A practicable laboratory flow-through exposure system for assessing the health effects of effluents in fish

Aquatic Toxicology 88 (2008) 164–172

Contents lists available at ScienceDirect

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

A practicable laboratory flow-through exposure system for assessing the health effects of effluents in fish Karen L. Thorpe a,∗ , Rachel Benstead c , Paul Eccles c , Gerd Maack a , Tim Williams b , Charles R. Tyler a a

School of Biosciences, Hatherly Laboratories, University of Exeter, Prince of Wales Road, Exeter, Devon EX4 4PS, UK Brixham Environmental Laboratory, AstraZeneca UK Limited, Freshwater Quarry, Brixham, Devon TQ5 8BA, UK c The Environment Agency, National Centre for Ecotoxicology and Hazardous Substances, UK b

a r t i c l e

i n f o

Article history: Received 20 February 2008 Received in revised form 8 April 2008 Accepted 10 April 2008 Keywords: Estrogen Wastewater treatment effluents Fathead minnow Exposure systems

a b s t r a c t The knowledge that exposure to estrogenic wastewater treatment work (WwTW) effluents induces a range of reproductive abnormalities in fish has highlighted the need to understand the wider health effects of effluents. Access to laboratory-based testing systems for WwTW effluents could greatly facilitate this endeavour. In this investigation, a laboratory-based test system was developed and applied for WwTW effluents using fathead minnows (Pimephales promelas). Sexually maturing fathead minnows were exposed, under flow-through conditions in the laboratory, for up to 21 days to graded concentrations of effluent from three different UK (temperate) WwTWs. The stability of the estrogenic component within the test system was assessed via measurements for estradiol and estrone concentrations in the effluent, and through determining estrogenic responses in an in vitro recombinant yeast estrogen screen (rYES) and in fish (plasma vitellogenin induction). The estrogen component of the effluents was stable within the holding system used (chilled <10 ◦ C) for up to 7 days and measured concentrations of estradiol and estrone were shown to differ by less than 20% between the first and final day of use for each batch of effluent. Total estrogenic activity as measured in the rYES was found to be more variable (up to 66% variance between measurements for the two time points) but there was no consistent trend for a reduction in estrogenic activity. Vitellogenin was induced in males in a concentration-dependent manner and the magnitude of the response observed was proportional to the average measured concentrations of estradiol and estrone in the exposure effluent. The system described, thus, provides a robust test method for evaluating the estrogenic effects of temperate WwTW effluents that could be further applied to assess wider health effects, including population-relevant endpoints such as reproduction, using model OECD warm-water fish species such as the fathead minnow. © 2008 Elsevier B.V. All rights reserved.

1. Introduction In fish living in rivers receiving high discharges of effluent from wastewater treatment works (WwTWs), a range of alterations related to reproductive development have been observed (Tyler et al., 1998; Jobling et al., 1998; Allen et al., 1999; Folmar et al., 2001; Vethaak et al., 2002). These effects have been attributed to estrogenic chemicals known to be present within WwTW effluents. Indeed, extensive laboratory-based studies have confirmed that estrogenic chemicals contained in WwTW effluents can induce many of the effects seen in effluent exposed fish (Tyler et al., 1998; Desbrow et al., 1998; Metcalfe et al., 2001; Seki et al., 2002; Nash et al., 2004; Parrott and Blunt, 2005; Jobling et al., 2006). Further-

∗ Corresponding author at: University of Basel, Programme MGU, Vesalgasse 1, CH-4051 Basel, Switzerland. Tel.: +41 61 267 0414; fax: +41 61 267 0409. E-mail address: [email protected] (K.L. Thorpe). 0166-445X/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.aquatox.2008.04.005

more, it has been shown that exposure to estrogenic chemicals can inhibit the reproduction of fish (Seki et al., 2002; Nash et al., 2004; Parrott and Blunt, 2005; Harries et al., 2000; Kang et al., 2003; Thorpe et al., 2007). Although, the concentrations of estrogenic chemicals typically measured in WwTW effluents are low, compared to those required to affect fish reproduction in short-term laboratory studies, there are still major concerns about long-term exposures to estrogenic effluents. This is because a prolonged exposure to some of these chemicals increases their level of effect and in one case this was shown to have population-level consequences (Kidd et al., 2007). Additionally, WwTW effluents contain mixtures of estrogenic chemicals and it is now well established that estrogenic chemicals can interact in an additive manner to induce effects on the reproductive physiology of fish at lower concentrations than those required individually (Thorpe et al., 2001, 2003; Brian et al., 2007). Specific studies that directly assess the effects of estrogenic effluents on population-relevant endpoints, such as egg production, are therefore, required to understand the consequence of expo-

