Introduction

The European eel (Anguilla anguilla L.) has its supposed spawning area in the Sargasso Sea which is 5,000–7,500 km away from its fresh water habitats in Europe. At the onset of gonadal maturation, eels start their migration back to their spawning grounds (van den Thillart et al. 2008). The process of maturation goes along with morphological and physiological changes for the fish: the transformation from so-called yellow eel into silver eel, including an increasing eye diameter and prolonged fins. Durif et al. (2005) used these and other externally visible characteristics to create a silvering index (SI), describing the maturation stages of the eel. Another important physiological change for maturing eels is that, while they migrate, they stop feeding and become reliant on their body energy reserves, mostly muscle lipid. It is assumed that larger females can use their lipid reserves more effectively and therefore have a better chance of successful migration and spawning (Clevestam et al. 2011). This is also supported by the observation that Anguilla rostrata tend to become larger and older at higher latitudes (Jessop 2010). Age and lipid have been reported to influence proposed spawning success (Palstra and van den Thillart 2010). Belpaire et al. (2009) reported decreasing lipid contents in yellow eels caught in Belgium, which the authors regarded as a crucial element for reproductive success.

The stock of the European eel continues to decline, reached a new minimum in 2011 and is considered to be “outside safe biological limits” (ICES 2011a). One hypothesis for the cause of this dramatic situation is low spawner quality due to contaminant effects (Palstra et al. 2006). Polycyclic aromatic hydrocarbons (PAHs) belong to the group of organic contaminants, are known to accumulate in sediments and suspended particulate matter (Keiter et al. 2008, Woelz et al. 2008), and have significant impact on the habitat of yellow and silver eels. In German rivers, the environmental quality standards were more often exceeded by chemicals belonging to the PAH group, than by other organic chemicals (Federal Environment Agency 2012). This underlines the importance of PAH for environmental quality assessment. PAH are rapidly metabolized and their metabolites are detectable in the bile. Because of this fast metabolization, chemical quantification of PAH in fish tissues provides only limited information. In marine monitoring, PAH metabolites in the bile of fish have been applied as indirect indicators of PAH exposure. Numerous studies concerning PAH metabolites in different fish species have been published before (Brinkmann et al. 2010, 2013; Tairova et al. 2012; Kammann 2007) as well as in the European eel (Ruddock et al., 2003; Nagel et al. 2012a). The main metabolite in fish bile is known to be 1-hydroxypyrene (1-OHPyr) which contributes up to 76 % of the sum of PAH metabolites. Other metabolites, detected in considerably lower levels in fish bile are 1-hxdroxyphenanthrene (1-OHPhen), 1-hydroxychrysene, and three metabolites of benzo(a)pyrene (Ruddock et al. 2003). PAH metabolites are prominent parameters in marine monitoring. They have recently been proposed in a suite of 13 core indicators to monitor hazardous substances and their effects in the Baltic Sea (HELCOM, 2012) and are part of marine monitoring programs (OSPAR 1998; Kammann et al. 2012). This background lets PAH metabolites become candidates do describe PAH exposure as possible indicators for the habitat quality for eels.

