Assessment of energetic costs of AhR activation by β-naphthoflavone in rainbow trout (Oncorhynchus mykiss) hepatocytes using metabolic flux analysis

Toxicology and Applied Pharmacology 271 (2013) 86–94

Contents lists available at SciVerse ScienceDirect

Toxicology and Applied Pharmacology journal homepage: www.elsevier.com/locate/ytaap

Assessment of energetic costs of AhR activation by β-naphthoflavone in rainbow trout (Oncorhynchus mykiss) hepatocytes using metabolic flux analysis Rance Nault a,⁎, Hiba Abdul-Fattah a, Gleb G. Mironov a, b, Maxim V. Berezovski a, b, Thomas W. Moon a a b

Ottawa-Carleton Institute of Biology, Department of Biology and Centre for Advanced Research in Environmental Genomics, University of Ottawa, Ottawa, Ontario, K1N 6N5, Canada Department of Chemistry, University of Ottawa, Ottawa, Ontario, K1N 6N5, Canada

a r t i c l e

i n f o

Article history: Received 21 August 2012 Revised 18 March 2013 Accepted 1 April 2013 Available online 29 April 2013 Keywords: Energetic costs Metabolic flux analysis β-Naphthoflavone Hepatocytes Rainbow trout

a b s t r a c t Exposure to environmental contaminants such as activators of the aryl hydrocarbon receptor (AhR) leads to the induction of defense and detoxification mechanisms. While these mechanisms allow organisms to metabolize and excrete at least some of these environmental contaminants, it has been proposed that these mechanisms lead to significant energetic challenges. This study tests the hypothesis that activation of the AhR by the model agonist β-naphthoflavone (βNF) results in increased energetic costs in rainbow trout (Oncorhynchus mykiss) hepatocytes. To address this hypothesis, we employed traditional biochemical approaches to examine energy allocation and metabolism including the adenylate energy charge (AEC), protein synthesis rates, Na+/K+-ATPase activity, and enzyme activities. Moreover, we have used for the first time in a fish cell preparation, metabolic flux analysis (MFA) an in silico approach for the estimation of intracellular metabolic fluxes. Exposure of trout hepatocytes to 1 μM βNF for 48 h did not alter hepatocyte AEC, protein synthesis, or Na+/K+-ATPase activity but did lead to sparing of glycogen reserves and changes in activities of alanine aminotransferase and citrate synthase suggesting altered metabolism. Conversely, MFA did not identify altered metabolic fluxes, although we do show that the dynamic metabolism of isolated trout hepatocytes poses a significant challenge for this type of approach which should be considered in future studies. © 2013 Elsevier Inc. All rights reserved.

Introduction Persistent organic pollutants (POPs) are ubiquitous environmental contaminants. Although there are a number of naturally occurring POPs such as polyaromatic hydrocarbons (PAHs) produced through incomplete combustion such as forest fires and brush fires (Gesto et al., 2009), human activity has led to dramatic increases in POP levels through sources such as industrial and urban sewage effluent, atmospheric deposition, and run-off and groundwater (Zhou et al., 2010). Properties of POPs including their lipophility, potential for long-range transport and persistence pose significant challenges to organisms that inhabit aquatic environments (Miranda et al., 2008; Schwarzenbach et al., 2010). Organisms to cope with exposure to these environmental contaminants, evolved inducible mechanisms to detect and respond to these chemicals by increasing cellular defense and detoxification systems (Bains and Kennedy, 2004; DuRant et al., 2007; Hahn, 2002; Pascussi et al., 2008). Among these mechanisms, the aryl hydrocarbon receptor (AhR) is involved in the detection and response to several POPs including dioxin-like polychlorinated biphenyls (PCBs) and PAHs (Reynaud and Deschaux, 2006; Schwarzenbach et al., 2010; Zhou et al., 2010). Activation of this cytosolic receptor leads to its dimerization with the ⁎ Corresponding author. Fax: +1 517 432 2310. E-mail address: [email protected] (R. Nault). 0041-008X/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.taap.2013.04.017

aryl hydrocarbon receptor nuclear translocator (ARNT) in the nucleus (Barouki et al., 2007) and the subsequent action of this complex as a transcription factor by binding to dioxin response elements (DRE's) on nuclear DNA. This results in the transcription of the AhR gene battery (Pascussi et al., 2008) including cytochrome P4501A1 (Cyp1a1), a phase I metabolic enzyme which acts as the first line of defense (Goksøyr and Förlin, 1992; Pesonen et al., 1992). Induction of Cyp1a1, measured through 7-ethoxyresorufin-O-deethylase (EROD) activities, is commonly used as a biomarker of exposure to AhR agonists. Activation of these detoxification mechanisms is thought to lead to significant energetic challenges to the intoxicated organism (Beyers et al., 1999; Handy et al., 1999; Levesque et al., 2002; Sherwood et al., 2000). A heterotroph gains energy from food which is allocated to cellular and organismal processes including maintenance metabolism, growth, activity, and reproduction (Beyers et al., 1999). Upon exposure to contaminants, induction of detoxification become necessary for survival, likely resulting in increased maintenance metabolic costs and decreased availability of energy for other physiological processes (Beyers et al., 1999). Evidence for such energetic costs is demonstrated as decreased growth rates in Dieldrin-exposed largemouth bass (Micropterus salmoides) (Beyers et al., 1999) and yellow perch (Perca flavescens) from toxic metal contaminated lakes (Levesque et al., 2002; Sherwood et al., 2000), reduced activity in copper exposed rainbow trout (Handy et al., 1999), and impaired reproductive success and growth rates associated

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

with depletion of glycogen and lipid stores in zebrafish (Danio rerio) exposed to effluent from an industrial plant (Smolders et al., 2003). However, these studies highlight the multiple physiological processes that may be affected by exposure due to a suite of behavioral, environmental and hormonal factors (Beyers et al., 1999; Handy et al., 1999) resulting in a complex model for the assessment of energetic costs. These complexities may be reduced by examining energetic costs at the cellular level. For example, increased oxygen consumption and decreased adenylate energy charge (AEC) is observed with induction of the xenobiotic transport protein P-glycoprotein by its substrates rhodamine and doxorubicin in rainbow trout hepatocytes (Bains and Kennedy, 2005; Hildebrand et al., 2009). Similarly, Rissanen et al. (2003) observed an increase in cellular respiration and glycolytic activity upon exposure to dehydroabietic acid (DHAA), a contaminant found in wood industry effluent. The objective of this study therefore is to test the hypothesis that activation of the AhR by the model agonist β-naphthoflavone (βNF) (Aluru and Vijayan, 2008) results in increased energetic costs in rainbow trout hepatocytes. Although βNF is not found in aquatic environments, but is regularly used as a surrogate of PAH exposure and to examine AhR-mediated toxic responses (Aluru and Vijayan, 2008; Gesto et al., 2009; Navas and Segner, 2000; Pesonen et al., 1992). It is predicted that if increased energetic costs are associated with activation of the AhR, hepatocytes would enter a catabolic state to produce additional energy thus affecting specific enzyme activities and the availability of cellular energy as defined by changes in the AEC. Additionally, we hypothesize that protein synthesis and Na +/K+-ATPase activities, two energetically costly processes (Krumschnabel et al., 1994; Pannevis and Houlihan, 1992; Wieser and Krumschnabel, 2001) will decrease to compensate for increased energy requirements upon βNF exposure. This study will also employ for the first time a metabolic flux analysis (MFA) approach in a fish cell preparation as a tool to examine reorganization of metabolic pathways. Therefore, we focused on a single model compound (βNF) at concentrations eliciting maximal response (AhR activation) to ensure detection of differences and due to limited numbers of hepatocytes preventing incorporation of multiple compounds. MFA, an in silico approach to estimate metabolic flux based on pathway stoichiometry and the assumption of pseudo-steady state fluxes using linear programming (Uygun et al., 2007; Varma and Palsson, 1994), is often used for biotechnological application (Edwards et al., 2001; Klamt and Stelling, 2002; Mavrovouniotis and Stephanopoulos, 1990). Recently, this approach has been applied to rat hepatocytes to examine the effects of the pesticide Triadimefon on metabolic fluxes (Iyer et al., 2010). The advantage of the MFA approach is that these pathways and their stoichiometry are well known and permits the examination of metabolic pathways as a network rather than individual fluxes (Lee et al., 1999).

