Introduction

There are reasons to believe that the Arctic marine ecosystem is being increasingly loaded with organic pollutants, both from industrial discharges transported from temperate latitudes, and due to human activities in the Arctic itself (Friedman and Selin 2012). Polycyclic aromatic hydrocarbons (PAHs) are a large group of widely spread environmental contaminants, reaching the Arctic through long-range atmospheric transport (Laender et al. 2011; Friedman and Selin 2012) as well as through discharges from large Canadian and Russian river systems (Yunker and MacDonald 1995; AMAP 1998; Dahle et al. 2003). As the Arctic holds important oil reserves, local sources of hydrocarbons from oil seepage are also believed to contribute significantly to the release of PAHs in Arctic environments (Doré 1995; Blasco et al. 2010; Foster et al. 2015). The discovery of recoverable oil reserves in the Arctic will probably result in petroleum production in the near future, further increasing pollutant loading in Arctic ecosystems. Thus, there is an urgent need for increased research effort regarding the fate and effects of PAHs in Arctic wildlife.

Polar cod (Boreogadus saida) is a cryo-pelagic key species in the Arctic shelf seas, also known from the Arctic Basin (Gradinger and Bluhm 2004). It is one of the most abundant fish species of the Arctic and constitutes a major link between lower and higher trophic levels (Hop and Gjøsæter 2013). A number of studies have investigated the biological effects of PAHs or other petroleum related products on this species (e.g. Nahrgang et al. 2009, 2010a, b; Dussauze et al. 2014; Geraudie et al. 2014); however, important gaps in knowledge remain concerning the toxicokinetics of PAHs in this species. In 2000, Ingebrigtsen and co-workers investigated the distribution and cellular binding of benzo(a)pyrene (BaP) in polar cod and showed an increased systemic availability of BaP through waterborne exposure compared with dietary exposure. More recently, Nahrgang et al. (2010b) suggested that polar cod exposed to dietary crude oil may show a potentially higher absorption of dietary PAHs compared to other fish species (James et al. 1996; van Veld et al. 1997; Couillard et al. 2009) due to slow gastric evacuation rates and high assimilation efficiency of lipids (Hop and Tonn 1998). Although of special importance for lipophilic contaminants, the dietary route of exposure has not been studied to the same extent as the waterborne exposure. Furthermore, studies have focused on a few model compounds, particularly BaP (James et al. 1991, 1996; Lemaire et al. 1992; Ingebrigtsen et al. 2000).

BaP, a five-ring PAH, is an ubiquitous contaminant that has been widely used as a model compound for pyrogenic-type PAH studies due to its carcinogenicity and mutagenicity, and risk of human and animal health (Phillips 1983; Stegeman and Lech 1991). On the other hand, the non-mutagenic tricyclic aromatic hydrocarbon phenanthrene (Phen) has received less attention, although it is highly abundant in crude oil comparatively to BaP. The aim of the present study was therefore to investigate and compare the tissue distribution of the model compound BaP with Phen in polar cod following oral administration.

Materials and methods

Chemicals

Uniformly labelled 14C-benzo(a)pyrene (BaP) dissolved in toluene (1 mg BaP/ml, specific activity 20–50 mCi/mmol, purity >97 %) and uniformly labelled 14C-phenanthrene (Phen) dissolved in ethanol (0.34 mg Phen/ml, specific activity 50–60 mCi/mmol, purity >97 %) were purchased from American Radiolabeled Chemicals, Inc.

Sampling and acclimation

Sexually mature fish of both sexes were collected in the waters off the Svalbard Archipelago (latitude 78°N) with a Campelen bottom trawl onboard R/V Helmer Hanssen in October 2012. The sampled specimens were transported to the biological station of UiT The Arctic University of Norway (latitude 70°N) and maintained in 300-l holding tanks with running seawater at a temperature of about three degrees Celsius (3 °C) and constant dimmed light until experimental start in November 2012. During acclimation, polar cod were fed until satiation three times weekly with one of its natural prey items, i.e. calanoïd copepods Calanus spp. (purchased frozen from Calanus AS, Tromsø). The characteristics of the included specimens are listed in Table 1.