PAH are metabolized by enzymes belonging to the group of cytochrome P450 monooxygenases (CYPs). Especially enzymes from the CYP1A sub-family are involved in phase I biotransformation of xenobiotics in many vertebrates. Substrates for CYP1A enzymes have to be of planar conformation: Hahn et al. (2005) showed that the expression of CYP1A mRNA is mainly regulated by binding of planar aromatic hydrocarbons to the cytosolic aryl hydrocarbon receptor (AhR). The enzymatic activity of CYP1A is typically assessed indirectly by measuring the O-deethylation of the substrate 7-ethoxyresorufin to the fluorescent product resorufin by ethoxyresorufin-O-deethylase (EROD). The second step of PAH metabolism is conducted by phase II biotransformation enzymes such as glutathione-S-transferases (GST), also controlled by the AhR. Both enzyme families, as well as their activities or corresponding mRNAs and proteins have been extensively studied in various fish species as biomarker of exposure to planar aromatic compounds, e.g., polychlorinated biphenyls (PCB), dioxins, and furans, as well as PAH (Schlenk et al. 2008) and also in European eel (Agradi et al. 2000; Bonacci et al. 2003; Buet et al. 2006; Fenet et al. 1996; Hewitt et al. 1998; Marohn et al. 2008; Pujolar et al. 2013, Teles et al. 2004; van der Oost et al. 1996). However, there are many confounding factors that influence the signal-to-noise ratio of the biomarkers (for review, see Whyte et al. 2000). Although the link between elevated enzyme activities and adverse effects in organisms is well established, induction of EROD or GST cannot be directly equated with toxicity. Therefore, it is important to compare the enzymatic activities with contaminant data.

The nutrition status of the fish is one important confounding factor for PAH metabolites in eel because cessation of feeding is a natural process for the eel, occurring during the silvering process. Starvation may mark the eels’ start of migration accompanied by regression of the digestive tract (Tesch 2003). It is known that, during periods of starvation, the amount of glucuronidated and sulphated PAH metabolites in bile increase (Beyer et al. 1997). Consequently, an increase in concentration of other bile contents such as bile pigments can be expected during starvation periods of fish (Richardson et al. 2004). In previous studies, we described the relation of PAH metabolite concentration in eel bile with maturation (Nagel et al. 2012a). We showed that this bias could be overcome when PAH metabolites were put into relation to bile pigments (Nagel et al. 2012b). Because of the fact that the concentration of PAH-metabolites and bile pigments in eel bile are influenced by cessation as part of the silvering process, bile pigments might be used to identify migratory status of eel. The condition of these fish is of special importance since healthy and well conditioned specimens are possibly favored for a successful reproduction. The eel management plans of the European Union allot that, for a successful restoration of the panmictic stock of the European eel, it has to be permitted that at least 40 % of the silver eel biomass can escape to the sea, relative to the best estimate of escapement that would have existed if no anthropogenic influences had impact on the stock (Council Regulation (EC) 1100/2007). For these reasons, it is of special importance to identify migrating eel to archive a better understanding of the mechanisms involved in silvering of eels and to determine their proportion in the local populations. Also, stocking of glass eel and elvers is a prevalent practice to support and sustain local fisheries. Nevertheless, no significant restoration of the population has been observed so far, suggesting that restoration plans are inefficient. Identification of migrating eel in the fresh water population might help to shed light on this problem.

The aim of the present study was (1) to prove PAH metabolites as well as GST and EROD activity as possible criteria for habitat quality in eels from German rivers and (2) to investigate the relation of bile pigment concentration and the migratory status of silver eels.

Materials and methods

Sampling of fish

All eels used in this study derived from commercial fishermen and were originally bought alive for stock assessment purposes within the EU Data Collection Framework, as defined by the European Commission (2008) and further specified for the presented time frame (European Commission 2010). Out of animal welfare considerations, it is of importance to the authors to mention that no additional eel had to be killed for the present study. In sum, 262 fish were caught between June and October 2011 in the German rivers Eider, Elbe, Rhine, Schlei/Trave, and Weser (Fig. 1). This time frame was intentionally chosen to lie in the expectable main feeding season of European eels in Germany (Tesch 2003).In the river Elbe, eel were sampled at five different locations from the estuary to 600 km upstream (Fig. 2). Fish were held in flow-through holding tanks for a maximum of 10 days until dissected for the collection of biological data. Eels were individually anesthesized by using diluted clove oil, weighed, measured, staged according to Durif et al. (2005), and then killed by decapitation. After decapitation, livers were excised, and the left distal lobe of each liver was directly transferred into liquid nitrogen for enzymatic analysis. Bile was directly extracted from the gallbladder by using a 1-mL disposable syringe with a hypodermic cannula, transferred to 1.5-mL cryovials, and stored at −20 °C until examination.