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

sure to estrogenically active WwTW effluents for fish populations. Such studies would also enable consideration of the possible contributory effects on reproductive health of other chemicals present within the complex matrix of WwTW effluents that could have the potential to target other biological pathways in the reproductive cascade. Such studies on population-relevant endpoints, however, have not yet been forthcoming. In this study we aimed to develop a practicable and easily transferable experimental system capable of testing for the estrogenic effects of effluents in fish in a controlled laboratory setting. Tests to evaluate effects of chemicals and their mixtures on egg production have been largely restricted to model species with a repetitive spawning strategy, a short life cycle and of a size suited to allow for high replication within a laboratory setting (Ankley and Johnson, 2004). The fathead minnow has been used extensively for chemical testing and considerable datasets are now available that clearly demonstrate the effect of estrogens and other chemicals on reproductive function in this species (Harries et al., 2000; Ankley and Johnson, 2004; Parrott and Blunt, 2005; Thorpe et al., 2007). For these reasons, together with the comparability of this species with other cyprinids inhabiting effluent-contaminated waters in the northern hemisphere (many indigenous European freshwater fish are close family members) the fathead minnow was the species of choice for the present study. Final effluents from three WwTWs were included within the study. Effluent from the first WwTW (Effluent I) had been previously shown to induce high levels of vitellogenin (VTG) in male fish and to feminise the male reproductive duct for exposures during early life (Rodgers-Gray et al., 2001) and was designated as a ‘potent’ estrogenic effluent. Effluents from the second and third WwTWs were selected to represent ‘intermediate’ (Effluent II) and ‘low’ (Effluent III) estrogenic potencies with estrogenic activities of between 2.2 and 3.0 ng estradiol equivalents (E2EQ) per litre, and between 0.9 and 1.2 ng E2EQ/L, respectively, based on the results of a UK-wide survey of WwTW effluents (Thorpe et al., 2006). A flow-through exposure system was developed in which effluent, obtained from the WwTW at regular intervals and stored under chilled conditions, was continuously delivered to the exposure system (exposure temperature of 25 ◦ C) to determine effects on the induction of VTG, gonad growth and the appearance of secondary sex characters in male and female fathead minnow. The stability of the estrogenic component of the effluent both within the storage and the dosing system was evaluated using a combination of analytical chemistry, to determine concentrations of the natural steroidal estrogens E2 and E1, and the recombinant yeast estrogen screen (rYES), to measure total estrogenic activity. The results from these investigations are used to demonstrate that the estrogenic potency of final effluents from temperate WwTWs can be reliably assessed using a laboratory-based exposure system employing warm-water fish species (e.g. fathead minnow).

2. Materials and methods 2.1. Test organisms Fathead minnow used in each experiment were bred from fish stocks originally held at Brixham Environmental Laboratory. Two weeks prior to the onset of each experiment, sexually maturing male and female fish (defined by the onset of the development of the secondary sex characters) were selected and separated into single sex tanks to prevent spawning activity. Body weights measured in representative subsets of 16 males and 16 females at the onset of each exposure were 1.52 ± 0.08 and 0.95 ± 0.05 g, respectively, in experiment I, 1.71 ± 0.15 and 1.31 ± 0.08 g, respectively,

165

in experiment II and 1.61 ± 0.07 and 0.82 ± 0.05 g, respectively, in experiment III. Throughout all acclimations and exposures fish were maintained under flow-through conditions in de-chlorinated water at 25 ± 1 ◦ C, with a 16 h light:8 h dark photoperiod. The fish were fed adult Artemia sp. twice daily and Ecostart 17, 1.0 mm fish food pellets (Biomar ltd., Brande, Denmark) once daily. 2.2. Water supply and test apparatus The supply of water to the laboratory dosing system was prepared using reverse-osmosis with the addition of salts as described in OECD Guideline 203. The conductivity of the test water ranged from 200 to 255 ␮S/cm. In all experiments, the tanks were gently aerated at the surface, using a glass pipette, to maintain dissolved oxygen concentrations at >80% of the air saturation value. Water temperatures were monitored daily and ranged between 24.1 and 25.5 ◦ C in all experiments, while pH levels were checked twice weekly and ranged between 6.6 and 7.8. Dilution water and test chemical flow-rates were checked twice weekly. Flow-rates (20 mL/min) to the individual aquaria (working volume of 20 L) provided a 75% replacement time of 24 h. 2.3. Wastewater treatment works effluent For experiment I, five batches of effluent (400 L per sampling occasion) were collected over 14 days during February 2005, using 10 L plastic carboys. For experiments II and III the effluent was collected at weekly intervals in batches of 2000 L during July and August 2005, respectively, using an industrial stainless steel tanker. Each batch of effluent was collected from the final effluent stream at the respective WwTW between the hours of 8 a.m. and 10 a.m. and immediately transported to the testing facility. For the first experiment, the effluent was stored outside at ambient environmental temperatures (February; <10 ◦ C) and transferred in batches to the testing laboratory, where it was acclimated to a temperature of 18 ◦ C, and then dosed to the test aquaria via glass mixing tanks in order to be diluted to the required concentration. For experiments II and III, the effluent was transferred into a fully enclosed stainless steel holding tank chilled to between 8 and 10 ◦ C. The effluent was pumped via a peristaltic pump from the storage tank to the test aquaria via the glass mixing tanks (Fig. 1). The low dosing rates used in each experiment allowed the effluent to slowly acclimate to the desired test temperature of 25 ◦ C before reaching the test aquaria. Conductivity, dissolved oxygen concentration, temperature and pH were checked for each batch of effluent on arrival. Conductivity ranged from 1155 to 1420 ␮S/cm for effluent I, from 650 to 882 ␮S/cm for effluent II, and from 1148 to 1258 ␮S/cm for effluent III. The pH values were comparable for all effluents tested and ranged from 7.6 to 8.4 and dissolved oxygen concentrations were above 80%. Temperatures of the individual batches of effluent collected ranged from 15.5 to 16.9, 18.4 to 24.6 and 20.7 to 22.0 ◦ C for effluents I, II and III, respectively, on arrival at the laboratory. In all experiments, the effluent storage system was fully drained and flushed with water immediately prior to each delivery to prevent a build-up of algae/bacteria and to remove any particulates that had precipitated from the effluent. 2.4. Estrogen positive controls For the estrogen positive controls, estradiol-17␤ (E2; 98% purity, Lot 103K1117, experiment I) and ethinylestradiol-17␣ (EE2; 98% purity, Lot 024K1196, experiments II and III) were purchased from Sigma, Poole, Dorset, UK. Solvent free stock solutions were prepared every three days by adding 1 mL of a concentrated stock solution of the steroidal estrogen, E2 or EE2 (prepared in HPLC grade acetone;

166

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

Fig. 1. Diagrammatic representation of the exposure system used for delivery of the effluent from the storage tank to the exposure tanks via the glass mixing vessels. A series of peristaltic pumps were used to deliver the effluent, dilution water and estrogen positive control to the respective mixing vessels and then from each mixing vessel to the replicate exposure tanks for each graded effluent treatment (25, 50 and 100% effluent) and the dilution water control (DWC) and positive control (PC). The treatments were randomly distributed within the exposure room, but are shown here ordered for ease of presentation.

