Identification of Synthetic Steroids in River Water Downstream from

Article pubs.acs.org/est

Identification of Synthetic Steroids in River Water Downstream from Pharmaceutical Manufacture Discharges Based on a Bioanalytical Approach and Passive Sampling Nicolas Creusot,†,‡ Selim Aït-Aïssa,*,† Nathalie Tapie,‡ Patrick Pardon,‡ François Brion,† Wilfried Sanchez,† Eric Thybaud,§ Jean-Marc Porcher,† and Hélène Budzinski‡ Institut National de l’Environnement Industriel et des Risques (INERIS), Unité Écotoxicologie in vitro et in vivo, Parc ALATA, BP2, F-60550 Verneuil-en-Halatte, France ‡ EPOC/LPTC, Université Bordeaux 1− UMR 5805 CNRS, F-33405 Talence, France § Institut National de l’Environnement Industriel et des Risques (INERIS), Pôle Dangers et Impacts sur le Vivant, Parc ALATA, BP2, F-60550 Verneuil-en-Halatte, France †

S Supporting Information *

ABSTRACT: A bioanalytical approach was used to identify chemical contaminants at river sites located downstream from a pharmaceutical factory, where reproductive alterations in wild fish have been previously observed. By using polar organic compound integrative samplers (POCIS) at upstream and downstream sites, biological activity profiles based on in vitro bioassays revealed the occurrence of xenobiotic and steroid-like activities, including very high glucocorticoid, antimineralocorticoid, progestogenic and pregnane X receptor (PXR)-like activities (μg standard-EQ/g of sorbent range), and weak estrogenic activity (ng E2-EQ/g of sorbent range). Chemical analyses detected up to 60 out of 118 targeted steroid and pharmaceutical compounds in the extracts. In vitro profiling of occurring individual chemicals revealed the ability of several ones to act as agonist and/or antagonist of different steroids receptors. Mass balance calculation identified dexamethasone, spironolactone, and 6-alpha-methylprednisolone as major contributors to corticosteroid activities and levonorgestrel as the main contributor to progestogenic activities. Finally, RP-HPLC based fractionation of POCIS extracts and testing activity of fractions confirmed identified compounds and further revealed the presence of other unknown active chemicals. This study is one of the first to report environmental contamination by such chemicals; their possible contribution to in situ effects on fish at the same site is suggested.

■

INTRODUCTION Occurrences of endocrine disruption in aquatic wildlife have been well documented in various environmental contexts around the world.1 In most cases, chemical contaminants have often been suspected, and sometimes proved, to be the cause of such ecotoxicological impacts. To date, most field investigations have focused on estrogenic compounds, such as steroid estrogens or surfactants, because these substances were among the first identified as endocrine disrupting compounds (EDCs), potently active to aquatic organisms and widely released from domestic wastewater treatment plant (WWTP) discharges.2 In addition to (xeno)estrogens, there is now recent evidence of the occurrence of other natural and synthetic steroids released from the effluents from WWTPs handling water from urban, industrial, or hospital activities.3−5 These contaminants include androgens, corticosteroids, and progestagens.6 Such hormonally active compounds can interfere with the regulation of endocrine systems, through the binding to steroid receptors, such as androgen (AR), glucocorticoid (GR), mineralocorticoid (MR) or progestagen receptors (PR), and © 2014 American Chemical Society

subsequent modulation of target genes involved in various essential physiological functions, such as reproduction, differentiation or stress response. Recent reviews have reported that aquatic organisms could be at risk when exposed to the predicted environmental concentrations of such compounds, that is, in the ng/L range.6,7 There is therefore a major need to better characterize the occurrence of these compounds and their effects on aquatic organisms and ecosystems. Wastewater effluents from pharmaceutical factories have recently been pointed out as potential sources of pharmaceuticals and other active compounds released into the aquatic environment, by contaminating receiving water bodies at concentrations ranging from the ng/L up to mg/L range and generating adverse effects on wildlife.8−11 In a previous study, we reported the occurrence of adverse reproductive effects such Received: Revised: Accepted: Published: 3649

April 27, February February February

2012 27, 2014 28, 2014 28, 2014

dx.doi.org/10.1021/es405313r | Environ. Sci. Technol. 2014, 48, 3649−3657

Environmental Science & Technology

Article

as vitellogenin induction, intersex, and sex-ratio alteration, in wild fish populations of gudgeon in French river sites located downstream from both pharmaceutical industry and urban WWTP discharges.12 In the present study, we aimed to assess which type of chemical contaminants were present downstream of WWTPs and identify possible candidates responsible for observed adverse effects. For this purpose, a bioanalytical approach based on a panel of in vitro bioassays and chemical analyses was implemented to establish contamination profiles in both sediments and water phase, at one upstream and two downstream sites. Because of the specific context of the studied sites, that is, a small mountain river with high seasonal variations in flow rate, we used integrative sampling (i.e., POCIS) instead of grab sample in order to minimize representativeness problems due to low and fluctuating concentrations.13 This approach also allowed water samples to be collected over a 6-month period (6 campaigns of one month each), thereby integrating over time possible variations linked to industry activity. Biological activity profiles based on in vitro activation of estrogen (ER), AR, GR, MR, PR, aryl hydrocarbon (AhR), and pregnane X receptors (PXR) revealed polar corticosteroids and progestogen compounds as the main water-active contaminants downstream from pharmaceutical industry. Major active compounds detected by the bioassays were partly identified on the basis of targeted chemical analyses and fractionation by reverse phase-high performance liquid chromatography (RPHPLC) coupled to bioassays. Their potential effects in fish are discussed in light of the effects observed at the same site.

