Differential gene expression and biomarkers in zebrafish (Danio rerio) following exposure to produced water components

Aquatic Toxicology 90 (2008) 277–291

Contents lists available at ScienceDirect

Aquatic Toxicology journal homepage: www.elsevier.com/locate/aquatox

Differential gene expression and biomarkers in zebrafish (Danio rerio) following exposure to produced water components T.F. Holth a,b,∗,1 , R. Nourizadeh-Lillabadi c,1 , M. Blaesbjerg d , M. Grung a , H. Holbech e , G.I. Petersen f , P. Aleström c , K. Hylland a,b a

Norwegian Institute for Water Research (NIVA), Gaustadalléen 21, N-0349 Oslo, Norway University of Oslo, Department of Biology, PO Box 1066 Blindern, N-0316 Oslo, Norway Norwegian School of Veterinary Science, Department of Basic Sciences and Aquatic Medicine, PO Box 8146 Dep, N-0033 Oslo, Norway d University of Copenhagen, Institute of Biology, Universitetsparken 13, DK-2100 Copenhagen, Denmark e University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark f DHI Water & Environment, Agern Allé 5, DK-2970 Hørsholm, Denmark b c

a r t i c l e

i n f o

Article history: Received 23 June 2008 Received in revised form 26 August 2008 Accepted 27 August 2008 Keywords: Produced water PAH Alkylphenol Biomarker Microarray Danio rerio

a b s t r a c t The main effluent from oil and gas production is produced water (PW), a waste that contains low to moderate concentrations of oil-derived substances such as polycyclic aromatic hydrocarbons (PAHs) and alkylphenols (APs). PW components may be present in seawater at low concentrations over large areas in the vicinity of oil and gas production facilities. In this study, zebrafish (Danio rerio) were exposed to control and three treatments (high-, pulsed-, low-dose) of a synthetic PW mixture for 1, 7 and 13 weeks. The aim was to investigate the development of transcriptome and biomarker responses as well as relationships between early responses and population-relevant effects. The synthetic PW contained a mixture of low-molecular-weight PAHs (<5 ring) and short-chain APs (C1–C4). The water-borne exposure levels (sum PAH) ranged from 0.54 ppb (low dose) to 5.4 ppb (high dose). Bile pyrene metabolites ranged from 17–133 ng g−1 bile in the control group to 23–1081 ng g−1 bile in the high exposure group. Similar levels have been observed in wild fish, confirming an environmentally relevant exposure. The expression of mRNAs of hepatic genes was investigated in the high exposure group using the Zebrafish OligoLibraryTM from Compugen. Functional clustering analysis revealed effects in the reproductive system, the nervous system, the respiratory system, the immune system, lipid metabolism, connective tissue and in a range of functional categories related to cell cycle and cancer. The majority of differentially expressed mRNAs of genes were down-regulated, suggesting reduction in gene transcription to be as relevant as up-regulation or induction when assessing biological responses to PW exposure. Biomarkers for effects of PAHs (cytochrome P450 1A) and environmental estrogens (vitellogenin) did not appear to be affected by the chronic exposure to low concentration of PW components. Effects at the population level included a reduction in condition factor in male fish from all exposed groups and spinal column deformations in the F1 generation of exposed groups. The different exposure regimes did not produce any significant differences in reproduction or recruitment. The results from this study demonstrate that environmentally relevant concentrations of PW affect gene expression and population-relevant endpoints in zebrafish, although links between the two were not obvious. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Discharges of produced water (PW) from offshore oil and gas production have been a growing concern for the last decades due to the potential long-term effects on biota. The composition and vol-

∗ Corresponding author. Tel.: +47 98227749; fax: +47 22185200. E-mail address: [email protected] (T.F. Holth). 1 These contributed equally to this study. 0166-445X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aquatox.2008.08.020

ume of discharged PW is highly variable depending on geological characteristics and the age of a field. In recent years, as the oil fields in the North Sea age, the volume discharge of PW has increased. The volume of PW discharged in the Norwegian sector of the North Sea was 162 million m3 in 2007 (OLF, 2008). Organic components in North Sea PW in 2006 were organic acids (94%), BTEX (benzene, toluene, ethylbenzene, xylene; 4.4%), alkylphenols (APs; 0.9%), phenols (0.5%) and polycyclic aromatic hydrocarbons (PAHs; 0.2%) (OLF, 2007), of which the environmental effects caused by PAHs and APs have been given the most attention.

278

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

PAHs and APs are two diverse groups of chemicals of natural origin. PAHs are composed of two or more fused benzene rings, both homo- and heterocyclic, which may be alkylated. Depending on their size, structure and solubility, PAHs can be divided into carcinogenic and non-carcinogenic compounds. The most studied are the carcinogenic PAHs, e.g. benzo[a]pyrene which consists of five fused benzene rings. Studies on effects of benzo[a]pyrene are numerous and it is often used as a model compound for effects of PAH exposure on fish (e.g. Beyer et al., 1997). Alkylphenols are phenols with an attached carbon-chain of varying length and structure. The major concern with APs has been their estrogenic properties, as demonstrated by the potent endocrine disruptors octyl- and nonylphenol and their ethoxylates (Nimrod and Benson, 1996). The concentration of APs in PW generally increases with decreasing length of their alkyl-chain (Boitsov et al., 2004). In recent studies, more effort has been put into investigating effects of shorter chained APs and their estrogenic properties (Gimeno et al., 1998a; Seki et al., 2003; Barse et al., 2006; Tollefsen et al., 2008). The use of early warning signals is based on a theoretical time-dependent relationship of responses from low to high levels of biological organization. Changes in gene expression will lead to changes in cellular processes, which may ultimately affect the organism and give rise to population level effects. Causal links between such early warning signals to chemical exposure and adverse organism or population level effects have not been widely documented, although a relationship in fish between PAH exposure and liver lesions have been indicated (Myers et al., 2003). The genomic tools available for North Sea fish species are not yet well developed (Douglas, 2006). There exists an expression array for flounder (Williams et al., 2003) as well as small cDNA libraries for cod (International Cod Genomic Consortium www.codgen.olsvik.info, Olsvik, P.A.; Atlantic Cod Genomics and Broodstock Development http://codgene.ca, Bowman, S. and Trippel, E.). In this study, in order to allow detection of genome wide responses, zebrafish (Danio rerio) was chosen as model species (Hill et al., 2005). The zebrafish genome has been widely studied (Teh et al., 2005) and more recently for changes in expression related to environmental stressors (Ton et al., 2003; van der Ven et al., 2005; Alestrom et al., 2006; Olsvik et al., 2007; Voelker et al., 2007). The Zebrafish OligoLibraryTM from Compugen (Alestrom zebrafish lab, http://microarray.no) was used to study differential mRNA expression of zebrafish genes. The metabolism of xenobiotics in animals is conducted primarily through the phase I–III metabolising systems. The purpose of these systems is to increase the hydrophilicity of endogenous or exogenous substances through hydrolysis, reduction, oxidation (phase I) and/or conjugation (phase II), and finally the efficiency of transport and excretion (phase III). The regulation and activity of specific phase I or II metabolising system components have commonly been used to quantify effects of xenobiotics in aquatic species. The effects of PAHs in aquatic organisms have been assessed using phase I enzyme activity (cytochrome P450 1A; CYP1A) and subsequent adverse effects such as DNA adduct formation and histopathological lesions (Myers et al., 2003). In addition, the exposure load of PAHs in fish may be quantified by measuring PAH metabolites in bile (Aas et al., 2000). Adverse effects of PAHs may be mediated through the aryl hydrocarbon receptor (AhR) pathway, activated by dioxin and dioxin-like compounds such as the carcinogenic PAHs. The nature of interactions of 2- and 3-ring PAHs with the AhR is, however, less clear, with some evidence indicating effects mediated by an AhR independent mode of action (Incardona et al., 2005). Also, fluoranthene (a 4-ring PAH) has been shown to inhibit CYP1A activity in fish (Willett et al., 2001). Clearly, there is a need for more information about the effects of low-molecular-weight PAHs on fish as they are the most abundant and bioavailable PAHs

in PW (Neff, 2002). The phase II metabolising system consists of conjugating enzymes such as sulfotransferases (SULT), glutathione S-transferases (GST) and UDP-glucuronosyltransferases (UDP-GT), increasing the excretion and transcellular transport of xenobiotics (Xu et al., 2005). However, phase II enzymes are generally less responsive than phase I systems. The combination of ocean currents, fish movement and fluctuating discharge volumes causes any PW exposure to be low and variable. The biological effects of PW discharges from offshore activities have been monitored the last decade, and sub-lethal effects have been observed in cod (Gadus morhua L.), haddock (Melanogrammus aeglefinus L.) and saithe (Pollachius virens L.) (Hylland et al., 2006a). Furthermore, adaptation to low level contaminants has been observed in flounder (Platichtys flesus L.) caught in contaminated areas (Eggens et al., 1996). Exposure studies using PW or PW components includes observations on gene regulation (Olsvik et al., 2007), DNA adduct formation (Beyer et al., 2001; Skadsheim, 2004), effects on membrane lipid composition (Meier et al., 2007a), physiological stress (Stephens et al., 2000) and reproductive effects (Meier et al., 2007b). PW has also been shown to contain both estrogenic and anti-androgenic components in in vitro bioassays (Tollefsen et al., 2007). In the present study, hepatic gene expression, biomarker responses and population-relevant endpoints were analysed in zebrafish exposed to a mixture of PAHs and APs. The study aimed to investigate the development of biological responses during long-term PW exposure and to screen for early warning signals to any observed population-relevant effects. 2. Materials and methods 2.1. Experimental setup Adult wild type zebrafish were purchased from a local fish dealer. The fish were sorted by sex then randomly distributed into 4 experimental groups in a total of 13 aquaria (n = 1690). The control group contained four replicate aquaria, while the exposure groups each contained three replicate aquaria. A total of 130 fish (65 males and 65 females) were held in each aquarium (25 l). The fish were acclimated to the exposure system for 1 week before start of the experiment. The flow of water was 75 l day−1 aquarium−1 and the aquaria were oxygenated by aeration. Water temperature was maintained at 26 ± 2 ◦ C and pH at 7.1 ± 0.3 (average ± range). Photoperiod was kept at a light/dark cycle of 14/10 h. Fish were fed three times a day with live brine shrimps (once a day during weekends). Brine shrimp cultures were maintained in a separate laboratory to avoid chemical contamination of the shrimp cultures. The fish were exposed for a total of 13 weeks to three different treatments of PAHs and APs combined in a cocktail, prepared according to median concentrations of the major classes of PAH and AP components in PW and estimated dilution factors (Table 1). The high dose had a nominal dilution factor 200 from PW, the low dose a nominal dilution factor 2000. Total nominal PAH and AP concentrations corresponded to 5.4 ppb PAH and 11.4 ppb AP in the high dose treatment and 0.54 ppb PAH and 1.14 ppb AP in the low dose treatment. To match the phenol content of PW, the treatments were also added 7 and 0.7 ppb phenol, respectively. A pulsed treatment was used to simulate the variation in a PW discharge. In the pulsed group, exposure alternated between high and control treatment with an interval of 1 week. The water was added salts according to ISO standard 7346-1 (1996), giving a concentration of 31.5 mg l−1 NaHCO3 (Merck KGaA, Damstad, Germany), 147 mg l−1 CaCl (Mallinckrodt Baker, Phillipsburg, NJ, USA), 61.7 mg l−1 MgSO4 (Mallinckrodt Baker) and 2.75 mg l−1 KCl (Merck KGaA). The solvent control group received acetone (5 ␮l l−1 ; Merck KGaA). Chemi-

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

279

Table 1 Concentrations of polycyclic aromatic hydrocarbons and alkylphenols present in the synthetic mixtures used in the experiment. Median values were based on discharge data from 16 Norwegian oil installationsa in the North Sea. Group

Component

Median PW concentration (␮g/l)

Nominal high concentration (ng/l)

Nominal low concentration (ng/l)

PAHs Naphthalene Acenaphtene Fluorene Anthracene Phenanthrene Dibenzothiophene Pyrene