Table 1 Description of experimental specimens: sex (F: females, M: male), body weight (g) and total body length (cm) at sampling point, amount of food administered (g) by forced feeding and equivalent doses of radiolabeled PAHs µCi/g fish
Full size table

Test substances and experimental design

The test substances (BaP and Phen) were homogenously mixed in the Calanus spp. food at a concentration of 4.16 µCi/g wet weight food (solvent concentration of 42 µl/g food). One week prior to the exposure, polar cod were transferred to two experimental tanks and feeding was stopped. On the exposure day, polar cod (total n = 24, mean (±SD) total wet weight 31.8 ± 9.3 g and mean body length 16.8 ± 9.6 cm) were individually weighed, and 1-ml tuberculin syringes (BD Plastipak™) were filled with food mixture corresponding to 2.8 ± 0.8 % body weight (see Table 1 for details). Polar cod were exposed to a single oral dose, through force-feeding, to one of the mixtures at a final dose of 0.12 ± 0.03 mCi/kg fish, equivalent to a concentration of 1.15 ± 0.36 mg BaP/kg fish and 0.40 ± 0.12 mg Phen/kg fish. The fish were then observed during 3 min to control for potential regurgitation of the food, and only one specimen regurgitated during the course of the experiment (Table 1). Polar cod were then transferred to two 300-l experimental tanks (BaP and Phen-exposed fish, respectively) under running seawater of temperature of 3 °C and constant dimmed light. During the experiment, polar cod were not fed. Three specimens from each group were sampled 2, 6, 12, and 30 days after administration. Briefly, polar cod were euthanised with Finquel MS 222 (tricaine methanesulfonate) and immediately frozen at −80 °C.

Autoradiography

The frozen fish were individually prepared for tape-section autoradiography according to the method of Ullberg (Ullberg 1954, 1977). The fish were embedded in a precooled (1 °C) aqueous gel of carboxymethylcellulose (1 %) and frozen on dry ice. From each fish, sagittal whole-body sections (20 μm) were obtained on tape (no. 821, 3 M, St. Paul, Minn.) in a cryo-microtome (PMV 450 MP, Palmstierna Mekaniska, Stockholm) at −20 °C. All sections were freeze-dried at −20 °C for 24 h. In order to avoid artefacts due to melting and diffusion of fat in the dehydrated sections, all handling and exposure took place below −20 °C. Selected freeze-dried sections were extracted successively with polar and non-polar solvents (5 % trichloroacetic acid for 1 min, 50 % methanol for 30 s, 100 % methanol for 30 s, heptane for 15 s, 100 % methanol for 30 s, 50 % methanol for 30 s and tap water for 5 min). Both native and extracted sections were exposed to X-ray film (Structurix D7 DW ETE, Agfa, Belgium) for 14 days (−20 °C). Autoradiograms from the freeze-dried sections were considered to represent unchanged mother compound (14C-BaP or 14C-Phen) plus its soluble metabolites, while autoradiograms from the freeze-dried solvent-extracted sections were considered to represent metabolites firmly bound to tissue macromolecules (Bergman 1979; Brandt and Brittebo 1989; Ingebrigtsen et al. 2000).

Scintillation counting

Liver and bile collected from all but one specimen (BaP, 2 days) were analysed through scintillation counting. Liver tissue was retrieved from the sectioned blocks and weighed (2.1–46.4 mg), 1 ml of Soluene was added, and the sample was allowed to digest overnight at 37 °C. Bile was collected from the native sections (20 µm thick) with a sharp, hollow probe (3 mm in diameter). From each fish, one to two bile samples were transferred to a vial and left to digest in 1 ml Soluene overnight at 37 °C. The following day, the samples were briefly whirl-mixed before 5 ml scintillation cocktail (Hionic fluor) was added each vial. A second round of whirl-mixing (30 s) was employed before the samples were left on the bench for at least 1 h to equilibrate. The counting took place in a Tri-Carb liquid scintillation analyser (1900CA, Packard) with a built-in quencher standard. Quantified radioactivity was presented as disintegrations per minute (dpm)/mg for liver and dpm/µl for bile.

Results

Autoradiograms from native and extracted sections are presented in Fig. 1, and quantitative measure of the presence of radiolabeled compounds in liver and bile, through scintillation counting, is presented in Fig. 2.

