Background
A great challenge in pharmacology and toxicology is to understand the molecular mechanisms behind how mixtures of compounds affect living organisms. This study focuses on two classes of substances, imidazoles and xenoestrogens, and how these chemicals alone and in combination affect hepatic drug-metabolizing hepatic cytochrome P450 (CYP) enzymes – specifically, CYP1A and CYP3A enzymes, in juvenile Atlantic cod (Gadus morhua).
Imidazoles and triazoles are used as fungicides both clinically as well as in horticulture and agriculture, posing a potential threat to wildlife. The triazole propiconazole has been detected in the aquatic environment [
1]. The azole antifungal effect resides in inhibition of CYP51 mediated ergosterol biosynthesis [
2]. In addition to disrupting key enzymes in fungus, azoles such as the imidazoles clotrimazole, ketoconazole, miconazole and prochloraz also cause endocrine disruption in vertebrates by inhibition of key enzymes in steroid homeostasis [
3‐
7]. Moreover, these fungicides inhibit drug-metabolizing CYP forms, including members of the CYP1, CYP2 and CYP3 gene families in vertebrates [
5,
8‐
13]. Effects on CYP forms may have adverse effects on metabolic clearance of endobiotics and xenobiotics. For example, in a study in fish, pre-exposure to clotrimazole resulted in increased bioaccumulation of the pro-carcinogen benzo [
a]pyrene in gizzard shad (
Dorosoma cepedianum) [
14].
Xenoestrogens comprise a wide variety of structurally diverse chemicals such as
o,p-DDT, ethynylestradiol, alkylphenols and bisphenol A. These substances are well-known or supposed to be endocrine disrupting substances in vertebrates and share in common that they activate the estrogen receptor (ER) and thereby elicit estrogenic responses [
15‐
17]. In addition to being estrogenic, these xenoestrogens interact with drug-metabolizing CYP forms, including members of the CYP1A and CYP3A subfamilies in vertebrates [
18‐
22].
Xenoestrogens are continuously released into the environment as a result of various anthropogenic activities. Induction of vitellogenesis in fish is a biomarker routinely used to assess the presence of estrogenic substances in the aquatic environment [
23,
24]. Induction of CYP1A-mediated ethoxyresorufin-
O-deethylase (EROD) activity is another established biomarker used to assess exposure to aromatic hydrocarbons. This response proceeds through activation of the aryl hydrocarbon receptor (AHR) by aromatic hydrocarbons including polyaromatic hydrocarbons, and planar polychlorinated biphenyls and dioxins [
25]. Some AHR agonists have been shown to be anti-estrogenic and cross-talk between AHR and ER has been suggested in vertebrates [
26‐
33].
In addition to activation of the ER, xenoestrogens also affect other steroid receptors. Nonylphenol up-regulated CYP3A1 gene expression in rat, through activation of the pregnane X receptor (PXR) [
34,
35]. We previously reported induction of CYP3A and CYP1A protein levels in Atlantic cod exposed to alkylphenols [
22].
Azole fungicides induce expression of multiple vertebrate CYP genes including members of the CYP1A, CYP2B and CYP3A subfamilies [
8,
9,
13,
36‐
38]. Clotrimazole activates the ligand-binding domain of the PXR, involved in CYP3A signalling,
in vitro from several mammalian species and zebra fish (
Danio rerio) [
39]. Both imidazoles and xenoestrogens inhibit drug-metabolizing enzymes, including members of the CYP1A and CYP3A subfamilies in vertebrates [
8‐
13,
18,
20,
22]. Thus, xenoestrogens and imidazoles conceivably share common routes for biotransformation. However, there is a lack of data regarding effects of combined exposure of imidazoles and xenoestrogens on these CYP forms in wildlife. Living organisms usually are exposed to mixtures of different classes of xenobiotics. Conceivably, exposure to mixtures may be more of a health threat than exposure to single compounds, as a result of interactions. Anthropogenic compounds may enter the environment through industrial activities and through the use of pharmaceuticals [
40]. Atlantic cod is an economically important species for fishery and a growing aquaculture industry, in addition to its ecological relevancy. Its distribution in the Northern Atlantic and the North Sea makes it vulnerable to effluents from on-shore and off-shore industries and from run-off entering the waters near highly industrialized and urbanized areas.