Naphthalene Acenaphtene Fluorene Anthracene Phenanthrene Dibenzothiophene Pyrene

310.0 1.7 12.4 0.7 16.0 3.8 0.7

1550 8.5 62 3.5 80 19 3.5

155 0.85 6.2 0.35 8 1.9 0.35

Alkylated PAHs C1-naphthalenes C2-naphthalenes C3-naphthalenes C1-dibenzothiophenes C2-dibenzothiophenes C1-phenanthrenes/anthracenes C2-phenanthrenes/anthracenes C3-phenanthrenes

2-Methylnaphthalene 2,6-Dimethylnaphthalene 2,3,6-Trimethylnaphthalene 4-Methyldibenzothiophene 4-Ethyldibenzothiophene 1-Methylphenanthrene/2-methylanthracene 1,5/1,7-Dimethylphenanthrene/9,10-dimethylanthracene 1,2,6-Trimethylphenanthrene

340.0 188.7 112.6 4.6 6.6 21.6/2.0 23.7/2.0 3.5

1700 943.5 563 23 33 108/10 118.5/10 17.5

170 94.4 56.3 2.3 3.3 10.8/1 11.9/1 1.8

Alkylphenols Phenol C1-alkylphenols C2-alkylphenols C3-alkylphenols C4-alkylphenols

Phenol p-Cresol 4-Ethylphenol 4-n-Propylphenol 4-tert-Butylphenol

1400 1500 600 100 40

7000 7500 3000 500 200

700 750 300 50 20

a

Oil fields: Ekofisk, Gullfaks, Oseberg, Snorre, Statfjord, Troll, Åsgard.

cal analysis of PAHs in the water was performed once a week as described below. 2.2. Chemicals and equipment PAHs and APs were purchased from Chiron AS (Trondheim, Norway) and Sigma–Aldrich (St. Louis, MO, USA). The purity of all compounds was at least 98% analytical grade. All stock solutions of selected PAHs and APs (Table 1) were kept at −20 ◦ C prior to use. Water quality parameters (salinity, hardness, alkalinity) were performed on a Dr. Lange ISIS 9000 photometer (Hach Lange GmbH, Düsseldorf, Germany). All tubings used were PTFE. Mixers were purchased from Heidolph Instruments Gmbh & Co. KG (Schwabach, Germany) and metering pumps from Iwaki Norge AS (Kolbotn, Norway). Flow pumps type 104 were purchased from Ole Dich Instrument makers (Hvidovre, Denmark). 2.3. Sampling design A random selection of eight male and eight female fish per aquarium was terminally sampled at four different time-points, at 1, 7 and 13 weeks of exposure in addition to a pre-start sampling. The fish were anaesthetised by a brief immersion in ice water and immediately euthanised by decapitation. The gall bladder was excised and frozen in liquid nitrogen for PAH-metabolite measurements. The liver was excised and frozen in liquid nitrogen for gene expression analysis (gene expression is defined as mRNA expression of genes) and measurements of CYP1A concentration. Due to small sample sizes it was not possible to analyse both parameters in the same fish. The head was frozen in liquid nitrogen for measurements of vitellogenin (VTG). Population parameters recorded during the exposure period were fecundity, calculated as the number of eggs per female per week, and egg sizes. Fertilisation success, hatching time and embryo survival were assayed during the last 3 weeks of the exposure. For assessment of fertilisation success, 100 eggs per aquarium were incubated over night and the number of fertilised eggs after 0 and 24 h were recorded. In total, 65 separate

experiments on fertilisation success were run. After the initial 24 h incubation, 20 eggs were randomly chosen for further incubation until hatching. After hatching, the hatching time and the number of dead or malformed larvae were recorded. The mortality during the entire experiment period was 8.5, 11.5, 14.6 and 16.2% in the control replicate aquaria, 0, 3.1 and 3.1% in the low dose replicates, 3.1, 3.8 and 16.2% in the pulsed dose replicates and 6.2, 10.0 and 17.7% in the high dose replicates. 2.4. Chemical analyses Bile was analysed for pyrene metabolites after 1 and 7 weeks of exposure. Preparation of hydrolysed bile samples was performed mainly as described by Krahn et al. (1992). Briefly, bile (0.2–1 ␮l) was added internal standard (triphenylamine; 0.4 g l−1 in methanol), diluted with demineralised water (10 ␮l) and hydrolysed with ␤-glucuronidase/aryl sulfatase (4 ␮l, from Helix pomatia, Sigma–Aldrich, USA) for 60 min at 37 ◦ C. Methanol (40 ␮l) was added to stop the reaction, the samples were centrifuged (4000 × g; 10 min) and the supernatant analysed by HPLC. The HPLC used was a Waters 2695 Separations Module with a 2475 fluorescence detector attached. The column was a Waters PAH C18 column with a Vydac 201TP5415 pre-column (5 ␮m; 4.6 mm × 250 mm). The mobile phase consisted of a gradient from 40%:60% (v/v) acetonitrile:water to 100% acetonitrile at a flow of 1 mL/min. The injection volume was 25 ␮l and column temperature 30 ◦ C. Water samples (1 l) were added internal standards, extracted with dichloromethane (150 mL), concentrated and analysed for selected PAHs using a gas chromatography–mass spectrometry (GC–MS) technique. The equipment consisted of an Agilent Technologies (Santa Clara, USA) 6890 GC linked to a 5973 mass selective detector which was used in selected ion monitoring (SIM) mode. The GC was equipped with a 30-m column with a stationary phase of 5% phenyl polysiloxane (0.25 mm internal diameter and 0.25 ␮m film thickness). The initial oven temperature was 50 ◦ C, which was raised stepwise to 290 ◦ C during a 20-min period. The injection was pulsed splitless. The actual concentrations of PAHs in water

280

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

Fig. 1. Sum PAH (acenaphthene, anthracene, phenanthrene, fluorene, naphthalene, pyrene) in water as measured by GC–MS. Nominal and 50% level concentrations of high and low dose treatments are indicated with dotted lines. Data is integrated over the duration of the experiment and presented as median values with quartiles and 10–90 percentiles. The scale is logarithmic. Different letters indicate significant differences at p ≤ 0.05.

approached 50% of the nominal concentrations, with median values about 10% of nominal concentrations (Fig. 1). The pulsed on exposure confirmed equal to high exposure situation, and the pulse off periods equal to control exposure situation. 2.5. Microarray strategy and RNA isolation Zebrafish livers from four experimental groups were used for the microarray experiments. The experimental groups were males after 1 week of exposure, females after 1 week of exposure, males after 7 weeks of exposure and females after 7 weeks of exposure, all from high dose exposures and their respective control groups. To determine differently expressed genes in exposed vs. control fish, triplicate liver samples were compared to pooled RNA samples from three (male) and five (female) control fish. Comparison was done by three dye swap experiments using a total of six slides for each experimental group. For RNA extraction the liver tissue was thawed on ice and resuspended in 1 mL of Trizol reagent (Invitrogen, Carlsbad, CA, USA). Homogenisation was carried out using Magnalyser Beads (Roche Diagnostics, Basel, Switzerland) and a Retsch MM301 homogenizer (Retsch GmbH, Haan, Germany). Total RNA isolation was performed according to the Trizol manufacturer’s protocol (Invitrogen), followed by a 15-min DNaseI treatment (Qiagen # 79254, Hilden, Germany) at 25 ◦ C. Further RNA purification was conducted by a RNAeasy mini kit (Qiagen # 74106). After purification, the samples were eluted in 40 ␮l of RNase-free water and aliquoted in duplicates for microarray and RT-PCR analyses. RNA yield and integrity was determined using NanoDrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, USA) and a 2100 Bioanalyzer instrument (Agilent Technologies, Santa Clara, CA, USA), respectively. None of the samples showed signs of degradation or impurities (260/280 and 260/230 > 1.8, RIN > 8.0). 2.6. Linear RNA amplification, target labelling and hybridization One microgram of total RNA was linearly amplified one round and labelled, using Amino Allyl MessageAmpTM II aRNA amplification kit (Ambion # 1753; Foster City, USA). Five micrograms of the resulting aRNA from the exposed and control groups were labelled either with Cy3-dUTP (Amersham # PA53022; Buckinghamshire, UK) or Cy5-dUTP (Amersham # PA55022). The labelled targets

were examined for amplification yield and incorporation efficiency by measuring the aRNA concentration at 260 nm, Cy3 incorporation at 550 nm and Cy5 at 650 nm using Nanodrop. A good aRNA probe had a labelling efficiency of 30–50 fluorochromes every 1000 bases. Two micrograms of each labelled aRNA target were mixed, 9 ␮l 25× fragmentation buffer (Agilent Technologies # 5183-5974) added, the final volume adjusted to 225 ␮l with RNAse-free H2 O followed by incubation for 30 min at 60 ◦ C. The hybridization solution was prepared by adding 220.5 ␮l of 2× hybridization buffer (Agilent Technologies # 5185-5973) and 4.5 ␮l sonicated herring sperm DNA (10 ␮g ␮l−1 ; Promega #D181A; Madison, WI, USA) to the labelled target aRNA. Microarray slides were pre-hybridized at 42 ◦ C for 60 min using 0.1% bovine serum albumin (BSA) Fraction V, 5× SSC, 0.1% SDS. Hybridization was performed at 60 ◦ C for 16 h using Agilent gasket slides G2534-60003, hybridization chamber (Agilent Technologies # G2534) and oven (Agilent Technologies # G2545A), according to Agilent 60-mer oligo microarray processing protocol. Microarray slides were then washed 3× 5 min in 0.5× SSC, 0.01% SDS (first wash at 42 ◦ C and next two at room temperature). Finally, slides were washed 3× in room temperature with 0.06× SSC and dried immediately with centrifugation at 1000 rpm for 1 min. 2.7. Data processing, statistical analyses and functional profiling Microarray slides were scanned using GenePix 4000B (Molecular Devices, Sunnyvale, CA, USA). Scanning was performed at a level just before saturation of most spots. Raw data generated from Genepix were imported into the Bioconductor package LIMMA and corrected for background (Smyth, 2004). For within and between array normalization print-tip Loess and scale were used, respectively (Smyth and Speed, 2003). An empirical Bayes moderated t-test (Smyth, 2004; Smyth et al., 2005) was applied to detect differentially expressed genes across treated and control samples. p-Values were corrected for multiple testing using Benjamini–Hochberg (BH) method (Benjamini and Hochberg, 1995) and b-values ≥0 were selected as differentially expressed genes. The generated gene list was further filtered for genes with low intensity and with small changes in expression. In the averaged normalized MA-Plot, the majority of genes were clustered in between M-values of ±0.5 (fold change ±1.4) selected to be threshold criteria for the differently expressed gene list. Empty or weak intensity spots, represented by A-values <6, were excluded from the final gene list. 2.8. Zebrafish oligo nucleotide library and microarray construction The arrays were constructed from a 65-mer oligo nucleotide probe set, the Zebrafish OligoLibraryTM (Compugen, Rockville, MD, USA) containing 16,399 oligonucleotide probes, originally representing 16,228 clusters (computed to represent unique genes), plus 171 positive control oligo nucleotide probes. Further 151 probes representing genes involved in toxicogenomics, bone metabolism and prion disease were added to the oligo arrays. The probes were selected from the 3 region of the gene sequences in order to cover a maximum number of splice variants. Oligo nucleotides were printed to CMT UltraGAPS slides at the Norwegian Microarray Consortium (NMC; www.microarray.no) with a BioRobotics Microgrid II (Genomic Solutions, Ann Arbor, MI, USA) and made accessible to the zebrafish community (http://ZFIN.org). Oligo arrays were frequently annotated with the Unigene build release, using Unigene & Gene Ontology Annotation Tool from Genome institute of Singapore (http://giscompute.gis.astar.edu.sg/∼govind/unigene db/). This tool reveals information