Fig. 1
figure 1

Native tissue sections (20 μm, first panel) with corresponding autoradiograms (second panel) and autoradiograms from extracted sections (third panel) of polar cod (Boreogadus saida) 2, 6 and 30 days following intragastric administration of 14C-benzo(a)pyrene (BaP) or 14C-phenanthrene (Phen). Autoradiograms from native sections represent both mother compound and its soluble metabolites. Non-extractable PAH-derived radioactivity, regarded to represent reactive intermediates bound to tissue macromolecules, is presented in the autoradiograms from extracted sections. Numbering of tissues; 1 gall bladder, 2 mid-intestine, 3 posterior intestine, 4 pyloric caeca, 5 liver, 6 anterior kidney, 7 posterior kidney, 8 urinary bladder, 9 central nervous system, 10 gonad. Scale bar (lower right) is 1 cm

Full size image
Fig. 2
figure 2

Radioactivity measured by liquid scintillation counting in liver (a, c) sampled from the sectioned blocks and bile (b, d) sampled from native sections of polar cod (Boreogadus saida). The fish were administered an oral dose of either BaP (1.15 ± 0.36 mg/kg fish) or Phen (0.40 ± 0.12 mg/kg fish) and three fish from each group was collected 2, 6, 12 and 30 days after administration. Black and open circles represent individual females and males, respectively

Full size image

Benzo(a)pyrene

Two days after administration, BaP-derived radioactivity was most prominent in the intestines and the gall bladder of the autoradiograms (Fig. 1). The same distribution pattern was observed 6 days after administration. Twelve days after administration, BaP-derived radiolabeling was very low in all tissues and all specimens, including bile and intestine, comparatively to the other time points. The quality of the autoradiograms was verified by the inclusion of standard microscales with known 14C concentrations, and there were no indications of any technical failure during film exposure or development for any of the autoradiograms. Although not observed, part of these low levels may be attributed to regurgitation of a portion of the administered dose due to the presence of toluene in the BaP food mixture (see Table 1). Due to the uncertainty of the validity of these autoradiograms, this group is not included in Fig. 1. After 30 days, radioactivity was still observed in the bile and the intestines. The high levels of radioactivity in the content of intestine and bile on the autoradiograms made it impossible to evaluate radioactivity in the intestinal and gall bladder walls, as the films were saturated. However, sections extracted with polar and non-polar solvents revealed the presence of BaP-derived radioactivity in the walls of the gall bladder and the intestines proportionate to the levels found in the content at all time points. Radioactivity in the liver was not observable on the autoradiograms but was confirmed in all groups by liquid scintillation counting (Fig. 2a). The median values were 15.3, 3.6, 2.5, and 6.8 dpm/mg liver 2, 6, 12, and 30 days after exposure, respectively. One individual in the first group (2 days after exposure) contained higher levels in the liver (252.1 dpm/mg liver) than all of the other BaP exposed fish. Thus this value was outside the range of the figure but was included in the statistical analysis. Liquid scintillation counting showed that the bile content of radioactivity (Fig. 2b) dropped from 3.6 × 103 to 0.8 × 103 dpm/μl from 2 to 6 days after exposure (median values). Thirty days after exposure, the median value increased to 7.2 × 103 dpm/μl. Furthermore, male polar cod seemed to have lower levels of radioactivity compared with females both on autoradiograms (not shown) and after scintillation counting (Fig. 2a, b). There were no significant differences (Kruskal–Wallis, p > 0.05) among time points (males and females combined) in either liver or bile.

Phenanthrene

Two days after administration, high levels of Phen-derived radioactivity were found in the intestines and gall bladder (Fig. 1). Comparatively, radioactivity was present to a visually lower degree in the liver, gonad, kidney and the urine. Furthermore, traces of radioactivity were observed in nervous tissue. Six days after administration, the levels of radioactivity were still prominent in the intestines and the bile and present in the liver, gonad, kidney and urine. After 12 days, radiolabeling as observed on autoradiograms was decreased in the liver, kidney and urine compared to the 6 days group, whereas levels in the bile and intestines seemed similar (not shown in Fig. 1). No traces of radioactivity were observable in the gonads at this time point. After 30 days, the levels of radioactivity were still prominent in the bile and the intestines, whereas the liver, kidney and urine did not contain observable amounts of radiolabeling. Liquid scintillation counting on liver showed a significant decline in Phen-derived radioactivity over time (Fig. 2c), confirming the trends observed on the autoradiograms. Furthermore, levels were remarkably higher for Phen-derived radioactivity compared with BaP-derived radioactivity (Fig. 2a). A drop in the median value from 782 to 217.2 dpm/mg tissue was observed in groups sampled two and six days after administration. The decline continued in groups sampled 12 and 30 days after administration to 58.5 and 11.2 dpm/mg tissue, respectively. Correspondingly, the levels in the bile (Fig. 2d) increased over time until 12 days after administration, from 2.3 × 103 via 9.4 × 103 to 19.3 × 103 dpm/µl bile (median values). The amount of radiolabeled compound 30 days after administration was almost identical to 12 days after administration (18.8 × 103 dpm/µl bile). Finally, sections extracted with polar and non-polar solvents contained traces of radioactivity in the walls of the gall bladder and the intestines at all time points. Apart from a significant difference between the liver content 2 and 30 days after exposure, there were no significant differences (Kruskal–Wallis, p > 0.05) among the time points (males and females combined) in either liver or bile.

