

(AH Diagnostics, Helsinki, Finland) for their invaluable help with quantitative PCR. References Abu-Bakar, A., Moore, M.R., Lang, M.A., 2005. Evidence for induced microsomal bilirubin degradation by cytochrome P450 2A5. Biochem. Pharmacol. 70, 1527–1535. Abu-Bakar, A., Lamsa, V., Arpiainen, S., Moore, M.R., Lang, M.A., Hakkola, J., 2007. Regulation of CYP2A5 gene by the transcription factor nuclear factor (erythroidderived 2)-like 2. Drug Metab. Dispos. 35, 787–798. Aitio, A., 1978. A simple and sensitive assay of 7-ethoxycoumarin deethylation. Anal. Biochem. 85, 488–491. Bakken, A.F., Thaler, M.M., Schmid, R., 1972. Metabolic regulation of heme catabolism and bilirubin production. I. Hormonal control of hepatic heme oxygenase activity. J. Clin. Invest. 51, 530–536. Bauer, M., Hamm, A.C., Bonaus, M., Jacob, A., Jaekel, J., Schorle, H., Pankratz, M.J., Katzenberger, J.D., 2004. Starvation response in mouse liver shows strong correlation with life-span-prolonging processes. Physiol. Genomics 17, 230–244. Bhalla, S., Ozalp, C., Fang, S., Xiang, L., Kemper, J.K., 2004. Ligand-activated pregnane X receptor interferes with HNF-4 signaling by targeting a common coactivator PGC1alpha. Functional implications in hepatic cholesterol and glucose metabolism. J. Biol. Chem. 279, 45139–45147. Boustead, J.N., Stadelmaier, B.T., Eeds, A.M., Wiebe, P.O., Svitek, C.A., Oeser, J.K., O'Brien, R.M., 2003. Hepatocyte nuclear factor-4 alpha mediates the stimulatory effect of peroxisome proliferator-activated receptor gamma co-activator-1 alpha (PGC-1 alpha) on glucose-6-phosphatase catalytic subunit gene transcription in H4IIE cells. Biochem. J. 369, 17–22. Cornelius, C.E., 1993. Fasting hyperbilirubinemia in Bolivian squirrel monkeys with a Gilbert's-like syndrome. Adv. Vet. Sci. Comp. Med. 37, 127–147. Damiola, F., Le Minh, N., Preitner, N., Kornmann, B., Fleury-Olela, F., Schibler, U., 2000. Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev. 14, 2950–2961. Ding, X., Lichti, K., Kim, I., Gonzalez, F.J., Staudinger, J.L., 2006. Regulation of constitutive androstane receptor and its target genes by fasting, cAMP, hepatocyte nuclear factor alpha, and the coactivator peroxisome proliferator-activated receptor gamma coactivator-1alpha. J. Biol. Chem. 281, 26540–26551. Finck, B.N., Kelly, D.P., 2006. PGC-1 coactivators: inducible regulators of energy metabolism in health and disease. J. Clin. Invest. 116, 615–622. Gilmore, W.J., Kirby, G.M., 2004. Endoplasmic reticulum stress due to altered cellular redox status positively regulates murine hepatic CYP2A5 expression. J. Pharmacol. Exp. Ther. 308, 600–608. Handschin, C., Lin, J., Rhee, J., Peyer, A.K., Chin, S., Wu, P.H., Meyer, U.A., Spiegelman, B.M., 2005. Nutritional regulation of hepatic heme biosynthesis and porphyria through PGC-1alpha. Cell 122, 505–515. Handschin, C., Spiegelman, B.M., 2006. Peroxisome proliferator-activated receptor {gamma} coactivator 1 coactivators, energy homeostasis, and metabolism. Endocr. Rev. 21, 770–783. Hirschfield, G.M., Alexander, G.J., 2006. Gilbert's syndrome: an overview for clinical biochemists. Ann. Clin. Biochem. 43, 340–343. Iber, H., Li-Masters, T., Chen, Q., Yu, S., Morgan, E.T., 2001. Regulation of hepatic cytochrome P450 2C11 via cAMP: implications for down-regulation in diabetes, fasting, and inflammation. J. Pharmacol. Exp. Ther. 297, 174–180. Jover, R., Bort, R., Gomez-Lechon, M.J., Castell, J.V., 2001. Cytochrome P450 regulation by hepatocyte nuclear factor 4 in human hepatocytes: a study using adenovirusmediated antisense targeting. Hepatology 33, 668–675. Kapitulnik, J., 2004. Bilirubin: an endogenous product of heme degradation with both cytotoxic and cytoprotective properties. Mol. Pharmacol. 66, 773–779.