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

about Unigene ID, Unigene description, Gene symbol, GO ontology, human and mouse homologue ID and descriptions. To obtain the annotation/putative gene descriptions, GenBank accession numbers were used as query. In the present study annotations were made using the latest gene build, Dr build #107-58. 2.9. Pathway analysis and functional profiling For predicted pathway and biological function analysis of significantly differentially transcribed genes, Ingenuity Pathways Analysis (IPA) (Ingenuity® Systems; www.ingenuity.com) was used. The lists with differentially expressed genes generated by the microarray analysis were translated into mammalian (human, mouse and rat) orthologs using the Unigene & Gene Ontology Annotation Tool and uploaded to IPA. The IPA software is an online exploratory tool with a curated database for over 20,000 (currently) mammalian genes and 1.9 million published literature references. IPAs database together with EntrezGene, Gene Ontology, etc., integrates transcriptomics data with mining techniques to predict and build gene networks, pathways and biological function clusters. The output results were p-values and given scores (−log p-value) which were computed based on the numbers of uploaded genes in the cluster or network and the size of cluster or network in the Ingenuity knowledge data base. Fisher’s exact test was used to determine the probability that a change of each predicted biological function was due to chance alone. They indicated likelihood of the genes in focus (genes uploaded to IPA) in a network being found together due to random chance. The scores of 2 or higher had at least a 99% likelihood of not being generated by chance alone. To support and evaluate hypothesis, the Comparative Toxicogenomics Database (CTD; http://ctd.mdibl.org) was used to search for known chemical-gene-disease interactions (Mattingly et al., 2006). 2.10. Quantitative real-time PCR (qRT PCR) In order to verify the microarray results, 15 differentially expressed genes were randomly selected for qRT PCR analyses. For this analysis, the same total RNA samples as used in the corresponding microarray experiment were used. PCR primers were designed with amplicon size <130 bp close to 3 end and spanning one intron using “Universal ProbeLibrary Assay Design Center/ProbeFinder version 2.40” (Mouritzen et al., 2005; http://www.roche-applied science.com/sis/rtpcr/upl/adc.jsp) and purchased from MWG-Biotech AG (Ebersberg, Germany). Primer sequences are summarized in Table 2. cDNA was produced from 1 ␮g total RNA using Superscript III Reverse Transcriptase (Invit-

281

rogen # 18080-044) according to protocol for oligo (dT)20 primed cDNA-synthesis. The RT-step was performed at 50 ◦ C for 1 h and the reaction was terminated by heating to 70 ◦ C for 15 min. The cDNA product was diluted 4×, aliquoted and stored at −80 ◦ C. QRT PCR was run on the LightCycler 2.0 instrument using LightCycler® FastStart DNA MasterPLUS SYBR Green I (Roche Diagnostics). All reactions were performed in duplicate with a master mix containing 0.5 ␮l of each primer (10 pmol) and 2.0 ␮l enzyme mix in a total reaction volume of 10 ␮l. The PCR conditions were 95 ◦ C for 10 min followed by 40 cycles of 10 s at 95 ◦ C, 10 s at 60 ◦ C and 6 s at 72 ◦ C. All control and target sample PCRs were performed with equal amount of cDNA, same master mix and in same LightCycler run for each gene. Melting curve analysis was performed to confirm formation of expected PCR products. Data analysis was performed using the LightCycler software version 4.0. The crossing point (Cp) was determined by use of the maximumsecond-derivative function on the LightCycler® Software, and the relative gene expression level obtained using Beta-actin as the internal reference gene. 2.11. Immunoassays 2.11.1. Vitellogenin Fish sampled in week 13 were analysed for vitellogenin concentrations. Heads were homogenised 10 times (w/w) in homogenisation buffer (50 mM Tris HCl pH 7.4; 1% protease inhibitor cocktail (Sigma–Aldrich # P8340)) using a teflon pistil and centrifuged at 4 ◦ C at 50,000 × g for 30 min. The supernatant was stored at −80 ◦ C until analysis for vitellogenin using a noncompetitive sandwich ELISA, as described by Holbech et al. (2001). 2.11.2. Cytochrome P450 1A Livers were homogenized in phosphate buffer (0.1 M K-PO4 pH 8.0; 0.15 M KCl) using a teflon pistil. The homogenized livers were then ultrasonicated for 1 s on low power before centrifugation at 2000 × g for 30 min to discard cell membranes and 50,000 × g for 2 h to separate cytosol and microsomes. Microsomes were resuspended in phosphate buffer (0.1 M K-PO4 ; 0.15 M KCl; 1 mM EDTA; 20% glycerol). Protein content was analysed by the method of Lowry et al. (1951). The CYP1A protein content was analysed by a semi-quantitative enzyme-linked immunosorbent assay (ELISA). In essence, samples were diluted to 10 ␮g protein mL−1 in carbonate–bicarbonate buffer (0.05 M; pH 9.6), applied to microtiter plates in quadruplicate and incubated at 4 ◦ C over night. After incubation, plates were washed with TTBS (20 mM Trisbuffer; 0.5 M NaCl; 0.05% Tween-20; pH 8.5) and incubated for

Table 2 Primers used for qRT PCR. All primers are listed from 5 to 3 . Accession

Gene

Forward

Reverse

AI721419 AF063446 NM 173222 BI890491 BI982951 BI704240 AF273479 AJ278269 BI890158 BM104515 AI545824 BI888920 BM104433 AF095457 BI866963

acyl-coA synthetase aryl hydrocarbon receptor 2 creatine kinase dead box polypeptide 39b dystonin opposite strand transcription unit to stag3 granulin 1 nothepsin nuclear autoantigenic sperm protein s100 calcium binding protein a1 transcribed locus ubiquitin-conjugating enzyme E2v2 hypothetical protein LOC793214 zona pellucida glycoprotein 3 zygote arrest 1

CCTCTGATCTGCTGTGAGGTC GCAGGAAGACCAGACTTCATTATT ACGCGGTATCGAGAGTCTGT TTCACGGTCTGCAGCAATAC CCTCAACGGCTTAGAACTGC GTCACCAATGGTTTCGTCAA GCTGCCATCGTCTTTCCA TCTTTGGCATCCATTACGG TTAATGGCAGCGACGACTC GCTTCAAGGGGAACTCAGTG AACCGATTCGTCACCTTCAG AGATGGGACGGTGAGCTG CCCAACATTGTCCTGATGCT CCACTGTTGCTGCAAGACAT TCATCAGCTGGATTGTCCAC

GATGAGGCTTGTCACGCTTT GACGGAGATGTTCCTCACCT AGGAAGTGGTCATCAATCAGC TTCACGAAGATCACCACCTG GCCTTGAATCACCACAACCT AGGCTGGTAACGCAGAACAT AGTCTCAGCGTGGGTCTGA AACACCTGATTCTGCACACG TCAGAAGATGCCCATCACG CTCCATCCCGGTTCTCATC GAGATTGAGCTTCCATCCGTA GCACTCCACTTTCAGGCTGT TAGCGGTCGTCCTGGTTG AGCAGTTGCAGGCTCTGTG GAGACTGTAGTGGCCGATGG

282

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

3.2. Gene expression

Fig. 2. Pyrene metabolites (1-OH-pyrene) in bile of zebrafish after 7 weeks of exposure. Data is presented as median values with quartiles and 10–90 percentiles. The scale is logarithmic. Different letters indicate significant differences at p ≤ 0.05.

30 min with buffered BSA (TTBS; 1% BSA) before addition of primary antibody (CP226, Biosense Laboratories, Norway; 1:1000 in TTBS containing 0.1% BSA). After incubation over night at 4 ◦ C, plates were washed in TTBS, secondary antibody added (Goat anti-rabbit IgG HRP conjugate) and incubated for 6 h at 4 ◦ C. Finally, plates were washed in TTBS, incubated with TMB Plus (Kem-en Tec Diagnostics, Taastrup, Denmark) and read in a spectrophotometer at 450 nm. 2.12. Statistical analysis of biomarker and population-relevant endpoints Homogeneity of variances and normal distribution of the data was evaluated by Levene’s test (Levene, 1960) before parametric tests were performed. If necessary, the data was log-transformed. Nested one-way analysis of variance (ANOVA) and Dunnett’s or Tukey’s post hoc test was performed to compare two or more group means, respectively (Sokal and Rohlf, 1994; Day and Quinn, 1989). When parametric tests were not applicable, Kruskal–Wallis nonparametric ANOVA was used (Sokal and Rohlf, 1994). The condition factor (CF) was calculated as (BW/length3 ) × 100. 3. Results 3.1. Pyrene metabolite levels in bile A 12-fold increase in bile pyrene metabolites was observed in the pulsed group after 7 weeks of exposure, as measured by HPLC (Dunnett test; d.f. = 39; p < 0.001) (Fig. 2). Also, a twofold increase in bile pyrene metabolites was observed in the high exposure group after 7 weeks of exposure (Dunnett test; d.f. = 39; p < 0.01). What appeared to be increased concentration and group variability of pyrene metabolites in bile was observed in both high- and pulsedexposure groups after 1 week, but levels were not significantly different to controls.

In each subset of genes there were between 250 (females, 1 week) and 400 (males, 7 weeks) genes that were significantly altered in expression (Table 3). A large portion of the genes were unique to each group (50–78%) and no single gene was differentially expressed in all groups. Typically, 30–40% of the genes were annotated. In males after 1 week of continuous high dose exposure, 31% of the genes were up-regulated and 78% were unique to the specific sex and time. After 7 weeks of continuous high dose exposure, 19% of the genes were up-regulated and 64% were unique to the specific sex and time. In females, only a small fraction of the genes was up-regulated, 8% after 1 week and 3% after 7 weeks of continuous high dose exposure. Among the differentially expressed genes in females, 58% were uniquely expressed after 1 week and 50% after 7 weeks. Due to the high number of differentially expressed genes (using the criteria stated in Section 2), lists containing the five genes with the most positive or negative fold-change values in addition to differentially expressed genes relevant for the further discussion in this paper are presented in Table 4 (males) and Table 5 (females). Full lists of differentially expressed genes are provided as supplementary material. Raw signal intensities and the complete list of channel ratios, log-ratios, adjusted p-values, t-statistics and B-statistics of all genes can be retrieved from the European Bioinformatics Institues ArrayExpress (http://www.ebi.ac.uk/), accessible through experiment accession number E-MEXP-1685. A total of 1132 genes were differentially expressed in the experimental groups analysed by microarrays. The human homologue ID’s (applicable for zebrafish) for more than 50% of these genes were obtained from the Genome Institute of Singapore (GIS) and submitted to IPA for functional clustering analysis. The comparative biological function analysis revealed the top six biological functions (p < 0.05) for each of three different categories, namely diseases and disorders, molecular and cellular functions, and physiological system development and function (Fig. 3). Ten of the 18 functional clusters could be related to effects on DNA, cellular or tissue level morphology, including categories such as genetic disorder, cell cycle and cancer. Two of the functional clusters were related to development and function of the nervous system, two were related to the immune system, one to respiration, one to reproductive system disease, one to lipid metabolism and one to function and development of connective tissue. The relative number of differentially expressed genes in the functional clusters was in general evenly balanced between the two sexes, but two clusters, the lipid metabolism and the reproductive system disease were sex-biased and will therefore be discussed in somewhat more detail. Genes involved in lipid metabolism included apolipoprotein B (apob), fatty acid synthase (fasn) and CCAAT/enhancer binding protein (cebpa). Of the 54 genes contained in this cluster, 42 were unique entries of which 86% were male-specific. The functional category of diseases and disorders revealed one female-biased cluster, the reproductive system disease, which included genes such as zona pellucida glycoprotein 3 (zp3) and S100 calcium binding protein A1 (s100a1). Another down-regulated gene closely associated to reproductive system disease, the 17ˇ-hydroxysteroid dehydrogenase (hsd17b12b),

Table 3 Specificity of the gene expression profiles. In each grid, the number of genes with significantly altered expression for the corresponding group is given.