Fisher Scientific), to a 10 L glass vessel. After evaporation of the acetone, at room temperature, 10 L of dilution water was added to the glass vessel and the solution stirred for approximately 2 h. The solvent free stock was then delivered to the glass mixing vessels via a peristaltic pump, to mix with the dilution water. 2.5. Experimental design Fathead minnow (8 males and 8 females in each of two replicate 20 L glass tanks for each treatment) were exposed for 14 days (experiment I) or 21 days (experiments II and III) to a dilution water control (DWC), estrogen positive control and graded effluent concentrations of 25, 50 and 100% (n = 32 fish/treatment). All experimental fish were sampled at the end of the exposure period.

efficiency of the extraction for each procedure spiked (0.05% of an estrogenic mixture containing 4 ␮g E2/L, 1.6 ␮g EE2/L, 8 ␮g E1/L and 800 ␮g NP/L prepared in methanol) dilution water and effluent samples were extracted under the same conditions. Dilution water spiked with 0.05% methanol was also extracted under the same conditions. The columns were eluted using methanol (5 mL/column) and the eluant stored at −20 ◦ C for subsequent measurement of estradiol and estrone concentrations (via GC–MS) and estrogenic activity (in the rYES). A full description of the procedures involved in processing the samples and performing the measurements is provided in Thorpe et al. (2006). Recoveries of total estrogenic activity in the rYES, and measured concentrations of E1 and E2 in the spiked effluent samples were 86, 89 and 93%, respectively, for effluent I, 97, 92 and 92%, respectively, for effluent II, and 81, 122 and 75%, respectively, for effluent III.

2.6. Measurement of estrogenic activity and concentrations of E2 and E1

2.7. Fish sampling

Composite water samples were collected from the replicate tanks for each treatment into solvent-cleaned flasks. For measurement of estrogenic activity in the recombinant yeast screen (rYES) a total of 700 mL was collected from each treatment on days 2, 4, 7, 10 and 14 (experiments I, II and III) and additionally on days 17 and 21 in experiments II and III. For the analytical determinants, a total of 2.5 L was collected from each treatment on days 4, 7, 14 (experiments I, II and III) and additionally on day 21 in experiments II and III. Samples were also collected on a daily basis from the effluent storage tank for analysis in the rYES and on the morning following delivery of each new batch of effluent for the analytical determinants. Immediately after collection, the samples were spiked with 0.05% methanol and extracted onto preconditioned Sep-Pak Classic C18 Cartridges (Waters Ltd, Hertfordshire, UK). To assess the

Fish were sacrificed in a lethal dose (500 mg/L) of MS222 (3-aminobenzoic acid ethyl ester, methanesulfonate salt; Sigma), buffered with 1 M NaOH to pH 7.4. Total lengths and wet body weight of the fish were recorded to the nearest 1 mm and 0.01 g, respectively, and the condition factor derived by expressing the cube of the total fish length as a percentage of the wet body weight. Blood was collected by cardiac puncture, using a heparinised syringe (1000 Units heparin/mL) then centrifuged (7000 × g; 5 min, 15 ◦ C) and the plasma removed and stored at −20 ◦ C until required for analysis of VTG using a carp ELISA (Tyler et al., 1999: limit of detection 5 ng/L). The gonads were removed, wet weighed to the nearest 0.01 mg and the GSI derived by expressing the gonad weight as a percentage of the total body weight. The numbers of tubercles on the snout of each fish were recorded and the dorsal fatpad

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

167

To investigate effects of the effluent exposure or positive controls, data were compared to the DWC using SPSS version 13.0. Data meeting the assumptions of normality and homogeneity of variance were analysed using one-way analysis of variance (ANOVA) followed by a comparison to the DWC using Dunnett. The VTG data were log 10 transformed prior to analysis to meet the assumptions of both normality and homogeneity of variance. For the rYES, estradiol equivalent concentrations (E2EQ/L) were calculated by determining the median effective concentration of E2 and then extrapolating across to determine the dilution of effluent concentrate required to produce the same level of effect. For less potent effluents, the concentration of E2 required to produce the same level of effect as that observed for the highest concentration of effluent (100% of a 700-fold concentrate) tested was determined.