Fig 1
figure 1

Sampling positions of 262 female eels caught in 2011 in the German rivers Eider (1), Elbe (2, a: Cuxhaven, b: Jork, c: Winsen, d: Gorleben, e: Hohengöhren), Ems (3), Rhine (4), Schlei/Trave (5), and Weser (6)

Full size image
Fig 2
figure 2

PAH metabolites 1-hydroxypyrene (light) and 1-hydroxyphenanthrene (dark) related to bile pigments (A380) in female eel caught in 2011 in German rivers; means and 95 % confidence intervals. Asterisks denote significant differences compared with animals from Elbe (Kruskal–Wallis ANOVA on ranks, Dunn’s post hoc test p ≤ 0.05)

Full size image

Lipid and age analysis

Muscle fat content was derived using the Distill Fish Fatmeter (Model FM-692, Distell.com), with the “EEL-2” preset (whole carcass including skin, head, tail, fins, and intestines). Measurements were done according to the manufacturer’s instructions with two exceptions: Fat content was determined on the left side of the fish only using a mean of four measurements, independent of fish length. In order to determine the accuracy of this method, lipid values of 51 eel were measured with both Fatmeter and a gravimetric method (Sühring et al. 2013) as reference. The average deviation of Fatmeter results to the reference values was −1.73 (standard deviation 4.44) percentage points, indicating that the Fatmeter rather produces estimates, which are, however, sufficient for the present study.

For ageing of individual fish, otoliths were cut out and prepared using the cutting and burning method (Graynoth 1999; Richards 1989; Todd 1980). Age readings were done according to a manual established by international experts (ICES 2009, 2011b).

PAH metabolites and bile pigments

PAH metabolites in bile samples were determined by a modified version of the method described by Kammann (2007) based on Krahn et al. (1984) but using slightly divergent high-performance liquid chromatography (HPLC) conditions: A volume of 25 μl bile was mixed with 95 μl water to which 5 μl of ß-glucuronidase/arylsulfatase solution (30/60 U/ml) were added and incubated for 2 h at 37°C on a heated shaker. The addition of 125 μl ethanol containing 5 mg/ml ascorbic acid stopped the reaction. The final solution represents a tenfold dilution of the bile sample and was centrifuged (5 min, 700×g). The clear supernatant was used for HPLC analysis immediately. The concentrations of PAH metabolites were determined using a LaChrom HPLC system (Merck Hitachi) comprising a quaternary pump (L-7100), an auto sampler (L-7200), and a fluorescence detector (L-7480). Standard solutions were diluted in acetonitrile with 5 mg/ml ascorbic acid. The column was a Nucleosil 100–3 C18 (3 × 125 mm) reverse phase and run at a flow of 0.55 ml/min. The initial mobile phase was acetonitrile/0.1% trifluoroacetic acid 50/50 (v/v). After 10 min, the solvent composition progressively changed to 60 % acetonitrile over 4 min and reached afterwards 100 % acetonitrile within 2 min. The excitation/emission wavelength pairs for 1-OHPyr and 1-OHPhen were 346/384 and 256/380 nm, respectively.

PAH metabolite concentrations were related to bile volume as well as to bile pigments measured as total absorbance at 380 nm (A380). For quality assurance of PAH metabolite analysis, each sample was processed twice. The limit of detection (LD) and the limit of quantification (LQ) were calculated according to DIN 32645 (DIN 1994) with a confidence level of 99 %. Considering the dilution during sample preparation, a LD of 3.4 (0.5) and a LQ of 22.5 (1.7) ng/ml bile were determined for 1-OHPyr (1-OHPhen). A fish bile sample as laboratory reference material was included in every sample batch to monitor the stability of the method (variation coefficient 15 % for 1-OHPyr). An intercalibration exercise of the method leads to z scores below +/− 2 (Kammann et al. 2013).