Figure 1. Map of river sites. Sediment and POCIS samples were collected upstream (Site A), between (Site B) and downstream (Site C) of pharmaceutical industry (P), and urban (U) WWTP discharges.

POCIS were deployed from June to November 2009 for six sampling campaigns of 1 month’s duration each. At each site, six POCIS were deployed per sampling date. POCIS preparation and extraction procedure were as previously described.14 Briefly, after collection, POCIS-sorbent (Oasis HLB phase) was extracted on an SPE system with sequential elution of 10 mL DCM, 10 mL DCM/methanol mixture (50:50 v/v) and 10 mL methanol. This mixture was evaporated under N2 flux until dryness and residues were redissolved in 200 μL of methanol for in vitro bioassays and chemical analyses. Fractionation Procedure. In order to isolate the active chemicals, POCIS extracts were fractionated by using a previously calibrated and validated protocol.15 In brief, fractionation was performed using a HPLC system (Prostar series, Agilent) equipped with a C18 column (Poursuit C18, 5 μm, 250 × 4,6 I.D, Agilent, Les Ulis, France). POCIS extract samples were separated at 25 °C at a flow rate of 1 mL/min using the following water:acetonitrile (v:v) gradient: 0−10 min (80:20), 10−60 min (80:20 to 55:45), 60−100 min (55:45 to 0:100), 100−120 min (0:100), 120−125 min (80:20). This fractionation procedure has been calibrated using a mixture of 35 chemicals that presented a broad range of hydrophobicity (−1 < log Kow < 7).15 The standard chemicals included some of the most abundant compounds detected in these POCIS extracts. A procedural blank (solvent only) was always run in parallel to check for possible contamination by the system. In Vitro Bioassays. Samples were tested in vitro for hormone and dioxin receptor agonists and antagonists by using a panel of seven permanent cell lines, namely MELN, MDAkb2, HG5LN-hPR, HG5LN-hMR, HG5LN-hPXR, and PLHC1, allowing the specific detection of ligands of ER, AR/GR, PR, MR, PXR, and AhR, respectively (cell lines summarized in SI Table S2). MDA-kb2 and PLHC-1 were obtained from the American Type Culture Collection. MELN and HG5LNderived cells were kindly provided by Dr Patrick Balaguer (INSERM Montpellier, France). Descriptions of cell lines, cell culture conditions and protocols for the assessment of EDCs in environmental extracts have been reported previously in detail.16−19 Briefly, cells were seeded in 96-well white opaque culture plates (Greiner CellStar; Dutscher, Brumath, France) at a density of 1 × 105 cells per well. After 24 h, cells were exposed to serial dilutions of reference ligands or environmental extracts in triplicates and left to incubate for 16 h. For antagonist activity assessment,

■

MATERIALS AND METHODS Chemicals and Reagents. Biological reagent grade methanol and dimethylsulfoxide (DMSO) were purchased from Sigma-Aldrich (St-Quentin-Fallavier, France); heptane and acetone (HPLC reagent grade) were purchased from VWR (France). 17β-Estradiol (E2), SR12813, spironolactone, flutamide and dexamethasone were purchased from SigmaAldrich (St-Quentin-Fallavier, France), and 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) was from LGC Standards (Molsheim, France). R1881, aldosterone and R5020 were a kind gift from Dr Patrick Balaguer (INSERM, France). All chemicals used as standards in analytical methods are listed in Supporting Information (Table S1). They were purchased either from Sigma-Aldrich or LGC Standards and were of analytical grade. Study Sites, Sampling, and Extraction Procedures. Location and environmental pressures of the investigated French river were as previously detailed.12 In the present study, three sites were sampled alongside a section of this river, from upstream to downstream (Figure 1): Site A was located 2 km upstream of a pharmaceutical industry effluent discharge; Site B was 0.1 km downstream of the industry discharge and 0.1 km upstream of an urban WWTP discharge; and Site C was 1 km downstream of the urban WWTP discharge. Sediments were sampled with a Van Veen grab sampler, sieved (mesh size, 1 mm), freeze-dried, homogenized, and stored in amber glass bottles at −20 °C before analysis. Using heptane/acetone (50:50 v/v), 5 g of dry sediment were extracted on an accelerated solvent extractor (ASE, Dionex, France) at 100 °C and 1500 PSI. Extracts were evaporated to 500 μL with a Turbovap II (Zymark) and then to dryness under N2. Residues were dissolved in 1 mL of methanol. TOC was calculated in accordance with the International Organization of Standardization recommendation (ISO 8245). 3650



Article

Figure 2. Biological activities detected in POCIS (Oasis HLB) extracts deployed for 1 month upstream (Site A), between (Site B) and downstream (Site C) the pharmaceutical and the urban WWTPs between June and November 2009. a: estrogenic activity in MELN cells; b: progestative activity in HELNPRB cells; c: glucocorticoidic activity in MDA-kb2 cells; d: antimineralocorticoidic activity in HG5LN-hMR cells; e: PXR-like activity in HG5LN-hPXR cells, f: dioxin-like activity in PLHC-1 cells.