activity were used to characterise the estrogenic content of the effluent. Measured concentrations of E1 and E2 were highly variable between the batches of effluent collected from each WwTW, but were consistently higher in effluent samples collected from the first WwTW (Effluent I). Concentrations of E1 ranged between 46.1 and 169.0 ng/L (mean 95.8 ± 26.0 ng/L) and for E2 between 4.3 and 12.1 ng/L (mean 7.1 ± 1.7 ng/L) in the batches of effluent I delivered. Measured concentrations of E1 and E2 in the batches of effluent II delivered ranged from 2.9 to 9.3 ng/L (mean 5.6 ± 1.9 ng/L) and from 0.3 to 3.3 ng/L (mean 1.5 ± 0.9 ng/L), respectively. Concentrations of E1 and E2 measured in the batches of effluent III were comparable to those detected in effluent II, ranging between 1.6 and 7.3 ng/L (mean 3.5 ± 1.9 ng/L) and between 0.6 and 2.3 ng/L (mean 1.2 ± 0.6 ng/L), respectively. The measurements of estrogenic activity conducted using the rYES also varied for the individual batches of effluent delivered; total estrogenic activity ranged from 18.8 to 54.8 ng E2EQ/L (mean 30.4 ± 6.3 ng E2EQ/L; n = 5) for effluent I, from 4.3 to 10.7 ng E2EQ/L (mean 7.4 ± 1.9 ng E2EQ/L; n = 3) for effluent II, and from 11.8 to 30.0 ng E2EQ/L (mean 22.8 ± 5.6 ng E2EQ/L; n = 3) for effluent III.

3. Results

3.2. Estrogenic content within the test system

3.1. Estrogenic content of the effluents

Measured concentrations of E1 and E2 in the effluent storage tank on the final day of use (day 3 for effluent I and day 7 for effluent II) were comparable to concentrations measured in the delivered effluent (day 1; Fig. 2). The total estrogenic activity determined in the rYES was more variable over the time periods given, but there

removed and wet weighed to the nearest 0.01 mg (a full description of the procedures involved in removing the dorsal fatpad is provided in Thorpe and Tyler, 2006). 2.8. Statistical analyses

A combination of analytical measurements for E1 and E2 (the two principal estrogens prevalent in WwTW effluents) and in vitro analysis (using the rYES) for measurement of total estrogenic

Fig. 2. Daily assessments of estrogenic activity (determined in the rYES, left) and measured concentrations of E1 (middle) and E2 (right) in the individual batches of effluent collected from WwTWs I (A), II (B) and III (C).

168

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

Fig. 3. Plasma vitellogenin (VTG) concentrations in male (A, C, E) and female (B, D, F) fathead minnow exposed to graded concentrations (0, 25, 50 and 100% effluent) of effluents I (A, B), II (C, D) and III (E, F) and an estrogen positive control. Each column represents the mean ± standard error of the mean (S.E.M.). Significant differences between the control and exposure groups are denoted as * P < 0.05.

were no clear trends for an increase or decrease in total estrogenic activity with time (Fig. 2). Concentrations of E1 and E2, and estrogenic activity (as E2EQs) measured in the undiluted effluent tanks were 23.8 ± 2.8, 2.9 ± 0.9 and 21.2 ± 2.9 ng/L, respectively, for effluent I, 23.6 ± 6.1, 4.0 ± 1.3 and 20.2 ± 4.3 ng/L, respectively, for effluent II and 12.5 ± 6.8, 3.9 ± 1.7 and 34.1 ± 7.1 ng/L, respectively, for effluent III. Measurable concentrations of E1 and E2, and total estrogenic activity were also detected in the control tanks (receiving dilution water only); 2.7 ± 0.9, 0.5 ± 0.2 and 2.5 ± 1.0 ng/L, respectively, in experiment I; 8.4 ± 1.5, 1.3 ± 0.3 and 17.6 ± 4.9 ng/L, respectively, in experiment II; and 3.8 ± 1.4, 0.8 ± 0.1 and 8.7 ± 3.9 ng/L, respectively, in experiment III. During each experiment, samples of the dilution water were also collected on three occasions, at the point of entry to the dosing system, to determine background concentrations of E1 and E2, and total estrogenic activity (as E2EQs) in the water; mean measured values were 3.9 ± 3.7, 0.2 ± 0.1 and 4.9 ± 2.4 ng/L, respectively, in experiment I; 1.2 ± 0.6, 0.4 ± 0.1 and 7.1 ± 1.4 ng/L, respectively, in experiment II; and 0.4 ± 0.2, 0.1 ± 0.1 and 7.1 ng/L, respectively, in experiment III (due to technical problems only one sample was measured in the rYES in experiment III). The mean measured concentrations of E2 in the positive control tanks in experiment I was 13.8 ± 3.6 ng/L (26.9 ± 8.4 ng E2EQs/L) and the concentration of EE2 in the positive control tanks in experiment II was 18.0 ± 2.0 ng/L (65.1 ± 3.1 ng E2EQs/L). Concentrations

of EE2 were not measured in experiment III, however, the higher E2EQ values measured with the rYES (69.2 ± 8.3 ng E2EQs/L) and the higher concentrations of VTG measured in both the male and female fish would appear to imply that concentrations of EE2 were higher than those measured in experiment II. 3.3. Biological effects of effluent exposure There was no evidence that exposure for up to 21 days to the WwTW effluents or the estrogen positive control affected survival, growth or condition of the male or female fish, or the appearance of the male secondary sex characters. Concentrated-related increases in plasma VTG were observed in males exposed to each effluent (P < 0.05), and effluent I showed the highest estrogenic potency, inducing a significant increase in male VTG concentrations, relative to the controls, at all dilutions tested (25, 50 and 100% effluent; P < 0.01; Fig. 3). Exposure to effluent I at full-strength (100%) resulted in a 5235-fold increase in plasma VTG concentrations in males, relative to the controls. For effluents II and III, increases in plasma VTG concentrations were only significant in males exposed to the 50 and 100% effluent (P < 0.05; Fig. 3) and the magnitudes of response were much smaller than for effluent I (15-fold increase for effluent II and a 215-fold increase for effluent III). Plasma VTG concentrations were also elevated in females, relative to the controls, exposed to 100% effluent I

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

169

Fig. 4. Gonadosomatic index (GSI) in male (A, C, E) and female (B, D, F) fathead minnow exposed to graded concentrations (0, 25, 50 and 100% effluent) of effluents I (A, B), II (C, D) and III (E, F) and an estrogen positive control. Each column represents the mean ± standard error of the mean (S.E.M.). Significant differences between the control and exposure groups are denoted as * P < 0.05.