Females 1 week (250 total) Females 7 weeks (400 total) Males 1 week (391 total) Males 7 weeks (359 total)

Females 1 week (250 total)

Females 7 weeks (400 total)

Males 1 week (391 total)

Males 7 weeks (359 total)

145 unique 78 24 10

78 201 unique 38 100

24 38 306 unique 36

10 100 36 229 unique

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

283

Table 4 Differentially expressed genes in males (n = 3) after 1 and 7 weeks of continuous high dose exposure identified by microarray analysis. For each time-point, the five genes with the most positive or negative fold-change values are presented in addition to genes relevant for the further discussion. Na = not available. GenBank accession

Description

Human/mouse homologue

Group

Fold-change

AI943057 BM101644 BI430233 AI558478 AY233269 BC066584 AI477068 U57656 BG883457 NM 194403 BI879986 AW116618 AI884148 AA495390 BI839678 AW076914 AI477589

wu:fb30e06 cdna clone IMAGE:7293246 transcribed locus wu:fb79f11 tumor protein p73 sult1A4 ppar lipoprotein lipase hypothetical LOC559097 mhc1 ufa cyp51 hsp4, like si:dkeyp-80d11.1 wu:fa01b01 Na zgc:158407 mhc1 uea

Apolipoprotein B Na Na Na Tumor Protein P73 SULT1A4 PPAR Lipoprotein lipase CYP7A1 Histocompatibility 2 CYP51A1 Heat shock 70 kDa protein 4 Na Na Na Histocompatibility 2 Histocompatibility 2

1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week

8.8 8.7 8.2 6.6 6.4 1.8 1.5 1.4 −1.6 −3.2 −4.0 −4.1 −15.8 −21.0 −23.0 −33.8 −124.3

AI397353 AI626507 BM026607 BG306356 AI942929 AW019036 BI671182 AI558603 BF938011 BF717548 AI477980 NM 131471 L35587 AF274877 AW279689 NM 194403 AW116064 BI880357 AI721517 BI671888 AW019265

Na Na Na transcribed locus proteasome-associated protein ECM29 zgc:92254 apolipoprotein A-IV 17ˇ-hsd gst, alpha-like apolipoprotein C-II apolipoprotein A-IV mhc1 uba hsp90b1 ahr interacting protein acetyl-coA carboxylase 1 mhc1 ufa zgc:56053 hypothetical LOC559001 wu:fc28f08 Na hypothetical protein LOC793403

Na Na Na Na KIAA0368 Glutathione S-transferase omega 1 Proline-rich nuclear receptor coactivator 2 17ˇ-HSD Transcribed locus Na Apolipoprotein A-IV Histocompatibility 2 HSP90B1 AhR interacting protein Acetyl-CoA carboxylase ˛ MHC1 UFA/Histocompatibility 2 Chitinase 1/Chitinase 3-like 1 Fatty acid synthase Na Na Na

7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks

was included in the category of endocrine system disorder (results not shown). To investigate the male-biased lipid metabolism gene clusters from week 1 and 7, the 43 molecules (40 unique entries) were resubmitted to IPA and reanalysed. The highest scored network from this analysis with a score of 48 contained 23 of the 40 unique entries to IPA (Fig. 4). The genes in this network coded for eight enzymes, four transcription factors, two cytokines/growth factors, three transporters, two kinases and one peptidase. Three of the gene products were either of unknown function (1) or clustered to larger functional groups (2). The network also revealed that 6 gene products were associated with extracellular space, 1 was membrane-associated, 10 were associated with cytosol and 4 were associated with the nucleus. Two gene products were of unknown cellular localisation. In addition to the network and function cluster analysis, a screen for potential biomarker candidates was performed using IPA (Table 6). This analysis was aimed to find genes differentially expressed in more than one experimental group. The male specific candidate molecular markers included fasn and cebpa, both temporally down-regulated and as revealed by the functional clustering, both involved in lipid metabolism. The female-specific candidate molecular markers included zp3 and s100a1, both included in the category of reproductive system disease. After 1 week of continuous high exposure to PW components, two genes were identified as candidate molecular markers, the acyl-CoA synthetase (acsl4) and phosphoinositide-3-kinase (pik3r1). The gene products of these

3.2 2.7 2.5 2.4 2.2 −1.4 −1.7 −1.7 −1.8 −2.1 −2.2 −2.3 −2.5 −2.6 −2.8 −3.4 −9.6 −10.0 −11.5 −15.8 −16.2

genes are involved in fatty-acid metabolism and intra-cellular signalling, respectively. After 7 weeks of continuous high exposure, 12 genes were identified as candidate molecular markers. The functions of these 12 genes related to immune system (c3), protein folding (hsp90ab1, pdia3), amino acid metabolism (abp1), lipid metabolism (fdps), transcriptional regulation (egr1, hipk1, cd2bp2), ribosome transport (npm1), vasoregulation (ace), negative control of cell growth and division (ppp2r5c) and extracellular matrix (vtn). 3.3. QRT PCR A subset of 15 genes with significantly altered microarray expression was chosen for validation by qRT PCR analysis. The data, normalized for beta actin expression, is presented in Table 7. All genes analysed by qRT PCR verified the microarray data with regard to the direction of up- or down-regulation, although the magnitude differed somewhat. 3.4. Population-relevant parameters and immunoassays There was an increase in male and female CF in all treatment groups throughout the experiment. The CF increase in males from exposed groups was 10% lower than in male control fish. After 12 weeks of exposure, a significant reduction in CF in all exposed groups was observed in male fish as compared to control (ANOVA, p = 0.024) (Fig. 5). In female fish, the CF of all exposed groups were

284

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

Table 5 Differentially expressed genes in females (n = 3) after 1 and 7 weeks of continuous high dose exposure identified by microarray analysis. For each time-point, the five genes with the most positive or negative fold-change values are presented in addition to genes relevant for the further discussion. Na = not available. GenBank accession

Description

Human/mouse homologue

Group

Fold-change

AI878254 BF717555 BI896378 AY050502 AY233269 AY024336 U57390 BC052971 BM103127 BM095607 AW134000 BM104433 BM101527

caveolin 2 nuclear protein 1 zgc:66313 guanylate cyclase activator 1B tumor protein p73 hsp70-interacting protein zona pellucida glycoprotein 3a.1 heat shock 70 kDa protein 5 zgc:77106 si:ch73-189n23.1 hypothetical LOC562043 hypothetical protein LOC793214 hypothetical LOC566620

Caveolin 2 Na Amylase Guanylate cyclase activator 1B Tumor protein p73 Hsp70-interacting protein Na Heat shock 70 kDa protein 5 Na Na Diaphanous homolog 3 Na Na

1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week 1 week

4.3 3.1 3.0 2.8 2.5 1.7 −1.9 −25.5 −31.1 −38.1 −58.1 −65.3 −164.3

AI667204 BI881961 BG884391 AY538257 BI982951 BI671182 AW466558 U55863 AI558603 BM184271 BM101590 U57390 AI396666 AI722494 AW019265 AW116064 BI671888

Na Na transcribed locus fibronectin 1b dystonin apolipoprotein A-IV zgc:56126 zona pellucida glycoprotein 3 b 17ˇ-hsd zona pellucida glycoprotein B zona pellucida protein C zona pellucida glycoprotein 3a.1 Na zgc:55941 hypothetical protein LOC793403 zgc:56053 Na

Na Na Na Fibronectin 1 Na Proline-rich nuclear receptor coactivator 2 Epoxide hydrolase 1 Na 17ˇ-HSD Na Zona pellucida protein C Na Na Chitinase 1/Chitinase 3-like 1 Na Chitinase 1/Chitinase 3-like 1 Na

7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks 7 weeks

reduced compared to control, but not sufficiently to cause them to be significantly different. Developmental defects in the F1 generation were observed in the exposed groups only (Fig. 6). The two endpoints observed were dead or deformed larvae. The deformations were apparent as spine malformations making the larvae unable to swim. Also, an increase in cumulative egg production was observed in all the exposed groups as compared to the controls, although the test strength was too small to detect any statistical difference (n = 3). Fertilisation success, egg sizes, hatching time and

2.7 2.4 2.4 1.8 1.7 1.5 1.5 −1.5 −1.5 −1.5 −1.6 −2.3 −5.7 −6.4 −6.7 −8.2 −11.2

embryo survival were assessed, but no differences between groups were observed. The analysis of VTG protein showed that in one replicate tank of the high exposure group, males had elevated mean VTG levels (5882 ng g−1 fish). All other replicates in all treatments were similar to VTG levels in control group males (789 ng g−1 fish). Due to the elevated levels in this one tank, the high exposure group was identified as significantly different from control in a Kruskal–Wallis ANOVA (H(3,N = 96) = 23,7; p < 0.05).

Fig. 3. Comparative functional cluster analysis by Ingenuity Pathways Analysis (Ingenuity® Systems; www.ingenuity.com). Of the total of 1132 genes, 571 were mapped to the corresponding genes in the IPA database. Top six significant functional clusters from three different categories are presented. D: diseases and disorders, M: molecular and cellular functions, and P: physiological system development and function. Numerals indicate the number of genes contained in each group.

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

285

Fig. 4. Lipid metabolism network in males after 1 (M1) and 7 (M7) weeks of exposure analysed by Ingenuity Pathways Analysis (Ingenuity® Systems; www.ingenuity.com). Direction and strength of gene regulation is indicated by colour (red: up-regulation; green: down-regulation) and colour intensity. Direct literature linkages between nodes are indicated by solid arrows. Indirect literature linkages are indicated by dotted arrows. The cellular location of the gene products is indicated by graphics.

Statistical analysis of CYP1A protein in males and females did not reveal any significant differences between treatment groups. 4. Discussion The present study describes responses in exposure and effect markers in zebrafish at several biological levels of organization following exposure to PW components. The water-borne exposure concentrations of PAH were confirmed by water analysis (GC–MS)

and bile pyrene metabolite levels (HPLC). Sum PAH in water was measured to 10% of the nominal concentration, which indicated an effective exposure concentration of 5.4 ppb in the high dose group and 0.54 ppb in the low dose group. The measured water concentration (sum PAH) included all 2-, 3- and 4-ring PAHs of which the 2-ring PAHs made up the bulk. Sum PAH may therefore have underestimated the relative compound contributions, especially the heavy weight compounds. The compounds also have different adsorption, uptake and breakdown kinetics, adding to

286

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

Table 6 Candidate biomarkers generated by IPA. Numbers indicate microarray fold-change values. Gene symbols are presented as mammalian identifiers. Gene symbol

Description

Females

FASN CEBPA BGN OCLN

fatty acid synthase cebp˛ biglycan occludin

ABP1 C3 HSP90AB1 PDIA3 EGR1 NPM1 FDPS HIPK1 CD2BP2 ACE PPP2R5C VTN

amiloride binding protein 1 complement component 3 hsp90˛b1 protein disulfide isomerase family A, member 3 early growth response 1 nucleophosmin farnesyl diphosphate synthase homeodomain interacting protein kinase 1 cd2 binding protein 2 angiotensin I converting enzyme protein phosphatase 2B vitronectin

ZP3 S100A1 NASP TUBB4 TBP AHR DAB1

zona pellucida glycoprotein 3 s100 calcium binding protein a1 nuclear autoantigenic sperm protein tubulin, beta 4 tata box binding protein aryl hydrocarbon receptor 2 disabled homolog 1

ACSL4 PIK3R1

acyl-coA synthetase phosphoinositide-3-kinase

1 week

Males 7 weeks

7 weeks −10.0 −2.1 −1.5 −1.6

−3.6 −3.8 −3.1 −2.7 −2.5 −2.3 −2.3 −2.1 −1.8 −1.8 −1.6 −1.5 −16.1 −7.2 −5.2 −1.9 −1.9 −1.8 −1.4

−3.4 −3.3 −2.5 −2.6 −2.0 −1.7 −1.9 −1.6 −1.5 −2.4 −1.8 −1.5

−3.5 −2.8 −2.3 −1.9 −1.5 −1.8 −1.8

−2.8 −2.8

the variability of the effective concentration estimate. The pyrene metabolites in bile describe a temporally integrated exposure load, and represent a quantification of the flux of pyrene through the fish. This measure may therefore describe the individual exposure situation more precisely. The pyrene metabolite levels in bile confirmed a dose-related exposure and were in the range of ecological significant levels observed in flounder (P. flesus) (Eggens et al., 1996; Kammann, 2007), dab (Limanda limanda) (Kammann, 2007), haddock (M. aeglefinus) and cod (G. morhua) from the North Sea (Hylland et al., 2006b; Aas et al., 2006). The variation of measured pyrene metabolites was larger in the high dose group, which may have been due to individual variation in metabolism as exposure load increased or to the feeding status of the fish (Collier and Varanasi, 1991).