fitting, top and slope curve values were determined and used as fixed parameters in the modeling of sample curves, in order to fulfill the requirement for curve parallelism. The hormone-like or dioxin-like activities in samples derived from bioassays were expressed as toxic-equivalents (bio-TEQs), which were determined as the ratio of the EC20 of the reference chemicals expressed as g/L to that of the sample expressed as equivalent gram of dry weight sediment per liter (g EQ/L) or equivalent gram of oasis HLB phase per liter for POCIS extracts. Toxicequivalents derived from chemical analyses (chem-TEQ) were determined according to the following equation: chem-TEQ = ∑ (Ci × REPi), where, for a given chemical i, Ci is the measured concentration in a sample and REPi is the inducing equivalent factor relative to the reference ligand (REPi = EC20 of reference compound/EC20 of test compound, on mass basis). Individual dose−response curves for all samples and individual chemicals that were used to determine toxicequivalents in each bioassay (bio-TEQ) are presented in the SI (Figures S1 to S4).

cells were coexposed to concentration ranges of environmental extracts and fixed concentrations of reference compounds that yield 80% of the maximal response. In the MDA-kb2 assay, which expresses both GR and AR, involvement of GR in agonist response induced by samples and individual chemicals was checked by coexposure with 1 μM RU-486, a GR antagonist. After exposure, the medium was removed and replaced by 50 μL per well of medium containing 0.3 mM luciferin. Luminescence signal was monitored in living cells with a microtiter plate luminometer (MicroBeta Wallac Luminometer). In the PLHC-1 bioassay, cells were exposed for 4 h (PAH-like activity) and 24 h (dioxin-like activity) and were then processed for 7-ethoxyresorufin-O-deethylase (EROD) activity assay in living cells as previously described.16 Chemical Analyses. Chemical analyses of POCIS extracts were targeted on pharmaceuticals including synthetic steroids, antibiotics, nonsteroidal anti-inflammatory drugs (NSAID) and other drug families. Targeted chemical analyses were performed using ultra high performance liquid chromatography instruments coupled to tandem mass spectrometry (RRLC-MS/MS from Agilent; UPLC-MS/MS from Waters) in multireaction monitoring (MRM) mode. List of analyzed chemicals and associated analytical methods is given in SI Table S2. Calculation of Bioassay- And Instrumentally Derived Toxic-Equivalents. Modeling of dose−response curves and determination of EC20 (i.e., concentration of extracts or chemical standards giving 20% of the maximum luciferase or EROD activities) was done with the Regtox 7.5 Microsoft Excel macro by using the Hill equation model (freely available at http://www.normalesup.org/∼vindimian/). The concentration−response curve for reference compound (e.g., E2 in the MELN assay) was first fitted to Hill equation. From that curve

■

RESULTS AND DISCUSSION Biological Activities in Sediments and POCIS Extracts. Bioassay analyses revealed different contamination profiles in sediments and POCIS. In sediments (SI Table S4), only estrogenic and PAH-like activities were detected at all sites and at similar levels, which were comparable to that reported in weakly to moderately contaminated sites in other field studies.18,20,21 In addition, no other in vitro activity could be detected, suggesting that, at these sites, EDCs did not accumulate in sediments. Such low activity levels could be 3651



Article

Table 1. In Vitro Endocrine Disrupting Profiles of the 28 Most Abundant Chemicals Detected in POCIS at Downstream Site Ba GR (MDAkb2)

Anti-GR (MDAkb2)

PR (HELNPRB)

MR (HG5LNMR)

Anti-MR (HG5LNMR)

PXR (HG5LNhPXR)

ER (MELN) reference compounds

estradiol

R1881

flutamide

dexamethasone

RU486

R5020

aldosterone

spironolactone

SR12813

EC20 (nM)

0.002

0.06

18.5

3.2

2.0

0.14

0.35

0.75

19

Active Compounds pregnenolone progesterone levonorgestrel spironolactone canrenone dexamethasone prednisone prednisolone 6αmethylPrednisolone cortisone hydrocortisone androstenedione testosterone dehydroepiandrosterone gemfibrozil carbamazepine Nonactive Compounds a

Anti-AR (MDAkb2)

AR (MDAkb2)

Relative potencies of test chemicals (REP = EC20 ref/EC20 chemical) 1.1 × 10−5 na 0.45 na na na na 0.048 2.4 × 10−3 7.5 × 10−7 na na na na 0.029 na 0.10 na 3.1 × 10−5 0.11 na na na 0.9 na 0.17 na na na 15.4 1.6 × 10−4 3 × 10−3 1.9 × 10−5 na 1 5.7 × 10−3 −4 −5 na na 10.6 na 4 × 10 6.8 × 10 na 0.3 na nd na na 1 na na na 0.27 2.1 × 10−4 na na na 1.6 × 10−4 7 × 10−4 na na 0.16 na nd na na 0.20 na na 1.3 × 10−2 0.16 na na na na 0.60 na na 1.4 × 10−5 0.45 na na na na na na na na 0.12 na na na na na na na na 0.28 × 10−4 na −7 −2 9.7 × 10 1.3 × 10 na na na na na 0.01 na na 2.8 × 10−2 na na na na na na na 2.9 × 10−4 na na na na na na na na 4.7 × 10−8 na na na na na na na 2.5 × 10−3 −5 na na na na na na na 3.7 × 10 6.0 × 10−4 Stavudine, sulfamethoxazole, propanolol, atenolol, sotalol, bisoprolool, clarythromycin, roxithromycin, caffein, theophyllin, paracetamol, bromazepam

EC20 and REP values are derived from dose-response curves presented in SI Figures S4 and S5. na: non active chemical.