(P < 0.01; Fig. 3) and 50 and 100% effluent II (P < 0.01; Fig. 3), but there was no evidence for an effect of effluent III (Fig. 3). Plasma VTG concentrations were significantly increased in the estrogen exposed males and females in all experiments (P < 0.05; Fig. 3). Increases in GSI, relative to the controls, were observed in males exposed to a 50% dilution of effluent I (P < 0.05; Fig. 4), but there was no evidence for an effect of the undiluted effluent (100%; P > 0.05; Fig. 4). Exposure to the full-strength effluent collected from WwTWs II and III was also observed to enhance testis size (increase male GSI; P < 0.05; Fig. 4). Exposure to E2 did not affect male GSI in experiment I (Fig. 4), but in experiments II and III, apparent decreases in GSI were observed in males exposed to EE2 (although this was only significant in experiment III, P < 0.05; Fig. 4). Significant reductions in the GSI were observed in all estrogen exposed females (P < 0.05; Fig. 4), but there was no evidence for an effect of any of the WwTW effluents on female GSI (Fig. 4).

4. Discussion This study describes the development and application of a flowthrough exposure system for assessing the estrogenic potency of UK (temperate) WwTW effluents in the laboratory under standardised conditions with the fathead minnow (a model OECD fish test

species). The stability of the test system for estrogens indicates it might equally be applied to evaluate the wider health effects of temperate WwTW effluents using this fish species.

4.1. Estrogenic content of the test WwTW effluents Measurement of the natural steroidal estrogens, E1 and E2, and in vitro assessment of total estrogenic activity, confirmed the presence of estrogens in each of the test effluents. The individual steroidal estrogens were detected in all batches of the effluent delivered, but their concentrations varied more than 3-fold between the batches collected from each WwTW. This supports other reports of temporal variations in the estrogenic content of WwTW effluents (Rodgers-Gray et al., 2000; Martinovic et al., 2008). Despite the variations observed, however, concentrations of E1 and E2 measured in effluent I were consistently high and were comparable with those previously reported for this effluent (Rodgers-Gray et al., 2000). Concentrations of E1 and E2, measured in effluents II and III, were also representative of concentrations previously measured for effluents from these two WwTWs (Thorpe et al., 2006). The results obtained using the rYES indicated that effluents I and III were similar in their total estrogenic activity but this varied from the expected activity profiles compared with the analytical chemistry for E1 and E2. Discrepancies between the results of the

170

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

rYES and measurements of the individual steroidal estrogens have previously been reported in investigations for WwTWs and are hypothesised to be due to the presence of other chemicals within the complex matrix of WwTW effluents that influence activity in the rYES (Thorpe et al., 2006). This has not been proven, however, and requires a detailed study to address this issue. It does, however, highlight the need for caution when relying only on analytical determinants and/or rYES data for assessments of the estrogenic content of WwTW effluents (and possibly for other complex chemical mixtures too). 4.2. Stability of the estrogenic component within the test system Investigations into the fate and behaviour of E1 and E2 in the aquatic environment have shown that half-lives for E1 and E2 (0.2–9 days in river water; Jurgens et al., 2002) can be increased 2fold when the incubation temperature is reduced from 20 to 10 ◦ C (Jurgens et al., 2002). The effluents used within this investigation were therefore held at a temperature of 8–10 ◦ C prior to being dosed to the exposure tanks, to minimise degradation. Concentrations of E1 and E2 measured at the time of delivery and on the final day of use (day 3 for effluent I; day 7 for effluent II) were shown to be highly comparable in the individual batches of effluents, thus confirming the effectiveness of these storage conditions for the natural estrogens. Further analysis using the rYES, however, indicated up to 3-fold fluctuations in the daily measurements of estrogenic activity for each batch of effluent. Although it is possible that this may be due to degradation of other chemicals within the effluent that affect the measurements of estrogenic activity in the rYES there was no evidence for consistent changes in estrogenic activity with time of storage. It is more likely that these fluctuations resulted, at least in part, from inter-sample variations in the efficiency of the extraction procedures and the rYES measurements. For all three WwTW effluents the measures of estrogen content (concentrations of E1 and E2, and estrogenic activity in the rYES) in the exposure (100% effluent) tanks differed from those measured in the effluent storage tanks. For effluent I, the lower E1 and E2 concentrations and estrogenic activity measured in the exposure tanks may have resulted, at least in part, from the exposure system employed which required the effluent to be transferred in bulk twice daily from the storage area to the testing facility. This system involved a prolonged period (up to 12 h) of holding the effluent at the test temperature (25 ◦ C) and this likely resulted in an increased degradation of the steroidal estrogens (Layton et al., 2000; Jurgens et al., 2002). For effluents II and III the exposure system was, therefore, adapted to enable us to continuously pump the effluent from the chilled effluent storage tank to the dosing system. This removed the need to hold the effluent for extended periods at the test temperature. Our ability to assess the stability of the estrogenic content, within the exposure tanks, when using this system, however, was compromised by the observation of elevated concentrations of E1 and E2, and total estrogenic activity in the exposure and the control tanks. To determine whether these elevated concentrations were due to background concentrations of estrogen in the dilution water supply, additional water samples were collected at the point of entry to the dosing system. Analysis of these samples did indicate the presence of both E1 and E2 in the dilution water itself however measured concentrations were low and were comparable to concentrations measured in the analytical blanks within each experiment. This implies that their detection was due, at least in part, to an artefact of the extraction and analytical procedures. Measures of total estrogenic activity (rYES) in the dilution water at the point of entry to the test system were higher (5 ng/L) than measured in the analytical blanks, suggesting that the dilution water was a contributing source to the enhanced total estrogenic activity