1 week −10.0 −2.0 −1.6 −1.5

−3.2 −2.0

A range of categories related to DNA repair, cell cycle, cellular interaction and proliferation and tissue morphology were identified by the gene functional clustering analysis, indicating that a variety of molecular mechanisms with possible adverse effects such as tissue lesions or cancer were altered. As biomarkers for the potential development of PAH-induced lesions or cancer, the phase I metabolising system has been frequently used (Myers et al., 2003). Of the known phase I metabolising system-related markers, an up-regulation of epoxide hydrolase and a down-regulation of aryl hydrocarbon receptor 2 (ahr2) were observed in females and a down-regulation of ahr-interacting protein was observed in males, but only ahr2 was differentially expressed over time and suggested as a candidate biomarker. Currently, two ahr isoforms have been well characterised in fish (ahr1 and ahr2), and a nega-

Table 7 Genes verified by qRT PCR. Values are presented as median and percentiles. Microarray fold-change values are presented as comparison. Accession

BI888920 BI890491 BI982951 AI545824 BI704240 AF273479 AI545824 AJ278269 BI890491 AI721419 BI866963 NM 173222 BI890158 BM104515 AF095457 BM104433 BI982951 AF063446 AI545824

Annotation

ubiquitin-conjugating enzyme E2v2 dead box polypeptide 39b dystonin transcribed locus opposite strand transcription unit to stag3 granulin 1 transcribed locus nothepsin dead box polypeptide 39b acyl-coA synthetase zygote arrest 1 creatine kinase nuclear autoantigenic sperm protein s100 calcium binding protein zona pellucida glycoprotein 3 hypothetical protein LOC793214 dystonin aryl hydrocarbon receptor 2 transcribed locus

Group

Male 1 week Male 1 week Male 1 week Male 1 week Male 7 weeks Male 7 weeks Male 7 weeks Female 1 week Female 1 week Female 1 week Female 1 week Female 1 week Female 1 week Female 1 week Female 1 week Female 1 week Female 7 weeks Female 7 weeks Female 7 weeks

Microarray

qRT-PCR

Fold-change

Median fold-change

1.5 −1.8 −1.8 −4.6 1.7 −2.1 −6.2 1.8 −2.6 −2.9 −2.9 −4.3 −5.2 −7.2 −16.1 −65.3 1.7 −1.8 −4.1

1.3 −1.5 −2.2 −6.7 1.3 −1.5 −7.6 2.3 −2.0 −1.8 −80 −15 −12 −109 −171 −275 1.7 −3.7 −3.6

25% percentile 1.0 −2.0 −3.8 −78 1.2 −5.2 −44 1.0 −3.7 −3.0 −113 −46 −23 −603 −1301 −2408 1.0 −5.4 −20

75% percentile 1.6 1.1 1.1 −4.1 2.1 2.8 1.4 2.8 −1.4 1.0 −9.3 −2.9 −2.9 −21 −40 −37 3.7 1.2 −1.3

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

Fig. 5. Condition factor of male and female zebrafish after 12 weeks of exposure. Data is presented as median values with quartiles and 10–90 percentiles. Statistical significance (p ≤ 0.05) is marked with *.

tive relationship in gene expression seems to exist in response to xenobiotic exposure (Arukwe et al., 2008). The down-regulation of ahr2 is in accordance with previous observations in PW exposed zebrafish (Arukwe et al., 2008). The up-regulation of epoxide hydrolase may indicate an excess of epoxides in the cells, known to form during cytochrome P450-dependent biotransformation of PAHs (Parkinson, 2001). Cytochrome P450 1A (CYP1A) gene and protein expression has been considered a sensitive biomarker for effects of PAHs (van der Oost et al., 2003), but neither hepatic gene expression nor protein expression of CYP1A seemed to be affected by exposure to PW components. Such a lack of CYP1A protein induction has also been observed in caged fish and natural fish populations near oil platforms (Förlin and Hylland, 2006; McIntosh et al., 2006) and in exposure studies using similar exposure conditions (Beyer et al., 2001). The PW potential of activating the AhR pathway has been shown to be highly variable, mainly due to the large variations in PW composition (Hurst et al., 2005). Low molecular weight PAHs (<5 rings) are not very potent inducers of CYP1A and some PAHs have even been shown to inhibit CYP1A expression (McKee et al., 1983; Willett et al., 2001). Also, potential antagonistic effects of PAHs and APs have been shown, e.g. receptor interactions such as the estrogen receptor (ER)-AhR crosstalk (Safe, 2001). PW has been shown to contain both ER agonists and androgen receptor antagonists (van Lipzig et al., 2005; Tollefsen et al., 2007), and decreased CYP1A expression has been observed in Atlantic cod (G. morhua) exposed to a combination of North Sea oil and APs (Sturve et al., 2006). In addition, other tissues than liver might be more responsive in zebrafish as observed by Olsvik et al. (2007). The results of this study suggests that regulation of the ahr2 gene may be more sensitive to low level contaminant exposure than the other phase I metabolising system components. Alternatively, it may indicate that transcriptional regulation of this

Fig. 6. Larval survival and deformities after 96 h post-fertilisation incubation. Control n = 160, low n = 380, pulse n = 340, high n = 420.

287

isoform is less susceptible for antagonistic effects such as receptor crosstalk. Metabolites generated by phase I metabolism may subsequently be conjugated by phase II enzymes. Gene regulations of some phase II enzymes were observed, such as an up-regulation of sulfotransferase (sult1a4) in males after 1 week and a down-regulation of GST in males after 7 weeks. An up-regulation of sult1a4 may, as expected, indicate increased conjugation of xenobiotics. Studies on GST activity in fish after xenobiotic exposure show conflicting results (van der Oost et al., 2003). The regulation of GST genes in mammals is generally mediated through the Ah-receptor (George, 1994), which render possible, as for CYP1A, interactions such as antagonistic receptor crosstalk which could provide an explanation for the down-regulation of GST genes. Of the 40 unique genes related to lipid metabolism in males, 23 genes were successfully mapped to a common network. This network contained extracellular lipid transport proteins such as apolipoprotein B, cytosolic lipid metabolising enzymes such as fatty acid synthase (fasn), lipoprotein lipase (lpl) and acetyl-coenzyme A carboxylase (acaca), microsomal CYP enzymes involved in cholesterol metabolism (cyp7a1 and cyp51a1) and nuclear transcription factors such as CCAAT/enhancer binding protein (cebpa). Two of these genes, fasn and cebpa, were down-regulated after both 1 and 7 weeks of exposure and suggested as candidate biomarkers. Another candidate biomarker related to lipid metabolism was farnesyl diphosphate synthase (fdps), of which the gene product is an enzyme involved in cholesterol synthesis (Goldstein and Brown, 1990). The gene products of fasn, a cytosolic enzyme (EC 2.3.1.85), and cebpa, a nuclear transcription factor, play key roles in the metabolism of fatty acid and the initiation of gluconeogenesis (Wakil, 1989; Wang et al., 1995). In mammals, the fasn gene is negatively regulated by ethinyl estradiol (Henriquez-Hernandez et al., 2007), a model estrogenic compound, and 2,3,7,8-tetrachlorodibenzo-paradioxin (Lakshman et al., 1989; Boverhof et al., 2006), a model compound for AhR-mediated responses. The down-regulation of fasn may therefore have been a response to both estrogen action mimicking chemicals, such as some APs, and chemicals acting through the AhR pathway, such as some PAHs. The regulation of the cebpa gene and/or protein has been shown to be affected by exposure to compounds such as ethinyl estradiol (Geier et al., 2003), cadmium (Ohba et al., 2007), tretinoin (Hashimoto et al., 2006) and ethanol (Harrison-Findik et al., 2006), whereas ethinyl estradiol and cadmium resulted in decreased Cebpa protein binding activity or cebpa gene expression, respectively, in mammals. The hepatic fatty-acid content is of importance with regard to energy metabolism, lipid-activated signal transduction and regulation of apoptosis (Shinomura et al., 1991; Scaglia and Igal, 2005). In fish, high plasticity in fatty-acid composition of membranes is of importance during cold-acclimation (Hsieh and Kuo, 2005; Hsieh et al., 2007). In addition to compound-specific effects on gene transcription, changes in membrane composition can affect the regulation of lipid metabolism genes (Sampath and Ntambi, 2005). Of the compounds present in the exposure mixture, APs have been shown to have affinity for biological membranes and their affinity increased with increasing length of the alkyl group (Meier et al., 2007a). The membrane fatty-acid profile of cod exposed to APs has been shown to contain more saturated fatty acids and less polyunsaturated fatty acids (PUFAs) (Meier et al., 2007a). Also, similar alterations in the membrane fatty-acid profile have been observed in haddock caught in oil production areas (Hylland et al., 2006a). The transcription of fasn and fdps is known to be suppressed by PUFAs (Le Jossic-Corcos et al., 2005; Teran-Garcia et al., 2007), but recent studies suggest that the fasn gene have several paralogues, of which some are robust to lipid-induced changes (Ducasse-Cabanot et al., 2007; Panserat et al., 2008). The lipid metabolism molecular network indicated

288

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

that processes in several cellular compartments were affected by the exposure, but only fasn and cebpa were maintained differentially expressed over time. As their gene products play key roles in the metabolism of lipids, they appear to be suitable as candidate biomarkers, but their response specificity to PW components and the potential interference of PUFAs needs to be investigated. One of the female-biased functional clusters was the reproductive system disease, containing the candidate biomarkers zona pellucida glycoprotein 3 (zp3) and S100 calcium binding protein A1 (S100a1), which were among the down-regulated genes found in females after both 1 and 7 weeks of exposure. In addition, several other genes belonging to the zp gene family were found differentially expressed in females. The Zp proteins are constituents of the extracellular glycoprotein matrix (zona pellucida) surrounding the oocyte, with Zp3 recognized as the sperm receptor (Hoodbhoy and Dean, 2004). The transcription of zp genes in the majority of teleosts is under estrogenic control (Larsson et al., 1994; Modig et al., 2006), although recent reports suggest that the regulation of zp gene and protein expression also may be influenced by other factors such as cortisol (Berg et al., 2004) or developmental stage (Liu et al., 2006). Due to its sensitivity, differential regulation of zp3 has been proposed as a marker for estrogen-exposure in combination with another molecular marker, a nucleoside diphosphate kinase (nm23) (Gunnarsson et al., 2007). The proposed marker nm23 was not available on the zebrafish array used in this study. The sensitivity of the zp3 gene to xenobiotics is clearly demonstrated in the current study, and the direction of regulation suggests an anti-estrogenic effect. The s100a1 has been linked to endometriosis in humans (Burney et al., 2007), hence clustering to the reproductive system disease, but no linkages to adverse effects have been made in fish. Another differentially expressed gene clustered to the resembling category of endocrine system disorder was the 17ˇ-hydroxysteroid dehydrogenase (17ˇ-hsd), coding for an enzyme (E.C. 1.1.1.62) known to be involved in androgen biosynthesis. A down-regulation of 17ˇhsd was observed in males and females after 7 weeks of exposure. Estrogens are known to inhibit the expression of androgen biosynthesis enzymes in fish (Baron et al., 2005; Filby et al., 2007), and PAHs (phenanthrene) inhibit steroidogenesis in vitro (Monteiro et al., 2000). The short- and medium-chained APs (C4–C7) found in PW are known to exert both estrogenic and anti-estrogenic properties (Gimeno et al., 1998b; Meier et al., 2007b). The down-regulation of 17ˇ-hsd, consistent with an estrogenic response, may have been a response to both short-chained APs and to specific PAHs such as phenanthrene. In addition, other candidate molecular biomarkers known to be regulated by ethinylestradiol were identified. Nasp, a histone transport protein (Alekseev et al., 2005), has been shown to be negatively regulated by ethinylestradiol in rats (Heneweer et al., 2007) and the immune system component C3 have been shown to respond in a tissue-specific manner to ethinylestradiol (Sundstrom et al., 1989). Neither of these genes have any known contaminant specific adverse effects, but this does not exclude their utility as markers of exposure to estrogenic compounds. Like these proposed markers, the commonly used biomarker for estrogenic effects, VTG, has no known adverse effects when induced by environmentally relevant concentrations of ER agonists. In this experiment, a difference in male plasma VTG between replicate tanks in the high exposure group was observed. Due to the extensive water monitoring and analytical procedures used in this study, technical errors seem unlikely. Such differences in VTG response have also been observed in sheepshead minnow (Cyprinodon variegatus) following exposure to a mix of alkylphenols (Beyer et al., 2001). As it is unlikely that all sensitive males were distributed to the same tank by chance, an alternative explanation might be a difference in sex ratio in the tanks at the time of sampling. Spawning females are known to excrete estrogenic compounds (own observations) and