related to the nature of the sediment, which had very low total organic carbon content (SI Table S3). The information on site contamination provided by POCIS sampling was different from that provided by sediment. Indeed, all investigated activities were detected at sites B and C, located downstream from the industrial and urban discharges (Figure 2). A remarkable finding was the significant glucocorticoid (11−180 μg dexamethasone-EQ/g sorbent), antimineralocorticoid (5−125 μg spironolactone-EQ/g sorbent) and progestagenic (0.4−3.5 μg R5020-EQ/g sorbent) activities that were quantified downstream of the pharmaceutical industry (Site B) and the urban WWTP (Site C), while nondetectable levels were recorded at the upstream site (Site A). PXR activity was detected at the upstream site but significantly increased at downstream sites, suggesting that PXR active chemicals were released by the WWTPs. The finding that similar in vitro profiles were observed at Sites B and C suggests that most of the detected compounds were mainly issued from the industrial discharge, with a minor contribution of the urban discharge. It is also noteworthy that in vitro activity profiles varied according to the period of sampling. For instance, the highest PR activity levels were measured in August (Figure 2b), while GR, anti-MR and PXR activities increased from August to November (Figure 2c−e); the lowest activity levels being measured in June and July. These different patterns of contamination over time might reflect qualitative changes in industrial discharges throughout the 6-month period of sampling. Some weak estrogenic (Figure 2a) and PAH-like (Figure 2f) activities were also detected in POCIS extracts, but the influence of discharges or time of sampling was not evidenced, as similar levels were found at upstream and downstream sites, the only exception being the peak that was observed for PAHlike activity in September at Site B. This suggested that the two WWTPs were not a major source of compounds responsible for

these activities. Together with activity profiles in sediment extracts, these data further show a low contamination of the river by ER and AhR active compounds at the investigated sites. Chemical Screening in POCIS Extracts. Bioassays showed that steroid-like compounds were present in high concentrations, downstream from the industrial discharge. To identify the nature of detected compounds, target chemical analyses were directed on natural and synthetic steroids (estrogens, androgens, progestagens and corticosteroids) and pharmaceutical compounds (antibiotics, nonsteroidal antiinflammatory drugs, antidepressants). Analyzed compounds were selected either as well-known activators of the detected in vitro activities and/or among the most commonly found pharmaceuticals in the environment. Out of the 118 target compounds, up to 60 were detected in POCIS from Sites B and C (SI Table S5). The contamination patterns revealed dexamethasone, spironolactone, 6-alpha-methylprednisolone, canrenone, hydrocortisone, prednisolone, prednisone, as well as cortisol and cortisone, as the most abundant compounds in the extracts. At the upstream site, none of these chemicals could be detected, while other pharmaceuticals such as caffeine, carbamazepin and roxithromycin were present (SI Table S5). Recent field studies have described steroid hormones such as glucocorticoids and progestagens as being continuously released by different sources such as hospital, industrial or urban discharges.3,5,22,23 Our data further suggest the potential contribution of pharmaceutical industry as a possible source for these compounds in the aquatic environment. Spironolactone is a mineralocorticosteroid antagonist drug that was the most prescribed in the United Kingdom in 20066 but no data are available on its environmental occurrence. To our knowledge, the present study is the first one that reports such a contamination of surface water by both this compound and its metabolite canrenone. Finally, the natural androgens testosterone and androstenedione were present only down3652



Article

Table 2. Comparison of Toxic-Equivalents (TEQ) Derived from Chemical (Chem-TEQ) and Bioassay Analyses (Bio-TEQ) in POCIS (Oasis HLB) at Site Ba

a

June

July

October

November

ER (ng E2-EQ/g Oasis HLB)

Bio-TEQ Chem-TEQ % chem/bio

2.9 0.0001 0.005%

1.0 0.0001 0.008%

2.4 0.03 1.3%

August

September 2.3 0.004 0.2%

3.2 0.2 4.8%

1.8 0.1 5.8%

PR (ng R5020-EQ/g Oasis HLB)


39.5 0.6 1.4%

32.5 0.1 0.4%

1556.0 888.7 57.1%

892.8 121.0 13.6%

381.5 0.7 0.2%

346.3 51.9 15.0%

GR (μg Dex-EQ/g Oasis HLB)


61.4 46.9 76.3%

11.1 9.3 83.9%

45.3 12.1 26.6%

96.3 32.0 33.2%

91.5 23.0 25.1%

180.3 107.7 59.7%

anti-MR (μg Spiro-EQ/g Oasis HLB)


14.4 34.2 237.1%

5.0 3.9 78.2%

11.7 12.4 106.1%

29.1 14.9 51.3%

62.4 25.5 40.9%

72.2 41.6 57.6%

PXR (μg SR-EQ/g Oasis HLB)


20.6 0.084 0.4%

8.8 0.005 0.1%

53.6 0.012 0.02%

76.0 0.002 0.003%

57.6 0.033 0.1%

99.6 0.01 0.01%

% chem/bio = ratio of Bio-TEQ to Chem-TEQ.