measured in the control and exposure tanks in experiments II and III. The source of this estrogenic activity within the dilution water is unknown, but may include alkylphenols, phthalates and other plasticizers that leached from the water storage and piping system. Irrespective of their source, the concentrations of estrogen detected in the dilution water could not completely account for the enhanced concentrations measured in the exposure tanks during the exposures to effluents II and III. It is likely that the enhanced estrogenic activity and concentrations of E1 and E2, in particular, in the exposure tanks originated from the adult females excreting endogenous estrogens into the water combined with the relatively slow tank water replacement times used. To further investigate the stability of E1 and E2 in the exposures for effluents II and III, we subsequently tested the system with pairs of adult fathead minnow only (rather than mixed groups of 8 male and 8 female fish). This was found to reduce background concentrations of both E1 and E2 to <2.1 and <0.9 ng/L, respectively, enabling us to subsequently establish a high stability for both E1 and E2 within the exposure system. As an example, for E1 in subsequent exposures the average concentrations in the 100% effluent exposure tanks were 92.7, 2.6 and 4.1 ng/L for effluents I, II and III, respectively, compared with average concentrations in the delivered effluents of 70.5, 1.6 and 3.8 ng/L. 4.3. Biological responses in fish exposed to the estrogenic WwTW effluents Exposure to each of the WwTW effluents induced a significant elevation of VTG in the plasma of male fish and the concentrations induced were proportional to the measured concentrations of E1 and E2 in the three effluents; concentrations of VTG were 142 ± 50, 25 ± 18 and 13 ± 3 ␮g/mL in males exposed to 100% effluents I, II and III, respectively. The observation of this relationship provides further evidence that these two estrogens are major contributors to the estrogenic potency of a WwTW effluent. Furthermore, it demonstrates that the daily fluctuations in estrogenic contents observed in the exposure tanks did not markedly affect the overall VTG response; the magnitude of the VTG response was determined by the average concentration of steroidal estrogens measured in the delivered effluent. Interestingly, however, no relationship was observed between the magnitude of the VTG response and the in vitro estimates of the overall estrogenic potency (rYES) for the three effluents. This again raises important questions regarding the use and robustness of this (and possibly other) in vitro assays for surveying the estrogenic potency of WwTW effluents to fish. The concentrations of VTG measured in males exposed to effluent I were also found to be consistent with those previously reported for in situ exposures to this effluent (Rodgers-Gray et al., 2001), supporting the fact that the laboratory exposure system provided a realistic exposure scenario for identifying the estrogenic activity of a representative potent effluent. Studies on VTG induction had not previously been conducted for effluents II and III, however, in these studies their biological potency (in terms of inducing a VTG response) was comparable to that previously reported for other WwTW effluents in the UK (Jobling et al., 2006). This demonstrates that the exposure system described can be used to evaluate the in vivo effects of weakly estrogenic effluents, in addition to more potent effluents. However, comparison of the results from the three exposures indicated that the test described was less effective when used as a tool to compare the estrogenic potency of each of the effluents; the effluent with high estrogenic potency (effluent I) was correctly identified but difficulties were encountered in attempting to distinguish between the two effluents with weaker estrogenic activity (effluents II and III) due to different background concentrations of VTG in the control males

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

in each test. Comparison of the concentrations of VTG in males exposed to 100% effluents II and III would imply that effluent III was the least potent effluent tested, however, the relatively high concentrations of VTG observed in the control males in experiment II means that caution should applied in making such a conclusion. The concentrations of VTG measured in control males from experiment II were 20-fold higher than those measured in experiments I and III and, although consistent with concentrations reported by others (<100–4000 ng/mL; Panter et al., 1998; Harries et al., 2000; Pawlowski et al., 2004; Brian et al., 2007; Watanabe et al., 2007), they were proportional to the higher concentration of estrogen measured in the control tanks for this experiment. Therefore, although effluent II induced VTG above control values, the possible contribution of background estrogen in the tanks, to this response in the effluent exposed fish, cannot be ignored. This highlights some of the difficulties involved in attempting to ascribe in vivo potency estimates for natural estrogens and their mixtures when the contribution of exogenous environmental sources of estrogen (e.g. other fish) cannot be completely eliminated. We would, therefore, recommend that other investigators adopting this or similar study designs that employ slow flow-rates or semi-static/static exposure regimes consider using either single-sexes in tanks, or males only, to minimise the influence of fish-derived estrogenic steroids as a confounding factor. Exposure to effluents I and II, but not effluent III, induced a significant elevation of VTG in the plasma of female fish, when compared to control values. The levels of induction in females, however, were small compared to those observed in the males, and there was no clear relationship between the induced VTG concentrations and the measured concentrations of estrogen in each effluent; concentrations of VTG in females exposed to the full-strength effluents were 729 ± 67, 117 ± 5 and 1112 ± 115 ␮g/mL, for effluents I, II and III, respectively. This contrasts with both the measured concentrations of E1 and E2 and the VTG responses observed for the males, implying that effluent III is the more potent effluent and effluent II the least potent. These conflicting results are a consequence of the high background concentrations of VTG in the plasma of mature females, reducing the sensitivity of the vitellogenic response in females (Harries et al., 2000; Pawlowski et al., 2004; Thorpe et al., 2007). In each of the three experiments, VTG concentrations also differed in the control females and thus resulted in different relative sensitivities to the effluents. These results highlight some of the complicating factors in the use of VTG in mature females as an endpoint for testing the estrogenic potency of effluents, compared with that in males. The suppressive effects on gonadal growth (lower GSI) of the estrogen in the positive control tanks in females (significant decrease in experiments I, II and II) and males (significant decrease in experiment III only) was consistent with expectation for fish at this stage of sexual development (Panter et al., 1998; Pawlowski et al., 2004). In contrast to the effects of the estrogen in the positive control tanks, however, there were no effects of the effluents on ovary growth in females and in males there was an enhanced growth of the testis. Previous investigations into the effects of estrogenic WwTW effluents have found both inhibitions (Diniz et al., 2005) and stimulations (Barber et al., 2007) of gonadal growth, and some have found no effects at all on gonadal development (RodgersGray et al., 2000; Robinson et al., 2003). These disparities in the effects of estrogenic WwTW effluents compared with individual estrogens would suggest either some sort of matrix effect due to other chemicals within the effluent reducing the potency for effects on gonadal growth, or the presence of chemicals that were stimulatory for gonadal development thus negating the suppressive effect of estrogen. The higher GSI in the effluent exposed males relative to the controls would suggest that the latter is a likely possibility.