87

Hepatocyte isolation, quantification and viability. Hepatocytes were isolated as described by Dugan and Moon (1998). The final two rinses used modified Hanks' medium (in mM: 0.8 MgSO4, 0.33 NaH2PO4, 0.44 KH2PO4, 136.7 NaCl, 5.4 KCl, 5.4 Hepes, 5.0 Hepes-Na, 5.0 NaHCO3, 1.5 CaCl2, 1.5% de-fatted BSA, pH 7.63). Initial hepatocyte viability was assessed using Trypan Blue exclusion and hepatocytes were used only when initial viabilities > 90%. Hepatocytes were resuspended in Hanks' culture media (modified Hanks' supplemented with antibiotic-antimycotic (Invitrogen 15240), 1× essential (Sigma M5550) and non-essential (Sigma M7145) amino acids, and 5 mM glucose) and aliquoted into culture plates (see below) or to plastic conical centrifuge tubes using a repeater pipette in duplicate. Hepatocytes in plates were pre-incubated for 24 h at 13 °C prior to exposures (see below). Culture conditions used were similar to those commonly used and optimized for rainbow trout primary hepatocytes (Hildebrand et al., 2009; Moon et al., 1985; Rissanen et al., 2003) and within a time-frame that is reported to maintain consistent enzyme activities (Pesonen and Andersson, 1991). Cells in tubes were centrifuged at 80 ×g for 2 min at 4 °C and rinsed twice with sodium phosphate buffer (PBS; pH 7.4) to remove BSA from the media. Cell pellets were sonicated in 6% PCA (perchloric acid) and centrifuged at 10,000 ×g for 30 s. Pellets were solubilized in 0.5 M NaOH and assayed for protein using the BCA assay (Sigma-Aldrich) with BSA in 0.5 M NaOH to generate a standard curve. Cell viability was determined by lactate dehydrogenase (LDH) leakage to the extracellular media using Triton-X 100 (final 1% v/v) as described by Bains and Kennedy (2004). Briefly, LDH was measured spectrophotometrically as the decrease of NADH measured at 340 nm in the presence of 1 mM pyruvate using a SPECTRAmax PLUS 384 microplate spectrophotometer (Molecular Devices, Sunnyvale, CA). Characterization of AhR activity. Hepatocytes (15 mg mL −1) were plated in 48 well plates and exposed to DMSO (control), or 0.01, 0.1, 1, 10 μM for 24 and 48 h. At the end of the exposure, the culture media was aspirated and plates were frozen at − 80 °C for no more than one week. EROD activities were assessed with modifications according to Kennedy et al. (1995). In short, 150 μL PBS was added to each well of the thawed 48 well plate. 7-Ethoxyresorufin (1 mM) in methanol was diluted to 0.63 μM in PBS immediately before use and 25 μL was added to each well. Plates were incubated in the dark for 5 min then reactions were initiated with the addition of 25 μL NADPH (1.2 mM) and read kinetically for 20 min at 530/590 nm excitation/emission filters using a SpectraMax GeminiXS fluorometer (Molecular Devices). Activity was estimated over the linear part of the kinetic reaction from a resorufin standard curve and normalized to protein content. βNF concentration eliciting the maximal EROD activity was chosen for all subsequent experiments.

Materials and methods Chemicals. L-4,5[ 3H]-leucine, antibiotic-antimycotic and de-fatted bovine serum albumin was purchased from PerkinElmer Inc. (Waltham, MA), Invitrogen Life Technologies (Carlsbad, CA) and MP Biomedical LLC (Solon, OH), respectively. βNF and all other chemicals, unless stated otherwise, were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). Fish. Female juvenile rainbow trout, Oncorhynchus mykiss weighing 150 to 300 g were obtained from Linwood Acres Trout Farm (Campbellcroft, ON) and maintained for at least 2 weeks in 1185 L tanks prior to experiments. Trout were fed commercial trout pellets and maintained under a 12L:12D photoperiod in 13 °C dechloraminated city of Ottawa tap water in the University of Ottawa Aquatic Care Facility. No trout possessed mature gonads. All experiments were done under the approval of the University of Ottawa Animal Care Protocol Review Committee and adhere to the guidelines established by the Canadian Council on Animal Care for the use of animals in research and teaching.

Adenylate energy charge. Hepatocyte ATP, ADP and AMP levels were determined as described by Hildebrand et al. (2009) using the commercial kit available from Lonza (Basel, Switzerland). Briefly, hepatocytes 10 mg mL−1) were lysed using the lysis buffer provided with the kit and AMP and ADP were converted to ATP using myokinase (100 U mL−1; AMP only) and pyruvate kinase (40 U mL−1) in Tricine buffer (in mM: 40 Tricine, 0.04 PEP, 10 KCl, pH 7.75). ATP was then measured according to manufacturer's instructions using an Lmax II 384 luminometer (Molecular Devices, Santa Clara, CA). Adenylate levels were calculated based on a standard curve and by correction for conversion to ATP; AEC was calculated as ([ATP] + 0.5[ADP]) / ([AMP] + [ADP] + [ATP]). Protein synthesis rates. Protein synthesis rates were assessed as L-4,5 [ 3H]-leucine incorporation into the trichloroacetic acid (TCA) insoluble fraction similar to the method described by Krumschnabel et al. (2000) and Kwast and Hand (1993). Briefly, L-4,5[3H]-leucine was added to a concentration of 0.5 μCi mL−1 in each well of a 48 well plate containing

88

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

25 mg mL−1 hepatocytes. Resuspended hepatocytes were sampled at 0, 2 and 4 h post [ 3H]-leucine addition by aliquoting 30 μL hepatocyte suspension onto G6 glass fibre filters (Fisher Scientific, Toronto, Canada) in duplicate. Filters were allowed to air-dry for several seconds then placed in ice-cold 10% TCA containing 5 mM unlabelled L-leucine for 10 min. This was followed by two 15 min washes in 5% TCA at room temperature and two 10 min washes in 99% ethanol. Filters were air-dried and added to scintillation vials for counting in 5 mL Ready-Safe scintillation fluid (Beckman Coulter, Brea, CA, USA) and counted using a Beckman LS6500 with internal quench correction. Rubidium uptake. Rb + uptake, an indicator of Na +/K+-ATPase activity, was measured according to Krumschnabel et al. (2000) in 48 well plates containing 25 mg mL−1 hepatocytes. Following a 10 min incubation in K +-free culture media (in mM: 0.33 Na2HPO4, 0.44 H3PO4, 0.8 MgSO4, 136.9 NaCl, 5.4 RbCl, 5 Hepes, 5 Na-Hepes, 1.5 CaCl2, 1.0% BSA, pH 7.63), 150 μL cell suspension was removed from each well, centrifuged in plastic conical tubes at 10,000 ×g for 2 min and the pellet washed 3-times in ice-cold MgCl2 medium (in mM: 95 MgCl2, 10 imidazole). The cell pellets were lysed in 5% PCA containing 0.05% Triton-X 100 and centrifuged for 2 min at 10,000 ×g. The resulting supernatant was diluted 10-times in KCl solution (26.8 mM) and analyzed for Rb+ using a Varian AA240 atomic absorption spectrometer (Agilent technologies, Santa Clara, CA) with acetylene carrier gas at 780 nm with a slit width of 0.2 nm using RbCl in KCl (26.8 mM) to generate a standard curve. Metabolite analysis. Hepatocytes (25 mg mL −1) were aliquoted into 6 well plates and pre-incubated for 24 h after which the media was removed and replaced with fresh media containing βNF or DMSO. One well containing only culture media with no hepatocytes was also included to calculate metabolite uptake or production. Following 24 h (0–24 h fluxes) exposure, culture media was collected and replaced with new culture media containing the same treatment for an additional 24 h. Culture media was again collected at 48 h (0–48 h fluxes). Metabolite fluxes were estimated by subtracting metabolite levels in culture media alone from culture media from hepatocyte exposures divided by 24 h and protein content to obtain fluxes in moles h −1 mg −1 protein. Extracellular metabolites were assessed spectrophotometrically (SPECTRAmax PLUS 384 microplate spectrophotometer) in triplicate based on a standard curve generated under the same conditions. Glucose was estimated enzymatically as described in Bergmeyer (1985). Lactate was assessed in deproteinized (6% PCA) and neutralized (5 M K2CO3) samples by following the production of NADH measured at 340 nm generated in the presence of LDH (1500 U mL−1), NAD+ (2.5 mM) and glycine (1:2 v/v) in a 96-well plate. Cholesterol and triglycerides were determined using commercially available kits from Teco Diagnostics (Anaheim, CA, USA) adjusted for 96-well plates and glycerol was measured using the free glycerol reagent (Sigma-Aldrich) adjusted for 96-well plates. Glutamine, glutamate, acetoacetate and β-hydroxybutyrate were assessed in deproteinized and neutralized samples. Glutamine was estimated by deamination of glutamate using glutaminase (4 U mL−1) in 50 mM acetate buffer incubated at 37 °C for 1 h. 25 μL of the deaminated glutamine samples and samples treated similarly in the absence of glutaminase were assayed in a 96-well plate by the addition of 25 μL assay medium (in mM; 8.0 NAD+, 2.67 ADP) and 50 μL 1:20 (v/v) hydrazine in Tris-EDTA buffer (in mM; 100 Tris–HCl, 2 EDTA), and read kinetically at 340 nm for 5 min to estimate background activities. 10 μL L-glutamic dehydrogenase (1235 U mL−1) was then added and the formation of NADH was followed kinetically at 340 nm. Glutamine levels were calculated by subtracting non-deaminated (glutamate) samples from deaminated samples (glutamine and glutamate). Acetoacetate and β-hydroxybutyrate were assessed as described by Bergmeyer (1985) in duplicate. Briefly, acetoacetate was enzymatically