The rationale of the present study was to identify possible sites of interactions between imidazoles and xenoestrogens. We hypothesise that combined exposure to these compounds may compromise the metabolic clearance not only of these xenobiotics themselves, but also of endobiotics such as circulating steroid hormones that share common routes of metabolism through hepatic CYP1A and CYP3A. Such endocrine disrupting effects may adversely affect the stability of wildlife populations.
The specific aim of our study was to examine interactions between two classes of compounds in livers of Atlantic cod. Thus, we investigated the effects of the model imidazole ketoconazole and of two types of xenoestrogens (nonylphenol and ethynylestradiol), as well as of a mixed exposure to ketoconazole and nonylphenol, on hepatic CYP1A and CYP3A protein expression and catalytic activities, and also on vitellogenesis and plasma levels of sex steroid hormones.
Conclusions
This study identifies, in Atlantic cod, interactions between ketoconazole and two different types of xenoestrogens on CYP1A and CYP3A. Ketoconazole acted as a non-competitive inhibitor of CYP1A and CYP3A activities and ketoconazole treatment also induced CYP1A protein expression. Ethynylestradiol acted as a non-competitive inhibitor of CYP1A and an uncompetitive inhibitor of CYP3A activities. In vitro studies revealed that nonylphenol was a non-competitive inhibitor of CYP1A; but it did not inhibit CYP3A. However, in vivo, nonylphenol synergistically impaired the ketoconazole-mediated inhibition of CYP3A activity, without affecting CYP3A protein levels. The study further illustrates that induction of CYP1A- and CYP3A gene expression can be partly or completely masked by inhibition of catalytic activities or vice versa. Taken together, the results indicate that CYP1A and CYP3A represent sites of interactions between those classes of xenobiotics. In future risk-assessment of, e.g., municipal effluents or produced water from oil platforms, that have been shown to contain xenoestrogens, it should be considered to identify other classes of substances, for example azoles that also interact with CYP1A and CYP3A. Our data may warn for ecotoxicological implications, as induction of EROD activity as well as plasma vitellogenin routinely are used as biomarkers to assess exposure to AHR and ER agonists in various biomonitoring programs in the aquatic environment.
Methods
Chemicals
The 4-nonylphenol and the 17α-ethynylestradiol, for the in vivo exposure experiment, were obtained from Fluka Chemie AG (Buchs, Switzerland). The 4-nonylphenol for the in vitro inhibition studies was from Berol Nobel (Stenungsund, Sweden). Dimethylsulphoxide (DMSO), 7-ethoxyresorufin, horseradish peroxidase- (HRP) conjugated goat-anti-mouse IgG, iodoacetamide, ketoconazole, ponceau-S, resorufin and tween-20 were obtained from Sigma Aldrich (Stockholm, Sweden). Reduced nicotinamide-adenine-dinucleotide-phosphate (NADPH) was from Roche Diagnostics (St Louis, MO, USA and Bromma, Sweden). Ready gels (12% continuous acrylamide in Tris:HCl), 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), precision protein standards (low range) and supported nitrocellulose membrane (0.45 μm) were purchased from BioRad (Sundbyberg, Sweden). The 17β-estradiol and testosterone enzyme immuno assay (EIA) kits were purchased from Cayman Chemical (Ann Arbor, MI, USA). The 11-keto-testosterone EIA kit and the Atlantic cod vitellogenin Enzyme Linked ImmunoSorbent Assay (ELISA) kit were obtained from Biosense Laboratories AS (Bergen, Norway). HRP-conjugated donkey-anti-rabbit IgG, the ECL™ Western blotting detection reagents and Immobiline™ DryStrip 7 cm ranging from pH 4 to 7 were from Amersham Biosciences (Uppsala, Sweden). Ampholytes for isoelectric focusing (Servalyt® Carrier ampholyt 3–10) was purchased from Serva Feinbiochemica (Heidelberg, Germany). Dithiothreitol (DTT), Kodak X-Omat AR-ray film, X-ray developer and fix were from VWR International (Stockholm, Sweden). The 7-benzyloxy-4-[trifluoromethyl] coumarin (BFC), 7-hydroxy-4-[trifluoromethyl] coumarin (HFC) and the CYP3A4 inhibition kit were from BD Biosciences Company, Gentest™ (Woburn, MA, USA). All other chemicals used were of the purest grade available in Sweden or Norway, from Sigma-Aldrich, BioRad and VWR international.