171

5. Conclusions The experiments reported here describe the development of an exposure system that uses of an effluent storage facility (chilled to 8 ◦ C) and a flow-through system to provide a realistic scenario for effluent exposures to fish. The estrogenic component of the effluents, although variable between batches of effluent, was shown to be stable within the exposure system and the magnitude of the VTG response in males exposed to the effluents was proportional to the measured concentrations of both E1 and E2. This indicates that the exposure system described could be used as a practicable and easily transferable experimental system to enable the estrogenic effects (and possibly wider health effects) of effluents to be determined in a controlled laboratory setting, using a standardised protocol. Controlled exposures to WwTW effluents and other environmentally relevant mixtures of chemicals are much needed to advance our understanding of the long-term consequences of WwTW effluent discharges for wild fish populations. Acknowledgements This work was co-sponsored by the UK Environment Agency and by AstraZeneca on grants awarded to CRT. References Allen, Y., Scott, A.P., Matthiessen, P., Haworth, S., Thain, J.E., Feist, S., 1999. Survey of estrogenic activity in United Kingdom estuarine and coastal waters and its effects on gonadal development of the flounder (Platichthys flesus). Environ. Toxicol. Chem. 18, 1791–1800. Ankley, G.T., Johnson, R.D., 2004. Small fish models for identifying and assessing the effects of endocrine-disrupting chemicals. ILAR J. 45, 469–483. Barber, L.B., Lee, K.E., Swackhamer, D.L., Schoenfuss, H.L., 2007. Reproductive responses of male fathead minnows exposed to wastewater treatment plant effluent, effluent treated with XAD8 resin, and an environmentally relevant mixture of alkylphenol compounds. Aquat. Toxicol. 82, 36–46. Brian, J.V., Harris, C.A., Scholze, M., Kortenkamp, A., Booy, P., Lamoree, M., et al., 2007. Evidence of estrogenic mixture effects on the reproductive performance of fish. Environ. Sci. Technol. 41, 337–344. Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P., Waldock, M., 1998. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening. Environ. Sci. Technol. 32, 1549–1558. Diniz, M.S., Peres, I., Pihan, J.C., 2005. Comparative study of the estrogenic responses of mirror carp (Cyprinus carpio) exposed to treated municipal sewage effluent (Lisbon) during two periods in different seasons. Sci. Total Environ. 349, 129– 139. Folmar, L.C., Denslow, N.D., Kroll, K., Orlando, E.F., Enblom, J., Marcino, J., et al., 2001. Altered serum sex steroids and vitellogenin induction in walleye (Stizostedion vitreum) collected near a metropolitan sewage treatment plan. Arch. Environ. Contam. Toxicol. 40, 392–398. Harries, J.E., Runnalls, T., Hill, E., Harris, C.A., Maddix, S., Sumpter, J.P., Tyler, C.R., 2000. Development of a reproductive performance test for endocrine disrupting chemicals using pair-breeding fathead minnows (Pimephales promelas). Environ. Sci. Technol. 34, 3003–3011. Jobling, S., Nolan, M., Tyler, C.R., Brighty, G.C., Sumpter, J.P., 1998. Widespread sexual disruption in wild fish. Environ. Sci. Technol. 32, 2498–2506. Jobling, S., Williams, R., Johnson, A., Taylor, A., Gross-Sorokin, M., Nolan, M., et al., 2006. Predicted exposures to steroidal estrogens in U.K. Rivers correlate with widespread sexual disruption in wild fish populations. Environ. Health Perspect. 114, 32–39. Jurgens, M.D., Holthaus, K.I.E., Johnson, A.C., Smith, J.J.L., Hetheridge, M., Williams, R.J., 2002. The potential for estradiol and ethinylestradiol degradation in English rivers. Environ. Toxicol. Chem. 21, 480–488. Kang, I.J., Yokota, H., Oshima, Y., Tsuruda, Y., Hano, T., Maeda, M., et al., 2003. Effects of 4-nonylphenol on reproduction of Japanese medaka, Oryzias latipes. Environ. Toxicol. Chem. 22 (10), 2438–2445. Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., Lazorchak, J.M., Flick, R.W., 2007. Collapse of a fish population after exposure to a synthetic estrogen. Proc. Nat. Acad. Sci. 104, 8897–8901. Layton, A.C., Gregory, B.W., Seward, J.R., Schultz, T.W., Sayler, G.S., 2000. Mineralization of steroidal hormones by biosolids in wastewater treatment systems in Tennessee, USA. Environ. Sci. Technol. 34, 3925–3931. Martinovic, D., Denny, J.S., Schmieder, P.K., Ankley, G.T., Sorensen, P.W., 2008. Temporal variation in the estrogenicity of a sewage treatment plant effluent and its biological significance. Environ. Sci. Technol. 42, 3421–3427. Metcalfe, C.D., Metcalfe, T.L., Kiparissis, Y., Koenig, B.G., Khan, C., Hughes, R.J., et al., 2001. Estrogenic potency of chemicals detected in sewage treatment plant