reduced to β-hydroxybutyrate by D-3-hydroxybutyrate dehydrogenase (3-HBDH; 15 U mL −1) in assay buffer (in mM; 100 Triethanolamine, 5 EDTA, 8 NADH, pH 7.5). Acetoacetate and β-hydroxybutyrate levels were then estimated by following the reduction of an Fe3+-BPS complex to Fe2+-BPS measured at 546 nm mediated by phenazine methosulfate. Acetoacetate levels were calculated by subtracting non-reduced (β-hydroxybutyrate) samples from reduced samples (β-hydroxybutyrate and acetoacetate). Media free fatty acids (FFA) were determined using gas chromatography similar to the method described by Vaillancourt et al. (2009) with the omission of lipid fraction separation due to the simplicity of the culture media. Briefly, lipids were extracted in chloroform: methanol (2:1 v/v; Folch medium) spiked with 9 μg 17:0 fatty acid. Following extraction, FFAs were methylated in 100 μL methanol, 1 mL dimethoxypropane and 40 μL HCl (12 N). Samples were dried under N2 gas and resuspended in isooctane and transferred to autosampler tubes. FFAs were analyzed using a 6890N network gas chromatograph system (Agilent technologies) and compared to pure fatty acid standards (Sigma-Aldrich) for retention time. Concentrations were determined as the area under individual peaks for all peaks that represented >1% of total free fatty acids and calculated based on the known concentration and area of the internal standard. Free amino acids (AA) were measured by capillary electrophoresis coupled to electrospray ionization mass spectrometry (CE-ESI-MS) using a PA800 Plus pharmaceutical analysis CE system (Beckman Coulter, Brea, CA, USA) and Waters SYNAPT G2 High Definition Mass Spectrometer (Waters, Milford, MA). Culture media samples and standards were filtered through a 0.22 μm filter (Fisher Scientific) prior to analysis. After filtrations samples were injected into a 90 cm fused silica capillary (50 μm ID) by pressure (1 psi 10 s). The 30 kV potential was used for separation with the positive polarity. 1% formic acid in water was used as a running buffer. The sheath liquid consisted of 50:50 methanol: water + 1% formic acid. MS voltages were as follows: capillary 3 kV, sampling cone 35 V, extraction cone 4 V. Purge gas was set to 5 L/h. Interface temperature was 100 °C. AA were quantified for [M + H]1+ ions with the mass window of 0.05 Da using QuanLynx software (Waters, Milford, MA) versus a standard curve of amino acid mixtures standards (0, 0.25×, 0.5×, 1× and 2× concentrations of a stock solution). Intracellular glycogen was assessed according to the method described by Keppler and Decker (1974). In short, hepatocytes were sonicated in ice-cold 6% PCA then centrifuged at 10,000 ×g for 5 min. Supernatants were transferred to a small glass tube and NaHCO3 (1 M) and amyloglucosidase (11.6 U mL−1) in acetate buffer (0.1 M) was added to the glass tube which was incubated at 35–40 °C for 2 h. At the end of the incubation, the reaction was stopped by the addition of 70% PCA, vortexed, and centrifuged at 6250 ×g. Glycogen was assayed according to Bergmeyer (1983) corrected for the presence of glucose. Metabolic flux analysis. Metabolic flux analysis is an in silico mathematical approach that optimizes flux through individual reactions when the metabolic network is in steady-state using known reaction stoichiometries (Iyer et al., 2010; Uygun et al., 2007; Varma and Palsson, 1994). The mathematical model for MFA was developed based on the pathways described by Iyer et al. (2010) and Yang et al. (2010). Using CellNetAnalyzer (Klamt et al., 2007) implemented in MATLAB (MathWorks Inc., Natick, MA), a stoichiometry matrix was developed by entering the known stoichiometry of individual reactions (Supplementary Table 1) and exported into the mathematical program GAMS integrated development environment version 26.6.3 (GAMS, Washington, DC). Constraints are used to limit upper and lower values of each reaction to optimize the fluxes. Flux constraints for measured metabolites (Supplementary Table 2) were determined by calculating maximum and minimum fluxes using the average measured flux plus or minus standard deviation, respectively. Flux ranges for glycerol, cholesterol