Animals and sampling
Hatchery reared juvenile Atlantic cod of both sexes with an average body weight (b.w.) around 400 g were supplied by Sekkingstad, Preserving AS, Hordaland, Norway. The fish were kept in 0.5 m
3 indoor glass fibre tanks, at Industrial Laboratory (ILAB), Bergen High Technology Centre (Bergen, Norway), provided with continuously flowing seawater at a temperature of 8 ± 0.5°C and a salinity of 3.4%. Throughout the experimental period, the fish were subjected to continuous 24 h artificial light (the regime the farm used for optimal fish growth). The fish were acclimated to these conditions for five days prior to the experimental period. During the experimental period the fish were starved and i.p. injected with either 12 mg ketoconazole/kg b.w. resuspended in peanut oil (2.5 mg/ml); 25 mg nonylphenol/kg b.w. dissolved in peanut oil (5.0 mg/ml); 5 mg ethynylestradiol/kg b.w. dissolved in peanut oil (1.0 mg/ml) or a mixture of ketoconazole and nonylphenol (12 mg ketoconazole + 25 mg nonylphenol/kg b.w. in peanut oil). Control fish were injected with 5 ml peanut oil/kg b.w. (vehicle). There were eight to nine fish in each treatment group. When designing the experiment, we could only test one combination due to limited fish numbers. We selected nonylphenol to combine with ketoconazole because a previous study showed that, in Atlantic cod, alkylphenols affect CYP1A/3A more strongly than the natural estrogen 17β-estradiol [
22]. The ketoconazole dose (12 mg/kg) was selected based on the results on CYP1A and CYP3A protein levels and enzyme activities from a previous dose-response study in rainbow trout [
13]. The nonylphenol dose (25 mg/kg) was selected as this dose is known to induce vitellogenesis in a number of fish species. In addition, in a previous study on the Atlantic salmon, this dose of nonylphenol also had effects on CYP1A and CYP3A [
19].
After five days exposure, the fish were sacrificed by a sharp blow to the head. Blood samples were collected from the dorsal vein using a heparinized syringe and the liver was quickly dissected out and placed in ice-cold homogenization buffer (0.1 M sodium phosphate buffer pH 7.4, containing 0.15 M KCl, 1 mM EDTA and 1 mM DTT). Liver microsomes were prepared according to the published protocol by Goksøyr [
50], and stored at -80°C. Total microsomal protein content was measured according to a published method by Bradford [
51], using bovine serum albumin as standard, and a SpectraFluor spectrophotometer from Tecan (Grödig/Salzburg, Austria). Blood plasma was isolated by centrifugation at 5,000 g for 10 min at room temperature and stored at -80°C. Ethical approval licence number of ILAB Bergen: 119. Experiment no. 0204.