For bile pigment determination, a volume of 25 μl bile was added to 475 μl water, and absorbance of 300 μl was recorded at 380 and 660 nm, respectively, using 96-well microplates and a UV/VIS microplate reader (Fluostar Optima, BMG Labtech, Offenburg, Germany). The concentration of biliverdin (nanograms per milliliter) was calculated from the absorbance at 660 nm using its molar extinction coefficient of 10,800 cm−1 mM−1 (Grossbard et al. 1987). Bile pigments are expressed as absorption units per milliliter (a.u./mL) in case of A380.

EROD and GST activity

Liver subcellular fractions were prepared according to the methods described by Bonacci et al. (2003). Briefly, pieces of liver samples were excised and added to 50 mM potassium phosphate buffer (pH 7.5) containing 0.5 mM dithiothreitol and 0.4 mM phenylmethylsulfonyl fluoride at a ratio of 1:10 (w/v) and homogenized using an electric disperser. Homogenates were centrifuged at 9,000×g and 4 °C for 20 min. The resulting supernatant (S9) was collected in fresh tubes. Samples were kept on ice during the whole procedure. Concentration of total proteins in S9 was measured following a dilution step in homogenization buffer using the bicinchoninic acid assay provided as kit (Sigma-Aldrich, Deisenhofen, Germany).

EROD activity was then measured according to Maria et al. (2005). In a quartz cuvette, 100 μl S9 were mixed with 1 ml 0.5 μM 7-ethoxyresorufin and 10 μl 10 mM reduced nicotinamide adenine dinucleotide phosphate solutions in 100 mM TRIS buffer (pH 7.4) containing 150 mM potassium chloride. Fluorescence of the reaction product resorufin was measured in 10-s intervals for 5 min in a spectrofluorometer (Jasco FP-750, Gross-Umstadt, Germany) with excitation and emission wavelengths of 530 and 585 nm, respectively. All samples were measured in duplicates. Blank measurements were performed to correct for spontaneous substrate conversion. A serial dilution reference curve for resorufin was recorded and used for interpolation of changes in product concentrations. Specific EROD activities were calculated and expressed as picomoles resorufin generated per minute reaction time and milligrams total protein.

The activity of GSTs in S9 was measured using to the method of Habig et al. (1974) according to the protocol recently published by Brinkmann et al. (2010). Briefly, 750 μl 100 mM sodium phosphate buffer (pH 6.5) was mixed with 30 μl 25 mM 1-chloro-2,4-dinitrobenzene (CDNB) solution in ethanol and dilutions of the S9 fraction in homogenization buffer in a cuvette. The reaction was initialized by addition of 75 μl 11 mM solution of reduced glutathione, and absorbance was recorded at 340 nm and 25 °C for 5-min in intervals of 5 s. All samples were measured in duplicates. Blank measurements were performed to correct for spontaneous substrate conversion, and the specific GST activity was expressed as nanomoles CDNB converted per minute reaction time and milligrams total protein.

Statistical methods and PCA

Since all datasets did not pass either the Barlett’s test for equal variances (p < 0.05) or the Kolmogorov–Smirnov test on Gaussian distribution (p < 0.05), they were analyzed by use of nonparametric Kruskal–Wallis ANOVA on ranks (p ≤ 0.001) or Spearman correlation test (p < 0.05). Dunn’s method or Mann–Whitney U test (p ≤ 0.05) was used to identify significant differences between sampling locations. Unless indicated, values are expressed as mean value 95 % confidence intervals. The principal component analysis (PCA) with Varimax rotation was performed with STATISTICA 6.0 (Stat. Soft, USA).