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

and triglyceride were set at − 10 to 10 nmol h −1 mg −1 protein as these were the smallest ranges that allowed a feasible model and may fall within the undetectable range of the assays used to measure these. Intracellular fluxes were constrained to infinite fluxes and irreversible fluxes were given a minimum flux rate of 0 nmol h −1 mg −1 protein (Supplementary Table 3). Because Cyp1a1 metabolism of βNF was not directly assessed, intracellular fluxes were modeled under Cyp1a1 constraints of 1, 10, 100 or 1000 nmol h −1 mg -1 protein with the control group constrained to 0 nmol h −1 mg −1 protein. Constraints were incorporated into the GAMS program and fluxes were then maximized and minimized using linear programming solved using CPLEX 12 solver (IBM Corporation, Armonk, NY) for each individual reaction assuming a steady-state system as described by Iyer et al. (2010). Enzyme activities. Hepatocytes (15 mg mL−1) were incubated in 48-well plates, exposed to βNF or DMSOfor 24 or 48 h, transferred to plastic centrifuge tubes, centrifuged and the medium replaced with homogenization buffer (1:15 w/v; in mM: 20 Tris–HCl, 5 EDTA, 5 MgCl2, 150 KCl, 5 β-mercaptoethanol, pH 7.4). Alanine aminotransferase (AAT; E.C. 2.6.1.2), glyceraldehyde-3-phosphate dehydrogenase (G-3PDH; E.C. 1.2.1.12), glucose-6-phosphate dehydrogenase (G6PDH; E.C. 1.1.1.49), β-hydroxyacyl CoA-dehydrogenase (HOAD; E.C. 1.1.1.35), phosphoenolpyruvate carboxykinase (PEPCK; E.C. 4.1.1.32), pyruvate kinase (PK; E.C. 2.7.1.40), isocitrate dehydrogenase (IDH; E.C. 1.1.1.42) and malic enzyme (ME; E.C. 1.1.1.40) activities were assessed by the appearance/disappearance of NADPH or NADH measured at 340 nm. Assay mixes consisted of: AAT (final in mM: 200 alanine, 50 imidazole, 0.2 NADH, 0.05 pyridoxal-5-phosphate, LDH 48 U mL−1, pH 7.4); G-3PDH (final in mM: 20 imidazole, 0.15 NADH, pH 7.6); G6PDH (final in mM: 50 imidazole, 7 MgCl2, 0.4 NADP, pH 7.4); HOAD (final in mM: 50 imidazole, 0.05 NADH, pH 7.4); PEPCK (final in mM: 80 Tris-HCl, 1 MgCl2, 1 MnCl2, 1.5 IDP, 0.17 NADH, 1 PEP, malate dehydrogenase 19 U mL−1, pH 7.4); PK (final in mM: 50 MOPS, 5 ADP, 100 KCl, 10 MgCl2, 0.1 F-1,6-P2, 0.15 NADH, LDH 80 U mL-1, pH 7.4); IDH (final in mM: 50 imidazole, 4 MgCl2, 0.4 NADP, pH 7.4); and ME (final in mM: 50 imidazole, 1 MgCl2, 0.4 NADP, pH 7.4). Plates (96-well) are measured kinetically for 5 min to observe background rates and reactions were then initiated by the addition of the substrate (AAT, 10 mM α-ketoglutarate;G-3PDH, 0.1 mM dihydroxyacetone phosphate; G6PDH, 1 mM G6P; HOAD, 0.1 mM acetoacetyl-CoA; PEPCK, 20 mM NaHCO3; PK, 5 mM PEP; IDH, 0.6 mM isocitrate; ME, 1 mM malate) and measured for 20 min. Citrate synthase (CS; E.C. 2.3.3.1) was estimated as the appearance of 5-thio-2nitrobenzoic acid measured at 412 nm. The sample was added to assay mix (final in mM: 50 Tris–HCl, 0.3 acetyl-CoA, 0,1,5,5′-dithiobis(2-nitrobenzoic acid, pH 8.0) in 96-well plate with and without oxaloacetate (0.5 mM) and measured kinetically. Rates in the absence of oxaloacetate were subtracted from those in presence of oxaloacetate.

89

among βNF concentrations or over time while all activities were significantly lower (p b 0.001) than the Triton X-100 positive control (Fig. 1B). Adenylate energy charge (AEC) Isolated hepatocytes exposed to βNF for 48 h exhibited no changes in AEC with values of 0.65 ± 0.02 in control and 0.65 ± 0.00 in βNF exposed hepatocytes (Fig. 2A). Moreover, levels of individual adenylates and total adenylates were also unchanged with levels of total adenylates in control cells of 1.24 ± 0.08 nmol mg−1 cells and 1.32 ± 0.08 nmol mg−1 cells in βNF treated cells (Fig. 2). Hepatocyte glycogen content Following 24 h exposure to βNF, no significant differences was observed in hepatocyte glycogen levels (Fig. 3A). Similarly, hepatocyte glycogen levels following 48 h exposure to βNF was not different from levels in control hepatocytes at either 0 h or 24 h. In contrast, glycogen levels are significantly lower (p b 0.05) in control hepatocytes at 48 h post-exposure. Protein synthesis and rubidium uptake Protein synthesis rates were determined as the rate of incorporation of L-4,5[ 3H]-leucine into the TCA-insoluble protein fraction. Protein synthesis rates showed no treatment or time-dependent effects

Statistical analysis. Statistical analyses were performed using Sigma Plot 11 (Systat Software Inc. San Jose, CA) using a two-way repeated measures ANOVA followed by a Holm–Sidak post hoc test when significant differences (p b 0.05) were identified. Results Characterization of AhR activation and cell viability Hepatocyte EROD activities increased with increasing βNF concentrations at both 24 and 48 h post-exposure reaching a maximum activity of 27.4 ± 4.0 pmol min−1 mg−1 protein at 24 h and 34.1 ± 4.1 pmol min−1 mg−1 protein at 48 h. Maximum activities were achieved at 1 μM βNF at both time points (Fig. 1A). EROD activity at 48 h post-exposure was also significantly higher (p b 0.05) than the 24 h post-exposure at concentrations of 0.1, 1 and 10 μM βNF. Furthermore, extracellular LDH activities were not significantly different

Fig. 1. EROD activities (A) and culture media LDH activities (B) for isolated rainbow trout hepatocytes at 24 h and 48 h incubation times and at different concentrations of βNF or at 1% Triton X-100 (TX) as a positive control for cell disruption. Data represent mean + S.E.M (n = 4). Statistical analysis consisted of a repeated measures two-way ANOVA for time and dose variables. Significant differences (p b 0.05) are represented by different letters within time points (dose dependent differences) while statistical differences within doses (time dependent differences) are indicated by asterisks.

90

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

Fig. 2. Adenylate energy charge (A) and levels of individual and total adenylates (B) in isolated hepatocytes exposed to βNF (1 μM) or DMSO vehicle control. Data represent mean + S.E.M. (n = 5). Statistical analysis consisted of a repeated measures two-way ANOVA; no significant differences were observed.

(Fig. 3B). Additionally, uptake of rubidium (Rb+) was assessed as an indicator of K + influx through Na+/K+-ATPase activity. Exposure to βNF showed no significant differences at any of the examined time points (Fig. 3C). However, Rb+ uptake was significantly different (p b 0.05) at 24 h for both control and βNF exposed hepatocytes. Although the Holm–Sidak post hoc test does not show significant differences at 48 h post-exposure, the less conservative Student–Newman–Keuls post hoc test indicates a statistically significant difference (p b 0.05) at this time point as well. Metabolic flux analysis Metabolite fluxes for all measured extracellular metabolites are shown on Table 1. The metabolite fluxes for ammonia, glycerol, triglyceride and cholesterol were undetectable at both time intervals preventing calculation of a metabolite flux. The flux for β-hydroxybutyrate and acetoacetate were detectable at low rates (in pmol h-1 mg-1 protein) compared with other metabolites (in nmol h−1 mg−1 protein) and showed no significant differences (p b 0.05). Similarly, amino acid and lactate fluxes did not show treatment or time-dependent significant differences. Alanine, arginine, histidine and proline showed the largest amino acid fluxes with alanine and proline being released to the extracellular media while arginine and histidine were taken up by the hepatocytes. Free fatty acid fluxes also showed no significant differences between treatment or duration of exposure. The absence of difference is likely due to large variability in fluxes (S.E.M values similar to means). On the other hand, glucose flux was significantly different (p b 0.05) at both time intervals assessed with glucose release occurring over the first 24 h period (0–24 h) and nearly an absence of flux in the second 24 h period (24–48 h). However, no differences were noted between control and βNF exposed hepatocytes. MFA was performed using the measured metabolite fluxes as constraints (see Materials and methods section) and the estimated flux

Fig. 3. Hepatocyte (A) glycogen content presented relative to control at 0 h (horizontal line; 2.73 ± 0.66 mg glycogen mg−1 protein), (B) L-4,5[3H]-leucine incorporation into the TCA-insoluble protein fraction relative to control at 0 h (8.0 ± 2.6 nmol h−1 mg−1 pro+ tein), and (C) Rubidium (Rb ) uptake relative to control at 0 h (0.6 ± 0.1 nmol Rb + min− 1 mg− 1 protein). Data represent mean + S.E.M for n = 4–5 and statistical analysis consisted of a repeated measured two-way ANOVA followed by Holm–Sidak post hoc test. Significant differences (p b 0.05) are indicated by different letters.