the female:male ratio in the specific replicate tank was 1.35 compared to 1.10 and 1.00 in the other two replicates. It can be deduced from the current study that gene responses to PW exposure include both estrogenic and anti-estrogenic modes of action. Some of the identified functional clusters and candidate biomarkers are novel markers in fish of which the specificity or sensitivity is not known, and possible links to adverse effects must be interpreted with caution. The gene products of the male-specific candidate biomarkers biglycan and occludin have been related to the process of bone mineralisation (Parisuthiman et al., 2005) and epithelial permeability (Chasiotis and Kelly, 2008), respectively, of which both processes have endpoints associated with adverse effects. Skeletal deformities have been observed in fish after contaminant exposure (Bengtsson, 1979; Warner and Jenkins, 2007), and in humans, biglycan have been found up-regulated in patients with scoliosis (Bertram et al., 2006). Spinal column deformations in the F1 generation of exposed groups were observed in the current study. The epithelial permeability of gill cells have been altered after contaminant exposure in several fish species (Stephens et al., 2000; Claireaux et al., 2004) and fusion of gill lamellae likely to impair ionoregulatory processes have been observed in fish after PW exposure (Stephens et al., 2000). Effects on respiration have also been observed in fish exposed to copper (De Boeck et al., 2007). These two genes could prove useful as markers for PWrelated skeletal deformities and effects on respiration, respectively. One gene identified as a short-term (1 week exposure) candidate biomarker, acyl-coa synthetase (acsl4), is involved in fatty-acid metabolism. It has been related to cardiac function in fish during temperature acclimation (Hicks et al., 1996) and it is likely to be positively regulated by peroxisome proliferator-activated receptors (PPARs) (Martin et al., 1997), known to be induced by, e.g. PAHs. In this study however, a negative correlation between PPAR␥ and acsl4 was observed in males after 1 week. Some chaperones were differentially expressed such as heat shock protein 90 (hsp90ab1), heat shock protein 70 (hsp70) variants and protein disulfide isomerase (pdia3). Of these genes, all coding for products involved in protein folding, hsp90ab1 and pdia3 were identified as candidate biomarkers. Although the Hsp90 protein has been used as an indicator for cellular stress, in particular stress induced by heavy metals, it is also responsive to non-chemical, environmental stressors (van der Oost et al., 2003) and the environmental relevance of these chaperones is not known. Several candidate biomarkers and functional gene clusters have either not been explored in fish or could not be directly associated with putative adverse effects. Some of these clusters were related to diverse functions such as development and function of connective tissue and development, function and disease of the immune system and the nervous system. Processes affecting the development and function of connective tissue may indicate the development of symptoms such as cirrhosis or fibrosis, as observed in fish collected in PAH contaminated areas (Henson and Gallagher, 2004). Two clusters indicated an effect in the immune system. Increased prevalence of diseases in fish has been monitored and correlated to contaminants in several areas, and whereas the prevalence of some diseases in North Sea fish seems to be decreasing, the prevalence of others is increasing (Hylland et al., 2006a). The down-regulation of several mhc1 genes in males in the current study may suggest a mechanism for disease susceptibility (Reynaud and Deschaux, 2006). Finally, the female candidate biomarker, dab1, has been shown to be related to development of the nervous system (Howell et al., 1997). Effects on nervous system components in fish have previously been observed, such as the inhibition of acetyl choline esterase (AChE) activity after PW exposure (Casini et al., 2006; own observations) or on nervous steroidogenic systems after exposure to crude oil water accomodated fraction (WAF) (Arukwe et al., 2008). Effects

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

on the nervous system might provide a part of the explanation for some observed behavioural responses in fish to PW exposure (e.g. Stephens et al., 2000). Other candidate marker genes not explored in fish were involved in amino acid metabolism (abp1), extracellular matrix (vtn), intra-cellular signalling (pi3k), microtubuli (tubb4), negative control of cell growth and division (ppp2r5c), ribosome transport (npm1), transcriptional regulation (cd2bp2, egr1, hipk1, tbp) and vasoregulation (ace). Tata box binding protein (tbp) and tubb2, a gene belonging to the ␤-tubulin gene family, have been suggested as RT PCR references in zebrafish (Filby and Tyler, 2007) and halibut (Fernandes et al., 2008), respectively. The results of this study show that these genes or gene families can be affected by contaminant exposure, and that they may be unsuitable as zebrafish reference genes. Interestingly, although the fish were exposed to two different exposure regimes (pulsed or continuous exposure), their population level responses were similar. The condition factor of male fish was equally reduced in all exposed groups relative to the control group. Exposure to dioxin, a well known AhR agonist, has been shown to negatively impact growth in several fish species (Elonen et al., 1998). Also, short-chained APs have been shown to affect growth in Atlantic cod (Meier et al., 2007b). One hypothesis explaining such a reduced growth is an increased activity level, as observed in turbot (Scophthalmus maximus) exposed to PW (Stephens et al., 2000). An increased activity level, leading to higher metabolism and energy expenditure, may also contribute to the regulation of lipid metabolism systems. The reduction in condition factor and regulation of lipid metabolism genes caused by PW components appeared to be positively correlated in the zebrafish males, making this system a potential early warning signal. Deformed or dead larvae were observed in the exposed groups only, at a low but similar rate. The malformations were apparent as spinal column flexures, which have been shown to be distinct from AhR-dependent effects and to be mediated by tricyclic PAHs (fluorene, dibenzothiophene, phenanthrene) (Carls et al., 1999; Incardona et al., 2004, 2005). In general, the down-regulation of all candidate biomarkers in this experiment suggests that inhibition of gene transcription or enzyme activity is as relevant as induction. The effects observed in condition factor and reproduction demonstrates that PW components may affect population-relevant parameters in zebrafish at environmentally relevant concentrations. Furthermore, differences between exposure regimes (pulsed or continuous) might not be discernible at higher effect levels. 5. Conclusions Zebrafish were exposed to a cocktail of light-weight PAHs and short-chained APs with a composition typical for North Sea PW. By measuring pyrene metabolites in bile (HPLC), the exposure was confirmed to be in the range of environmentally relevant levels. Two population level effect parameters were affected: condition factor was reduced in male fish and developmental defects were observed in the F1 generation of exposed groups. Interestingly, no significant differences in these gross indices were observed between pulsed or continuous exposure regimes. Functional clustering analysis of differentially expressed genes in PW exposed fish revealed exposure-related differences in the reproductive system, the nervous system, the respiratory system, the immune system, lipid metabolism, connective tissue and in a range of functional categories related to cell cycle and cancer. Of the phase I metabolising system markers, a down-regulation of ahr2 was observed. No induction was observed in CYP1A quantity, which may be a result of AhR independent effects of low-molecular-weight PAHs or to differences in tissue and/or species sensitivity to CYP1A induction. The most distinct candidate biomarker genes were the fatty acid

289

synthase and the zona pellucida 3, involved in lipid metabolism and reproductive processes, respectively. In general, the downregulation of all candidate biomarkers identified suggested that inhibition of gene transcription or enzyme activity is as relevant as induction when assessing the effects of PW exposure. Overall, this study shows that environmentally relevant concentrations of PW components may lead to changes in gene expression and population-relevant endpoints, although no specific links between the different biological levels of effect were obvious. Acknowledgements We would like to thank Anna Osypchuk, Eirin Sva Stomperudhaugen, Eivind Bøe, Helene Øverås, Ingeborg Rønning, Inger Katharina Gregersen, Kine Martinsen, Marte Rindal Jacobsen, Mette Albrektsen and Silje Røysland for assistance with zebrafish sampling and dissections. We would also like to thank Anja Julie Nilsen and Inger Katharina Gregersen for preparation of bile samples. This study was funded by the Research Council Norway (RCN) “The Oceans and Coastal Areas”-programme (project 164419) and NIVA (internal research projects). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.aquatox.2008.08.020. References Aas, E., Baussant, T., Balk, L., Liewenborg, B., Andersen, O.K., 2000. PAH metabolites in bile, cytochrome P4501A and DNA adducts as environmental risk parameters for chronic oil exposure: a laboratory experiment with Atlantic cod. Aquat. Toxicol. 51 (2), 241–258. Aas, E., Jonsson, G., Sundt, R., Westerlund, S., Sanni, S., 2006. Monitoring of PAH metabolites and metals in bile from caged Atlantic Cod (Gadus morhua) and wild Pelagic Fish along expected contaminant gradients in the North Sea. In: Hylland, K., Lang, T., Vethaak, D. (Eds.), Biological Effects of Contaminants in Marine Pelagic Ecosystems. SETAC Press, Brussels, pp. 263–276. Alekseev, O.M., Widgren, E.E., Richardson, R.T., O’Rand, M.G., 2005. Association of NASP with HSP90 in mouse spermatogenic cells—stimulation of ATPase activity and transport of linker histones into nuclei. J. Biol. Chem. 280 (4), 2904–2911. Alestrom, P., Holter, J.L., Nourizadeh-Lillabadi, R., 2006. Zebrafish in functional genomics and aquatic biomedicine. Trends. Biotechnol. 24 (1), 15–21. Arukwe, A., Nordtug, T., Kortner, T.M., Mortensen, A.S., Brakstad, O.G., 2008. Modulation of steroidogenesis and xenobiotic biotransformation responses in zebrafish (Danio rerio) exposed to water-soluble fraction of crude oil. Environ. Res. 107 (3), 362–370. Baron, D., Fostier, A., Breton, B., Guiguen, Y., 2005. Androgen and estrogen treatments alter steady state messengers RNA (mRNA) levels of testicular steroidogenic enzymes in the rainbow trout, Oncorhynchus mykiss. Mol. Reprod. Dev. 71 (4), 471–479. Barse, A., Chakrabarti, T., Ghosh, T.K., Pal, A.K., Jadhao, S.B., 2006. One-tenth dose of LC50 of 4-tert-butylphenol causes endocrine disruption and metabolic changes in Cyprinus carpio. Pestic. Biochem. Physiol. 86 (3), 172–179. Bengtsson, B.E., 1979. Biological variables, especially skeletal deformities in fish, for monitoring marine pollution. Philos. Trans. R. Soc. B 286 (1015), 457–464. Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate–a practical and powerful approach to multiple testing. J. R. Stat. Soc. B: Met. 57 (1), 289–300. Berg, A.H., Westerlund, L., Olsson, P.E., 2004. Regulation of Arctic char (Salvelinus alpinus) egg shell protein and vitellogenin during reproduction and in response to 17␤-estradiol and cortisol. Gen. Comp. Endocrinol. 135, 276–285. Bertram, H., Steck, E., Zimmermann, G., Chen, B.H., Carstens, C., Nerlich, A., Richter, W., 2006. Accelerated intervertebral disc degeneration in scoliosis versus physiological ageing develops against a background of enhanced anabolic gene expression. Biochem. Biophys. Res. Commun. 342 (3), 963–972. Beyer, J., Bechmann, R.K., Taban, I.C., Aas, E., Reichert, W., Seljeskog, E., Sanni, S., 2001. Biomarker measurements in long term exposures of a model fish to produced water components (PAHs and alkylphenols). Report AM-01/007, Akvamiljø. p. 28. Beyer, J., Sandvik, M., Skåre, J.U., Egaas, E., Hylland, K., Waagbø, R., Goksøyr, A., 1997. Time- and dose-dependent biomarker responses in flounder (Platichthys flesus L.) exposed to benzo[a]pyrene, 2,3,3 ,4,4 ,5-hexachlorobiphenyl (PCB-156) and cadmium. Biomarkers 2 (1), 35–44. Boverhof, D.R., Burgoon, L.D., Tashiro, C., Sharratt, B., Chittim, B., Harkema, J.R., Mendrick, D.L., Zacharewski, T.R., 2006. Comparative toxicogenomic analysis of the