The determination of REPs allowed toxic equivalent values to be derived from chemical analyses (chem-TEQ) that were further compared to bioassay derived toxic equivalents (bioTEQs) in POCIS extracts from Site B. With this approach, analyzed target chemicals were well explicative of measured GR and anti-MR activities. Chem-TEQs explained 25%−90% (except for anti-MR in June, see Discussion below) of bioTEQs, depending on type of activity and date of sampling (Table 2). Examination of the contribution of individual compounds identified dexamethasone and 6α-methyl-prednisolone as the main contributors to GR activity (SI Table S6). These two compounds, together with spironolactone and, to a lesser extent, cortisone, prednisone and prednisolone, were also contributors to anti-MR activity measured by the bioassay (SI Table S7). Our findings are in line with previous reports of a significant contribution of dexamethasone and predisolone in the prediction of GR activity of industrial and hospital effluents.4 Some differences between bio- and chem-TEQs in POCIS extracts were, however, noted. For instance, an overprediction of anti-MR activity was observed in June, which could be explained by the presence of a mixture of both MR agonists and antagonists leading to a reduced anti-MR activity of the mixture. This was indeed confirmed by mixture fractionation and testing of fractions (see below). Such a masking effect in a mixture containing both receptor agonists and antagonists has already been reported for ligands of the androgen receptor detected in sediment samples.29 The other measured activities, that is, ER, PR, and PXR activities, were only weakly explained by analyzed chemicals (Table 2). For instance, dehydroepiandrosterone (DHEA) was found as the main estrogenic compound in October and November, although it explained at best 5% of overall activity of the extract (SI Table S8). Levonorgestrel (0−987 ng/g of sorbent) was identified as the main progestogenic compound, accounting for up to 50% in the August sample but much less in other campaigns (SI Table S9). Spironolactone and progesterone also contributed weakly to the progestogenic activity (SI Table S9). Overall, the weak

stream of the industrial discharge, whereas natural and synthetic estrogens were not detected at any of the examined sites. Such an unusual contamination pattern is notable since estrogens are generally detected in urban effluents, together with androgen contamination.19,23,24 In addition, this seems at odds with the occurrence of abnormal synthesis of vitellogenin in individual male fish in the river.12 Overall, these data raise the question of endocrine disrupting effects of steroids other than (xeno)estrogens in fish. In Vitro Activity of Individual Contaminants and Mass Balance Analysis. In an attempt to evaluate the contribution of detected chemicals to in vitro biological activities in POCIS extracts, the potency of individual compounds to induce a biological response was first determined by establishment of dose−response curves (SI Figure S4) and calculation of their relative potencies (REPs) in each bioassay (Table 1). Out of 28 compounds tested, 16 were able to induce a positive response in at least one bioassay and 12 in at least two bioassays (Table 1). For instance, GR agonists were also potent MR ligands, which is in accordance with the common evolution of glucocorticoid and mineralocorticoid receptors that share overlapping physiological functions and several ligands.25−27 The same is observed for PR activators, which were also antiMR ligands. In addition, certain compounds were able to exert multiple nuclear receptor-mediated activities. For instance, spironolactone, a well-known MR antagonist, also exerted progestogenic, (anti)glucocorticoid, and PXR-like activities in our assays. In MDA-kb2 cells, where both AR and GR are expressed and modulate luciferase expression, spironolactone was able to induce a partial GR-mediated agonistic response as well as a GR and AR antagonistic response after coexposure with dexamethasone and dihydrotestosterone, respectively (Table 1, SI Figure S5). This compound has been recently reported as an androgenic compound in fathead minnow (fhm) in vivo, as well as in vitro using a transient reporter gene assay based on fhmAR expression in CV-1 cells.28 In that study, as in the present one, it was found only weakly active in MDA-kb2 cells. 3653



Article

Table 3. Biological Activities in RP-HPLC Fractions of POCIS Extract from the Site C (Sampled in October 2009)

consistency between chemical and biological analyses for these activities could be due to the fact that chemical analyses targeted a limited set of specific chemicals, while other active compounds including metabolites of identified steroids were likely present. Moreover, although mass balance analysis based on addition concentration has proven a useful model to describe in vitro quantification of mixtures of steroid-like compounds such as (xeno)estrogens18 or GR agonists,4 it cannot be excluded that mixture effects such as synergy or antagonism could have contributed to deviations between predicted and measured activity in the present case study. Isolation of Known and Unknown Active Compounds in POCIS Extract. In an attempt to isolate active compounds from mixtures, the POCIS extract was fractionated using RPHPLC to yield 40 subfractions that were further tested in the

different bioassays. The fractionation protocol was calibrated using a mixture of standard chemicals, including both the most abundant ones detected in samples and a panel of environmental EDCs with logKow comprised between −1 and 7.15 RPHPLC profiles for the different activities are reported in Table 3. Overall, these results confirmed GR, anti-MR and PR activities as predominant ones compared to ER or PXR activities. Moreover, it not only confirmed assumptions from mass balance analysis regarding the involvement of target compounds but also highlighted the presence of other, as yet unidentified, active compounds. Four major peaks of GR activity were observed in F11−12, F15, F17−18, and F22. On the basis of the retention times of standard chemicals which was confirmed by targeted LC-MS/ MS analysis of the fractions (data not shown), 6α3654