Calculation for the use of bile pigments as indicator for pre-migratory stage of eel

To elucidate the hypothesis that bile pigments could be an indicator for pre-migratory stage, the eel with higher bile pigment concentrations were distinguished from those individuals which were clearly not migrating and most probably still feeding. Therefore, in a first step, the 119 eels in stages 1 and 2 (yellow eels) shown in Table 1 were chosen as a group of non-migrating eels. Three individuals with SI of 2 and lipid contents below 11 % were excluded from the whole dataset because they were obviously not feeding due to a fishing hook in the intestine or due to injury. The remaining fish comprised the group of non-migrating eels with A380 < 106 a.u./mL and biliverdin < 1554 ng/mL. These two threshold values were used in a second step to separate migrating and non-migrating eels in the whole data set to check the hypothesis described above. The ranges of biological characteristics of the resulting groups are presented and compared with other studies in Table 2.

Table 1 PAH metabolites 1-hydroxypyrene (1-OHPyr) and 1-hydroxyphenanthrene (1-OHPhen) related to bile volume or bile pigments (A380) respectively; ethoxyresorufin-O-deethylase (EROD), glutathione-S-transferases (GST), age, lipid, and length
Full size table
Table 2 Biological indicators of female silver eel in pre-migratory stage from German rivers selected by either A380 or biliverdin thresholds respectively and compared with published data for migrating and mature eel
Full size table

Results

PAH metabolites in all rivers

Mean values and standard deviations of the PAH metabolites 1-OHPyr and 1-OHPhen in bile, concentration of bile pigments, and biological parameters of 262 individually analyzed eels are given in Table 1. Data are grouped for rivers and SI, respectively. The mean values of 1-OHPyr (1-OHPhen) cover a broad range from 323 to 3,806 (110 to 699) ng/mL or 2.5 to 38.8 (0.9 to 8.8) ng/A380, respectively. Highest mean values of 1-OHPyr and 1OHPhen were analyzed in fish from the river Rhine, which is the case for both: volume-related [nanograms per milliliter] and bile pigment-related concentrations [nanograms per A380]. Mean bile pigment concentration varied from 35.1 to 205.3 a.u./mL. Highest concentrations of bile pigments were found in fish with an SI of 3 or higher in all rivers. Concentrations of 1-OHPyr in fish bile differed significantly (p ≤ 0.05) between most rivers. Only samples from Schlei/Trave and Rhine showed no significant difference from each other. Even when the maturation stages are regarded separately (data in Table 1), fish from Schlei/Trave, Eider, and Rhine appear to be more highly contaminated. Regarding PAH metabolite concentration related to bile pigments in eel, regional differences are visible (Fig. 2): While eel from Rhine, Schlei/Trave, and Eider showed the highest means in 1-OHPyr, fish from Elbe and Weser tend to provide lower concentrations. The influence of maturation and the linked nutrition status are predominantly ruled out with the relation to bile pigments (A380). The second metabolite under investigation, 1-OHPhen, was found in lower concentrations than 1-OHPyr in all samples. Only in the relatively low contaminated fish from the Weser were the concentrations of the two metabolites close (Fig. 2, Table 1). The lower concentration of 1-OHPhen compared with 1-OHPyr is typical for fish bile (Ruddock et al. 2003; Kammann 2007; Kammann and Gercken 2010).

PAH metabolites EROD and GST in the River Elbe

Mean values and standard deviations of the activities of EROD and GST and biological parameters of individually analyzed eel (n = 232) are given in Table 1. The maximum mean activity of EROD (GST) of 7.3 pmol/mg*min (236 nmol/mg*min) were determined in samples from the River Elbe. Both PAH metabolites relative to bile pigments, as well as EROD activity in eels sampled at different sites along the lower Elbe River showed a similar spatial pattern (Fig. 3). The lowest values for both biomarkers from the Elbe catchment were found in eels from Cuxhaven, located directly at the estuary. Compared with this location, EROD activities were significantly higher in eels from Gorleben, Winsen, and Jork. Animals from Jork showed both the highest EROD activities (5.4-fold higher compared with Cuxhaven) and biliary PAH metabolite concentrations (2.8-fold higher, Kruskal-Wallis ANOVA on ranks, Dunn’s post hoc test, p ≤ 0.05). A significant correlation between EROD activity and SI could be detected regarding eel from all rivers (n = 232, r = −0.135, p = 0.039, Spearman rank correlation) but was not present in the data subset from the river Elbe (n = 153, r = −0.085, p = 0.293).