ranges are presented in Supplementary Fig. 1. Maximum and minimum values can be found in Supplementary Table 4. Estimated flux ranges showed no difference between control (Cyp1a1 activity constrained to 0 nmol h−1 mg−1 protein) and βNF exposed samples (Cyp1a1 activity constrained to 1, 10, 100 or 1000 nmol h−1 mg−1 protein) as all estimated fluxes for individual reactions showed overlap in flux ranges (Fig. 4 and Supplementary Fig. 1). This is also true for time-dependent comparisons of individual fluxes. Enzymatic activities Maximum enzyme activity was assessed for AAT, CS, G-3PDH, G6PDH, HOAD, IDH, ME, PEPCK and PK. Among all enzymes examined,

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

91

Table 1 Metabolite flux measurements assessed in the medium of control or 1 μM βNF-exposed hepatocytes. Metabolite (nmol h−1 mg−1 protein)

Exposure Time (h)

Treatment Control

β-naphthoflavone

Ammonia

0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48 0–24 24–48

ND ND −0.22 (0.40) 0.11 (0.20) 0.70 (0.55) −0.73 (0.47) −0.06 (0.17) −0.06 (0.1) 0.28 (0.27) 0.30 (0.18) 37.23 (9.36)a 1.79 (2.78)b 19.91 (8.07) 14.65 (6.32) ND ND ND ND ND ND 23.11 (14.12) −17.50 (16.25) 11.82 (5.50) 7.00 (2.88) −101.91 (41.32) −91.33 (57.80) −1.13 (0.54) −0.86 (0.73) 0.05 (0.30) 0.07 (0.33) 6.93 (4.69) 1.97 (1.54) −27.31 (8.81) −18.5 (12.61) 3.12 (2.64) 4.19 (2.84) 1.53 (1.06) 5.4 (3.44) −3.8 (1.40) −3.09 (1.18) 8.04 (4.52) 15.62 (17.4) −12.34 (4.99) −8.55 (5.47) −0.16 (0.06) −0.15 (0.06) 2.23 (1.88) 3.77 (2.13)

ND ND −0.07 (0.38) 0.02 (0.34) 0.02 (0.31) −0.51 (0.39) 0.08 (0.08) 0.10 (0.07) −0.00 (0.01) −0.04 (0.03) 32.56 (7.40)a −4.71 (3.08)b 18.12 (7.30) 11.23 (4.88) ND ND ND ND ND ND 14.69 (10.45) −17.94 (5.15) 10.51 (4.85) 5.82 (2.25) −104.2 (38.87) −92.46 (56.66) −1.15 (0.46) −1.13 (0.75) 0.04 (0.37) −0.01 (0.27) 6.15 (5.04) 2.07 (1.1) −28.85 (6.33) −23.17 (12.55) 2.82 (3.20) 2.35 (2.32) 1.18 (1.42) 3.79 (2.13) −3.78 (1.15) −3.01 (1.13) 4.89 (3.00) 9.79 (11.36) −11.43 (4.58) −8.23 (5.16) −0.17 (0.05) −0.15 (0.06) 2.31 (2.22) 3.46 (1.89)

Glutamate Glutamine D-3-β-Hydroxybutyrate‡ Acetoacetate‡ Glucose Lactate Glycerol Triglyceride Cholesterol Free fatty acids Alanine Arginine Asparagine Aspartate Cysteine Histidine Isoleucine/Leucine Lysine Phenylalanine Proline Threonine Tyrosine Valine

Values presented are means (±S.E.M) for n = 4–5. Negative values represent metabolite uptake. Significant differences (p b 0.05), determined by a repeated measures two-way ANOVA, are indicated by different letters among exposure times. No significant differences were observed within treatments. ‡ Rates are expressed as pmol h−1 mg−1 protein.

Fig. 4. Metabolic flux ranges estimated by MFA for citrate synthase (CS) (flux 13) and alanine aminotransferase (AAT) (flux 21) flux during the initial 24 h and the second 24 h exposure periods using constraints determined in control hepatocytes and βNF-exposed hepatocytes with Cyp1a1 activity constrained to 1, 10, 100 and 1000 nmol h−1 mg−1 protein. Flux numbers on the x axis represent reactions described in Supplementary Table 1.

to assess maximal energetic costs. This is consistent with previous studies that show maximal EROD activities in rainbow trout hepatocytes exposed to βNF at concentrations of 0.59 and 1.59 μM for 48 h (Navas and Segner, 2000) and EROD activities increased continually over 48 h in the presence of 0.36 μM βNF (Pesonen et al., 1992). Additionally, the estimated EC50 (~46 nM; results not shown) is within the range of previously reported EC50s ranging from 14 to 164 nM for βNF (Navas and Segner, 2000; Scholz and Segner, 1999). Furthermore, the absence of significant changes in cell membrane integrity indicates that the concentrations used are not cytotoxic. Although βNF is not found in the environment, EROD activities ~7 times higher in rainbow trout caged in the brook Vallkärrabäcken (Sweden) near an old landfill site containing 2–3 times more PAH-metabolites than a reference site (Hanson and Larsson, 2011) suggests that the βNF concentration used in this study is relevant as a surrogate for environmental AhR ligand exposure. Energetic costs were assessed by examining the AEC following a 48 h exposure to βNF, and assessing protein synthesis and Na+/K+-ATPase activities, two of the most energetically costly processes in fish hepatocytes (Krumschnabel et al., 1994; Pannevis and Houlihan, 1992; Wieser and Krumschnabel, 2001) and rat thymocytes (Buttgereit and Brand, 1995). Although the validity of using the AEC has been debated (Atkinson, 1977; Fromm, 1977), studies using trout (Hildebrand et al.,

Table 2 Hepatocyte enzyme activities. Enzyme (nmol min−1 mg−1 protein)

Exposure time (h)

Treatment

Alanine aminotransferase

24 48 24 48 24 48 24 48 24 48 24 48 24 48 24 48 24 48

48 52.21 165.3 158.78 16.02 15.99 78.36 79.25 10.71 10.68 44.06 45.48 6.79 6.8 13.68 25.19 16.99 17

Citrate synthase

only AAT and CS showed significant activity differences (p b 0.05) (Table 2). AAT activities were lower in βNF exposed hepatocytes following 24 h exposure and returned to control values 48 h post-exposure. Conversely, CS activity was significantly higher in βNF-treated hepatocytes following 48 h exposure compared to control hepatocytes at that same time point.

Glyceraldehyde-3-phosphate dehydrogenase Glucose-6-phosphate dehydrogenase β-Hydroxyacyl CoA dehydrogenase Isocitrate dehydrogenase

Discussion

Malic enzyme

This study examined the energetic costs of AhR activation in rainbow trout hepatocytes and any potential metabolic changes linked to these costs. βNF concentration and time-dependent EROD activation was assessed to identify the maximal tissue response. This analysis showed a maximal EROD response at 48 h to a 1 μM βNF concentration (Fig. 1), conditions chosen for subsequent experiments in order

Phosphoenolpyruvate carboxykinase Pyruvate kinase

β-Naphthoflavone

control (5.53) (5.07) (22.19) (14.83) (2.11) (1.82) (16.12) (17.3) (1.86) (1.84) (3.09) (3.38) (1.59) (1.55) (1.5) (8.43) (2.61) (3.24)

38.78 54.13 183.19 190.87 15.99 16.28 79.57 80.53 10.7 10.52 43.52 45.82 6.64 6.95 17.09 20.16 16.61 17.44

(3.97)⁎ (7.43) (28.34) (20.52)⁎ (2.18) (2.03) (16.77) (17.93) (1.99) (1.99) (3.34) (3.97) (1.57) (1.56) (1.65) (6.85) (3.18) (3.4)

Values presented are means (±S.E.M) for n = 4. Significant differences (p b 0.05), determined by a repeated measures two-way ANOVA, are indicated by an asterisk between treatments. No time-dependent significant difference was observed.