Article

river sites subjected to pharmaceutical industry release.8,9,11 At our study site, major reproductive abnormalities have been reported in wild fish,12 which further raises the question of the toxicological effects of such compounds and the risk to aquatic species. Unlike estrogens, and to a lesser extent androgens, there is very limited information on the effects of corticosteroids and progestogens on freshwater fish.6 Yet, due to the ability of these molecules to interfere with multiple signaling pathways through their interaction with different steroid receptors, they can potentially alter physiological processes mediated by these receptors, such as reproductive or developmental functions.31 Some recent studies reported that corticoids and progestogens could significantly affect reproduction and induce masculinization in fish as demonstrated for spironolactone (≥5 μg/L),28,32 norethindrone (≥0.025 μg/L),33 or the cortisol metabolite 5αandrostan-3,11,17-trione (≥1 μg/L).34 The effective aqueous concentrations of spironolactone reported in these studies are in the same order, or nearly above, as those extrapolated in the river from POCIS in our present study. Also, levonorgestrel, a synthetic progestin that interfered with multiple steroid receptors in our study (Table 1), has been shown to cause alteration of ER-regulated genes in the brain of zebrafish embryos,35 inhibition of fathead minnow reproduction at low ng/L range, as well as androgenic effect at higher concentrations.36 In our study, levonorgestrel was measured in three sampling campaigns, hence suggesting chronic exposure of wild fish inhabiting the river. Levonogestrel concentrations in the water were 33, 4, and 2 ng/L (approximated concentrations using a median Rs of 0.1 L/day) in July, August, and October samples, respectively. These values are in the range of effect concentrations reported in previous studies.35,36 Overall, it appears more than likely that the steroids and steroid-like compounds detected and identified in the present study have significantly contributed to fish reproduction alterations observed in the river. Their occurrence as mixtures capable of interfering with multiple steroid receptor pathways makes the accurate prediction of their impacts on fish health more difficult, and need further investigations. In summary, this study stresses the need to increase knowledge of the effects of corticosteroids and progestagens on aquatic organisms. Furthermore, it underscores the need to better characterize and identify such environmental contaminants that interfere with crucial signaling pathways involved in development and reproduction in various environmental contexts. As shown in our study, multireceptor bioassay-based monitoring coupled with effect-directed analysis can usefully serve this purpose.

methylprednisolone and dexamethasone were present in F11 and F12 respectively and were very likely responsible for GR activity in these fractions; however, other unknown glucocorticoid compounds were present in F15, 17−18, and 22 and need to be identified. This fractionation step also further confirmed the involvement of 6α-methylprednisolone (F11), dexamethasone (F12), spironolactone (F18) and androstenedione (F19) in the strong anti-MR activity detected. As for GR activity, the strong antiMR activities detected in F15−16 and F21−23 revealed the presence of unknown active compounds. Assessment of PR activity confirmed the possible involvement of spironolactone in F19 but also highlighted that other unknown PR active compounds were also present in F15−16 and F20−21. Further studies using LC-Q-TOF are under way to identify these unknown active compounds isolated in RP-HPLC fractions (Gardia-Parege, Creusot, Aı̈t-Aı̈ssa and Budzinski, in preparation). Interestingly, the fractionation step also revealed the occurrence of agonistic MR activity in F9 and F11, which was not detected in the global mixture. This stresses the importance of unravelling complex mixtures to reveal active chemicals that could not be detected initially due to matrix or mixture effect. This MR activity was mediated, at least partly, by prednisolone and 6α-methylprednisolone that were coeluted in these fractions. Risks for Aquatic Organisms. This study aimed to characterize chemical contamination of river water at sites where fish reproduction alteration has been observed. Overall, we showed river water contamination by a mixture of steroids and steroid-like compounds and identified GR and PR agonists and MR antagonists as the most abundant compounds in POCIS. However, concentration in POCIS do not quantitatively reflect actual concentration in the water phase because of different sampling rates of individual chemicals, which can vary from 0.01 to >1 across chemical classes.13,14 To our knowledge, sampling rates (Rs) of these identified chemicals are not described. By using a median Rs of 0.1 L/day,13 however, the range of theoretical concentrations in water extrapolated from measured concentrations in POCIS extracts was 0.05−1.26 μg/ L (6α-methylprednisolone), 0.02−0.5 μg/L (spironolactone) and 0.001−2.9 μg/L (dexamethasone) across sampling campaigns at Site B. In addition, some of the compounds measured in POCIS samples were also measured in industrial effluent sampled in October 2009 at the outlet of the WWTP. Dexamethasone, spironolactone and prednisone were found to be present at relatively high concentrations, that is, 23, 10, and 0.3 μg/L, respectively. This range of concentration is in line with extrapolated water concentrations at downstream Site B considering a dilution (approximately by >50-fold) of the effluent in the river. Altogether, these results confirmed the industry sewage water as a source of contamination of the studied river and suggested that the high levels measured in POCIS extracts reflected significant concentrations (i.e., up to the μg/L range as extrapolated concentrations) in the water phase. Such high levels measured just downstream of industrial discharge may render the study site exceptional in light of environmental levels that are mostly reported for steroids and corticoids, that is, in the ng/L to tens of ng/L range in surface waters and effluents.6,30 But this site may not be a unique case and could reflect an emerging environmental concern as ecotoxicological disorders have been recently reported in other

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +33 344 556 511; fax +33 344 556 767; e-mail: selim. [email protected]. Notes