290

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291

hepatotoxic effects of TCDD in Sprague–Dawley rats and C57BL/6 mice. Toxicol. Sci. 94 (2), 398–416. Boitsov, S., Meier, S., Klungsoyr, J., Svardal, A., 2004. Gas chromatography–mass spectrometry analysis of alkylphenols in produced water from offshore oil installations as pentafluorobenzoate derivatives. J. Chromatogr. A 1059 (1–2), 131–141. Burney, R.O., Talbi, S., Hamilton, A.E., Vo, K.C., Nyegaard, M., Nezhat, C.R., Lessey, B.A., Giudice, L.C., 2007. Gene expression analysis of endometrium reveals progesterone resistance and candidate susceptibility genes in women with endometriosis. Endocrinology 148 (8), 3814–3826. Carls, M.G., Rice, S.D., Hose, J.E., 1999. Sensitivity of fish embryos to weathered crude oil. Part I. Low-level exposure during incubation causes malformations, genetic damage, and mortality in larval Pacific herring (Clupea pallasi). Environ. Toxicol. Chem. 18 (3), 481–493. Casini, S., Marsili, L., Fossi, M.C., Mori, G., Bucalossi, D., Porcelloni, S., Caliani, I., Stefanini, G., Ferraro, M., di Catenaja, C.A., 2006. Use of biomarkers to investigate toxicological effects of produced water treated with conventional and innovative methods. Mar. Environ. Res. 62, S347–S351. Chasiotis, H., Kelly, S.P., 2008. Occludin immunolocalization and protein expression in goldfish. J. Exp. Biol. 211 (10), 1524–1534. Claireaux, G., Desaunay, Y., Akcha, F., Auperin, B., Bocquene, G., Budzinski, F.N., Cravedi, J.P., Davoodi, F., Galois, R., Gilliers, C., Goanvec, C., Guerault, D., Imbert, N., Mazeas, O., Nonnotte, G., Nonnotte, L., Prunet, P., Sebert, P., Vettier, A., 2004. Influence of oil exposure on the physiology and ecology of the common sole Solea solea: experimental and field approaches. Aquat. Living Res. 17 (3), 335–351. Collier, T.K., Varanasi, U., 1991. Hepatic activities of xenobiotic metabolizing enzymes and biliary levels of xenobiotics in english sole (Parophrys vetulus) exposed to environmental contaminants. Arch. Environ. Contam. Toxicol. 20 (4), 462–473. Day, R.W., Quinn, G.P., 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monogr. 59 (4), 433–463. De Boeck, G., Van der Ven, K., Meeus, W., Blust, R., 2007. Sublethal copper exposure induces respiratory stress in common and gibel carp but not in rainbow trout. Comp. Biochem. Physiol. C 144 (4), 380–390. Douglas, S.E., 2006. Microarray studies of gene expression in fish. OMICS 10 (4), 474–489. Ducasse-Cabanot, S., Zambonino-Infante, J., Richard, N., Medale, F., Corraze, G., Mambrini, M., Robin, J., Cahu, C., Kaushik, S., Panserat, S., 2007. Reduced lipid intake leads to changes in digestive enzymes in the intestine but has minor effect on key enzymes of hepatic intermediary metabolism in rainbow trout (Oncorhynchus mykiss). Animal 1 (9), 1272–1282. Eggens, M.L., Opperhuizen, A., Boon, J.P., 1996. Temporal variation of CYP1A indices, PCB and 1-OH pyrene concentration in flounder, Platichthys flesus, from the Dutch Wadden Sea. Chemosphere 33 (8), 1579–1596. Elonen, G.E., Spehar, R.L., Holcombe, G.W., Johnson, R.D., Fernandez, J.D., Erickson, R.J., Tietge, J.E., Cook, P.M., 1998. Comparative toxicity of 2,3,7,8tetrachlorodibenzo-p-dioxin to seven freshwater fish species during early life-stage development. Environ. Toxicol. Chem. 17 (3), 472–483. Fernandes, J.M.O., Mommens, M., Hagen, O., Babiak, I., Solberg, C., 2008. Selection of suitable reference genes for real-time PCR studies of Atlantic halibut development. Comp. Biochem. Physiol. B 150 (1), 23–32. Filby, A.L., Thorpe, K.L., Maack, G., Tyler, C.R., 2007. Gene expression profiles revealing the mechanisms of anti-androgen-and estrogen-induced feminization in fish. Aquat. Toxicol. 81 (2), 219–231. Filby, A.L., Tyler, C.R., 2007. Appropriate ‘housekeeping’ genes for use in expression profiling the effects of environmental estrogens in fish. BMC Mol. Biol. 8 (10). Förlin, L., Hylland, K., 2006. Hepatic cytochrome P4501A concentration and activity in Atlantic cod caged in two North Sea pollution gradients. In: Hylland, K., Lang, T., Vethaak, D. (Eds.), Biological Effects of Contaminants in Marine Pelagic Ecosystems. SETAC Press, Brussels, pp. 253–262. Geier, A., Dietrich, C.G., Gerloff, T., Haendly, J., Kullak-Ublick, G.A., Stieger, B., Meier, P.J., Matern, S., Gartung, C., 2003. Regulation of basolateral organic anion transporters in ethinylestradiol-induced cholestasis in the rat. BBA-Biomembranes 1609 (1), 87–94. George, S.G., 1994. Enzymology and molecular biology of phase II xenobioticconjugating enzymes in fish. In: Malins, D.C., Ostrander, G.K. (Eds.), Aquatic toxicology; Molecular, Biochemical and Cellular perspectives. Lewis Publishers, CRC Press, pp. 37–85. Gimeno, S., Komen, H., Gerritsen, A.G.M., Bowmer, T., 1998a. Feminisation of young males of the common carp, Cyprinus carpio, exposed to 4-tert-pentylphenol during sexual differentiation. Aquat. Toxicol. 43 (2–3), 77–92. Gimeno, S., Komen, H., Jobling, S., Sumpter, J., Bowmer, T., 1998b. Demasculinisation of sexually mature male common carp, Cyprinus carpio, exposed to 4-tert-pentylphenol during spermatogenesis. Aquat. Toxicol. 43 (2–3), 93– 109. Goldstein, J.L., Brown, M.S., 1990. Regulation of the mevalonate pathway. Nature 343 (6257), 425–430. Gunnarsson, L., Kristiansson, E., Förlin, L., Nerman, O., Larsson, D.G.J., 2007. Sensitive and robust gene expression changes in fish exposed to estrogen—a microarray approach. BMC Genomics 8, 149. Harrison-Findik, D.D., Schafer, D., Klein, E., Timchenko, N.A., Kulaksiz, H., Clemens, D., Fein, E., Andriopoulos, B., Pantopoulos, K., Gollan, J., 2006. Alcohol metabolismmediated oxidative stress down-regulates hepcidin transcription and leads to increased duodenal iron transporter expression. J. Biol. Chem. 281 (32), 22974–22982. Hashimoto, K., Sonoda, Y., Yamakado, M., Funakoshi-Tago, M., Yoshida, N., Rokudai, A., Aizu-Yokota, E., Kasahara, T., 2006. C/EBP alpha inactivation in FAK-

overexpressed HL-60 cells impairs cell differentiation. Cell. Signal. 18 (7), 955–963. Heneweer, M., Houtman, R., Poortman, J., Groot, M., Maliepaard, C., Peijnenburg, A., 2007. Estrogenic effects in the immature rat uterus after dietary exposure to ethinylestradiol and zearalenone using a systems biology approach. Toxicol. Sci. 99 (1), 303–314. Henriquez-Hernandez, L.A., Flores-Morales, A., Santana-Farre, R., Axelson, M., Nilsson, P., Norstedt, G., Fernandez-Perez, L., 2007. Role of pituitary hormones on 17 alpha-ethinlyestradiol-induced cholestasis in rat. J. Pharmacol. Exp. Ther. 320 (2), 695–705. Henson, K.L., Gallagher, E.P., 2004. Glutathione S-transferase expression in pollutionassociated hepatic lesions of brown bullheads (Ameiurus nebulosus) from the Cuyahoga River, Cleveland, Ohio. Toxicol. Sci. 80 (1), 26–33. Hicks, J.M.T., Bailey, J.R., Driedzic, W.R., 1996. Acclimation to low temperature is associated with an increase in long-chain acyl-CoA synthetase in rainbow trout (Oncorhynchus mykiss) heart. Can. J. Zool. 74 (1), 1–7. Hill, A.J., Teraoka, H., Heideman, W., Peterson, R.E., 2005. Zebrafish as a model vertebrate for investigating chemical toxicity. Toxicol. Sci. 86 (1), 6–19. Holbech, H., Andersen, L., Petersen, G.I., Korsgaard, B., Pedersen, K.L., Bjerregaard, P., 2001. Development of an ELISA for vitellogenin in whole body homogenate of zebrafish (Danio rerio). Comp. Biochem. Physiol. C 130 (1), 119–131. Hoodbhoy, T., Dean, J., 2004. Insights into the molecular basis of sperm–egg recognition in mammals. Reproduction 127 (4), 417–422. Howell, B.W., Gertler, F.B., Cooper, J.A., 1997. Mouse disabled (mDab1): a Src binding protein implicated in neuronal development. EMBO J. 16 (1), 121–132. Hsieh, S.L., Hu, C.Y., Hsu, Y.T., Hsieh, T.J., 2007. Influence of dietary lipids on the fatty desaturase expression in hybrid tilapia acid composition and stearoyl-CoA (Oreochromis niloticus × O. aureus) under cold shock. Comp. Biochem. Physiol. B 147 (3), 438–444. Hsieh, S.L., Kuo, C.M., 2005. Stearoyl-CoA desaturase expression and fatty acid composition in milkfish (Chanos chanos) and grass carp (Ctenopharyngodon idella) during cold acclimation. Comp. Biochem. Physiol. B 141 (1), 95–101. Hurst, M.R., Chan-Man, Y.L., Balaam, J., Thain, J.E., Thomas, K.V., 2005. The stable aryl hydrocarbon receptor agonist potency of United Kingdom Continental Shelf (UKCS) offshore produced water effluents. Mar. Pollut. Bull. 50 (12), 1694–1698. Hylland, K., Beyer, J., Berntssen, M., Klungsoyr, J., Lang, T., Balk, L., 2006a. May organic pollutants affect fish populations in the North Sea? J. Toxicol. Environ. Health A 69 (1–2), 125–138. Hylland, K., Ruus, A., Børseth, J.F., Bechmann, R.K., Barsiene, J., Grung, M., Tollefsen, K.E., Myhre, L.P., 2006b. Biomarkers in monitoring—a review. NIVA-Report 5205, Norwegian Institute for Water Resarch, Oslo. p. 106. Incardona, J.P., Collier, T.K., Scholz, N.L., 2004. Defects in cardiac function precede morphological abnormalitites in fish embryos exposed to polycyclic aromatic hydrocarbons. Toxicol. Appl. Pharmacol. 196 (2), 191–205. Incardona, J.P., Carls, M.G., Teraoka, H., Sloan, C.A., Collier, T.K., Scholz, N.L., 2005. Aryl hydrocarbon receptor-independent toxicity of weathered crude oil during fish development. Environ. Health Perspect. 113 (12), 1755–1762. Kammann, U., 2007. PAH metabolites in bile fluids of dab (Limanda limanda) and flounder (Platichthys flesus): spatial distribution and seasonal changes. Environ. Sci. Pollut. Res. 14 (2), 102–108. Krahn, M.M., Burrows, D.G., Ylitalo, G.M., Brown, D.W., Wigren, C.A., Collier, T.C., Chan, S.L., Varanasi, U., 1992. Mass spectrometric analysis for aromatic compounds in bile of fish sampled after the Exxon Valdes oil spill. Environ. Sci. Technol. 26, 116–126. Lakshman, M.R., Chirtel, S.J., Chambers, L.L., Coutlakis, P.J., 1989. Effects of 2,3,7,8tetrachlorodibenzo-para-dioxin on lipid-synthesis and lipogenic enzymes in the rat. J. Pharmacol. Exp. Ther. 248 (1), 62–66. Larsson, D.G.J., Hyllner, S.J., Haux, C., 1994. Induction of vitelline envelope proteins by estradiol-17-beta in 10 Teleost species. Gen. Comp. Endocrinol. 96 (3), 445–450. Le Jossic-Corcos, C., Gonthier, C., Zaghini, I., Logette, E., Shechter, I., Bournot, P., 2005. Hepatic farnesyl diphosphate synthase expression is suppressed by polyunsaturated fatty acids. Biochem. J. 385, 787–794. Levene, H., 1960. In: Olkin, I., et al. (Eds.), Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling. Stanford University Press, Stanford, CA, pp. 278–292. Liu, X.J., Wang, H., Gong, Z.Y., 2006. Tandem-repeated zebrafish zp3 genes possess oocyte-specific promoters and are insensitive to estrogen induction. Biol. Reprod. 74 (6), 1016–1025. Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193, 265–275. Martin, G., Schoonjans, K., Lefebvre, A.M., Staels, B., Auwerx, J., 1997. Coordinate regulation of the expression of the fatty acid transport protein and acyl-CoA synthetase genes by PPAR alpha and PPAR gamma activators. J. Biol. Chem. 272 (45), 28210–28217. Mattingly, C.J., Rosenstein, M.C., Davis, A.P., Colby, G.T., Forrest, J.N., Boyer, J.L., 2006. The Comparative Toxicogenomics Database: a cross-species resource for building chemical–gene interaction networks. Toxicol. Sci. 92 (2), 587–595. McIntosh, A.D., Hylland, K., Gowland, B.T.G., Davies, I.M., 2006. Assessment of hepatic 7-ethoxyresorufin O-deethylase activity and CYP1A concentration in Herring (Clupea harengus) and Saithe (Pollachius virens) from two areas of the North Sea. In: Hylland, K., Lang, T., Vethaak, D. (Eds.), Biological Effects of Contaminants in Marine Pelagic Ecosystems. SETAC Press, Brussels, pp. 93–102. McKee, M.J., Hendricks, A.C., Ebel, R.E., 1983. Effects of naphthalene on benzo[A]pyrene hydroxylase and cytochrome-P-450 in Fundulus-Heteroclitus. Aquat. Toxicol. 3 (2), 103–114.