Fig. 3
figure 3

EROD (ethoxyresorufin-O-deethylase) activity (white bars) and concentration of 1-hydroxypyrene relative to A380 (black bars) at different sampling locations along the river Elbe given as means and their 95 % confidence intervals (n > = 13). Asterisks denote significant differences compared with animals from Cuxhaven (Kruskal–Wallis ANOVA on ranks, Dunn’s post hoc test p ≤ 0.05)

Full size image

PCA

The PCA in Fig. 4a explains 54.7 % of the variance with the first two factors. Factor 1 explains 31.1 % of the variance and refers mainly to PAH metabolites 1-OHPyr and 1-OHPhen related to bile pigments (factor loadings (FL) < = −0.93). Factor 2 explains 23.6 % of the total variance and is dominated by SI (FL-0.71) and the inversely related variables GST and lipid (FL = 0.60 and −0.56) as well as EROD (FL = 0.44). In Fig. 4b, the first two factors explain 52.9 % of the variance. Factor 1 represents 32.5 % of the variance and is dominated by the two PAH metabolites (FL < = 0.78) and by EROD (FL = 0.56). Factor 2 stands for 20.3 % of the variance and explains mainly the variables SI and lipid (FL < = 0.61). GST, however, shows weaker relations to the first two factors. An overview on all factor loadings is given in Table 3.

Fig 4
figure 4

Principal component analysis of female European eel from German rivers considering fish from all rivers (a) or from the river Elbe exclusively (b): PAH = PAH metabolites 1-hydroxypyrene and 1-hydroxyphenanthrene [nanograms per A380]; SI = Silvering Index according to Durif et al. (2005); Lipid [%], EROD = ethoxyresorufin-O-deethylase activity [pmol/(mg protein*min)]; GST = glutathione-S-transferase activity [nmol/(mg protein*min)]

Full size image
Table 3 Factor loadings for the first two factors (F1, F2 with variance levels) of a principal component analysis of female European eel from German rivers considering fish from all rivers or from the river Elbe exclusively: PAH metabolites = 1-hydroxypyrene and 1-hydroxyphenanthrene [nanograms per A380]
Full size table