92

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

2009) and mammalian (Matsui et al., 1994) hepatocytes as well as the recent evidence of AMPK as an energy sensing protein (Oakhill et al., 2011), support AEC as a viable indicator of the energetic status of cells. In this study, βNF exposure did not significantly change AEC or individual adenylate levels, although AEC values reported here (0.65, Fig. 2A) were lower than the 0.7–0.9 reported for ‘healthy’ organisms (Cattani et al., 1996) or trout hepatocytes (Hildebrand et al., 2009). While the extended incubation period may account for the low AEC values, Caldwell and Hinshaw (1994) did report whole liver AEC values in the range of 0.4 to 0.75 when rainbow trout were held under hypoxic, normoxic and hyperoxic conditions. AEC values were assessed in whole livers of trout used in this study and values were 0.6 ± 0.04 (data not presented) suggesting that AEC is maintained post-isolation and for the duration of incubation. Similarly, protein synthesis rates showed no treatment or timedependent significant differences. It was predicted that protein synthesis would decrease to conserve energy although activation of the AhR is typically followed by an increase in specific protein levels such as Cyp1a1 of the AhR gene ‘battery’ (Pesonen et al., 1992). The absence of differences may result from compensation of protein synthesis induced following gene induction (Aluru and Vijayan, 2008) and repression for energy conservation. Na +/K +-ATPase activity, on the other hand, showed significant time-dependent differences but no treatment-dependent changes. The decreased Rb + influx, a direct estimate of Na +/K +-ATPase (Krumschnabel et al., 2001) may be a response by the cells to reduce ATP consumption and to maintain a stable AEC for the duration of the incubation period. Such an energy conservation mechanism is observed in trout hepatocytes exposed to hypoxia (Bogdanova et al., 2005; Krumschnabel et al., 1996). Interestingly, Rb + influx at 48 h returned towards control levels implicating a trade-off between maintenance of ion regulation and energy conservation. Indeed, reduction in Na +/K +-ATPase is reported to lead to an ion imbalance in trout hepatocytes (Krumschnabel et al., 1996). Although βNF exposure did not result in changes in AEC or alter energetically costly processes, hepatocyte metabolism may be reorganized to support cellular processes and detoxification mechanisms. To test this possibility, changes in metabolite fluxes were assessed using metabolic flux analysis (MFA) which as far as we are aware is the first time this has been done in a fish hepatocyte model. We followed a similar method described for cultured rat hepatocytes by Iyer et al. (2010). Extracellular metabolite fluxes were used to generate constraints for the mathematical model. Among measured fluxes, only glucose showed significant differences based upon time, not βNF, going from release to nearly no flux. FFA fluxes, although not significantly different, also showed apparent time-dependent changes going from release to uptake. However, the observed large S.E.M values may mask any significant changes and may be due to differences in intracellular lipid stores between fish. Rissanen et al. (2003) reported glycolysis increased with increased energy demands in rainbow trout hepatocytes exposed to DHAA, a wood industry contaminant. However, the unchanged flux of lactate, and the absence of changes in PK and G-3PDH activities do not support an increased glycolytic flux with βNF exposure. Krumschnabel et al. (2001) did report that trout hepatocyte glycolysis accounted for only about 6% of total ATP production and likely glycolysis has little impact on energetics which may explain the absence of observed changes. The initial release of glucose is expected in isolated rainbow trout hepatocytes as they are in negative glycogen balance and continuously breaking down glycogen (Mommsen, 1986). Interestingly, glycogen stores were maintained in the initial 24 h interval suggesting an increased use of glucogenic substrates such as amino acids, glycerol, or lactate (French et al., 1981; Renaud and Moon, 1980; Vijayan and Moon, 1992) despite a net release of the highly glucogenic amino acid alanine and lactate to the extracellular media. Glycerol has been reported to be a preferred glucogenic substrate in American eel (Anguilla rostrata) and a possible source for glycogen sparing (Renaud and Moon, 1980). The

use of glycerol for glucogenesis may also explain undetectable levels in the media. This is supported by the observed release of FFAs to the extracellular media which may occur from the breakdown of intracellular triglyceride stores serving to release glycerol as is believed to occur in starving fish (Suarez and Mommsen, 1987). This would eventually lead to depletion which could explain the apparent switch from release to uptake and variability in flux rates (reactions 38a–b; Supplementary Fig. S1). Furthermore, depletion of triglyceride stores may explain the change in glycogen stores at 48 h post-exposure in control cells, but suggests that βNF prevents glycogen breakdown by an unknown mechanism. No studies have examined intracellular triglyceride stores in rainbow trout hepatocytes as a test of this hypothesis. Consistent with unchanged metabolites fluxes, MFA also did not identify any effect of Cyp1a1 induction. This was also true over time despite time-dependent changes in glucose and FFA fluxes. The lack of changes in intracellular flux may be related to the lack of changes in adenylate content, AEC or AMPK regulating minor fluctuations in catabolic and anabolic metabolism (Long and Zierath, 2006; Oakhill et al., 2011). Additionally, the variation of extracellular fluxes used to constrain the reactions, which are likely a result of variability in energy stores among individual fish, may have resulted in large estimated intracellular flux ranges as the variability will accumulate in reactions producing or using metabolites that are precursors for multiple endpoint metabolites. This is particularly evident within the TCA cycle which receives and/or produces metabolites for most, if not all, measured fluxes (e.g. precursors for glucose, lipid, and amino acid metabolism) and demonstrates the largest flux ranges of all reactions (Supplementary Fig. S1 and Supplementary Table 4). In order to overcome the accumulation of variability within these reactions, future studies could use isotopic tracers to accurately measure flux of central intracellular reactions. Analysis of enzyme activities, however, showed a decrease in AAT from βNF exposure in the first 24 h of exposure and an increase in CS following βNF exposure for 48 h. This is also not confirmed by the MFA as flux through AAT (Flux 21) and CS (Flux 13) did not change under increased Cyp1a1 activity constraints. Estimated enzyme activities represent maximal rates and do not necessarily represent actual fluxes, however, a decrease in AAT could suggest a decrease in the interconversion of pyruvate and alanine and may result in reduced breakdown of glycogen to pyruvate for its subsequent conversion to alanine which could explain glycogen sparing in βNF exposed hepatocytes. Reduction in AAT has also been observed in livers of Arctic char (Salvelinus alpinus) exposed to polychlorinated biphenyls which are also AhR agonists (Vijayan et al., 2006). On the other hand, increased CS would suggest an increased oxidative flux which may explain the apparent increase in uptake of extracellular FFAs, possibly for its incorporation into the TCA cycle and generation of ATP (Gagnon, 2002). PCB exposed Arctic char are reported to increase β-oxidation as measured by HOAD activities (Vijayan et al., 2006) which would support this hypothesis, although HOAD activity did not change in this study. The absence of significant changes in AEC and metabolic flux reported here is contrary to previous reports of increased energetic costs following exposure to toxicants. For example, an altered AEC was observed upon exposure to xenobiotic transporter P-glycoprotein substrate (Hildebrand et al., 2009) and ATP depletion was observed upon DHAA exposure (Rissanen et al., 2003) in rainbow trout hepatocytes. However, these toxicants are very different in their mode of action; P-glycoproteins consume ATP to transport xenobiotics (Hildebrand et al., 2009) while DHAA is believed to result in an ionic imbalance by interacting with the membrane, leading to increased activity of Na+/K+-ATPase (Rissanen et al., 2003), another direct consumer of ATP. AhR activation is not reported to directly influence ATP consumption which may explain the absence if AEC changes and metabolic fluxes in our study. In fact, this study suggests that ATP consuming processes are attenuated and similarly, energetic costs associated may also be attenuated to conserve energy.