141

Kim, S.K., Novak, R.F., 2007. The role of intracellular signaling in insulin-mediated regulation of drug metabolizing enzyme gene and protein expression. Pharmacol. Ther. 113, 88–120. Kobayashi, K., Sueyoshi, T., Inoue, K., Moore, R., Negishi, M., 2003. Cytoplasmic accumulation of the nuclear receptor CAR by a tetratricopeptide repeat protein in HepG2 cells. Mol. Pharmacol. 64, 1069–1075. Lavery, D.J., Lopez-Molina, L., Margueron, R., Fleury-Olela, F., Conquet, F., Schibler, U., Bonfils, C., 1999. Circadian expression of the steroid 15 alpha-hydroxylase (Cyp2a4) and coumarin 7-hydroxylase (Cyp2a5) genes in mouse liver is regulated by the PAR leucine zipper transcription factor DBP. Mol. Cell Biol. 19, 6488–6499. Liu, C., Li, S., Liu, T., Borjigin, J., Lin, J.D., 2007. Transcriptional coactivator PGC-1alpha integrates the mammalian clock and energy metabolism. Nature 447, 477–481. Martinez-Jimenez, C.P., Castell, J.V., Gomez-Lechon, M.J., Jover, R., 2006. Transcriptional activation of CYP2C9, CYP1A1 and CYP1A2 by HNF4{alpha} requires coactivators PGC1{alpha} and SRC1. Mol. Pharmacol. 70 (5), 1681–1692. Mayo, K.E., Miller, L.J., Bataille, D., Dalle, S., Goke, B., Thorens, B., Drucker, D.J., 2003. International Union of Pharmacology. XXXV. The glucagon receptor family. Pharmacol. Rev. 55, 167–194. Nebert, D.W., Dalton, T.P., 2006. The role of cytochrome P450 enzymes in endogenous signalling pathways and environmental carcinogenesis. Nat. Rev. Cancer 6, 947–960. Puigserver, P., Rhee, J., Donovan, J., Walkey, C.J., Yoon, J.C., Oriente, F., Kitamura, Y., Altomonte, J., Dong, H., Accili, D., Spiegelman, B.M., 2003. Insulin-regulated hepatic gluconeogenesis through FOXO1-PGC-1alpha interaction. Nature 423, 550–555. Rhee, J., Inoue, Y., Yoon, J.C., Puigserver, P., Fan, M., Gonzalez, F.J., Spiegelman, B.M., 2003. Regulation of hepatic fasting response by PPARgamma coactivator-1alpha (PGC-1): requirement for hepatocyte nuclear factor 4alpha in gluconeogenesis. Proc. Natl. Acad. Sci. U. S. A. 100, 4012–4017. Salonpaa, P., Pelkonen, O., Kojo, A., Pasanen, M., Negishi, M., Raunio, H., 1994. Cytochrome P4502A5 expression and inducibility by phenobarbital is modulated by cAMP in mouse primary hepatocytes. Biochem. Biophys. Res. Commun. 205, 631–637. Schmid, R., Hammaker, L., 1963. Metabolism and disposition of C14-bilirubin in congenital nonhemolytic jaundice. J. Clin. Invest. 42, 1720–1734. Sidhu, J.S., Omiecinski, C.J., 1995. cAMP-associated inhibition of phenobarbital-inducible cytochrome P450 gene expression in primary rat hepatocyte cultures. J. Biol. Chem. 270, 12762–12773. St Pierre, J., Drori, S., Uldry, M., Silvaggi, J.M., Rhee, J., Jager, S., Handschin, C., Zheng, K., Lin, J., Yang, W., Simon, D.K., Bachoo, R., Spiegelman, B.M., 2006. Suppression of reactive oxygen species and neurodegeneration by the PGC-1 transcriptional coactivators. Cell 127, 397–408. Tirona, R.G., Kim, R.B., 2005. Nuclear receptors and drug disposition gene regulation. J. Pharm. Sci. 94, 1169–1186. Ulvila, J., Arpiainen, S., Pelkonen, O., Aida, K., Sueyoshi, T., Negishi, M., Hakkola, J., 2004. Regulation of Cyp2a5 transcription in mouse primary hepatocytes: roles of hepatocyte nuclear factor 4 and nuclear factor I. Biochem. J. 381, 887–894. Viitala, P., Posti, K., Lindfors, A., Pelkonen, O., Raunio, H., 2001. cAMP mediated upregulation of CYP2A5 in mouse hepatocytes. Biochem. Biophys. Res. Commun. 280, 761–767. Woodcroft, K.J., Hafner, M.S., Novak, R.F., 2002. Insulin signaling in the transcriptional and posttranscriptional regulation of CYP2E1 expression. Hepatology 35, 263–273. Woodcroft, K.J., Novak, R.F., 1999. The role of phosphatidylinositol 3-kinase, Src kinase, and protein kinase A signaling pathways in insulin and glucagon regulation of CYP2E1 expression. Biochem. Biophys. Res. Commun. 266, 304–307. Yoon, J.C., Puigserver, P., Chen, G., Donovan, J., Wu, Z., Rhee, J., Adelmant, G., Stafford, J., Kahn, C.R., Granner, D.K., Newgard, C.B., Spiegelman, B.M., 2001. Control of hepatic gluconeogenesis through the transcriptional coactivator PGC-1. Nature 413, 131–138.