172

K.L. Thorpe et al. / Aquatic Toxicology 88 (2008) 164–172

effluents as determined by in vivo assays with Japanese medaka (Oryzies latipes). Environ. Toxicol. Chem. 20, 297–308. Nash, J.P., Kime, D.E., Van der Ven, L.T.M., Wester, P.W., Brion, F., Maack, G., et al., 2004. Long-term exposure to environmental concentrations of the pharmaceutical ethynylestradiol causes reproductive failure in fish. Environ. Health Perspect. 112, 1725–1733. Panter, G.H., Thompson, R.S., Sumpter, J.P., 1998. Adverse reproductive effects in male fathead minnows (Pimephales promelas) exposed to environmentally relevant concentrations of natural oestrogens, oestradiol and oestrone. Aquat. Toxicol. 42, 243–253. Parrott, J.L., Blunt, B.R., 2005. Life-cycle exposure of fathead minnows (Pimephales promelas) to an ethinylestradiol concentration below 1 ng/L reduces egg fertilization success and demasculinizes males. Environ. Toxciol. 20, 131–141. Pawlowski, S., van Aerle, R., Tyler, C.R., Braunbeck, T., 2004. Effects of 17␣-ethinylestradiol in a fathead minnow (Pimephales promelas) gonadal recrudescence assay. Ecotox. Environ. Saf. 57, 330–345. Robinson, C.D., Brown, E., Craft, J.A., Davies, I.M., Moffat, C.F., Pirie, D., et al., 2003. Effects of sewage effluent and ethynyl estradiol upon molecular markers of oestrogenic exposure, maturation and reproductive success in the sand goby (Pomatoschistus minutus, Pallas). Aquat. Toxicol. 62, 119–134. Rodgers-Gray, T.P., Jobling, S., Kelly, C., Morris, S., Brighty, G., Waldock, M.J., et al., 2001. Exposure of juvenile roach (Rutilus rutilus) to treated sewage effluent induces dose-dependent and persistent disruption in gonadal duct development. Environ. Sci. Technol. 35, 462–470. Rodgers-Gray, T.P., Jobling, S., Morris, S., Kelly, C., Kirby, S., Janbakhsh, A., et al., 2000. Long-term temporal changes in the estrogenic composition of treated sewage effluent and its biological effects on fish. Environ. Sci. Technol. 34, 1521– 1528. Seki, M., Yokota, H., Matsubara, H., Tsuruda, Y., Maeda, N., Tadokoro, H., Kobayashi, K., 2002. Effect of ethinylestradiol on the reproduction and induction of vitel-

logenin and testis-ova in medaka (Oryzias latipes). Environ. Toxicol. Chem. 21, 1692–1698. Thorpe, K.L., Benstead, R., Hutchinson, T.H., Tyler, C.R., 2007. Associations between altered vitellogenin concentrations and adverse health effects in fathead minnow (Pimephales promelas). Aquat. Toxicol. 85, 176–183. Thorpe, K.L., Cummings, R.I., Hutchinson, T.H., Scholze, M., Brighty, G., Sumpter, J.P., Tyler, C.R., 2003. Relative potencies and combination effects of steroidal oestrogens in fish. Environ. Sci. Toxicol. 37, 1142–1149. Thorpe, K.L., Gross-Sorokin, M., Johnson, I., Brighty, G., Tyler, C.R., 2006. An assessment of the model of concentration addition for predicting the estrogenic activity of chemical mixtures in wastewater treatment works effluents. Environ. Health Perspect. 114, 90–97. Thorpe, K.L., Hetheridge, M.J., Hutchinson, T.H., Scholze, M., Sumpter, J.P., Tyler, C.R., 2001. Assessing the biological potency of binary mixtures of environmental estrogens using vitellogenin induction in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Sci. Toxicol. 35, 2476–2481. Thorpe, K.L., Tyler, C.R., 2006. Biological effect measures for endocrine active chemicals. Environment Agency Science Report No. SC00043/SR. Tyler, C.R., Jobling, S., Sumpter, J.P., 1998. Endocrine disruption in wildlife: a critical review of the evidence. Crit. Rev. Toxicol. 28, 319–361. Tyler, C.R., van Aerle, R., Hutchinson, T.H., Maddix, S., Trip, H., 1999. An in vivo testing system for endocrine disruptors in fish early life stages using induction of vitellogenin. Environ. Toxicol. Chem. 18, 337–347. Vethaak, A.D., Lahr, J., Kuiper, R.V., Grinwis, G.S.M., Rankouhi, T.R., Giesy, J.P., et al., 2002. Estrogenic effects in fish in The Netherlands: some preliminary results. Toxicology 181, 147–150. Watanabe, K.H., Jensen, K.M., Orlando, E.F., Ankley, G.T., 2007. What is normal? A characterization of the values and variability in reproductive endpoints of the fathead minnow, Pimephales promelas. Comp. Biochem. Physiol. C 146, 348–356.