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

Conclusions We hypothesized that activation of the AhR by βNF would increase tissue energetic costs leading to the reorganization of metabolism in rainbow trout hepatocytes. Energetic costs were assessed by examining the AEC, the rate of protein synthesis and Na+/K+-ATPase, and metabolic reorganization was assessed using the MFA and enzyme activities. Although no changes were observed in AEC or energetically costly processes with βNF exposure, this study demonstrates that rainbow trout hepatocytes undergo metabolic changes over the duration of the incubation. This is consistent with Segner et al. (1994) who reported a significantly decreased incorporation of [ 14C]-acetate into lipids 48 h after preparing trout hepatocytes. Such significant time-dependent changes may explain the absence of observed changes using MFA. This model is based on the assumption of pseudo steady-state (no accumulation of intracellular metabolites) but the apparent switch in glucose and FFA fluxes suggests that intracellular stores may be more important than originally appreciated. Future studies using MFA in this or similar model should employ labeled metabolites to estimate intracellular fluxes (Matsuoka and Shimizu, 2010). In conclusion, βNF exposure does lead to a change in metabolism in rainbow trout hepatocytes, shifting from glycogen breakdown to the metabolism of fatty acids by β-oxidation. However, the energetic demands required for maintaining a stable AEC for the duration of the exposure exceeded the demands encountered by βNF exposure. Moreover, this study has used for the first time a MFA approach in fish cells and provides valuable insight for future studies looking to use this approach in a novel in vitro model. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.taap.2013.04.017. Conflict of interest statement The authors declare that there are no conflicts of interest. Acknowledgments This study was made possible by two operating grants from the Natural Sciences and Engineering Research Council (NSERC) of Canada to MVB and TWM as well as funding from uOttawa in support of studentship to RN and GGM, and a co-op grant to A-FH. References Aluru, N., Vijayan, M.M., 2008. Brain transcriptomics in response to beta-naphthoflavone treatment in rainbow trout: the role of aryl hydrocarbon receptor signaling. Aquat. Toxicol. 87 (1), 1–12. Atkinson, D.E., 1977. Daniel Atkinson—adenylate energy charge is a key factor. Trends Biochem. Sci. 2, N198–N200. Bains, O.S., Kennedy, C.J., 2004. Energetic costs of pyrene metabolism in isolated hepatocytes of rainbow trout, Oncorhynchus mykiss. Aquat. Toxicol. 67 (3), 217–226. Bains, O.S., Kennedy, C.J., 2005. Alterations in respiration rate of isolated rainbow trout hepatocytes exposed to the P-glycoprotein substrate rhodamine 123. Toxicology 214 (1–2), 87–98. Barouki, R., Coumoul, X., Fernandez-Salgueroc, P.M., 2007. The aryl hydrocarbon receptor, more than a xenobiotic-interacting protein. FEBS Lett. 581 (19), 3608–3615. Bergmeyer, H.U., 1983. Methods of Enzymatic Analysis. Academic Press, New York. Bergmeyer, H.U., 1985. Methods of Enzymatic Analysis. Verlag Chemie, Weinheim. Beyers, D.W., Rice, J.A., Clements, W.H., Henry, C.J., 1999. Estimating physiological cost of chemical exposure: integrating energetics and stress to quantify toxic effects in fish. Can. J. Fish. Aquat. Sci. 56 (5), 814–822. Bogdanova, A., Grenacher, B., Nikinmaa, M., Gassmann, M., 2005. Hypoxic responses of Na+/K+ ATPase in trout hepatocytes. J. Exp. Biol. 208 (10), 1793–1801. Buttgereit, F., Brand, M.D., 1995. A hierarchy of ATP-consuming processes in mammalian-cells. Biochem. J. 312 (1), 163–167. Caldwell, C.A., Hinshaw, J.M., 1994. Nucleotides and the adenylate energy-charge as indicators of stress in rainbow-trout (Oncorhynchus mykiss) subjected to a range of dissolved-oxygen concentrations. Comp. Biochem. Physiol. B 109 (2–3), 313–323. Cattani, O., Serra, R., Isani, G., Raggi, G., Cortesi, P., Carpene, E., 1996. Correlation between metallothionein and energy metabolism in sea bass, Dicentrarchus labrax, exposed to cadmium. Comp. Biochem. Physiol. C 113 (2–3), 193–199.

93

Dugan, S.G., Moon, T.W., 1998. Cortisol does not affect hepatic α- and β-adrenoceptor properties in rainbow trout (Oncorhynchus mykiss). Fish Physiol. Biochem. 18, 343–352. DuRant, S.E., Hopkins, W.A., Talent, L.G., 2007. Energy acquisition and allocation in an ectothermic predator exposed to a common environmental stressor. Comp. Biochem. Physiol. C 145 (3), 442–448. Edwards, J.S., Ibarra, R.U., Palsson, B.O., 2001. In silico predictions of Escherichia coli metabolic capabilities are consistent with experimental data. Nat. Biotechnol. 19 (2), 125–130. French, C.J., Mommsen, T.P., Hochachka, P.W., 1981. Amino acid utilization in isolated hepatocytes from rainbow trout. Eur. J. Biochem. 113, 311–317. Fromm, H.J., 1977. Herbert Fromm—control by energy charge is an untenable theory. Trends Biochem. Sci. 2, N198–N200. Gagnon, M.M., 2002. Metabolic disturbances in fish exposed to sodium pentachlorophenate (NaPCP) and 3,3′,4,4′,5-pentachlorobiphenyl (PCB126), individually or combined. Comp. Biochem. Physiol. C 132 (4), 425–435. Gesto, M., Tintos, A., Rodriguez-Illamola, A., Soengas, J.L., Miguez, J.M., 2009. Effects of naphthalene, beta-naphthoflavone and benzo(a)pyrene on the diurnal and nocturnal indoleamine metabolism and melatonin content in the pineal organ of rainbow trout, Oncorhynchus mykiss. Aquat. Toxicol. 92 (1), 1–8. Goksøyr, A., Förlin, L., 1992. The cytochrome-p-450 system in fish, aquatic toxicology and environmental monitoring. Aquat. Toxicol. 22, 287–311. Hahn, M.E., 2002. Aryl hydrocarbon receptors: diversity and evolution. Chem. Biol. Interact. 141 (1–2), 131–160. Handy, R.D., Sims, D.W., Giles, A., Campbell, H.A., Musonda, M.M., 1999. Metabolic trade-off between locomotion and detoxification for maintenance of blood chemistry and growth parameters by rainbow trout (Oncorhynchus mykiss) during chronic dietary exposure to copper. Aquat. Toxicol. 47 (1), 23–41. Hanson, N., Larsson, A., 2011. Biomarker analyses in caged and wild fish suggest exposure to pollutants in an urban area with a landfill. Environ. Toxicol. 26 (3), 315–324. Hildebrand, J.L., Bains, O.S., Lee, D.S.H., Kennedy, C.J., 2009. Functional and energetic characterization of P-gp-mediated doxorubicin transport in rainbow trout (Oncorhynchus mykiss) hepatocytes. Comp. Biochem. Physiol. C 149 (1), 65–72. Iyer, V.V., Ovacik, M.A., Androulakis, I.P., Roth, C.M., Ierapetritou, M.G., 2010. Transcriptional and metabolic flux profiling of triadimefon effects on cultured hepatocytes. Toxicol. Appl. Pharmacol. 248 (3), 165–177. Kennedy, S.W., Jones, S.P., Bastien, L.J., 1995. Efficient analysis of cytochrome P4501A catalytic activity, porphyrins, and total proteins in chicken embryo hepatocyte cultures with a fluorescence plate reader. Anal. Biochem. 226 (2), 362–370. Keppler, D., Decker, K., 1974. Glycogen determination with amyloglucosidase. In: Bergmeyer, H.U. (Ed.), Methods of Enzymatic Analysis. Academic Press, New York, pp. 1127–1131. Klamt, S., Stelling, J., 2002. Combinatorial complexity of pathway analysis in metabolic networks. Mol. Biol. Rep. 29 (1–2), 233–236. Klamt, S., Saez-Rodriguez, J., Gilles, E., 2007. Structural and functional analysis of cellular networks with Cell NetAnalyzer. BMC Syst. Biol. 1, 2. Krumschnabel, G., Schwarzbaum, P.J., Wieser, W., 1994. Coupling of energy supply and energy demand in isolated goldfish hepatocytes. Physiol. Zool. 67 (1), 438–448. Krumschnabel, G., Biasi, C., Schwarzbaum, P.J., Wieser, W., 1996. Membrane-metabolic coupling and ion homeostasis in anoxia-tolerant and anoxia-intolerant hepatocytes. Am. J. Physiol. Regul. Integr. Comp. Physiol. 270 (3), R614–R620. Krumschnabel, G., Biasi, C., Wieser, W., 2000. Action of adenosine on energetics, protein synthesis and K+ homeostasis in teleost hepatocytes. J. Exp. Biol. 203 (17), 2657–2665. Krumschnabel, G., Manzl, C., Schwarzbaum, P.J., 2001. Importance of glycolysis for the energetics of anoxia-tolerant and anoxia-intolerant teleost hepatocytes. Physiol. Biochem. Zool. 74 (3), 413–419. Kwast, K.E., Hand, S.C., 1993. Regulatory features of protein-synthesis in isolatedmitochondria from Artemia embryos. Am. J. Physiol. 265, R1238–R1246. Lee, K., Berthiaume, F., Stephanopoulos, G.N., Yarmush, M.L., 1999. Metabolic flux analysis: a powerful tool for monitoring tissue function. Tissue Eng. 5 (4), 347–368. Levesque, H.M., Moon, T.W., Campbell, P.G.C., Hontela, A., 2002. Seasonal variation in carbohydrate and lipid metabolism of yellow perch (Perca flavescens) chronically exposed to metals in the field. Aquat. Toxicol. 60 (3–4), 257–267. Long, Y.C., Zierath, J.R., 2006. AMP-activated protein kinase signaling in metabolic regulation. J. Clin. Invest. 116 (7), 1776–1783. Matsui, Y., Kitade, H., Kamiya, T., Kanemaki, T., Hiramatsu, Y., Okumura, T., Kamiyama, Y., 1994. Adenylate energy charge—charge of rat and human cultured-hepatocytes. In Vitro Cell. Dev. Biol. Anim. 30A (9), 609–614. Matsuoka, Y., Shimizu, K., 2010. Current status of 13C-metabolic flux analysis and future perspectives. Process Biochem. 45 (12), 1873–1881. Mavrovouniotis, M.L., Stephanopoulos, G., 1990. Computer-aided synthesis of biochemical pathways. Biotechnol. Bioeng. 36 (11), 1119–1132. Miranda, A.L., Roche, H., Randi, M.A.F., Menezes, M.L., Ribeiro, C.A.O., 2008. Bioaccumulation of chlorinated pesticides and PCBs in the tropical freshwater fish Hoplias malabaricus: histopathological, physiological, and immunological findings. Environ. Int. 34 (7), 939–949. Mommsen, T.P., 1986. Comparative gluconeogenesis in hepatocytes from salmonid fishes. Can. J. Zool. 64 (5), 1110–1115. Moon, T.W., Walsh, P.J., Mommsen, T.P., 1985. Fish hepatocytes: a model metabolic system. Can. J. Fish. Aquat. Sci. 42 (11), 1772–1782. Navas, J.M., Segner, H., 2000. Antiestrogenicity of β-naphthoflavone and PAHs in cultured rainbow trout hepatocytes: evidence for a role of the arylhydrocarbon receptor. Aquat. Toxicol. 51 (1), 79–92. Oakhill, J.S., Steel, R., Chen, Z.P., Scott, J.W., Ling, N., Tam, S., Kemp, B.E., 2011. AMPK is a direct adenylate charge-regulated protein kinase. Science 332, 1433–1435. Pannevis, M.C., Houlihan, D.F., 1992. The energetic cost of protein synthesis in isolated hepatocytes of rainbow trout (Oncorhynchus mykiss). J. Comp. Physiol. B 162 (5), 393–400.