Discussion

Spatial differences

The concentrations of PAH metabolites in fish bile exhibited spatial differences between the rivers: Eel from Schlei/Trave, Eider, and Rhine exhibited the highest contamination with PAH metabolites (Fig. 2).These findings are in accordance with Nagel et al. (2012a) who detected elevated concentrations of PAH metabolites in eel caught in the river Trave and described eels from the Elbe as “medium contaminated”. However, Nagel et al. (2012b) did not relate their results to bile pigments. Investigations on PAH metabolites in European eel caught in UK estuaries revealed 1-OHPyr concentrations in bile ranging from 120 to 7,600 ng/ml (Ruddock et al. 2003), which is twice as high as the mean values presented in Table 1 but quite close of the range of individual results (97–6,609 ng/ml). Sühring et al. (2013) analyzed up to three times higher concentrations of brominated flame retardants in eel from the Rhine compared with eel from the Elbe. This is in accordance with the findings of the present study: a twofold higher value of 1-OHPyr in eels from the Rhine compared with the Elbe (Fig. 2). It is also in accordance with the poorer chemical status of the Rhine compared with the river Elbe (Federal Environment Ministry 2010). However, the high 1-OHPyr levels in Schlei/Trave and Eider, as found in the present study, are not reflected in the assessments cited above. The results suggest that PAH metabolites can contribute to an assessment of habitat quality for eel in German rivers. On the other hand, it has to be mentioned that EROD activity in eel is at least partly influenced by the maturation because of the significant correlation between EROD and SI regarding fish from all rivers which is in accordance with Whyte et al. (2000). However, in a subset of the data, comprising fish from the river Elbe, no such correlation could be detected. It can thus be assumed that EROD activity in fish from the river Elbe was not strongly influenced by sexual maturation, and the observed spatial differences can be related to a possible influence of pollutants. Therefore, this data subset has been chosen for a closer look on the spatial EROD pattern: PAH metabolites relative to bile pigments and EROD activity in eels sampled at different sites along the lower Elbe River in shown in Fig. 3. The sampling site Jork at the Elbe River with tidal influence is situated in only approximately 20 km distance downstream from the harbor of the city Hamburg. The sampling site Winsen is located approximately 20 km upstream of Hamburg. An influence of migrating eels from this heavily polluted industrial area on PAH metabolites and EROD activities is very likely. Many studies have demonstrated that PAH and PCB concentrations decreased in suspended particulate matter with increasing stream kilometer, with the highest concentrations close to the Czech border (BFG 2008; Heise et al. 2005). Dioxins and furans in freshly deposited sediments also show such spatial trend, although the highest concentrations are typically found close to the tributary Mulde. In the vicinity of the sampling site Gorleben, elevated dioxin/furan concentrations of about 50 pg WHO-TEQ/g dw sediment (compared with 20 pg WHO-TEQ/g dw at Cuxhaven) have been measured (Stachel et al. 2011), being one potential explanation for the elevated EROD activities. The results suggest that PAH among other contaminants may contribute to the enhanced EROD activities in eels from the river Elbe. However, the high EROD activities in eel from the river Elbe compared with the other rivers under investigation are not reflected in concentrations of PAH metabolites (Table 1). This fact suggests that PAHs are not likely to be always the major cause for EROD activities in eel.

PAH metabolites, EROD, and GST linked to eel physiology

Besides contaminants, the physiology of the fish may influence enzyme activities. Two PCAs were performed to elucidate the main linkages between six selected variables from Table 1: Two PAH metabolites (contaminants), silvering index and lipid content (physiology) as well as the enzyme activities EROD and GST. These variables were chosen to obtain a balanced PCA approach with equal numbers of variables from the three groups: contaminants, physiology, and enzymatic effects. The known linkage of GST and lipid-related processes in fish (Leaver and George 1998) is reflected in Fig. 4a. It has been shown in laboratory experiments before that EROD or CYP1A levels respectively are influenced by the hormonal status and sexual maturity (here represented by SI) of the fish (Whyte et al. 2000), which is in accordance with Fig. 4a. There is no evidence from Fig. 4a that EROD is influenced by the exposure of the eel to PAHs because of the fact that EROD and PAH metabolites are related to two different factors (nearly 90° angle between the variables). Therefore, we doubt that EROD and GST generally reflect effects related to PAH exposure in the samples. However, in fish from the river Elbe, PAH metabolites and EROD showed very similar spatial patterns (Fig. 3). In addition, the PCA of eel from the river Elbe (Fig. 4b) shows a close relation of EROD and the two PAH metabolites. Thus, we cannot exclude that EROD reflects pollution with organic contaminants at least in some samples. On the other hand, GST activities did not differ significantly between the rivers or the investigated sampling sites (results not shown). Van der Oost et al. (2003) reported that GST activity in fish does not respond always to xenobiotics: In only one third of laboratory studies, a significant GST activity increase was reported after exposure to organic pollutants.

Both enzyme activities, GST and EROD, appear to be strongly linked to physiological status of the eel (SI and lipid, respectively). But, for EROD, there is evidence that in some cases a pollution effect, e.g., of PAH or other AhR binding contaminants, might be reflected by this enzyme activity in eel. When EROD and GST are used as biological effect markers, eel should represent one maturation stage only (e.g., yellow eel) to avoid interference with physiology as described above.