The authors declare no competing financial interest. 3655



■

Article

compounds in multi-contaminated sediment: Identification of novel ligands of estrogen and pregnane X receptors. Anal. Bioanal. Chem. 2013, 405 (8), 2553−2566. (16) Louiz, I.; Kinani, S.; Gouze, M. E.; Ben-Attia, M.; Menif, D.; Bouchonnet, S.; Porcher, J.-M.; Ben-Hassine, O. K.; Aït-Aïssa, S. Monitoring of dioxin-like, estrogenic and anti-androgenic activities in sediments of the Bizerta lagoon (Tunisia) by means of in vitro cellbased bioassays: Contribution of low concentrations of polynuclear aromatic hydrocarbons (PAHs). Sci. Total Environ. 2008, 402, 318− 329. (17) Creusot, N.; Kinani, S.; Balaguer, P.; Tapie, N.; LeMenach, K.; Maillot-Maréchal, E.; Porcher, J. M.; Budzinski, H.; Aït-Aïssa, S. Evaluation of an hPXR reporter gene assay for the detection of aquatic emerging pollutants: Screening of chemicals and application to water samples. Anal. Bioanal. Chem. 2010, 396 (2), 569−583. (18) Kinani, S.; Bouchonnet, S.; Creusot, N.; Bourcier, S.; Balaguer, P.; Porcher, J. M.; Aït-Aïssa, S. Bioanalytical characterisation of multiple endocrine- and dioxin-like activities in sediments from reference and impacted small rivers. Environ. Pollut. 2010, 158, 74−83. (19) Bellet, V.; Hernandez-Raquet, G.; Dagnino, S.; Seree, L.; Pardon, P.; Bancon-Montiny, C.; Fenet, H.; Creusot, N.; Aït-Aïssa, S.; Cavailles, V.; Budzinski, H.; Antignac, J.-P.; Balaguer, P. Occurrence of androgens in sewage treatment plants influents is associated with antagonist activities on other steroid receptors. Water Res. 2012, 46 (6), 1912−1922. (20) Brack, W.; Blaha, L.; Giesy, J. P.; Grote, M.; Moeder, M.; Schrader, S.; Hecker, M. Polychlorinated naphthalenes and other dioxin-like compounds in Elbe River sediments. Environ. Toxicol. Chem. 2008, 27 (3), 519−528. (21) Hurst, M. R.; Balaam, J.; Chan-Man, Y. L.; Thain, J. E.; Thomas, K. V. Determination of dioxin and dioxin-like compounds in sediments from UK estuaries using a bio-analytical approach: Chemical-activated luciferase expression (CALUX) assay. Mar. Pollut. Bull. 2004, 49 (7− 8), 648−658. (22) Labadie, P.; Budzinski, H. Development of an analytical procedure for determination of selected estrogens and progestagens in water samples. Anal. Bioanal. Chem. 2005, 381 (6), 1199−1205. (23) Chang, H.; Wan, Y.; Hu, J. Y. Determination and source apportionment of five classes of steroid hormones in urban rivers. Environ. Sci. Technol. 2009, 43 (20), 7691−7698. (24) Thomas, K. V.; Hurst, M. R.; Matthiessen, P.; McHugh, M.; Smith, A.; Waldock, M. J. An assessment of in vitro androgenic activity and the identification of environmental androgens in United Kingdom estuaries. Environ. Toxicol. Chem. 2002, 21 (7), 1456−1461. (25) Arriza, J. L.; Weinberger, C.; Cerelli, G.; Glaser, T. M.; Handelin, B. L.; Housman, D. E.; Evans, R. M. Cloning of human mineralocorticoid receptor complementary-DNAStructural and functional kinship with the glucocorticoid receptor. Science 1987, 237 (4812), 268−275. (26) Prunet, P.; Sturm, A.; Milla, S. Multiple corticosteroid receptors in fish: From old ideas to new concepts. Gen. Comp. Endrocrinol. 2006, 147 (1), 17−23. (27) Rashid, S.; Lewis, G. F. The mechanisms of differential glucocorticoid and mineralocorticoid action in the brain and peripheral tissues. Clin. Biochem. 2005, 38 (5), 401−409. (28) LaLone, C. A.; Villeneuve, D. L.; Cavallin, J. E.; Kahl, M. D.; Durhan, E. J.; Makynen, E. A.; Jensen, K. M.; Stevens, K. E.; Severson, M. N.; Blanksma, C. A.; Flynn, K. M.; Hartig, P. C.; Woodard, J.; Berninger, J. P.; Norberg-King, T. J.; Johnson, R. D.; Ankley, G. T. Cross species sensitivity to a novel androgen receptor agonist of potential environmental concern, spironolactone. Environ. Toxicol. Chem. 2013, 32, 2528−2541, doi.org/10.1002/etc.2330 (29) Weiss, J. M.; Hamers, T.; Thomas, K. V.; van der Linden, S.; Leonards, P. E. G.; Lamoree, M. H. Masking effect of anti-androgens on androgenic activity in European river sediment unveiled by effectdirected analysis. Anal. Bioanal. Chem. 2009, 394 (5), 1385−1397. (30) Tolgyesi, A.; Verebey, Z.; Sharma, V. K.; Kovacsics, L.; Fekete, J. Simultaneous determination of corticosteroids, androgens, and

ACKNOWLEDGMENTS We thank William Sremski (Office National de l’Eau et des Milieux Aquatiques, Délégation Inter-Régionale AuvergneLimousin, France) for his precious help with field sampling. This work was supported by the French Ministry of Ecology and Sustainable Development (P181-DRC43) via INERIS. We also thank the Région Aquitaine, FEDER and Europe for financial support regarding mass spectrometry equipment facilities. CPER A2E is more specifically acknowledged for the purchase of the LC/MS/MS system.