T.F. Holth et al. / Aquatic Toxicology 90 (2008) 277–291 Meier, S., Andersen, T.C., Lind-Larsen, K., Svardal, A., Holmsen, H., 2007a. Effects of alkylphenols on glycerophospholipids and cholesterol in liver and brain from female Atlantic cod (Gadus morhua). Comp. Biochem. Physiol. C 145 (3), 420–430. Meier, S., Andersen, T.E., Norberg, B., Thorsen, A., Taranger, G.L., Kjesbu, O.S., Dale, R., Morton, H.C., Klungsoyr, J., Svardal, A., 2007b. Effects of alkylphenols on the reproductive system of Atlantic cod (Gadus morhua). Aquat. Toxicol. 81 (2), 207–218. Modig, C., Modesto, T., Canario, A., Cerda, J., von Hofsten, J., Olsson, P.E., 2006. Molecular characterization and expression pattern of zona pellucida proteins in gilthead seabream (Sparus aurata). Biol. Reprod. 75 (5), 717–725. Monteiro, P.R.R., Reis-Henriques, M.A., Coimbra, J., 2000. Polycyclic aromatic hydrocarbons inhibit in vitro ovarian steroidogenesis in the flounder (Platichthys flesus L.). Aquat. Toxicol. 48 (4), 549–559. Mouritzen, P., Noerholm, M., Nielsen, P.S., Jacobsen, N., Lomholt, C., Pfundheller, H.M., Tolstrup, N., 2005. ProbeLibrary: a new method for faster design and execution of quantitative real-time PCR. Nat. Methods 2 (4), 313–316. Myers, M.S., Johnson, L.L., Collier, T.K., 2003. Establishing the causal relationship between polycyclic aromatic hydrocarbon (PAH) exposure and hepatic neoplasms and neoplasia-related liver lesions in English sole (Pleuronectes vetulus). Hum. Ecol. Risk Assess. 9 (1), 67–94. Neff, J.M., 2002. Polycyclic aromatic hydrocarbons in the ocean. In: Neff, J.M. (Ed.), Bioaccumulation in Marine Organisms. Elsevier, Amsterdam, pp. 241–318. Nimrod, A.C., Benson, W.H., 1996. Environmental estrogenic effects of alkylphenol ethoxylates. Crit. Rev. Toxicol. 26 (3), 335–364. Ohba, K., Okawa, Y., Matsumoto, Y., Nakamura, Y., Ohta, H., 2007. A study of investigation of cadmium genotoxicity in rat bone cells using DNA microarray. J. Toxicol. Sci. 32 (1), 107–109. OLF, 2008. Environmental Report 2007. The Norwegian Oil Industry Association, Stavanger, p. 62. OLF, 2007. Environmental Report 2006. The Norwegian Oil Industry Association, Stavanger, p. 56. Olsvik, P.A., Lie, K.K., Stavrum, A.K., Meier, S., 2007. Gene-expression profiling in gill and liver of zebrafish exposed to produced water. Int. J. Environ. Anal. Chem. 87 (3), 195–210. Panserat, S., Ducasse-Cabanot, S., Plagnes-Juan, E., Srivastava, P.P., Kolditz, C., Piumi, F., Esquerre, D., Kaushik, S., 2008. Dietary fat level modifies the expression of hepatic genes in juvenile rainbow trout (Oncorhynchus mykiss) as revealed by microarray analysis. Aquaculture 275 (1–4), 235–241. Parisuthiman, D., Mochida, Y., Duarte, W.R., Yamauchi, M., 2005. Biglycan modulates osteoblast differentiation and matrix mineralization. J. Bone Miner. Res. 20 (10), 1878–1886. Parkinson, A., 2001. Biotransformation of xenobiotics. In: Klaassen, C.D. (Ed.), Casarett and Doull’s Toxicology. The basic Science of Poisons. McGraw-Hill, New York, pp. 133–224. 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. Safe, S., 2001. Molecular biology of the Ah receptor and its role in carcinogenesis. Toxicol. Lett. 120 (1–3), 1–7. Sampath, H., Ntambi, J.M., 2005. Polyunsaturated fatty acid regulation of genes of lipid metabolism. Annu. Rev. Nutr. 25, 317–340. Scaglia, N., Igal, R.A., 2005. Stearoyl-CoA desaturase is involved in the control of proliferation, anchorage-independent growth, and survival in human transformed cells. J. Biol. Chem. 280 (27), 25339–25349. Seki, M., Yokota, H., Matsubara, H., Maeda, M., Tadokoro, H., Kobayashi, K., 2003. Fish full life-cycle testing for the weak estrogen 4-tert-pentylphenol on medaka (Oryzias latipes). Environ. Toxicol. Chem. 22 (7), 1487–1496. Shinomura, T., Asaoka, Y., Oka, M., Yoshida, K., Nishizuka, Y., 1991. Synergistic action of diacylglycerol and unsaturated fatty-acid for protein-kinase-C activation—its possible implications. Proc. Natl. Acad. Sci. U.S.A. 88 (12), 5149–5153. Skadsheim, A., 2004. Cod stocks exposed to crude oils: uptake, metabolism and biomarker effects. Report AM-2004/005, RF-Akvamiljø, Randaberg, p. 192. Smyth, G.K., Michaud, J., Scott, H.S., 2005. Use of within-array replicate spots for assessing differential expression in microarray experiments. Bioinformatics 21 (9), 2067–2075.

291

Smyth, G.K., 2004. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat. Appl. Genet. Mol. Biol. 3 (1), 3. Smyth, G.K., Speed, T.P., 2003. Normalization of cDNA microarray data. Methods 31, 265–273. Sokal, R.R., Rohlf, F.J., 1994. Biometry. W.H. Freeman & Company, New York. Stephens, S.M., Frankling, S.C., Stagg, R.M., Brown, J.A., 2000. Sub-lethal effects of exposure of juvenile turbot to oil produced water. Mar. Pollut. Bull. 40 (11), 928–937. Sturve, J., Hasselberg, L., Falth, H., Celander, M., Förlin, L., 2006. Effects of North Sea oil and alkylphenols on biomarker responses in juvenile Atlantic cod (Gadus morhua). Aquat. Toxicol. 78, S73–S78. Sundstrom, S.A., Komm, B.S., Poncedeleon, H., Yi, Z., Teuscher, C., Lyttle, C.R., 1989. Estrogen regulation of tissue-specific expression of complement-C3. J. Biol. Chem. 264 (28), 16941–16947. Teh, C., Parinov, S., Korzh, V., 2005. New ways to admire zebrafish: progress in functional genomics research methodology. Biotechniques 38 (6), 897– 906. Teran-Garcia, M., Adamson, A.W., Yu, G., Rufo, C., Suchankova, G., Dreesen, T.D., Tekle, M., Clarke, S.D., Gettys, T.W., 2007. Polyunsaturated fatty acid suppression of fatty acid synthase (FASN): evidence for dietary modulation of NF-Y binding to the Fasn promoter by SREBP-1c. Biochem. J. 402, 591–600. Tollefsen, K.E., Eikvar, S., Finne, E.F., Fogelberg, O., Gregersen, I.K., 2008. Estrogenicity of alkylphenols and alkylated non-phenolics in a raindow trout (Oncorhyncus mykiss) primary hepatocyte culture. Ecotoxicol. Environ. Saf. 71 (2), 370– 383. Tollefsen, K.E., Harman, C., Smith, A., Thomas, K.V., 2007. Estrogen receptor (ER) agonists and androgen receptor (AR) antagonists in effluents from Norwegian North Sea oil production platforms. Mar. Pollut. Bull. 54 (3), 277–283. Ton, C., Stamatiou, D., Liew, C.C., 2003. Gene expression profile of zebrafish exposed to hypoxia during development. Physiol. Genomics 13 (2), 97–106. van der Oost, R., Beyer, J., Vermeulen, N.P.E., 2003. Fish bioaccumulation and biomarkers in environmental risk assessment: a review. Environ. Toxicol. Pharmacol. 13 (2), 57–149. van der Ven, K., De Wit, M., Keil, D., Moens, L., Van Leemput, K., Naudts, B., De Coen, W., 2005. Development and application of a brain-specific cDNA microarray for effect evaluation of neuro-active pharmaceuticals in zebrafish (Danio rerio). Comp. Biochem. Physiol. B 141 (4), 408–417. van Lipzig, M.M.H., Vermeulen, N.P.E., Gusinu, R., Legler, J., Frank, H., Seidel, A., Meerman, J.H.N., 2005. Formation of estrogenic metabolites of benzo[a]pyrene and chrysene by cytochrome P450 activity and their combined and supra-maximal estrogenic activity. Environ. Toxicol. Pharmacol. 19 (1), 41–55. Voelker, D., Vess, C., Tillmann, M., Nagel, R., Otto, G.W., Geisler, R., Schirmer, K., Scholz, S., 2007. Differential gene expression as a toxicant-sensitive endpoint in zebrafish embryos and larvae. Aquat. Toxicol. 81 (4), 355–364. Wakil, S.J., 1989. Fatty-acid synthase. A proficient multifunctional enzyme. Biochemistry 28 (11), 4523–4530. Wang, N.D., Finegold, M.J., Bradley, A., Ou, C.N., Abdelsayed, S.V., Wilde, M.D., Taylor, L.R., Wilson, D.R., Darlington, G.J., 1995. Impaired energy homeostasis in C/EbpAlpha knockout mice. Science 269 (5227), 1108–1112. Warner, K.E., Jenkins, J.J., 2007. Effects of 17 alpha-ethinylestradiol and bisphenol a on vertebral development in the fathead minnow (Pimephales promelas). Environ. Toxicol. Chem. 26 (4), 732–737. Willett, K.L., Wassenberg, D., Lienesch, L., Reichert, W., Di Giulio, R.T., 2001. In vivo and in vitro inhibition of CYP1A-dependent activity in Fundulus heteroclitus by the polynuclear aromatic hydrocarbon fluoranthene. Toxicol. Appl. Pharmacol. 177 (3), 264–271. Williams, T.D., Gensberg, K., Minchin, S.D., Chipman, J.K., 2003. A DNA expression array to detect toxic stress response in European flounder (Platichthys flesus). Aquat. Toxicol. 65 (2), 141–157. Xu, C.J., Li, C.Y.T., Kong, A.N.T., 2005. Induction of phase I, II and III drug metabolism/transport by xenobiotics. Arch. Pharmacol. Res. 28 (3), 249–268.