Bile pigments and pre-migratory stage of eel

As mentioned before, concentrations of PAH metabolites and bile pigments increase when fish stop feeding. For bile pigments, this is a well known fact for decades (McCormick and Podoliak 1984). Consequently, bile pigment concentration may indicate the duration of starvation process in eel, which coincides with maturation. We hypothesize that the concentration of bile pigments measured as A380 (a.u. per milliliter) or biliverdin (nanograms per milliliter) in bile indicates starvation and may therefore act as an additional indicator for the migratory status of silver eels. Figure 5 shows that high A380 values coincide in most cases with a small range in lipid content and with the SI stages 3, 4, and 5, which are typical for migrating eel. However, some fish with lower SI also exhibit higher concentrations of bile pigments, indicating that maturation is not the only cause for cessation of feeding. On the other hand, not all eels staged as SI 3, 4, or 5 are starving (Fig. 5). This observation indicates that not all mature eel (according to SI) in German rivers have stopped feeding or at least not doing so at once. To elucidate the hypothesis described above that bile pigments could be an indicator for pre-migratory stage, potential migratory eel were selected by higher bile pigment concentrations (for details compare the “Materials and methods” section) and compared with migratory eel selected by other criteria in Table 2. Clevestam et al. (2011) investigated a larger number of female silver eel (n = 387) caught during autumn in Danish waters near the Öresund Strait. Those eels are considered to be migrating because they were caught while leaving the Baltic Sea towards their spawning regions in the Sargasso Sea. Pankhurst (1982) used the eye diameter related to the body length to distinguish non-migration yellow eel from migrating silver eel. Durif et al. (2005) used external characteristic of eel including length, weight, eye diameter, and fin length to stage female eel from 1 to 5 with increasing maturity. Most biological values from the present study, especially age, SI, and length show good accordance to the cited studies (Table 2). However, the lipid contents of migrating eel are lower in the present study than described by Clevestam et al. (2011). Although eels with higher lipid content up to 37.6 % have been detected in the present study (Fig. 5), they were apparently not starving. The large number and different origins of eels in the study of Clevestam et al. (2011) are causes for the broader-ranging biological parameters compared with the present study.

Fig 5
figure 5

A380 (absorption at 380 nm) related to lipid content and maturation expressed as SI (Silvering Index, Durif et al. 2005). Dashed line: threshold of potential migrating eel characterized by A380 >106 a.u./ml. Asterisks: fish starving due to fishing hooks or injury

Full size image

This is a first attempt to show that the cessation of feeding indirectly measured by bile pigments (either A380 or biliverdin) could be an additional indicator to enlighten the physiological processes during eel maturation in fresh water habitats. However, the threshold values might differ in other regions because they are influenced by the source or the overall level of nutrition. Eel identified as migratory due to analysis of bile pigments have been caught in all investigated rivers at different dates and by different fishermen together with non-migratory yellow eel. Therefore, it is not likely that artificial starvation during sampling has influenced the results. However, the proof of this assumption is still missing, and we can therefore not exclude a possible bias in the data caused by the sampling process. A380 or biliverdin could be new indicators, preferably applied in combination with other parameters, like those listed in Table 2, to identify migrating silver eel and to give additional insight in maturation. This hypothesis has to be confirmed in the future by controlled laboratory experiments and by enhancing the number of data and comparing the findings presented in this study to fish from other regions.

Conclusion

We conclude that the PAH metabolites 1-OHPyr and 1-OHPhen in the bile as well as EROD activity in the liver of European eel can be used to describe the habitat quality in German rivers. Bile pigments can be new indicators contributing to identify the proportion of migratory eel, which is crucial for a fresh water habitat in the light of the European eel management. In search for suitable stocking areas for glass eels, PAH metabolites and EROD can provide valuable information, even if the general benefit of stocking can be discussed (Marohn et al. 2013). Healthy and well conditioned silver eel growing in suitable habitats have the best prepositions for successful reproduction, which is in accordance with the goals of the eel management plans of the European Union.