94

R. Nault et al. / Toxicology and Applied Pharmacology 271 (2013) 86–94

Pascussi, J.M., Gerbal-Chaloin, S., Duret, C., Daujat-Chavanieu, M., Vilarem, M.J., Maurel, P., 2008. The tangle of nuclear receptors that controls xenobiotic metabolism and transport: crosstalk and consequences. Annu. Rev. Pharmacol. 48, 1–32. Pesonen, M., Andersson, T., 1991. Characterization and induction of xenobiotic metabolizing enzyme activities in a primary culture of rainbow trout hepatocytes. Xenobiotica 21 (4), 461–471. Pesonen, M., Goksøyr, A., Andersson, T., 1992. Expression of P4501A1 in a primary culture of rainbow trout hepatocytes exposed to β-naphthoflavone or 2,3,7,8tetrachlorodibenzo-p-dioxin. Arch. Biochem. Biophys. 292 (1), 228–233. Renaud, J.M., Moon, T.W., 1980. Characterization of gluconeogenesis in hepatocytes isolated from the American eel Anguilla rostrata LeSueur. J. Comp. Physiol. B 135 (2), 115–125. Reynaud, S., Deschaux, P., 2006. The effects of polycyclic aromatic hydrocarbons on the immune system of fish: a review. Aquat. Toxicol. 77 (2), 229–238. Rissanen, E., Krumschnabel, G., Nikinmaa, M., 2003. Dehydroabietic acid, a major component of wood industry effluents, interferes with cellular energetics in rainbow trout hepatocytes. Aquat. Toxicol. 62 (1), 45–53. Scholz, S., Segner, H., 1999. Induction of Cyp1a in primary culture of rainbow trout (Oncorhynchus mykiss) liver cells: concentration-response relationship of four model substances. Ecotoxicol. Environ. Saf. 43 (3), 252–260. Schwarzenbach, R.P., Egli, T., Hofstetter, T.B., von Gunten, U., Wehrli, B., 2010. Global water pollution and human health. Annu. Rev. Environ. Resour. 35, 109–136. Segner, H., Blair, J.B., Wirtz, G., Miller, M.R., 1994. Cultured trout liver-cells—utilization of substrates and response to hormones. In Vitro Cell. Dev. Biol. Anim. 30A (5), 306–311. Sherwood, G.D., Rasmussen, J.B., Rowan, D.J., Brodeur, J., Hontela, A., 2000. Bioenergetic costs of heavy metal exposure in yellow perch (Perca flavescens): in situ estimates with a radiotracer (137Cs) technique. Can. J. Fish. Aquat. Sci. 57 (2), 441–450.

Smolders, R., De Boeck, G., Blust, R., 2003. Changes in cellular energy budget as a measure of whole effluent toxicity in zebrafish (Danio rerio). Environ. Toxicol. Chem. 22 (4), 890–899. Suarez, R.K., Mommsen, T.P., 1987. Gluconeogenesis in teleost fishes. Can. J. Zool. 65 (8), 1869–1882. Uygun, K., Matthew, H.W.T., Huang, Y., 2007. Investigation of metabolic objectives in cultured hepatocytes. Biotechnol. Bioeng. 97 (3), 622–637. Vaillancourt, E., Haman, F., Weber, J.M., 2009. Fuel selection in Wistar rats exposed to cold: shivering thermogenesis diverts fatty acids from re-esterification to oxidation. J. Physiol. 587 (17), 4349–4359. Varma, A., Palsson, B.O., 1994. Metabolic flux balancing—basic concepts, scientific and practical use. Biotechnology 12 (10), 994–998. Vijayan, M.M., Moon, T.W., 1992. Acute handling stress alters hepatic glycogenmetabolism in food-deprived rainbow trout (Oncorhynchus mykiss). Can. J. Fish. Aquat. Sci. 49 (11), 2260–2266. Vijayan, M.M., Aluru, N., Maule, A.G., Jorgensen, E.H., 2006. Fasting augments PCB impact on liver metabolism in anadromous arctic char. Toxicol. Sci. 91 (2), 431–439. Wieser, W., Krumschnabel, G., 2001. Hierarchies of ATP-consuming processes: direct compared with indirect measurements, and comparative aspects. Biochem. J. 355 (2), 389–395. Yang, H., Ierapetritou, M.G., Roth, C.M., 2010. Effects of amino acid transport limitations on cultured hepatocytes. Biophys. Chem. 152 (1–3), 89–98. Zhou, H.L., Wu, H.F., Liao, C.Y., Diao, X.P., Zhen, J.P., Chen, L.L., Xue, Q.Z., 2010. Toxicology mechanism of the persistent organic pollutants (POPs) in fish through AhR pathway. Toxicol. Mech. Methods 20 (6), 279–286.