For
in vitro inhibition studies, feral Atlantic cod of both sexes were caught off the West coast of Sweden and placed in concrete basins provided with recirculating aerated seawater at 10 ± 2°C and a salinity of 3.0% and alternative light/dark photoperiods of 12 hours. Prior to sampling, the animals were starved and acclimated to these conditions for three weeks. Eight fish were injected i.p. with β-naphthoflavone (BNF), 50 mg/kg b.w. dissolved in peanut oil (5.0 mg/ml). The fish were placed in a 100 l glass aquarium provided with aerated seawater (above) and 30% of the water volume was replaced each day. To eliminate visual stress, the sides of the aquaria were covered with black plastic sheets. After 3 days exposure, the fish were sacrificed. Livers were quickly dissected out and placed in ice-cold homogenization buffer. Livers were pooled from twenty untreated Atlantic cod and from eight BNF treated Atlantic cod, respectively. Microsomal fractions were isolated (above) and stored in aliquots at -80°C. Ethical approval from the Ethical committee of Göteborg license number (99–2003). The duration of exposure was decided according to results from previous time-course studies showing maximal CYP1A protein and EROD activities in rainbow trout and in the marine viviparous blenny (
Zoarces viviparous), 3 days post-injection with either the prototypical CYP1A inducers BNF or 3-methylcholanthrene [
52‐
54].
CYP1A- and CYP3A protein blot analyses
Western blot analyses of 40 μg hepatic microsomal protein were carried out using enhanced chemoluminescence (ECL), based on the protocol previously described [
55] and PAb raised against rainbow trout CYP1A and CYP3A [
41,
55,
56]. The intensity of each protein band was determined by densitometry on scanned fluorograms using Labview 7.0 from National Instruments (Austin, TX, USA).
The 2D gel electrophoresis was performed using immobilised pH gradient gels with linear gradient from pH 4 to 7. The samples were concentrated by acetone precipitation and pellets dissolved in rehydration buffer (8 M urea, 2 M thiourea, 20 mM DTT, 4% CHAPS, 0.5% Triton X-100, 0.5% ampholyte 3–10 and <0.02% bromophenolblue) to a final protein concentration of 20 μg/μl or 80 μg/μl. The samples were rehydrated overnight followed by isoelectric focusing for 2.5 h. The rehydration was passive and carried out overnight in an Immobiline Dry Strip reswelling tray (Amersham Biosciences). First-dimension isoelectric focusing (IEF) was performed on a Multiphor II unit (Amersham Biosciences) at 20°C using a MultiDrive XL power supply (Pharmacia LKB). Settings for IEF were 30 min at 100 V and 3 h at 3500 V for a total of 10,520 Vh. Amperage and wattage were set to 2 mA and 5 W. The proteins were resolved on 9% continuous acrylamide gel in Tris:HCl, including sodium dodecyl sulphate polyacrylamide using a mini-gel apparatus from BioRad at 200 V for 45 min. Each sample consisted of pooled liver microsomes from eight to nine fish from each treatment group. Gels loaded with 25 μg microsomal protein were initially stained with 0.1 % (w/v) Coomassie brilliant blue, and then destained, followed by silver staining. The latter was performed according to Heukeshoven and Dernick [
57]. Stained gels were scanned and analyzed using the PDQUEST 7.1 software (BioRad). Gels loaded with 100 μg microsomal proteins were electrotransferred to nitrocellulose membrane and immunoblotted for CYP3A, as described above.
Catalytic assays
The CYP1A activity was determined as EROD activity, using resorufin as standard in a SpectraFluor plate reader according to the protocol provided by Nilsen
et al. [
58]. The CYP3A catalytic activity was measured as BFCOD activity, using HFC as standard. The BFC assay was performed based on a published protocol by Miller
et al. [
59] and optimized for rainbow trout liver microsomes (T. Hegelund and M. Celander, unpublished data). The reaction mixture consisted of 200 μM BFC, bovine serum albumin (1.6 mg/ml), 2 μM NADPH and 10 μl liver microsomes in a total volume of 200 μl in 0.2 M potassium phosphate buffer pH 7.4 in a 96-multiwell plate using a VICTOR™ 1420 Multilabel Counter from Wallac Sverige AB (Upplands Väsby, Sweden).