Discussion

The high level of radioactivity in the bile clearly shows that absorption of both compounds had taken place within 2 days following administration. This is in accordance with previous studies of fish exposed intragastrically to BaP (James et al. 1991; McElroy et al. 1991; Lemaire et al. 1992) and Phen (Solbakken et al. 1979; Solbakken and Palmork 1980; Solbakken et al. 1982, 1983).

Furthermore, our results suggest that a larger fraction of Phen was subjected to systemic distribution than BaP. Indeed, the tissue distribution of BaP-derived radioactivity was confined to the intestinal tract and the biliary system. Liquid scintillation counting revealed the presence of radiolabeling in the liver at all time points, although levels of radioactivity were significantly reduced compared with those of Phen. The main load of Phen-derived radioactivity was present in the bile and intestines. However, unlike BaP, radioactivity derived from Phen was also observed on autoradiograms in the liver, kidney and urine through day 12, and traces were detected in nervous tissue and gonads. Several factors may influence differently the uptake and toxicokinetics of dietary-borne BaP and Phen, including molecular structure and size, the composition of the diet and transport across intestine (reviewed by Varanasi et al. 1989 and Ramesh et al. 2004). Dietary fats, for instance, have been shown to increase the bioavailability of lipophilic PAHs such as BaP in both mammals (Laurent et al. 2001) and fish (Vetter et al. 1985). The presence of bile has also been shown to play a crucial role in the uptake of large lipophilic PAHs such as BaP, but not for Phen (Rahman et al. 1986).

Hepatic and intestinal metabolism is an important factor in the elimination of PAHs. The low levels of radioactivity detected in the livers of specimens exposed to BaP compared to Phen could indicate a more efficient hepatic metabolism of BaP than Phen. Low levels of radioactivity in the liver have also been reported in sea bass (Dicentrarchus labras) force fed BaP (Lemaire et al. 1992). A more efficient metabolism of BaP compared to Phen is also supported by Niimi and Palazzo (1986) who showed that the half-life (t 1/2) of BaP was significantly shorter (<2 days) in rainbow trout (Oncorhynchus mykiss) intragastrically fed BaP in comparison to Phen (t 1/2 = 9 days). Although Phen is metabolised by CYP450 isoenzymes, it has been shown not to induce CYP450 activity in several fish species (e.g. Goksøyr et al. 1986; Billiard et al. 2004), potentially leading to a reduced biotransformation efficiency of this compound compared to BaP. This was further supported by a lower rate of metabolism of Phen compared to BaP in liver microsomes of brown bullhead (Ameriurus nebulosus; Pangrekar et al. 1995). Although only based on a few specimens, females seemed to show higher tissue levels of BaP compared to males after both autoradiographical examination and scintillation counting. These gender differences are not yet understood but deserve further investigation. In the present study, polar cod was likely in an advanced gonadal developmental stage (Hop et al. 1995) at the time of the experiment (November), potentially showing gender differences in liver P450 content or activity and thus gender-specific bioaccumulation potential for lipophilic compounds. This is in accordance with data from channel catfish, Ictalurus punctatus, where general gender differences in P450 isoforms as well as gender differences in CYP1A induction after treatment with β-naphthoflavone have been described (Perkins and Schlenk 1998). Furthermore, there is a known cross-talk between the aryl hydrocarbon receptor and the estrogen receptor (Gräns et al. 2010), also potentially playing a role in this gender effect. Increased CYP1A induction in males would indicate a faster metabolism and thus lower retained amounts of the investigated substance. Since CYP1A is unlikely induced by Phen, also the lack of gender difference in the Phen-exposed group can be explained by this model.