■

REFERENCES

(1) Sumpter, J. P. Endocrine disrupters in the aquatic environment: An overview. Acta Hydrochim. Hydrobiol. 2005, 33 (1), 9−16. (2) Desbrow, C.; Routledge, E. J.; Brighty, G. C.; Sumpter, J. P.; Waldock, M. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening. Environ. Sci. Technol. 1998, 32 (11), 1549−1558. (3) Chang, H.; Hu, J. Y.; Shao, B. Occurrence of natural and synthetic glucocorticoids in sewage treatment plants and receiving river waters. Environ. Sci. Technol. 2007, 41 (10), 3462−3468. (4) Schriks, M.; van Leerdam, J. A.; van der Linden, S. C.; van der Burg, B.; van Wezel, A. P.; de Voogt, P. High-resolution mass spectrometric identification and quantification of glucocorticoid compounds in various wastewaters in The Netherlands. Environ. Sci. Technol. 2010, 44 (12), 4766−4774. (5) Van der Linden, S. C.; Heringa, M. B.; Man, H. Y.; Sonneveld, E.; Puijker, L. M.; Brouwer, A.; Van der Burg, B. Detection of multiple hormonal activities in wastewater effluents and surface water, using a panel of steroid receptor CALUX bioassays. Environ. Sci. Technol. 2008, 42 (15), 5814−5820. (6) Runnalls, T. J.; Margiotta-Casaluci, L.; Kugathas, S.; Sumpter, J. P. Pharmaceuticals in the aquatic environment: Steroids and antisteroids as high priorities for research. Hum. Ecol. Risk Assess. 2010, 16 (6), 1318−1338. (7) Besse, J. P.; Garric, J. Human pharmaceuticals in surface waters implementation of a prioritization methodology and application to the French situation. Toxicol. Lett. 2008, 176 (2), 104−123. (8) Carlsson, G.; Orn, S.; Larsson, D. G. J. Effluent from bulk drug production is toxic to aquatic vertebrates. Environ. Toxicol. Chem. 2009, 28 (12), 2656−2662. (9) Gunnarsson, L.; Kristiansson, E.; Rutgersson, C.; Sturve, J.; Fick, J.; Forlin, L.; Larsson, D. G. J. Pharmaceutical industry effluent diluted 1:500 affects global gene expression, cytochrome P450 1a activity, and plasma phosphate in fish. Environ. Toxicol. Chem. 2009, 28 (12), 2639−2647. (10) Phillips, P. J.; Smith, S. G.; Kolpin, D. W.; Zaugg, S. D.; Buxton, H. T.; Furlong, E. T.; Esposito, K.; Stinson, B. Pharmaceutical formulation facilities as sources of opioids and other pharmaceuticals to wastewater treatment plant effluents. Environ. Sci. Technol. 2010, 44 (13), 4910−4916. (11) Gilbert, N. Drug waste harms fish. Nature 2011, 476 (7360), 265−265. (12) Sanchez, W.; Sremski, W.; Piccini, B.; Palluel, O.; MaillotMarechal, E.; Betoulle, S.; Jaffal, A.; Ait-Aissa, S.; Brion, F.; Thybaud, E.; Hinfray, N.; Porcher, J. M. Adverse effects in wild fish living downstream from pharmaceutical manufacture discharges. Environ. Int. 2011, 37 (8), 1342−1348. (13) Harman, C.; Allan, I. J.; Vermeirssen, E. L. M. Calibration and use of the polar organic chemical integrative samplerA critical review. Environ. Toxicol. Chem. 2012, 31 (12), 2724−2738. (14) Tapie, N.; Dévier, M. H.; Soulier, C.; Creusot, N.; Le Menach, K.; Aït-Aïssa, S.; Vrana, B.; Budzinski, H. Passive samplers for chemical substance monitoring and associated toxicity assessment in water. Water Sci. Technol. 2011, 63 (10), 2418−2426. (15) Creusot, N.; Budzinski, H.; Balaguer, P.; Kinani, S.; Porcher, J. M.; Ait-Aissa, S. Effect-directed analysis of endocrine-disrupting 3656



Article

progesterone in river water by liquid chromatography-tandem mass spectrometry. Chemosphere 2010, 78 (8), 972−979. (31) Leatherland, J. F.; Li, M.; Barkataki, S. Stressors, glucocorticoids and ovarian function in teleosts. J. Fish Biol. 2010, 76 (1), 86−111. (32) Howell, W. M.; Hunsinger, R. N.; Blanchard, P. D. Paradoxical masculinization of female western mosquitofish during exposure to spironolactone. Prog. Fish-Cult. 1994, 56 (1), 51−55. (33) Paulos, P.; Runnalls, T. J.; Nallani, G.; La Point, T.; Scott, A. P.; Sumpter, J. P.; Huggett, D. B. Reproductive responses in fathead minnow and Japanese medaka following exposure to a synthetic progestin, Norethindrone. Aquat. Toxicol. 2011, 99 (2), 256−262. (34) Grillitsch, B.; Altmann, D.; Schabuss, M.; Zornig, H.; Sommerfeld-Stur, I.; mostl, e. mammalian glucocorticoid metabolites act as androgenic endocrine disruptors in the medaka (Oryzias latipes). Environ. Toxicol. Chem. 2010, 29 (7), 1613−1620. (35) Brion, F.; Le Page, Y.; Piccini, B.; Cardoso, O.; Tong, S. K.; Chung, B. C.; Kah, O. Screening estrogenic activities of chemicals or mixtures in vivo using transgenic (cyp19a1b-GFP) zebrafish embryos. Plos One 2012, 7 (5), 10. (36) Zeilinger, J.; Steger-Hartmann, T.; Maser, E.; Goller, S.; Vonk, R.; Lange, R. Effects of synthetic gestagens on fish reproduction. Environ. Toxicol. Chem. 2009, 28 (12), 2663−2670.

3657


Identification of Synthetic Steroids in River Water Downstream from

Recommend Documents