In vitroinhibition of CYP1A and CYP3A
In vitro inhibition studies were carried out in 96-multiwell plates using a VICTOR™ 1420 Multilabel Counter. The IC50 values were determined for nonylphenol, ethynylestradiol, ketoconazole and the ketoconazole:nonylphenol (1:5) mixture on CYP1A and CYP3A activities. The substances were dissolved in DMSO and diluted with ethanol. The final concentrations never exceeded 0.01% (v/v) DMSO and 0.001% ethanol (v/v). For CYP1A and CYP3A inhibition studies, pooled liver microsomes from BNF-treated and from untreated Atlantic cod, respectively, were used. For comparison, the IC50 values for ketoconazole, nonylphenol and ethynylestradiol were determined in cDNA expressed human CYP3A4 baculovirus supersomes using the CYP3A4 inhibition kit from BD Gentest.
In vitroincubation studies
Pooled liver microsomes from untreated Atlantic cod were pre-incubated for 10, 30 and 60 min, at room temperature, with ethynylestradiol and ketoconazole following CYP3A western blot analysis. The reaction mixture consisted of microsomes (2.5 or 5.0 mg protein/ml) and various concentrations of ethynylestradiol (35, 50 and 100 μM) or ketoconazole (0.3 and 1.0 μM) ± 3 μM NADPH in a total volume of 50 μl in homogenization buffer, containing 20% (v/v) glycerol. The CYP3A western blot analysis was performed as described above. Ethynylestradiol and ketoconazole were dissolved in acetonitrile (vehicle) and the final acetonitrile concentration in the reaction mixture was 0.02% (v/v).
Plasma vitellogenin analysis
Plasma levels of vitellogenin protein were determined using a non-competitive sandwich ELISA kit and employing rabbit PAb against Atlantic cod vitellogenin from Biosense Laboratories AS (Bergen, Norway) [
58]. Each plasma sample was diluted (1:20, 1:15,000 and 1:50,000) and 100 μl was analyzed and compared to purified Atlantic cod vitellogenin protein standards (ranging between 0.12 and 2,000 ng/ml). The signal was detected at A
492 after 15 min incubation with substrate solution, using a VICTOR™ 1420 Multilabel Counter.
Plasma sex steroid hormone analyses
Plasma levels of 17β-estradiol and testosterone were determined using competitive EIA kits from Cayman Chemical (Ann Arbor, MI, USA). Plasma levels of 11-keto-testosterone were analyzed using a competitive EIA kit, from Biosense Laboratories AS (Bergen, Norway). Plasma from each fish was concentrated (2:1) by extraction once with six volumes diethyl ether and 50 μl was analyzed and compared to purified standard substances. The signals were detected at A405 after 40 min (17β-estradiol), 60 min (testosterone) or 80 min (11-keto-testosterone) incubation with substrate solution, using a VICTOR™ 1420 Multilabel Counter.
Statistics
Data were tested for homogeneity of variances using the Levene's test. When there was homogeneity of variances we used a parametric one-way ANOVA, followed by Scheffé post hoc test. When there was no homogeneity of variances we used the non-parametric Kruskal-Wallis ANOVA, followed by the two-tailed Mann-Whitney U test. No values were log transformed. Data are presented as means (n = 6–9 fish) accompanied with the standard deviations (SD). The significance level (α) was set at 0.05. The statistical analyses were performed using SPSS 11.0 software, from SPSS Sweden AB (Sundbyberg, Sweden).
Acknowledgements
We thank Åsa Berglund and Susan Westerberg, Department of Zoophysiology, Göteborg University and Christina Tolfsen Department of Molecular Biology, University of Bergen for excellent technical assistance. The Improving Human Potential Programme from the European Union, through Contract No HPRI-CT-2002-00188 to LH and MC, has funded access to installations from the University of Bergen. Financially supported by the Faculty of Science, Göteborg University, and grants from Swedish EPA (ReproSafe), C.F. Lundström Foundation and His Swedish Royal Majesty Carl XVI Gustaf's 50-Anniversary Foundation to MC.
Authors' contributions
LH performed most of the analyses, participated in fish exposure, sampling, experimental design and drafted the manuscript. BEG assisted with experimental design, 2D-analysis and writing. AG participated in experimental design and writing. MCC rose funding, coordinated, participated in fish exposure, sampling, experimental design and writing.