Tissue-bound residues of radioactivity from both BaP and Phen were detected in the intestinal mucosa and most probably reflect site-specific CYP1A-catalysed metabolism and covalent binding of reactive intermediates. These observations are in accordance with previous studies with rainbow trout (Sandvik et al. 1998) and polar cod (Ingebrigtsen et al. 2000) where bound residues of BaP-derived radioactivity were detected in the intestinal mucosa. CYP1A activity has been reported in the intestinal tract of several fish species, localised in the epithelial cells of the mucosa as well as the vascular endothelium in the lamina propria (Stegeman et al. 1991; Smolowitz et al. 1991, 1992). In the present study, the tissue-bound BaP-derived radioactivity in the intestinal wall was far higher than the Phen-derived radioactivity. This observation strongly indicates that reactive intermediates of BaP are formed to a far greater extent compared with Phen in the intestinal mucosa of polar cod.

The present study suggested that the main excretory route for both compounds and their related metabolites was via the bile. The absence of BaP-related radioactivity in the urine and the low levels of urinary radioactivity related to Phen (through day 12) indicated that urinary excretion played a secondary role. Biliary excretion of PAHs is well documented in fish (e.g. Solbakken and Palmork 1981; James et al. 1991; Lemaire et al. 1992; Van Veld et al. 1997), and PAH metabolites in fish bile are widely used as biomarker of PAH exposure (Beyer et al. 2010). PAHs taken up via the diet will primarily be transported to the liver via the hepatic portal vein and undergo first path metabolism, where their metabolites will be secreted to the bile (Varanasi et al. 1989; Kleinow et al. 2008). This pathway was supported for polar cod in the present study as radioactivity was detected in both liver and bile after administration of both compounds. Nevertheless, some renal elimination seemed to take place for Phen but not for BaP. This reflects important toxicokinetic differences between these two compounds and should be further investigated. Finally, polar cod as well as other polar fish species possess aglomerular kidneys (Christiansen et al. 1996), an adaptation to retain small antifreeze proteins (Eastman and DeVries 1986). As earlier suggested, this adaptation strengthens the hypothesis that elimination of PAHs and their related metabolites is primarily occurring via the bile in polar cod (Christiansen et al. 1996; Ingebrigtsen et al. 2000).

The high levels of radioactivity in the bile and intestinal tract throughout the study were most probably a result of enterohepatic circulation. Furthermore, the higher levels of Phen-derived radioactivity compared to BaP-derived hepatic radioactivity indicated a higher metabolic rate of the latter. Enterohepatic circulation implies that hepatic metabolites excreted in bile are reabsorbed from the intestine and transported to the liver via portal blood. Accordingly, the retention time of the metabolites may be prolonged depending on the extent of reabsorption. Furthermore, enterohepatic circulation implies that the liver will be continuously exposed to the mother compound and its metabolites during the entire elimination period. This mechanism is regarded to be important for PAH compounds undergoing metabolic bioactivation, such as BaP, because reactive intermediates readily bind to macromolecules, such as proteins and nucleic acids which may in turn result in cellular damage. To our knowledge, the significance of enterohepatic circulation for the kinetics of organic contaminants has not been as well studied in fishes as it has been in mammals (Dutczak et al. 1991; Ramesh et al. 2004; Jandacek and Tso 2007).

In conclusion, the present study showed a more important systemic distribution of Phen compared to BaP. This finding may be of significance for an Arctic species potentially at risk of exposures to PAHs of petrogenic origin. Further work is necessary to investigate the toxicokinetics of both single and mixtures of PAHs and the factors contributing to their bioavailability and transport across the intestines in dietarily exposed fish (Erickson et al. 2008). Ongoing studies will complement these results by evaluating the comparative toxicokinetics of Phen between water and dietary exposure. A comprehensive understanding of the mechanisms controlling uptake, distribution and elimination of this complex and vast group of organic contaminants is still largely lacking for fish species. This knowledge is important for both the evaluation of the risk these pollutants may represent to wild life species and for interpretation of their toxicity, especially for Arctic species where little data are yet available.